This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.
Mashnik, Stepan G. (2000). "On Solar System and Cosmic Rays Nucleosynthesis and Spallation Processes". arXiv:astro-ph/0008382. {{cite arXiv}}: |class= ignored (help); Unknown parameter |month= ignored (help)
Can you give me a full list of the possible forms of URL that should be converted to arXiv parameters, noting any similar-looking URLs that should NOT be converted? Thanks, Martin(Smith609 – Talk) 15:08, 7 March 2011 (UTC)
Urls of the form http://arxiv.org/abs/SOMETHING or http://arxiv.org/pdf/SOMETHING should be converted to |arxiv=SOMETHING — due to recent changes in the {{cite journal}} and {{citation}} templates |arxiv= is now a separate parameter so we don't need to go through the |id= parameter. Also www.arxiv.org and xxx.lanl.gov are the same as arxiv.org without the www. The SOMETHING part may have a couple of different formats: either yymm.nnnn or archive/identifier, but I think it's easier just to treat it as a atomic unit. There are some rarer access paths (two of them in here) and some other mirror sites but that should at least get most of them safely. Urls that do not have the "/abs/" or "/pdf/" parts in them should be avoided. —David Eppstein (talk) 23:47, 10 March 2011 (UTC)
The regex I used to match them is \|(\s*)?url(\s*)?=(\s*)?http://(www\.)?(|au\.|br\.|cn\.|de\.|es\.|fr\.|il\.|in\.|jp\.|ru\.|tw\.|ul\.|aps\.|lanl\.|xxx\.)?(arxiv|lanl)\.(org|gov)/(abs|eprint|pdf)/.Headbomb {talk / contribs / physics / books} 18:20, 19 March 2011 (UTC)
Likewise a bare <ref>http://arxiv.org/abs/0710.4523</ref> or <ref>[http://arxiv.org/abs/0710.4523]</ref> should be converted to <ref>{{cite arxiv|eprint=0710.4523}}</ref> Headbomb {talk / contribs / physics / books} 10:46, 23 March 2011 (UTC)
Done A {{cite web}} with such a link should remain a {{cite web}}
Bare references (<ref>http://www.amazon.com/dp/0123456789</ref> or <ref>[http://www.amazon.com/dp/0123456789]</ref>) should be converted to <ref>{{ASIN|0123456789}}</ref> (keeping in mind the country logic for non-.com URLs) Headbomb {talk / contribs / physics / books} 12:12, 20 March 2011 (UTC)
Bare references (<ref>http://articles.adsabs.harvard.edu/full/1998MNRAS.301..787L</ref> or <ref>[http://articles.adsabs.harvard.edu/full/1998MNRAS.301..787L]</ref>) should be converted to <ref>{{cite journal|bibcode=1998MNRAS.301..787L}}</ref>
Bare references (<ref>http://www.jstor.org/stable/1424736</ref> or <ref>[http://www.jstor.org/stable/1424736]</ref>) should be converted to <ref>{{cite journal|jstor=1424736}}</ref>
Headbomb {talk / contribs / physics / books} 12:38, 20 March 2011 (UTC)
Not sure whether converting bare URLs to references would be covered by existing bot request? If not, please feel free to make a request for this function. Martin(Smith609 – Talk) 23:45, 27 March 2011 (UTC)
{{resolved}}
mojibake on dashes/accented characters from citation bot button
User input needed. This bug has been marked as needing user verification.
Please check that the suggested fix works.
If it does, set |status={{resolved}}.
If not (and you're using the latest version of the bot), please indicate where the problem is still occurring.
Thanks!
Please let me know precisely how to replicate this bug. Martin(Smith609 – Talk) 03:28, 17 March 2011 (UTC)
Also see for example [1] (towards the end) and this diff --- seems to mangle first non-Latin character in "journal" field, but not other fields like "title". cab (call) 08:20, 29 March 2011 (UTC)
Never mind, this seems not entirely related, I'll file a separate report below. cab (call) 12:40, 31 March 2011 (UTC)
Here's an example: {{ref doi|10.1007/BF00394819}} Martin(Smith609 – Talk) 19:02, 22 April 2011 (UTC)
This example is of a bug in the crossRef database. As far as I know, everything within my remit was Fixed in GitHub Pull 346. Martin(Smith609 – Talk) 04:35, 19 June 2011 (UTC)
Not sure that it's possible to convert from Title Case to Sentence Case, or desirable to convert in the opposite direction. Thoughts? Martin(Smith609 – Talk) 23:44, 27 March 2011 (UTC)
Bot does not upload at the correct location... yet again
User input needed. This bug has been marked as needing user verification.
Please check that the suggested fix works.
If it does, set |status={{resolved}}.
If not (and you're using the latest version of the bot), please indicate where the problem is still occurring.
Thanks!
Are any of these edits associated with a version of the bot greater than 265? Martin(Smith609 – Talk) 04:32, 18 March 2011 (UTC)
I haven't found version in the edit summaries, but these are recent creations and the lists seems to keep growing. For example [2] was made on March 14. Headbomb {talk / contribs / physics / books} 18:12, 19 March 2011 (UTC)
Handling improved in r343. Let me know if any more appear. Martin(Smith609 – Talk) 02:16, 12 April 2011 (UTC)
This should be useful for the bot. I chose the "endnote", since it returns the full name of the journal, but there are other formats. Headbomb {talk / contribs / physics / books} 11:21, 15 March 2011 (UTC)
Sounds great! Obviously useful for bibcodes; presumably its DOI database is not as extensive as CrossRef's. What other uses do you forsee it being put to? Martin(Smith609 – Talk) 05:00, 17 March 2011 (UTC)
CrossRef probably has a bigger doi database true. However, many (older) publications do not have DOIs, but they do have bibcodes. The ADSABS database will also returns many identifiers (arxiv, bibcode, and doi are returned when available), and you can query it via any of them, meaning that you can further cross check a {{cite arxiv}} to see if it should be updated to a {{cite journal}}. The journal "issue" also tend to be better documented on the ADSABS database, at least in my experience. Headbomb {talk / contribs / physics / books} 07:15, 17 March 2011 (UTC)
Working on it: bibcodes are supported in r290. Do you have an example of an instance with an issue number? Martin(Smith609 – Talk) 04:07, 18 March 2011 (UTC)
Because arxiv use :s in their XML, the bot couldn't extract the journal or DOI. I've worked around this in r289. Martin(Smith609 – Talk) 03:49, 18 March 2011 (UTC)
Were you using r289 or greater? You'll need to replace "citation-bot" with "DOI_bot" in the URL to use the latest version of the bot. The current stable version is r273. Martin(Smith609 – Talk) 18:16, 19 March 2011 (UTC)
Done: The CrossRef database is now consulted to recover the journal name. Martin(Smith609 – Talk) 16:19, 27 March 2011 (UTC)
Likewise here [5] it converts, but it's a bit half-assed. I notice that there's no DOI in the arxiv, so it can't use that and there's not much it could do. But the main problem when converting to a cite journal is that the bot keeps the year from the cite arxiv which it shouldn't do. It's very common that publication year differ between a preprint and the official publication, so whenever it upconverts, the year/date should be discarded. Headbomb {talk / contribs / physics / books} 20:11, 19 March 2011 (UTC)
This point improved in r313. Martin(Smith609 – Talk) 16:05, 27 March 2011 (UTC)
Awesome. I didn't test it yet, but would it be possible to assign priorities to databases? Like CrossRef > ADSABS> arXiv.org? As in, if you find a doi in the arXiv.org database, query CrossRef with the DOI first, if that fails, query ADSABS with the DOI, and only then use the other information from Arxiv.org? I'm mentionning this because data from ArXiv.org tends to be the crappiest, and CrossRef seems to give the best. Headbomb {talk / contribs / physics / books} 13:04, 28 March 2011 (UTC)
That'd be quite a bit of work; can you show me a few examples where this would be beneficial, so that I can get my head in? Martin(Smith609 – Talk) 04:29, 19 June 2011 (UTC)
Note that http://www.jstor.org/pss/197008, which the new link resolves to if you take the bracket away, goes to a completely different article in a completely different journal.
Bot now (r306) won't interpret the SICI as a JSTOR ID. It should be able to determine the JSTOR ID from the sici, though. I'll implement that anon. Martin(Smith609 – Talk) 18:00, 22 March 2011 (UTC)
When {{cite journal}}/{{citation}}/any {{cite xxx}} (except {{cite web}}) do not have URLs (post cleanup), the accessdate should be removed. It's not displayed, and is pretty useless.
Not sure what you're getting at... If there's a {{citation}} without a url, it's pretty safe to assume that people aren't citing a web page. No? Headbomb {talk / contribs / physics / books} 18:17, 22 March 2011 (UTC)
Could there ever be an archiveurl but no url, for example? Just making sure that we've thought of everything before I code. Martin(Smith609 – Talk) 02:12, 23 March 2011 (UTC)
The only scenario I can think of where that would happen is when someone (bot, script, human) archived a url such as |url=http://articles.adsabs.harvard.edu/full/1998MNRAS.301..787L (which I've seen a few times), and then someone (bot, script, human) later cleaned up the url into |bibcode=1998MNRAS.301..787L.
Not sure what should be done in this situation. When there's an archive url and no url, the template gives an error, so it would be cleaned up pretty fast anyway. Maybe citation bot can check if the archived url is the same as the url it just cleaned up, and remove everything (url, accessdate, archiveurl, and archivedate). Or maybe it's safe to assume that if the url resolves to an identifier, so does the archive url and again everything can be removed (url, accessdate, archiveurl, archidate). Headbomb {talk / contribs / physics / books} 10:45, 23 March 2011 (UTC)
I always wondered what the point was of retaining dead urls when archiveurls were in place. There should be a better way. LeadSongDogcome howl! 14:10, 23 March 2011 (UTC)
Is there any guarantee that a citation with a URL and an archiveurl will also have fields such as title, author, publisher, journal, etc. filled in? If not, the dead URL should be retained in case the archive goes away, because the publisher can often be identified from the URL. Jc3s5h (talk) 15:53, 23 March 2011 (UTC)
It's not like the article history is being deleted, the dead url is still there anyhow. Also, I'm not aware of us having a history of problems with archives that vanish. After all, that's what they're for. If they don't make provisions for ongoing availability they won't get the fonds in the first place. Still, if there's a real, valid concern for this, we can simply keep the dead url in hidden text.LeadSongDogcome howl! 18:13, 23 March 2011 (UTC)
The problem with doing this sort of change, according to wikipedia bot policies currently in place and according to several very vocal folks, is that its not allowed because it doesn't render any changes to the page. So since it doesn't render any changes to the page (just like nested, attention=no, invalid parameters on templates, and a wide array of other things) it cannot be deleted via a bot. If someone wants to try and change this policy I would be glad to support it but till then we can't do this type of change unless something else is being changed on the article at the same time. --Kumioko (talk) 18:40, 23 March 2011 (UTC)
We've gone a bit off topic; the citation bot should NOT remove URLs just because they are dead, nobody should be doing this (need to look for correct URL if moved / mirror / archive link but not just drop URL). Removal of |accessdate= when a URL is converted to an identifier is fine and a sensible housekeeping step. Rjwilmsi 14:08, 24 March 2011 (UTC)
Yeah it should definitely not remove dead urls, although this never was about that. Just cleaning up citations with no urls but with accessdates (for whatever reason). Headbomb {talk / contribs / physics / books} 20:54, 25 March 2011 (UTC)
digression
I agree that its sensible, the problem is according several users inluding CBM and Xeno in multiple venues (including AWB recently causing the revising of one of the AWB rules of use) if the change does not render a change to the page or a category it cannot be done alone and must only be done if another more major change is being made at the same time. I don't agree with it and have repeatedly fought against it but thats the way it is so if somone wants to do this change they either have to limit it to only articles where another more significant change is being made at the same time or change the rule. --Kumioko (talk) 14:58, 24 March 2011 (UTC)
Who here suggested that this would be done as the only change in an article? Making these edits are perfectly within the bot policy of uncontroversial maintenance. Headbomb {talk / contribs / physics / books} 21:29, 24 March 2011 (UTC)
Its been said many times over the last few months by several users most notably CBM, Xeno and a couple of other that changes that do not render changes to the article or that do not affect a category cannot be done alone and that is one of the prime reasons why they modified the rules of use recently for AWB. You can see these discussions on the talk pages for the Wikiproject Council, AWB, the ANI that CBM opened up against me when he revoked my AWB access and a variety of other places. Even my talk page archives, Rich Farmourough and a others. All because they said that the "consensus" was that these changes are disruptive, fill wathclists, waste resources, etc. I fought hard against it and lost. I even lost my access to AWB because I told CBM that 1 editor not liking a change didn't constitute a lack of consensus. Rather than open a discussion he revoked my AWB and opened an ANI against me. Ok fine its water under the bridge and Wikipedia is poorer for it (obviously I'm still a little pissed about that). In the end I can do nothing to stop this, nor do I want to and aside from the fact that it apparently breaches a policy I never agreed with is perfectly fine. But I wanted to present the warning so that if you do go ahead with it you were aware that someone may stop the bot in protest or block it completely from editing. Because a stupid policy exists that prevents us from removing garbage fields and deprecated parameters from articles because they don't render changes. --Kumioko (talk) 20:01, 25 March 2011 (UTC)
The problem with your edits was that you were doing edits of very little value, which was effectively spamming watchlists with bot-like edits for no real reason (and unlike other low-value edits with bot flags, you cannot filter them out of your watchlist). Removing something like "|nested=yes" did not prevent damage, nor did it cleanup anything other than the edit window. This was raised many times, and you kept doing it against consensus. Removing |accessdate=2005-06-28 from an article without a url however, it not such an edit. If there's no url, but there's an accessdate, when another editor comes along and adds a URL without noticing or updating the existing accessdate field, you have a false accessdate that is displayed. That is why accessdate should be cleaned up, unlike the |nested=yes which got your AWB access revoked. Were it up to me, I'd have sent a bot (and not a human with AWB, per the watchlist thing) to nuke the |nested=yes from talk pages templates, however regardless of my personal position on the issue, the community doesn't like these low-value edits when done on their own.
My point in all that statement really boils down to them being the same type of edit. Regardless of whether a future change will cause it to render removing it at this time does not render any changes and therefore according to policy as has been quoted to me many times it can't be done. For what its worth though nested was not why my access was revoked. It was for doing things like this and it was this edit with no nested parameter at all where I removed about a thousand characters of crap that could cause future rendering problems just as you point out above, slow page load times and confuse editors not to mention making it a whole lot harder to deal with for the bots having to mine through it to make whatever change their there for, potentially causing them to break something. So I go back to my original comment of, if someone wants to do this type of change then the unnecessarily restrictive rule that is being used to brow beat editors trying to make the 'pedia better needs to be changed. Otherwise you are more the welcome to make the changes to articles were other more significant changes are being made. With all that said I am not an admin, I have absolutely no other power except for that of the pen, or keyboard as it were so if you go ahead with it I am not going to stop you but its setting a bad example when we use a policy as a mallet on one editor and then push it aside for another because we don't want to follow it. --Kumioko (talk) 01:41, 26 March 2011 (UTC)
Its not a digression when a concern is raised about policy. Just because you don't like the policy doesn't mean you don't have to follow it. But I won't post here again because its obvious that knowone cares. So it seems policy is only policy when we want it to be and perhaps this is a WP:IAR scenario. --Kumioko (talk) 04:12, 26 March 2011 (UTC)
Is there consensus on a solution that can be implemented, or shall I close this thread? Martin(Smith609 – Talk) 02:26, 12 April 2011 (UTC)
Also [9] for Nature DOI & Bibcode. Rjwilmsi 09:39, 26 March 2011 (UTC)
Investigated further: broken between 289 and 294. All edits with 294 or later need to be checked, I've already had to revert over 50. Rjwilmsi 11:58, 26 March 2011 (UTC)
I've requested a temporary bot block on WP:ANI. Rjwilmsi 12:09, 26 March 2011 (UTC)
The bot searched the AdsAbs database using the specified title. However AdsAbs often returned very fuzzy matches. As of r307 the bot now checks that the returned title is a good match for the input title. Sorry about the inconvenience caused in the interim; I'll be doing more thorough testing after each edit from the nonce. Martin(Smith609 – Talk) 13:18, 27 March 2011 (UTC)
I'm not convinced that a match on title alone will be sufficiently reliable. Are you cross validating other criteria such as volume, year, journal and/or pages etc.? Rjwilmsi 20:46, 27 March 2011 (UTC)
Bot adding URLs of articles reviewing books to the books themselves
Oh, I've just noticed that this issue has already been reported earlier. But it doesn't appear to have been resolved. — Cheers, JackLee–talk– 08:20, 31 March 2011 (UTC)
That edit was tagged "([Pu316]Misc citation tidying.)" by the bot. Does it still make that error after the recent code revisions?LeadSongDogcome howl! 15:41, 31 March 2011 (UTC)
I haven't encountered any other problems with the articles on my watchlist so far. — Cheers, JackLee–talk– 08:40, 7 April 2011 (UTC)
Reviews share much semantic information with books. Any ideas on how to avoid false positives? Martin(Smith609 – Talk) 02:27, 12 April 2011 (UTC)
I could be flip and say "Don't use {{citation}} where {{cite book}} is intended", but I won't. If a citation has an ISBN the bot should not add data pertaining to a publication in a serial, with a possible exception where that citation appears in an article that is in Category:Books. If the input data provides ambiguos key data (such as title and author only) but searches find both a target with an ISBN and a target with a serial title or ISSN then it would not hurt to explicitly flag the bot-added wikitext for human consideration. Another option is to show {{cite book}} (reviewed at {{cite journal}}), though there could be objections to that on wp:SAYWHEREYOUGOTIT grounds.LeadSongDogcome howl! 17:12, 12 April 2011 (UTC)
In an article about the book, or about the author of the book, it's useful to include reviews, but as separate sources, not as additions to the book's publication data. As for {{citation}} vs {{cite book}}, they have different formats: {{cite book}} goes with {{cite journal}} and is incompatible with {{citation}}. In any case, some books really do have dois or urls for the whole book that would be helpful for citation bot to add, so it is not as simple as "don't add urls to books". In this case, though, the presence of a "review author" in the ADS data is a giveaway that this is a book review, and the fact that the review author is not the same as the author of the {{citation}} is an indication that the original citation refers to the book itself and not to its review. —David Eppstein (talk) 17:51, 12 April 2011 (UTC)
Format consistency (a MOS issue) should never be allowed to trump accuracy of attribution (a V issue). That way lies madness. If converting to all-{{citation}} or all-{{cite xxx}} breaks the correctness of the attribution then the conversions should stop. Readers can tolerate a misplaced comma or period far better than they can a citation of the wrong source. When the source metadata is fully identified the conversion might then be conducted more safely. LeadSongDogcome howl! 19:08, 12 April 2011 (UTC)
There's nothing inaccurate or unverifiable about books formatted using {{citation}}. —David Eppstein (talk) 02:45, 13 April 2011 (UTC)
The example given for this bug was
{{citation|last=Shirley|first=John W[illiam]|title=Thomas Harriot: A Biography|location=Oxford|publisher=[[Oxford University Press|Clarendon Press]]|year=1983|isbn=978-0-19-822901-8}}
which the bot changed to
{{citation|last=Shirley|first=John W[illiam]|title=Thomas Harriot: A Biography|location=Oxford|journal=Journ. History of Astronomy V.17|volume=17|pages=71|publisher=[[Oxford University Press|Clarendon Press]]|year=1983|isbn=978-0-19-822901-8|bibcode=1986JHA....17...71D}}
If instead the entry had been
{{cite book |last=Shirley|first=John W[illiam]|title=Thomas Harriot: A Biography|location=Oxford|publisher=[[Oxford University Press|Clarendon Press]]|year=1983|isbn=978-0-19-822901-8}}
the ambiguity would not have been there in the first place.LeadSongDogcome howl! 04:22, 13 April 2011 (UTC)
It might be worth mentioning that {{citation}} and {{cite book}} take almost the same parameters as each other, and display similarly-named parameters in almost the same manner (punctuation being the main difference). As regards specifics, {{cite book}} allows both |trans_title= and |trans_chapter= whereas {{citation}} doesn't, conversely, citation has a parameter for the journal name (|journal= and its four synonyms) which cite book doesn't. It is the presence of any one of |journal=, |periodical=, |newspaper=, |magazine= or |work= which causes {{citation}} to use the "journal" format; if all five are absent, it uses the "book" format. A smaller matter is that if you want harvard referencing to link, and you're not intending to use a custom link (ie |ref={{harvid|...}}), you must provide an explicit |ref=harv to cite book, whereas that is the default action for citation. Apart from those threetwo areas, {{citation}} and {{cite book}} primarily differ in some very minor ways such as the separator between the various items of information. --Redrose64 (talk) 11:30, 13 April 2011 (UTC) amended Redrose64 (talk) 13:15, 13 April 2011 (UTC)
That's correct as far as it goes, but it misses a key factor. A non-entry in {{citation}} of |journal= or its synonyms is not the same as an active entry saying "this is a book I'm citing". Use of {{cite book}} provides that additional bit of information. The problem can also be seen as the ambiguity of |title= which may be used to mean an entire book or a single letter to the editor of a newspaper in the same {{citation}}.LeadSongDogcome howl! 20:22, 13 April 2011 (UTC)
This ambiguity is a good thing, and one of the big reasons why I prefer {{citation}} to {{cite book}} etc. The reason is that when we use {{citation}} we don't need to decide whether we should count a technical report or a senior thesis or a doctoral dissertation or a booklet or a set of course lecture notes or a 100-page online preprint or a single-article issue of a journal is really a "book", and we don't need to set up and maintain separate {{cite report}} and {{cite thesis}} and {{cite booklet}} and {{cite course}} and {{cite preprint}} templates; we just tell citation what types of information are available for this citation and it does the formatting. For similar reasons, when we're using {{citation}}, we don't need to decide whether a periodical that we're citing something in is a journal or a magazine or a newsletter or a newspaper or a catalog, and we don't need separate {{cite journal}} and {{cite newspaper}} and {{cite magazine}} and {{cite newsletter}} and {{cite catalog}} templates. Especially in the context of Citation bot, a piece of software that is not necessarily good at making these sorts of subtle distinctions between different types of publication, it's helpful to avoid having to make these distinctions at all. —David Eppstein (talk) 20:59, 13 April 2011 (UTC)
I can't agree that it is a good thing to leave to the bot a determination that humans have difficulty making. Why should we then turn around and complain that the bot gets it wrong? LeadSongDogcome howl! 21:58, 13 April 2011 (UTC)
The error in question had little or nothing to do with the fact that the citation in question was to a book. It was an error of not understanding that the ADS "review author" field implies that the ADS entry is to a review of a source and is not an entry for the source itself. The same error could equally easily have been made for a published review of a journal article. So the fact that the broken citation was a {{citation}} template rather than a {{cite book}} template is a complete red herring. As for why {{citation}} is a good thing, in relation to what citation bot understands or doesn't understand: it's because using {{citation}} allows its users (including citation bot) to omit irrelevant information (like whether it really is a book vs a booklet or whatever), freeing them from having to understand that irrelevant information. —David Eppstein (talk) 23:08, 13 April 2011 (UTC)
Are you perhaps confusing the needs of readers with those of editors? Many of both groups consider peer review, journal rankings, and even finding physical access to the dead trees to be relevant factors when assessing how much credence to give the source. In any case, the "review author" field is not universally supported by all serial indices, so while using that might be a solution for ADS, it is not a general fix for all the databases. If available we would do better to implement a "publicationtype" parameter (as on PubMed) to explicitly distinguish reviews, systematic reviews, original articles, letters, consensus reports, etc. This could also enormously facilitate article reviews by highlighting the use of primary sources. LeadSongDogcome howl! 00:50, 14 April 2011 (UTC)
Does this point need further action from me? Martin(Smith609 – Talk) 04:15, 19 June 2011 (UTC)
[10] "GéRaldine". The CrossRef API gives the correct capitalization ("Géraldine").
We can't proceed until
Bot operator's feedback on what is feasible
That edit was many versions ago, see the edit comment: "11:56, 27 March 2011 Citation bot 2 (talk | contribs) m (618 bytes) ([cw310]Misc citation tidying.)" It looks to have been done during the creation of the cite doi subpage, though the edit comment doesn't reflect that. Pubmed also has the correct spelling for the forename. LeadSongDogcome howl! 20:59, 13 April 2011 (UTC)
Another example (this one using version 305). In this case, CrossRef has no data. Ucucha 23:25, 18 April 2011 (UTC)
Is it this fix that now causes existing condensed page ranges to be expanded? This is I believe against the wikiproject medicine preference to display the page ranges in the condensed format. Rjwilmsi 10:30, 20 June 2011 (UTC)
Curly quotes
Get rid of curly quotes (WP:MOS#Quotation marks). You can put $title = preg_replace(array('/[`‘’]/u', '/[“”]u/'), array("'", '"'), $title); or something somewhere in your code. Ucucha 11:42, 16 June 2011 (UTC)
That doesn't seem to work, and neither does $str = preg_replace('~̵[679];|[\x{2039}\x{203A}\x{2018}-\x{201B}]|&[rl]s?[ab]?quo;~u', "'", $str);. I'm stumped. Suggestions welcome! Martin(Smith609 – Talk) 23:43, 18 June 2011 (UTC)
Where exactly is the relevant code? Perhaps it has to do with different character encodings somewhere. Does preg_last_error() return anything? Ucucha 23:48, 18 June 2011 (UTC)
The regex in r358 fails to compile: there's a closing parenthesis missing in '~&([rlb][ad]?quo;~' (line 2010 of DOItools.php). Not sure exactly what you want there, but you should either add or remove some parentheses there. Ucucha 00:03, 19 June 2011 (UTC)
Ugh. Well spotted. Thank you! This brought a couple of other bugs to the surface too, so performance and edit summaries are both now improved. Fixed in GitHub Pull 362
I get an empty page at http://en.wikipedia.org/w/index.php?title=Seppuku&action=submit with the "Warning: An administrator deleted this page since you started editing it. Please check the deletion log to see the reasoning. " reasoning. Show changes indicates that the bot would save an empty page.
Replication instructions
Try running it on the same page?
We can't proceed until
User
-->
Requested action from maintainer
fix the bug? Medium priority, I guess - the bot simply does not work on some pages
I can't reproduce this. Does it work now? Martin(Smith609 – Talk) 22:11, 11 June 2011 (UTC)
Adding additional authors messes up citations using "et al".
If one of the authors in the citation has "et al" after his name (which is standard practice, to indicate many additional authors), citation bot will add all of the secondary authors, but keep "et al" in the original name.
I suggest that Citation bot does not add additional authors when there is the string "et al" in an existing author field. (Note: I do not believe that it is within Citation bot's remit to change the citation format with multiple authors from an "et al" convention to a full list of the authors' names)
Fix severe bug. Something has gone seriously wrong here.
Those diffs seem to be the result of near-simultaneous edits by the bot and a human (note the timestamps), which MediaWiki doesn't always handle well. I don't think much can be done about this problem. Ucucha 03:27, 10 May 2011 (UTC)
Yet every time the bot uses that edit summary it makes the same mistake [14][15], so it's more than just a coincidental edit conflict. I have yet to find an example of that particular edit summary that didn't wind up making a content change.... Sailsbystars (talk) 03:39, 10 May 2011 (UTC)
The reason for that is that when the bot makes that edit, it normally does not change the article text, and therefore the edit is not recorded. So the only edits that are recorded are those that go wrong. Ucucha 03:48, 10 May 2011 (UTC)
Thanks for the explanation, that makes sense. Sailsbystars (talk) 03:59, 10 May 2011 (UTC)
I've updated the edit summary to highlight the likelihood that it needs reverting. Fixed in GitHub Pull 375Martin(Smith609 – Talk) 21:54, 11 June 2011 (UTC)
Speaking of that, this edit does look like a preventable error. Presumably the page size exceeds the bot's buffer. Ucucha 13:46, 10 May 2011 (UTC)
Expand the memory allotted to the bot? I'm not sure. It did this on two of those list pages; one was 143,889 bytes and got truncated to 1,078; the other went from 79,654 to 64,791. I can't find any other instances where the bot truncated an article, although it did this same thing on larger articles (e.g., Hockey stick controversy, 104,001 bytes). One possibility I can think of is that the number of citation templates in these pages somehow caused the problem. User:Ucucha/List of mammals/Primates currently has an incomplete cite doi; perhaps you can run the bot on that page to check whether it is still doing this. Ucucha 14:25, 14 June 2011 (UTC)
Perhaps the phase of the moon was better this time. I guess that means you fixed the bug in one way or another. Ucucha 04:28, 19 June 2011 (UTC)
Could you point me to the approval for this task? I certainly do not see any kind of approval for the bot to make edits that "SHOULD PROBABLY BE REVERTED". It seems the bot is simply making null edits, and causing problems when running into edit conflicts? Please disable this task unless you obtain approval for it. –xenotalk 15:58, 7 July 2011 (UTC)
It appears to be a very rare use-case, where the bot detects an edit conflict and so cannot determine if in fact it is making a non-null edit. I'm not sure why the bot doesn't handle edit conflicts by simply re-starting with the newer version of the page (after a delay). That would seem to be more robust. LeadSongDogcome howl! 16:28, 7 July 2011 (UTC)
The bot should probably not be making null edits in the first place - and certainly not without approval to do so. The job queue is there for a reason. –xenotalk 16:33, 7 July 2011 (UTC)
No comment on the appropriateness, but if a null-edit ever must be performed, do so with an empty appendtext parameter instead of the page content in text. That way, you have no problems with edit conflicts. Amalthea 16:54, 7 July 2011 (UTC)
... and I note that it's not MediaWiki at fault here, it seems that edit conflict detection is disabled in the bot. Amalthea 16:57, 7 July 2011 (UTC)
Thanks for pointing this out Amalthea; I've enabled this, which should resolve the issue. Martin(Smith609 – Talk) 22:20, 16 July 2011 (UTC)
No, the bot operator—sorry for the confusion. Ucucha 08:14, 8 June 2011 (UTC)
Done; I needed an mb_convert_encoding. Martin(Smith609 – Talk) 21:39, 11 June 2011 (UTC)
This is not fixed; see this recent edit.
nb - seems to work fine using IE on a windows box, but not using Firefox 3.6.18 on Mac OS 10.6.7 pablo 13:46, 16 June 2011 (UTC)
Thanks for the pointer; I hope that it works now! Martin(Smith609 – Talk) 02:48, 19 June 2011 (UTC)
Seems fine now, thanks! pablo 13:54, 19 June 2011 (UTC)
Cochrane Database Syst Rev converts to title, chapter?
I'm trying to capitalize on the "series" information in CrossRef, but this doesn't always mean the same thing. I've implemented your suggestion of not using the data when |journal= is set. Martin(Smith609 – Talk) 01:04, 19 June 2011 (UTC)
See this edit, and look at the footnotes before and after the edit. The bot removed the body of several named footnotes that it believed were defined elsewhere. They were not, apparently, because errors now appear. Still researching what happened.
The bot combines refs, but renames the ref name to have extra quotemarks -- example: <ref name= "Playfair90"> to be <ref name=" "Playfair90""/>. And not always, see Playfair84 at the diff link. It appears to happen more frequently when there is a space after the name= text. Ref code then reads the " " as the ref name, kicking out an error. AnomieBOT then sweeps up behind and removes the original ref name, leaving <ref name=" "/>, compounding the issue.
What should happen
Can the bot be made to not add the extra quotes? such as, checking for existing quotemarks on the name first?
I didn't realise that placing spaces between "name = 'Playfair'" was legit. The bot now allows for this behaviour. Martin(Smith609 – Talk) 23:54, 18 June 2011 (UTC)
Clean-up diff. The bot added a stray period after some authors that had only one first name. Also, per Template:Cite doi#Formatting, there should be no space between the initials.
We can't proceed until
Bot operator's feedback on what is feasible
Thanks for fixing. Ucucha 00:03, 19 June 2011 (UTC)
{{resolved}}
The bot says DOI 10.1111/j.1096-0031.2009.00255.x at Mollusca is "broken", but it works perfectly. Please fix the bot and rmv the linked "edit" note. --Philcha (talk) 21:52, 19 June 2011 (UTC)
Might have been due to the line break between the DOI and the }} (changed here). If so, that's a bug in the bot. Ucucha 22:05, 19 June 2011 (UTC)
I cannot replicate this. Perhaps it was a temporary glitch with Crossref? Unless I hear otherwise I'll mark this as resolved. Martin(Smith609 – Talk) 21:50, 16 July 2011 (UTC)
The bot made a bit of a mess of my citations on 2 April. It erased about 50 citations and credited them all to one source. Could you revert what it did, and I'll redo the edits I made today? Thanks, Yoninah (talk) 16:15, 20 June 2011 (UTC)
OK, I figured out how to undo what the bot did. Yoninah (talk) 16:20, 20 June 2011 (UTC)
Incomplete removal of second DOI from id breaks formatting of citation
Deleterious: Human-input data is deleted or articles are otherwise significantly affected. Many bot edits require undoing.
What happens
The citation by Lopez and Law in path decomposition has two DOIs (yes really), one of which was formatted as a DOI: wikilink in the id section. Citation bot removed the second DOI but retained the two brackets at the start of the wikilink, causing the entire citation template's formatting to break. (Note, this was back in May, but I only just noticed it recently. So it's possible it's a dup of an already-fixed bug.)
Agreed: in such a case I would add another reference to cover the alternate source. --Redrose64 (talk) 18:59, 14 July 2011 (UTC)
Regardless of whether you like or dislike the existing formatting, Citation bot should never break the syntax of a citation template to the point that it shows up in the article as garbled plain text. Re-opening because two different problems (the nonstandard id and the garbled template) do not cancel out to make zero problems. —David Eppstein (talk) 21:33, 16 July 2011 (UTC)
What solution do you propose? Martin(Smith609 – Talk) 21:51, 16 July 2011 (UTC)
That the bot check that the [ ] and { } in the parameters of the modified citation are all properly balanced, and that it abort making a change to that citation if they are not. —David Eppstein (talk) 01:36, 17 July 2011 (UTC)
What about citations that are meant to have unbalanced braces? Martin(Smith609 – Talk) 02:03, 17 July 2011 (UTC)
You mean like ones in <nowiki> tags? If you know how to parse them and ignore them, fine. If not, and the change causes a very small number of valid citation tags with intentional unbalanced braces to be uneditable by the bot and require manual editing instead, is that such a great price to pay for not breaking things? —David Eppstein (talk) 02:10, 17 July 2011 (UTC)
Sounds like it's a choice between breaking some things and breaking others. In such cases, I tend to assume that the format that doesn't match the specifications is the one to try to amend. Martin(Smith609 – Talk) 00:54, 20 August 2011 (UTC)
A specific edit to the bot's code is requested below.
Requested action from maintainer
The bot needs to respect existing format styles. The above examples are extremely common on medical articles that use Diberri's tool for generating cite journal refs from PMIDs and the abbreviated forms typical in medical journals. WP:CITEVAR does not allow such style changes without gaining consensus per-article. An article that is altered by this bot will be much harder to maintain by editors using Diberri's tools because, in order to be consistent, they would have to manually fix the page ranges and to tediously enter all the author parameters.
I have not tried to read the source code, but from this page and the actions of the bot, I infer that the logic for handling authors is to count how many authors were specified by the editor (if any), count how many authors are specified in the database, and examine the editor-specified authors for a few special cases, such as a trailing "et al." It is presumed that if the editor specified fewer authors than the database, and didn't specify "et al.", he must have forgot and more authors should be added, and further, that the first editor-specified author matches the first database-specified author, etc.
I regard this behavior as dangerous. The bot is drawing conclusions about parameters it can't actually interpret (the editor-specified authors). If the number of authors specified by the editor does not match the number specified by the database, and the last author does not contain "et al.", the bot could add a category and comment indicating there is a citation that needs manual attention, but should not try to "fix" it. The only entries that should be fixed are those that have no creators listed (no authors, editors, translators, or any similar designation). Jc3s5h (talk) 11:34, 17 July 2011 (UTC)
Will do when time permits. I'm soon to be away for a while. Martin(Smith609 – Talk) 20:40, 8 August 2011 (UTC)
Bug?
I have no knowledge of bots, but this edit just made things worse. Cheers, Huldra (talk) 03:09, 26 July 2011 (UTC)
That is a definite bug. In all SGML-type markup languages (which includes XHTML), the <!-- --> markers enclose comments. It is incorrect to replace the pair of keyboard hyphen-minus characters with any other horizontal line, and especially not with a single em dash. --Redrose64 (talk) 13:20, 26 July 2011 (UTC)
I think that this has been fixed; can it be reproduced using a recent version of the bot? Martin(Smith609 – Talk) 21:20, 26 July 2011 (UTC)
Hi there. Someone pointed me to this edit, which I am assuming is an edit-conflict error? Anyhow, just thought you should be aware. Cheers, 28bytes (talk) 20:34, 27 July 2011 (UTC)
Almost certainly an EC. Three edits in three minutes.LeadSongDogcome howl! 22:57, 27 July 2011 (UTC)
That citation is for PMID16062575, which is a response to McCrae's PMID16058406. The earlier McCrea paper had been published on pages 21-4. Somehow that page range got picked up by the bot for the de Vaucouleurs & Wertz letter commenting on it. LeadSongDogcome howl! 20:28, 16 August 2011 (UTC)
Now I see, the cite had the wrong PMID inserted way back in this edit from 2008. Then this edit by Citation bot in 2009 trimmed out the correct PMID, leaving the wrong one in place. LeadSongDogcome howl! 20:42, 16 August 2011 (UTC)
There was a bug in a regexp that italicized species names; it saw the 'sp' in spine and thought that it was an abbreviation for 'species'. Now Fixed in GitHub Pull 401. Martin(Smith609 – Talk) 19:05, 17 September 2011 (UTC)
A specific edit to the bot's code is requested below.
Requested action from maintainer
distinguish refnames when same webcyclopedia referred
Unless the references are identical, they won't be combined: the bot cannot tell what differences might be significant. If this doesn't answer your point then perhaps you could spell out more clearly the difference between what happens and what should happen. Martin(Smith609 – Talk) 21:17, 25 September 2011 (UTC)
I see now that the ref that the bot named Rynasiewicz already had the same material as an existing ref that had already been misnamed as Rynasiewicz. I have corrected manually--JimWae (talk) 20:14, 27 September 2011 (UTC)
Will look at page ranges. I've tried to ensure that the visual output is unchanged as far as authors go. Adding the extra makes it easier to export the citation to an external program. Let me know if I have missed something with the formatting. Martin(Smith609 – Talk) 03:27, 20 June 2011 (UTC)
Wikiproject medicine contributors may go mad if you switch from the use of |author= even if the displayed result is unchanged and meta data becomes available. Rjwilmsi 10:38, 20 June 2011 (UTC)
Please describe what happens and what should happen, so that I can understand the problem. Martin(Smith609 – Talk) 21:27, 16 July 2011 (UTC)
If you follow the diff given, and then go to the references section, there is a red error message at ref no. 25, which wasn't there before the bot edit. Somehow it's broken how <ref name="Holland1909" /> links. --Redrose64 (talk) 22:06, 16 July 2011 (UTC)
The full template prior to the bot edit was as follows:
{{cite journal
|last=Dossena
|first=G
|authorlink=
|year=1954
|month=January
|title=Quello che la medicina deve ad Agostino Bassi
|language=Italian
|trans_title=Debt of medicine to Agostino Bassi
|journal=Rivista d'ostetricia e ginecologia pratica
|volume=36
|issue=1
|pages=43–53
| publisher = | location = | issn =
| pmid = 13168166
| bibcode = | oclc =| id = | url = | language = | format = | accessdate = | laysummary = | laysource = | laydate = | quote =
}}
The |language= parameter occurs twice, but only the first instance has a value. It is a feature of MediaWiki template expansion that should any named parameter occur more than once, all are ignored except the last one. So, |language=Italian|language= is exactly equivalent to |language= and therefore the rendered appearance before and after the bot edit would be the same. --Redrose64 (talk) 19:17, 27 June 2011 (UTC)
True, though surely the correct fix would be to remove the blank parameter only. Rjwilmsi 20:41, 27 June 2011 (UTC)
The bot is programmed not to affect the visible output of the citation unless necessary. This way articles are not damaged for readers. Editors are responsible for previewing their content before saving. Martin(Smith609 – Talk) 21:16, 16 July 2011 (UTC)
The documentation for {{Cite book}} contains the example "pages=100–110" while the documentation for {{Citation}} contains the example "pages=153–61". There does not appear to be any standard about how to format such page ranges, and it is wrong for the bot to create a standard where none exists. Jc3s5h (talk) 15:56, 11 July 2011 (UTC)
It seems all that edit did was reflect the existing page range as shown on Pubmed. Is there some reason to think something more systematic is at work?LeadSongDogcome howl! 20:51, 11 July 2011 (UTC)
I don't know why it did what it did, but there is no standard expressed in the template documentation as to whether the page range is the page range occupied by the article in the journal, or the page range that supports the claim in the article. Therefore it is improper for the bot owner to unilaterally and without broad consensus decide that it means the page range occupied by the article in the journal. Jc3s5h (talk) 21:11, 11 July 2011 (UTC)
Since the purpose of referencing is to aid verification, it is better to be precise rather than vague when giving a source, so that a reviewer can be directed to the exact position in the source of the claimed fact. We cannot expect such people to read the whole of the source; many academic papers may run into dozens of pages in one journal. Therefore, to my mind, it's the second of these "the page range that supports the claim in the article". --Redrose64 (talk) 22:11, 11 July 2011 (UTC)
Redrose64, I agree that's what the parameters ought to mean, and I encourage you to edit the documentation of the various and sundry templates to reflect that interpretation. Jc3s5h (talk) 22:15, 11 July 2011 (UTC)
I think you are both missing the issue here (quite unrelated to the issue you are discussing), which is that the bot changed a full range of the form "100–110" to a partial range of the form "100–10". Those are both valid styles to refer to exactly the same thing. I prefer the former, but agree that the bot shouldn't be changing one to the other. Ucucha 22:17, 11 July 2011 (UTC)
If you look closely at Template:Citation/doc you will see that the way page ranges are shown is different for journals than from books. In this case, the journal citation used the same page range style as used on Pubmed, which was entirely appropriate. That said, I can think of few details that strike me as less significant. How about getting the substantive fixes done before worrying about this? LeadSongDogcome howl! 14:49, 12 July 2011 (UTC)
The problem is not whether "its trivial or not", the problem is that the bot should not be changing this because it overides human editorial decisions. Headbomb {talk / contribs / physics / books} 14:56, 12 July 2011 (UTC)
Is the bot ignoring exclusion? Or are you expecting it to not edit at all? I note in the contribs that other edits by the bot using the same notation 381 just before and just after the one it made to Kazimierz Fajans inserted page ranges that were fully elaborated, not abbreviated. It's pretty clearly just following the source. LeadSongDogcome howl! 15:28, 12 July 2011 (UTC)
I don't understand what LeadSongDog's comment is supposed to mean (before the edit conflict).
In the case of journal articles, there are two different ways to use the pages parameter. One is in articles with just a <references/> section, in which case the specific pages supporting the claim should be given. The other case is an article that uses shortened citations, in which case the specific pages that support the claim should be given in the shortened citation and the full citation (which might be a cite xxx or citation template) would give the full page range that the journal article occupies.
Since the bot is not designed to detect whether an article uses shortened citations, and is incapable of detecting whether the shortened citations have been fully implemented or are a "work in progress", the bot should not alter page ranges. Jc3s5h (talk) 15:33, 12 July 2011 (UTC)
Whether articles use just a "<references/>" or "shortened citations" is completely irrelevant here. The only thing relevant is that citation bot should not change the citation style used by an article. This applies to things like changing "Smith, J." to " Smith J" just as well as changing "pp. 100–110" to "pp. 100–10". Headbomb {talk / contribs / physics / books} 16:22, 12 July 2011 (UTC)
I was referring to abbreviated page ranges, as Headbomb indicated, and as the original bug report showed. The "shortened citations" bit seems to be a red herring. Again, the bot did not alter the page ranges, it only altered the style of page range representation (in this case from 402–404 to 402–4). Frankly, I'm much more concerned about a biography with a single reference than I am with how many digits are shown in the page range representation of that reference, but that's just me. LeadSongDogcome howl! 16:31, 12 July 2011 (UTC)
LeadSongDog, you stated "It's pretty clearly just following the source." If by that you mean it is using the page range as stated in Pubmed, or some other database, then it's doing the wrong thing, and should leave any page or pages parameter entered by an editor alone. That's both a matter of style, and a matter of substance, because it may not be appropriate to list all the pages that the article occupies. Jc3s5h (talk) 17:37, 12 July 2011 (UTC)
Jack, it is solely a matter of style, not substance, because the substantial value of the parameter was not changed. The edit did add the cited article's title, which had previously been omitted. Both representations denote the same three pages, as would "402, 403, 404" if someone wanted to use that odd form. It was originally the bot which added that page range as "402–4" over two years ago, and it used the form found on Pubmed. Headbomb revised it here. It is not entirely clear if this citationbot edit was derived from the version Headbomb committed just one minute earlier or based on the one Yobot committed an hour before, but it appears to be based on Headbomb's version. (It might be helpful in cases like this if the bot's edit comment explicitly stated the oldid.) Do you really think it would be a reasonable use of resources to have the bot trawl through edit histories for every parameter change in order to see whether a bot or a human editor inserted the parameter value? Or did you have some other way in mind for it to distinguish? LeadSongDogcome howl! 18:53, 12 July 2011 (UTC)
Maybe the particular edit you pointed out was harmless, but it could be a warning of harmful edits that have already, or will in the future, happen. If the logic that the bot is following is to take a page range from a database and overwrite a page (or range) provided to an editor, then sooner or later it will make harmful edits. So the logic should be simple; if the page or pages parameter is present and non-empty the bot should not edit it. Jc3s5h (talk) 19:42, 12 July 2011 (UTC)
If we stop every bot from working on the premise that it might get something wrong, there would not be any point in having them, would there? If we have examples of actual problems, they can be addressed. Of course the bot code is available if someone wants to do a code review. LeadSongDogcome howl! 00:13, 13 July 2011 (UTC)
This is not a matter of "right" and "wrong" this is a matter of WP:BOTPOL. Bots should not annoy editors and override perfectly legitimate editorial decisions. Period. Headbomb {talk / contribs / physics / books} 04:13, 13 July 2011 (UTC)
I see nothing to that effect in the current revision of BOTPOL, perhaps it has changed? It does seem an excellent aspirational goal for all editors, though not always practicable. LeadSongDogcome howl! 05:45, 13 July 2011 (UTC)
carefully adheres to relevant policies and guidelines
That's something of a reach to "not annoy editors". Or am I looking at the wrong set of bullets?
Anyhow, I just quickly reread all the approvals, and the discussion of what to do with the basic citation parameters doesn't seem to have been very explicit. You seems to have spent more attention on the basic questions such as whether bare urls should be replaced. Still, you certainly seem to feel strongly about this, for whatever reason. I'll just drop the stick and move away. LeadSongDogcome howl! 13:39, 13 July 2011 (UTC)
The bot tries to complete incomplete page ranges. It considers any page range without a dash to be incomplete. It turns out that there are more ways of specifying an &endash; than I was aware of, so the bot didn't recognize certain en-dash characters in WP. Hopefully all the possibilities are taken into account as of r392 and no further abbreviations should be made. Sorry that it took me so long to get round to fixing this. Martin(Smith609 – Talk) 00:51, 20 August 2011 (UTC)
Does that mean that if the bot encounters the "pages" parameter with only one digit group and no dash, it looks in a database to try to find the page range. On the other hand, if it finds the "page" parameter, and a digit group, it leaves it alone? Jc3s5h (talk) 01:09, 20 August 2011 (UTC)
Never mind, I performed an experiment and confirmed that if a single page is given with the "page" parameter, and the "pages" parameter is absent, the bot will leave it alone. Jc3s5h (talk) 15:14, 20 August 2011 (UTC)
The fourth author was missing from the ref. Instead the 5th author was given as 4th author. Citation bot added the 5th author but did not check it was identical to the 4th. -- Crowsnest (talk) 07:10, 15 July 2011 (UTC)
The fourth author was specified using |first4=S. |last4=Jarabo. The bot doesn't modify human-entered authors. It correctly added a missing |author5=. Martin(Smith609 – Talk) 21:12, 16 July 2011 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
the bot unfixed a typo in a citation that I had corrected. It replaced "Energy Drinks" with "Energy Dinks". another editor reverted the bot's edit back to mine.
I don't know if this happens frequently or not, but it might be worth looking into why the bot did this and whether it can be prevented from doing it.
Your edit and the bot's edit were only two minutes apart and there is no obvious reason why the bot would change "drinks" to "dinks", since the word was not even in a citation template, so I think this was a race condition of some sort (for example, the bot may have started its edit before you saved), leading the bot to overwrite your edit. There may be a way to fix this, but as far as I know MediaWiki's handling of edit conflicts is fairly bad in general, so that glitches like this happen. Ucucha 11:55, 18 July 2011 (UTC)
There's a pretty clear trend that the bot is not handling such edit conflicts well. I'd suggest that there needs to be some sort of watch over the bot's contribs that checks for likely race conditions (say any other edit within 3 minutes?) and gets a human to check the results.LeadSongDogcome howl! 17:55, 18 July 2011 (UTC)
Sure? The diff was from midnight 17/18 July. Your fix was before that, wasn't it? I'm thinking that there might be a special kind of race condition that crops up at midnight... LeadSongDogcome howl! 22:13, 19 July 2011 (UTC)
The revision number should be listed in the edit summary; you can check which revision was running. I don't tend to roll out 'bleeding-edge' revisions until I'm comfortable that they're not (too) buggy.Martin(Smith609 – Talk) 19:36, 20 July 2011 (UTC)
I undid your edit because it said you fixed the problem I pointed out, but it wasn't fixed. Maybe you can try again. Arkmanda (talk) 04:19, 3 August 2011 (UTC)
This seems to refer to this edit. The generated <ref name=2009factbook> may have been at issue. I note this was another midnight edit. I'm growing convinced there's some kind of timing problem at work.LeadSongDogcome howl! 04:50, 3 August 2011 (UTC)
This is another case of a bug that has been fixed in the current version of the bot, but an older version of the script is being run. I shall update the stable version imminently.
In that Sextans B edit, there is a perfectly duplicated defining entry for a named ref in wp:LDR form. The bot deleted both, replacing them each with a callout of the named ref instead. What it should have done is delete one entirely and leave the other unchanged. Gotta give the little guy props for finding it, though. LeadSongDogcome howl! 21:14, 3 August 2011 (UTC)
{{cite journal |bibcode=1997A&A...322..709K}} is not processed. the ampersand has to be replaced with %26 like {{cite journal |bibcode=1997A%26A...322..709K}} before the bot will do anything with it.
We can't proceed until
Agreement on the best solution
The bot will need to query with A&26%A, but the ampersand (A&A) follows the correct template usage. Headbomb {talk / contribs / physics / books} 20:18, 16 August 2011 (UTC)
I just noticed the same problem. It should be a simple enough change to convert the ampersand into &26% before querying the URL, right? Modest Geniustalk 19:48, 18 August 2011 (UTC)
Actually, I don't think that it does this any more. I've updated the function list to be a little pithier. Thanks for pointing this out. Martin(Smith609 – Talk) 13:57, 20 August 2011 (UTC)
I've tested things pretty extensively, and this regex cleans absolutely everything flawlessly. I highly-recommend you add this to an article "pre-cleaning" logic that would kick in before doing queries and whatnots.
One problem with using a regexp on the entire page is that it affects all templates, which might produce adverse effects. Also the regexp that you suggest doesn't handle URLs that lack the http:// or have a "pss" format. But I agree that a 'pre-expansion formatting sweep' could catch cases like this; I'll look into it. Martin(Smith609 – Talk) 13:47, 20 August 2011 (UTC)
Well then it's a simple matter of having (http://)? and (stable|pss). You probably have ways to restrict the regex to citation templates too.
My point is that there are often other formats that the regexp won't catch. Are you sure that the regexps I'm already using are not adequate? See function get_identifiers_from_url() in expandFns.phpMartin(Smith609 – Talk) 15:23, 20 August 2011 (UTC)
I'm getting a 500 (server error) from google. I don't know what regex you are using, but it could be improved to cover cases like this. Headbomb {talk / contribs / physics / books} 16:21, 20 August 2011 (UTC)
|jstor= is not listed as a parameter for Template:Cite news. I've tweaked the regexp so that this format will be okay in {{cite journal}} (etc.) templates. Martin(Smith609 – Talk) 18:02, 20 August 2011 (UTC)
It's not listed but it works. If we want to be even better, those cases should be converted to a {{{cite journal}}. Headbomb {talk / contribs / physics / books} 03:46, 21 August 2011 (UTC)
Done. Are there other templates besides Cite Web that should be examined in a similar way (i.e. converted to Cite Journal)? I decided not to include Cite Encyclopaedia as I can imagine Encyclopaedia entries being archived in JSTOR. Martin(Smith609 – Talk) 16:12, 21 August 2011 (UTC)
Afraid I'm not going to venture into the Vcite world. One citation format is quite enough to keep me occupied. Martin(Smith609 – Talk) 16:59, 21 August 2011 (UTC)
Alright then I guess that's everything that would need to be converted. Headbomb {talk / contribs / physics / books} 17:20, 21 August 2011 (UTC)
unknown, other than to restore the article to using that approach
We can't proceed until
A specific edit to the bot's code is requested below.
Requested action from maintainer
comments would be useful; maybe you've seen this before.
The Princess Louise, Duchess of Argyll page has a lot of messy references on it, so it is possible that something else is triggering the deletion of the London Gazette template (note; it's not doing them all; 8 out of 11. Of course, I may be missing something in what I was trying to do.
Example that I should have checked is Epidemiol Infect. 1992 August; 109(1): 1–22. PMCID: PMC2272232, which isn't an antique PDF only PMC target and has
citation_pmid as 1499664 in the meta tags. Existence of a PMCID doesn't guarantee the existence of a PMID and Diberri's template filler at least requires finding the PMID to populate the data. Note that this case happens mainly on older articles — 1960s and earlier. RDBrown (talk) 05:33, 22 August 2011 (UTC)
Sample had two LDR entries "ilo" and "ilo2", otherwise identical. Former had two inline callout instances [2a, 2b], latter had one [6]. Bot deleted both LDR definitions and converted [6] to [2c], eg <ref name "ilo" /> LeadSongDogcome howl! 18:57, 22 September 2011 (UTC)
{{resolved}}
Redundant chapter when contribution already present
the bot code probably needs to have an adjustment made.
{{reflist
| group = lower-alpha
| refs =
{{#tag:ref
| Experimental data for BB-1 and BB-2 was lumped together and the rounded average calculated.<ref name="Bryan1901" />
| name = range
}}
}}
{{reflist
| refs =
<ref name="Bryan1901">
{{cite doi|10.1111/j.1559-3584.1901.tb03372.x}}
</ref>
}}
in the above, the bot changed the entire ref-definition using {{cite doi}} to <ref name="Bryan1901" /> leaving the citation undefined and spilling red in the article. This edit took things in another direction—moved cite def out of reflist and invoked it with {{sfn}}—and the bot isn't offering to change the page at all, now (which is fine). —Portuguese Man o' War 07:07, 22 September 2011 (UTC)
It looks like this was an attempt to use wp:REFNEST that went wrong, and the bot just made it worse. LeadSongDogcome howl! 12:56, 22 September 2011 (UTC)
Compare
The sentence being supported.{{#tag:ref
|Experimental data for BB-1 and BB-2 was lumped together and the rounded average calculated.<ref name="Bryan1901b">
{{cite doi |10.1111/j.1559-3584.1901.tb03372.x }}</ref>
|group="foots"
}}
== Bot incorrectly combines multiple references ==
{{bot bug
| title = Bot incorrectly combines multiple references
| status = {{tl|resolved}}
| reported by = [[User:Nick Thorne|<font color = "darkblue">'''Nick Thorne'''</font>]] [[User talk:Nick Thorne|<font color = "darkblue"><sup>''talk''</sup></font>]] 04:47, 3 October 2011 (UTC)
| type of bug = Deleterious<!-- Please select one of:
Improvement: The bot would be much better if ...
Cosmetic: The bot modifies appearance (e.g. capitalization) but not content
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
Deleterious: Human-input data is deleted or articles are otherwise significantly affected. Many bot edits require undoing.
Catastrophical: The bot should stop editing immediately
-->
| what happens = Bot incorrectly combines separate inline references to different page ranges within a source to only one page range. This makes complete nonsense of the references.
<!-- and/or: --> | what should happen = Leave these references alone
| link showing what happens = [http://en.wikipedia.org/w/index.php?title=Timeline_for_aircraft_carrier_service&curid=18976493&diff=453554448&oldid=444016821 here]
| how to replicate the bug =
| waiting for = Consensus<!--
User: Input from editors
Consensus: Agreement on the best solution
Operator: Bot operator's feedback on what is feasible
Maintainer: A specific edit to the bot's code is requested below.
-->
| action required from maintainer = <!--specific details-->
}}
<!-- Discussion starts below this line -->
The problem here is that the same ref name was in use for two different page ranges:
*<code>21 February — {{HMS|Implacable|R86|6}} laid down.<ref name="ches-128-129">Chesneau (1998), pp.128–129</ref></code>
*<code>December — HMS ''Indefatigable'' decommissioned.<ref name="ches-128-129">Chesneau (1998), pp.134–139</ref></code>
--[[User:Redrose64|<span style="color:#d30000; background:#ffeeee; text-decoration:inherit">Red</span>rose64]] ([[User talk:Redrose64|talk]]) 08:51, 3 October 2011 (UTC)
== Error in journal article title ==
{{bot bug
| title = Error in journal article title<!-- Brief description to display in section header -->
| status = {{tl|Resolved}}
| reported by = [[User:Tryptofish|Tryptofish]] ([[User talk:Tryptofish|talk]]) 22:40, 12 October 2011 (UTC)
| type of bug = Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
| what happens = The bot changed the title of a journal article to the title of a ''different'' journal article that was cited on the same page.
| link showing what happens = [https://en.wikipedia.org/w/index.php?title=Opioid_receptor&action=historysubmit&diff=455282882&oldid=455277152]
| how to replicate the bug =
| waiting for = Consensus<!--
User: Input from editors
Consensus: Agreement on the best solution
Operator: Bot operator's feedback on what is feasible
Maintainer: A specific edit to the bot's code is requested below.
-->
| action required from maintainer = <!--specific details-->
}}
<!-- Discussion starts below this line -->
:The cite for Ingoglia had the wrong PMID and the PMC and DOI for the wrong PMID, which I've now fixed for you. The bot picked up the title off those identifiers. [[User talk:Rjwilmsi|<font color="darkgreen">'''''Rjwilmsi'''''</font>]] 07:11, 13 October 2011 (UTC)
::Thanks! That certainly explains it. --[[User:Tryptofish|Tryptofish]] ([[User talk:Tryptofish|talk]]) 20:37, 13 October 2011 (UTC)
== Enacted small enhancements ==
*<s>Missed reference combination opportunity ([http://en.wikipedia.org/w/index.php?title=Grisellatheca&diff=prev&oldid=425207413 example]) [[User:Smith609|Martin]] '''<small>([[User:Smith609|Smith609]] – [[User_talk:Smith609|Talk]])</small>''' 18:51, 21 April 2011 (UTC)</s> {{fixedin|351}}
*<s>NB seems to break when trying to expand endnote templates, for instance http://en.wikipedia.org/w/index.php?title=Template:Molluscan_diets&oldid=425855201. [[User:Smith609|Martin]] '''<small>([[User:Smith609|Smith609]] – [[User_talk:Smith609|Talk]])</small>''' 17:31, 25 April 2011 (UTC)</s>
*<s> Some URLs are crashing the bot, in articles [[Sexually_dimorphic_nucleus]], [[Chronic prostatitis/chronic pelvic pain syndrome ]], and [[Recent_African_origin_of_modern_humans]]. [[User:Smith609|Martin]] '''<small>([[User:Smith609|Smith609]] – [[User_talk:Smith609|Talk]])</small>''' 01:02, 14 June 2011 (UTC)</s> {{Fixedin|355}} [[User:Smith609|Martin]] '''<small>([[User:Smith609|Smith609]] – [[User_talk:Smith609|Talk]])</small>''' 01:24, 16 June 2011 (UTC)
*Pipes in titles with cite web? e.g. {{cite web | url=http://www.meteorologynews.com/2009/10/29/cloud-streets-photographed-over-gulf-of-mexico/ | title=Cloud Streets Photographed over Gulf of Mexico: Gallery of Cloud Streets Images | Meteorology News | accessdate=2009-10-29}} [[User:Smith609|Martin]] '''<small>([[User:Smith609|Smith609]] – [[User_talk:Smith609|Talk]])</small>''' 16:04, 13 June 2011 (UTC)
**To me, that should be formatted as {{para|title|Cloud Streets Photographed over Gulf of Mexico}} {{para|work|Meteorology News}}. --[[User:Redrose64|<span style="color:#d30000; background:#ffeeee">Red</span>rose64]] ([[User talk:Redrose64|talk]]) 20:12, 13 June 2011 (UTC)
*(Also e.g. http://en.wikipedia.org/w/index.php?title=Trevor_Linden&diff=prev&oldid=434161461 and http://en.wikipedia.org/w/index.php?title=Turn_Left_(Doctor_Who)&diff=prev&oldid=434585064)
{{fixedin|374}}
* In press page ranges may be denoted | pages = n/a–n/a . See {{Cite journal | last1 = Leliaert | first1 = F. | last2 = Verbruggen | first2 = H. | last3 = Zechman | first3 = F. W. | title = Into the deep: New discoveries at the base of the green plant phylogeny | doi = 10.1002/bies.201100035 | journal = BioEssays | volume = 33 | issue = 9 | pages = 683-692 | year = 2011 | pmid = 21744372| pmc = }}.[[Special:Contributions/192.75.204.31|192.75.204.31]] ([[User talk:192.75.204.31|talk]]) 15:50, 11 July 2011 (UTC)
* "| Volume: 12 " and "| Issue: 12" are replaced by "Volume = : | unused_data=12". [[User:Smith609|Martin]] '''<small>([[User:Smith609|Smith609]] – [[User_talk:Smith609|Talk]])</small>''' 14:06, 12 July 2011 (UTC)
*NB still rearranging citation parameters: http://en.wikipedia.org/w/index.php?title=Template%3A+Cite+doi%2F10.2307.2F2400629&diff=prev&oldid=425351739 [[User:Smith609|Martin]] '''<small>([[User:Smith609|Smith609]] – [[User_talk:Smith609|Talk]])</small>''' 15:48, 22 April 2011 (UTC): This was in r338; cannot replicate now. [[User:Smith609|Martin]] '''<small>([[User:Smith609|Smith609]] – [[User_talk:Smith609|Talk]])</small>''' 22:13, 16 July 2011 (UTC)
*Remove leading zero ({{diff2|424890924|example}}). [[User talk:Ucucha|Ucucha]] 18:01, 19 April 2011 (UTC)
*:{{fixedin|374}} [[User:Smith609|Martin]] '''<small>([[User:Smith609|Smith609]] – [[User_talk:Smith609|Talk]])</small>''' 13:17, 19 June 2011 (UTC)
== Suggested small enhancements ==
*NB http://toolserver.org/~verisimilus/Bot/DOI_bot/category.php?cat=Refs_for_Citation_Bot is down. [[User:Smith609|Martin]] '''<small>([[User:Smith609|Smith609]] – [[User_talk:Smith609|Talk]])</small>''' 19:01, 21 April 2011 (UTC)
== Dead bot ==
{{bot bug
| title = The bot is dead
| status = {{tl|resolved}}
| reported by = [[Special:Contributions/70.137.134.91|70.137.134.91]] ([[User talk:70.137.134.91|talk]]) 09:48, 1 November 2011 (UTC)
| type of bug = <!-- Please select one of:
Improvement: The bot would be much better if ...
Cosmetic: The bot modifies appearance (e.g. capitalization) but not content
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
Deleterious: Human-input data is deleted or articles are otherwise significantly affected. Many bot edits require undoing.
Catastrophical: The bot should stop editing immediately
-->
| what happens =
<!-- and/or: --> | what should happen =
| link showing what happens =
| how to replicate the bug =
| waiting for = Consensus<!--
User: Input from editors
Consensus: Agreement on the best solution
Operator: Bot operator's feedback on what is feasible
Maintainer: A specific edit to the bot's code is requested below.
-->
| action required from maintainer = <!--specific details-->
}}
<!-- Discussion starts below this line -->
Above bug report was a mess, but it seems to be regarding a (now remedied) toolserver account issue, not a bot bug. [[User:LeadSongDog|LeadSongDog]] <small>[[User talk:LeadSongDog#top|<font color="red" face="Papyrus">come howl!</font>]]</small> 01:25, 3 November 2011 (UTC)
== RFC on identifiers ==
There is an [[Wikipedia:Village_pump_(policy)#RFC_on_the_bot-addition_of_identifier_links_to_citations|RFC on the addition of identifier links to citations by bots]]. Please comment. <span style="font-variant:small-caps; whitespace:nowrap;">[[User:Headbomb|Headbomb]] {[[User talk:Headbomb|talk]] / [[Special:Contributions/Headbomb|contribs]] / [[WP:PHYS|physics]] / [[WP:WBOOKS|books]]}</span> 15:50, 15 August 2011 (UTC)
== Fails on certain jstor ==
{{bot bug
| title = <!-- Brief description to display in section header -->
| status = new bug
| reported by = <span style="font-variant:small-caps; whitespace:nowrap;">[[User:Headbomb|Headbomb]] {[[User talk:Headbomb|talk]] / [[Special:Contributions/Headbomb|contribs]] / [[WP:PHYS|physics]] / [[WP:WBOOKS|books]]}</span> 16:42, 21 August 2011 (UTC)
| type of bug = <!-- Please select one of:
Improvement: The bot would be much better if ...
Cosmetic: The bot modifies appearance (e.g. capitalization) but not content
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
Deleterious: Human-input data is deleted or articles are otherwise significantly affected. Many bot edits require undoing.
Catastrophical: The bot should stop editing immediately
-->
| what happens = Fails on certain jstor like {{jstor|4494763}} (or a different jstor to the same article {{jstor|10.1086/519028}}. See [http://en.wikipedia.org/w/index.php?title=Principle_of_transformation_groups&oldid=446007535] and [http://en.wikipedia.org/w/index.php?title=Principle_of_transformation_groups&oldid=446008178]. The bot works fine when the doi is given, however [http://en.wikipedia.org/w/index.php?title=Principle_of_transformation_groups&oldid=446008351][http://en.wikipedia.org/w/index.php?title=Principle_of_transformation_groups&diff=prev&oldid=446008389].
<!-- and/or: --> | what should happen =
| link showing what happens =
| how to replicate the bug =
| waiting for = Consensus<!--
User: Input from editors
Consensus: Agreement on the best solution
Operator: Bot operator's feedback on what is feasible
Maintainer: A specific edit to the bot's code is requested below.
-->
| action required from maintainer = <!--specific details-->
}}
<!-- Discussion starts below this line -->
Funny, the JSTOR API seems not to be returning results. The bot plugs the JSTOR identifier into the dc.identifier search term accessible via http://dfr.jstor.org/sru/? ; in the past this has worked, today it doesn't seem to. I wonder whether this is a temporary glitch - I hope so! [[User:Smith609|Martin]] '''<small>([[User:Smith609|Smith609]] – [[User_talk:Smith609|Talk]])</small>''' 16:57, 21 August 2011 (UTC)
==References==
{{reflist|name="Bryan1901b"}}
Things were working fine in the version before the bot-run, and I'm happy with the sfn-route I took things in. NB: the quotes around 'foots' are problematic (and I prefer the lower-alpha' presentation). Anyway, the bot seems to have improperly parsed things due to the nesting and/or multiple reflist instances. (see also; Indiana) —Portuguese Man o' War 13:49, 22 September 2011 (UTC)
Is this a problem that would be avoided if conventional ref tags were used? Martin(Smith609 – Talk) 21:22, 25 September 2011 (UTC)
Sorry I didn't notice this for a week. I expect it may be due to using the #tag:ref. That article has moved away from that approach, both because of this and because it is better to stick to the usual syntax when possible. At some point ref-tags may be able to be nested ;) It seems to me that (past) limitations of MediaWiki are responsible for millions of references being fleshed-out inline instead of kept in discrete sections. The current inability to nest effectively is similarly forcing practice in other than the best direction. The costs of late software are huge. The cost of developing quality software are huge, too. —Portuguese Man o' War 05:01, 3 October 2011 (UTC)
Error code: BADMD5: The supplied MD5 hash was incorrect.
Writing to De_Sitter_invariant_special_relativity ... Edit may have failed. Retrying: xxx Still no good. One last try: Failed. Error code: BADMD5: The supplied MD5 hash was incorrect.
Run on a page containing (certain?) special characters. The character string passed to the md5-hasher must be in a different encoding to the string received by Wikimedia (?)
Confusing results when no changes are made by the bot
For new users of this bot, such as myself yesterday, the results when no changes are made are baffling. I almost gave up on the bot thinking it was another broken or developing Wikipedia function. One sees this at the top of the page:
Latest revision _______ Your text.
Followed by the edit window. When one saves the page hoping for some improved citations one sees no changes, and nothing shows up in the revision history.
Fortunately, I used it on a page where it made some changes and now I see how it works. I suggest some kind of indication be added to the results saying something like this:
"Citation bot sees nothing it can fix, and no changes have been made."
Thanks for this bot. It is very useful. --Timeshifter (talk) 21:19, 7 October 2011 (UTC)
Messes with author... again
Status
Duplicate of Expanding author lists truncated with et al.
I think it's pretty much established by now that the bot should not be editing the author fields for any reason. It's removing et al., cluttering citations with additional parameters, etc... This has been raised several times in the past, and I'm re-raising this again because for some reason these changes were re-enabled. The bot does not have consensus to perform changes related to authors fields, so please stop the bot from making them. Headbomb {talk / contribs / physics / books} 09:41, 20 June 2011 (UTC)
Can you please clarify how the displayed output is changed? Martin(Smith609 – Talk) 21:03, 20 June 2011 (UTC)
The et al. is removed, indicating that the named author did it all by himself, when in fact he has unnamed coauthors. The name of a collaboration is given after the author in the case of this particular work, but it is unclear if that is an indication of coauthors, or just affiliation information. Since citation templates have no manual of style, there is no way to determine the meaning of the phrase in parenthesis after the author. Readers must not be expected to understand how Citation bot works, or even that it exists, so I preemptively reject any explanation about how the bot looks it up in some database. The rules for presenting authorship in any such database are not incorporated in the description of the citation templates, and thus are not available to readers.Jc3s5h (talk) 22:12, 20 June 2011 (UTC)
I have looked at the diff link above, and whilst there are several changes, they mostly follow the same pattern. Ignoring irrelevances such as issue numbers and page ranges, the main problem seems to be centred around a misapprehension of how the |display-authors= parameter works. For example, in the first change, the |author=R. Brandelik et al. (TASSO collaboration) is changed to |author=R. Brandelik (TASSO collaboration) and two parameters are added: |author-separator=,|display-authors=1. This would work as intended if either |author2= or |last2=|first2= were provided, but they aren't. |display-authors= only operates if there are more authors than the figure specified. As in:
Doe, John (2011). "A Paper". A Journal. {{cite journal}}: Invalid |display-authors=1 (help)
I seem to recall that in the past the bot would fill in those missing author parameters. Why is it not doing so here? If it were, the "et al." would be triggered. --Redrose64 (talk) 11:26, 21 June 2011 (UTC)
I've made a start with a fix here. I am not convinced that "J.D. Smith et al (JADE collaboration)" is a valid value for the author parameter, so am not sure how to handle it. If it is the norm in particle physics, then I can add a special case. If it is confined to a couple of articles then it will be quicker to fix by hand. Please advise. Martin(Smith609 – Talk) 21:44, 16 July 2011 (UTC)
It's indeed the norm, but it also doesn't change the fact the the bot should not mess with the author fields in the first place. Having a bot adding lists of unseen authors for the sake of metadata has never been approved, and whenever a bot's been messing with the authors fields, it always comes up here so I doubt it qualifies as uncontroversial changes. Headbomb {talk / contribs / physics / books} 14:43, 17 July 2011 (UTC)
Will you stop the bot from making those stupidly annoying changes? [28]. I don't know how many of them I reverted in the last few days. I've asked for a blocked until this is resolved. Headbomb {talk / contribs / physics / books} 14:42, 29 September 2011 (UTC)
A specific edit to the bot's code is requested below.
Requested action from maintainer
Looks like there is consensus to translate "id=ISSN xxxx-xxxx" to "issn=xxxx-xxxx", but not to add or remvoe an issn parameter otherwise. Correct me if wrong.
I don't think so... Template:Cite journal/doc doesn't state when not to provide an |issn= - it merely states that |publisher= is only necessary in the absence of |issn= or |doi= - this does not mean that in the presence of a |publisher=, the |issn= should be removed. Oh BTW, here's another. --Redrose64 (talk) 19:09, 21 September 2011 (UTC)
Actually, Template:Cite journal/doc says "Only include if ISSN and DOI are unavailable." That means the publisher should be omitted if those parameters are present, and says nothing about whether the publisher should be named if ISSN and DOI are absent. Indeed, most scholarly citations I have seen do not name the publisher. Jc3s5h (talk) 19:38, 21 September 2011 (UTC)
The problem was not about the 'publisher' field. The problem is that if ISSN is presented like e.g. 'id = ISSN 0208-189X', the bot does not replace it with 'issn = 0208-189X', but just remove it. The latest removal I discovered happened just 30 minutes ago. Removing the ISSN field, even if incorrectly formatted, is not the correct solution. Beagel (talk) 16:44, 6 October 2011 (UTC)
I tend to think that issn's are almost always useless and should be omitted, but that's a human editorial decision that shouldn't be left up to the bot. I agree that deleting id=ISSN rather than converting it into issn= is the wrong thing for the bot to do. —David Eppstein (talk) 18:23, 6 October 2011 (UTC)
We've been over this ground before, iirc. In any case where there is no more specific article identifier available, the ISSN is useful in finding the serial, though an OCLCn would often be equal or better. The template can and should suppress display of the ISSN when such article-specific ID is provided (e.g. DOI, PMID, JSTOR, etc) but there's no reason the ISSN shouldn't be included in the COINS metadata. Converting "|id=ISSN " to "|issn=" is a harmless and helpful bot task. LeadSongDogcome howl! 19:20, 6 October 2011 (UTC)
Using the Citation bot at the Hebrew Wikipedia
Greetings,
Could Citation bot run at the Hebrew Wikipedia as well? We have translated Template:Cite doi to he:תבנית:ציטוט DOI and the relevant data should be saved to subpage /מזהה DOI. The reference itself should look the same as in the English Wikipedia, preferrably with the aid of a divstyle alignment (i.e. <div style="direction: ltr;">To the left</div>). Many thanks, ליאור • Lior (talk) 16:11, 19 September 2011 (UTC)
That would be great. Past requests for internationalization have petered out when I've enquired whether someone is willing to put in a bot request at their native wiki. If you can get approval for the bot to run, I'll start working on the coding.
Frequently, the {{citation}} template is followed by a period, outside the template. When the bot converts it to cite journal (to match the majority of other citations in the article), the bot adds a postscript= to suppress the period generated by the template, so that we don't get doubled periods. This causes the correct appearance, but increases the complexity of the template's parameters. It would be simpler and better to simply remove the trailing period instead of using a postscript.
^ Niemitz, C. (2010). "The evolution of the upright posture and gait--a review and a new synthesis.". Die Naturwissenschaften 97 (3): 241–263. Bibcode 2010NW.....97..241N. doi:10.1007/s00114-009-0637-3. PMC 2819487. PMID20127307. edit
04:55, 10 June 2010 Citation bot 1
citation 37 is empty
We can't proceed until
Agreement on the best solution
Have you been able to replicate this bug since 2010? Martin(Smith609 – Talk) 03:45, 8 October 2011 (UTC)
The date format did not match the existing format of day first dates, as it is not a full date should not use ISO format and should use the "month" & "year" parameters.
Often the pipe is part of the preceding parameter. In either case its usage is incorrect. Any suggestions about how to improve the bot's guesswork?
Regarding the main bug, what fix do you propose? Martin(Smith609 – Talk) 03:44, 8 October 2011 (UTC)
Might be part of the preceding parameter, but I don't think I ever saw a pipe in a URL. Headbomb {talk / contribs / physics / books} 04:43, 8 October 2011 (UTC)
You can't have an unescaped pipe in the URL as the template parser always interprets it as a separator, but there are some URLs that do use pipes that must be escaped (as the CB is doing). In the pipe character example I think it would be better if the bot didn't try to fix it, and maybe added a 'bad cite' tag. I've been working on a script to fix such errors (parameter name and = missing), but it's not easy and often requires human review. Rjwilmsi 07:19, 8 October 2011 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
On Template talk:Cite journal, a number of examples crafted to exhibit variation in parameters, or the difference between {{cite journal}} and {{citation}}, were altered so that the points under discussion were no longer suitably illustrated.
Please leave discussion pages alone. If I recall correctly, that this has come up before, some months ago.
I've tagged the page with {{bots|deny=citation bot}}. Martin(Smith609 – Talk) 03:41, 8 October 2011 (UTC)
Author names again
Please do design some short field to stop the bot from changing author lists in references; et al. could be ideal. The bot is very useful, but it is adding long author lists and other code such as "author separator", "display authors= " and "<Please add first missing authors to populate metadata.>". This is a real issue with overloading long articles as it adds so much junk code. The bot is currently designed to care about the article output ignoring the code length. Another undesirable feature is "|postscript = ..; inconsistent citations" which adds so much code just to care about a full stop at the end of the reference. We do love this bot and do not want to disable it on specific articles. Just a little flexibility in its operation (don't change when asked not to) will solve this longstanding issue. Materialscientist (talk) 00:59, 9 October 2011 (UTC)
WTF?
Whast on earth is going on here? There are edits removing the whole content of the page and leaving nothing but fragments of code [29], leaving the following edit summary: "Touching page to update categories. ** THIS EDIT SHOULD PROBABLY BE REVERTED ** as page content will only be changed if there was an edit conflict." Paul B (talk) 10:53, 5 October 2011 (UTC)
This is a recurring problem: User talk:Citation bot/Archive1#bizarre deletion and addition of content. The bot does not have approval for null edits, whether they are done properly or not. Accordingly, I have blocked Citation bot 2 from editing pending confirmation that the bot will no longer make these edits improperly, and no longer make them without approval. –xenotalk 16:40, 5 October 2011 (UTC)
The earlier case was an edit conflict. This case looks like an unavailable server, which likely was a problem with the MW 1.18 rollout. Similar edit comments gave the impression the error was the same. LeadSongDogcome howl! 01:25, 6 October 2011 (UTC)
The task should not have been active at all; it has no approval to make null edits. –xenotalk 14:34, 11 October 2011 (UTC)
DOI: 10.1007/3-540-63238-7_38 could not be auto-filled; rerunning the bot after manually filling the template inserted an incorrect author name for the reference.
Thank you for the extremely helpful bot! A pity it didn't work for this reference. Maybe this can be helpful for improving the parser; I don't "expect" any action, I just wanted to share this DOI with you, since it didn't work. --Chire (talk) 14:47, 8 October 2011 (UTC)
Incidentally, {{cite journal}} is the wrong template to use here — LNCS is not a journal, it's a book series. But that was not a bot error as the bot didn't choose that template. I changed the one in the example to {{cite conference}} instead. —David Eppstein (talk) 18:56, 8 October 2011 (UTC)
Thank you! Makes sense. cite journal was the pregenerated template when I clicked "fill manually". --Chire (talk) 09:42, 9 October 2011 (UTC)
It looks like a server is down. Here is an example of an edit summary:
"[cw408] Touching page to update categories. ** THIS EDIT SHOULD PROBABLY BE REVERTED ** as page content will only be changed if there was an edit conflict."
Here is an excerpt from an edit:
"<h1>Sorry! This site is experiencing technical difficulties.</h1><p>Try waiting a few minutes and reloading.</p><p><small>(Can't contact the database server: <span dir="ltr">Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2) (localhost)</span>)</small></p><hr /><div style="margin: 1.5em">You can try searching via Google in the meantime.<br />
AKA pure clutter. Metadata is a secondary concern. It should not trump desires for clean edit windows. et al. means "shut up" about the other authors, there's too many to list or care for. Headbomb {talk / contribs / physics / books} 20:15, 17 September 2011 (UTC)
<indent>The issue might be more general. I perceive it as currently the bot tends to not care about bloating the reference templates with unused text, whereas this has at least two issues (i) the article code becomes hard to work with, especially for newcomers; (ii) reaching parsing limits on some articles. Apart from expanding author lists, the bot adds "author-separator" field and a long line of |postscript= and "{{inconsistent citations}} .." note. Can the bot not do that please? Citations are often inconsistent for numerous other (than full stop at the end) reasons which the bot can't assess and which are sorted out manually at WP:GAN or WP:FAC. Cheers. Materialscientist (talk) 23:04, 17 September 2011 (UTC)
Given the above I've modified the bot (in r408) so that it only adds the 'bloat' if this is already the editorial preference on the page (i.e. "|display_authors=" is set in other citations). I would suggest that someone modifies Template:Citation/doc, which I had been reading to imply that it was incorrect to specify more than one author in the 'author' parameter. I just re-read Template:Cite journal/doc, says something different. Cheers, Martin(Smith609 – Talk) 14:34, 1 October 2011 (UTC)
This user's unblock request has been reviewed by an administrator, who accepted the request.
This bug led to a block request of User:Citation bot 1 per [34] . Because Citation bot 1 does not have a separate user talk page (it redirects here) I am asking for it to be unblocked here. Note that User:Citation bot is not blocked, but User:Citation bot 1is.
Accept reason:
See below.
As noted below, this request pertains to User:citation bot 1, which doesn't have a separate talk page. Martin(Smith609 – Talk) 16:09, 1 October 2011 (UTC)}}
Seems to me that if the bot did not expand the author lists whenever there's an et\s+al\.? in the author/first/last/coauthor/etc... fields, it would solve this issue. Headbomb {talk / contribs / physics / books} 04:40, 8 October 2011 (UTC)
Agree that the duplication is not useful, though I think it would be better to replace the JSTOR link with the DOI link. Rjwilmsi 07:13, 13 October 2011 (UTC)
If find duplication very useful. If I click on "doi" I don't know where I'll land. I'll know where I'll land if I click on the JSTOR link. Plus, we also have offline version to think of, and these greatly benefit from having both the doi and the jstor given. Some databases are queriably via JSTOR, others via DOI. Both offer complementary information. Headbomb {talk / contribs / physics / books} 07:47, 13 October 2011 (UTC)
Duplication of links in "cite journal" might be bad, but there are subtleties: (i) Access. Most institutions either have access to jstor or haven't; this is unpredictable for other links (doi/pmid/etc). (ii) Visibility. jstor always provides the first page with the abstract, but it is not plain text and can't be copy/pasted; only some doi (wiley) lead to a similar first-page access. Most doi/pmid lead to plain-text abstracts, but some just to the title. Materialscientist (talk) 07:56, 13 October 2011 (UTC)
If my first comment wasn't clear I was only talking about the strict case where there is a jstor link e.g. |jstor=11345 and a DOI with the JSTOR in it e.g. |doi=10.2307/11345 (not real values), since both resolve to the same place. If there is a jstor and a DOI to a different publisher then both are certainly useful and are not duplication. Rjwilmsi 08:29, 13 October 2011 (UTC)
When most citations are cite xxx, citation templates are converted, but the ref=harv parameter is never included, so linking from Harvard citations or short footnotes to the full bibliographic information is broken.
What should happen
Always include ref = harv on conversions from citation to cite xxx.
I have tried it out on Seniors centre and got the message "Educational Activity Programs for Older Adults Already has an ISBN. Book reference assessed in 0 secs."
What should happen
It should report ISBN missing and/or try to fix it
Just a quick note that I reverted an edit by this bot. I saw no improvements, and arguably three changes which are all making the article worse. (The citation to cite book change is perhaps arguable, but I have noticed in the past that cite book sometimes does not work with harvtxt templates.)--Andrew Lancaster (talk) 14:58, 14 October 2011 (UTC)
By default, {{cite book}} doesn't work with {{harvtxt}} (or any of the other Harvard-group templates); but it can be made to work by the simple addition of |ref=harv to the {{cite book}}, something that is built in to {{citation}}. --Redrose64 (talk) 16:29, 14 October 2011 (UTC)
Thanks for that information, but what is the advantage of doing that?--Andrew Lancaster (talk) 21:41, 14 October 2011 (UTC)
The advantage of changing {{citation}} to {{cite book}} is for consistency: previous to the bot edit, the article had one each of {{citation}}, {{cite book}}, {{cite journal}} and {{cite web}}, which is a ratio of 3:1 of {{cite xxx}} over {{citation}}, and you shouldn't mix the two styles in the same article.
The advantage of adding |ref=harv to a {{cite book}} is so that {{harvtxt}} will link to it. I now see that in Causality, {{harvtxt}} isn't actually used (nor are any other Harvard-group templates, or {{sfn}}), so the absence of |ref=harv isn't a problem. --Redrose64 (talk) 10:50, 15 October 2011 (UTC)
I performed a test case and found that the bot does not add |ref=harv when converting from citation to cite xxx, even when there is a short footnote pointing to the reference, and even when other cite xxx templates in the article are using |ref=harv. This is a bug. A crude solution would be to always add |ref=harv. A fancy solution would be to detect whether there is a short footnote or parenthetical note pointing to what the anchor would be, and add the parameter when necessary. Jc3s5h (talk) 13:51, 15 October 2011 (UTC)
@Redrose, I do not understand the importance of using only one template on an article, but anyway, if it is important what would be wrong with just using citation templates on an article?--Andrew Lancaster (talk) 11:50, 17 October 2011 (UTC)
I didn't say that only one template should be used. I hoped to say that only one style should be used. The three templates {{cite book}}, {{cite journal}} and {{cite web}} (collectively known as {{cite xxx}}) are all the same style (known as Help:Citation Style 1); {{citation}} is a different style, see Wikipedia:Citation templates#Introduction, first list, items 2 & 3. There is no harm in using {{citation}} exclusively, but if the article is already using the {{cite xxx}} templates, new citations should use those, or others in the same group.
The bot selects one title as "authorative" where humans have incorrectly used title for chapter in book citations; this deletes vital citation content: the chapter of works cited in collection
What should happen
The bot leaves multiple title parameters in book citations for manual resolution
We can't proceed until
Agreement on the best solution
I have a hard enough time indicating to editors that chapters in collections need to be cited individually without this occurring where humans get it half right. This needs to be left for manual resolution. Fifelfoo (talk) 23:58, 15 October 2011 (UTC)
removing needed italics from journal article title
In this example, the bot removed the italics from the Latin species name Smilodon fatalis in a journal article title, which of course are appropriate and which are present in the actual title.
Amazon/asin
Hi
I just saw this edit by the CitationBot and wanted to know why the bot is removing access dates. I consider them very helpful if I want to verify when the information was obtained, since it could have changed in the mean time. --Maitch (talk) 13:32, 20 October 2011 (UTC)
Ugh! Why was that using {{cite web}} instead of {{cite video}} anyhow? The idea of Amazon as a video "publisher" would be a scary one if it were accurate. But accessdates are for volatile urls, not for permalinks. An ASIN, like an ISBN, is supposed to describe exactly one edition. Do you have reason to think that isn't the case? LeadSongDogcome howl! 15:25, 20 October 2011 (UTC)
I'm not entirely sure what you mean. I did not write that article, but Amazon is sometimes used to prove that a DVD simply exists or to document a release date. That part of Amazon is not user edited and therefore can still be considered a RS. "Cape Feare" is a featured article that also uses Amazon as a source. --Maitch (talk) 23:06, 20 October 2011 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
In attempting to combine named refs which happen to be malformed with one quote (not a pair) e.g. <ref name="socialnomics.net>...</ref>, two extra double quotes are introduced <ref name=""socialnomics.net" />, rather than just one <ref name="socialnomics.net" />
Please stop making this edit. It adds nothing useful. When an editor has been reverted, they should not make the same edit again. The edit removes the italics from a scientific name, wrongly, and repeats two pieces of information (the issue number / page number in one instance – many online journals don't fully distinguish the two, and a DOI beginning 10.2307, when a JSTOR link is already included). --Stemonitis (talk) 10:22, 21 October 2011 (UTC)
Citation bot autoimplements Harvard style?
Can someone explain why did the bot add "| ref =harv" to "cite journal"
here? Materialscientist (talk) 12:51, 21 October 2011 (UTC)
I guess Smith609 has taken note of the bug described in #Impairs function of Harvard and short citations and fixed it in a very simple way. I infer that if the bot has to change any {{citation}} to a {{cite xxx}} it sticks ref=harv into every template it touches on that run. Since in the diff provided by Materialscientist, the citation for the book by Neil Nodding was converted from citation to cite xxx, every template in that edit had ref=harv stuck in. Jc3s5h (talk) 13:37, 21 October 2011 (UTC)
Please note that the addition of |ref=harv does not change the referencing style: all it does is to create an anchor to facilitate the linking of parenthetical refs or shortnotes, such parenthetical refs or shortnotes having already been created by means of templates in the {{harv}} family, including {{sfn}}. The {{citation}} template has a |ref=harv built in by default - the {{cite xxx}} family do not.
In the article Education, ref no. 42 is created using one of these templates: to be precise, <ref>{{Harvnb|Noddings|1995|pages=1–6}}</ref>. If you click on the bluelink in that ref, it doesn't go anywhere. You can reveal the broken link as a red error message by following the technique described at User:Ucucha/HarvErrors. The addition of |ref=harv to the {{cite book}} in ref 40 will fix that broken link. --Redrose64 (talk) 16:40, 21 October 2011 (UTC)
I understand that. My point is the bot creates code clutter, and adding simple logical rules can prevent it from doing so. See, e.g., subsection "Author names again" above. With all the benefits the bot brings, I often resort to disabling it on long articles. Materialscientist (talk) 00:20, 22 October 2011 (UTC)
Bot filled missing information into a journal citation, but placed "111011113923008" as the page number. I have not been able to find this page number stated in journal places online, so I suspect it is some sort of parsing error. The number is excessive anyway, suggesting an invalid page number anyway.
What should happen
Refrain from inserting this again. Instead leave alone or find correct page number.
I would suspect rerunning the bot through the Folding@home article would add back in this page number, as I have undone that particular work.
We can't proceed until
Bot operator's feedback on what is feasible
Requested action from maintainer
Please investigate the cause, possible making a special case to leave this particular journal citation alone. A better solution would be to have the bot find the correct page number and use that instead.
or cite doi any more. comes back with "the user account at toolserver.org expired" if click on "jump queue"
70.137.134.91 (talk) 09:52, 1 November 2011 (UTC)
Gives error message
403: User account expired
The page you requested is hosted by the Toolserver user verisimilus, whose account has expired. Toolserver user accounts are automatically expired if the user is inactive for over six months. To prevent stale pages remaining accessible, we automatically block requests to expired content.
If you think you are receiving this page in error, or you have a question, please contact the owner of this document: verisimilus [at] toolserver [dot] org. (Please do not contact Toolserver administrators about this problem, as we cannot fix it—only the Toolserver account owner may renew their account.)
HTTP server at toolserver.org - ts-admins [at] toolserver [dot] org
70.137.134.91 (talk) 10:14, 1 November 2011 (UTC)
I will second this -- I was trying to use the cite DOI template a few minutes ago, and it appears to be broken. Wingman4l7 (talk) 23:59, 2 November 2011 (UTC)
Well, it seems that the bot did expand it, just not as soon as expected. Perhaps the toolserver account has been sorted? LeadSongDogcome howl! 01:16, 3 November 2011 (UTC)
Yip. Sorry: my home computer has been out of action for a while so I missed the chance to renew my account. All should be back in order now! Martin(Smith609 – Talk) 01:27, 3 November 2011 (UTC)
Here. Search for "last2" to find the duplications. I thought the ampersand in the "author=" list tricked the bot; thus replaced ampersands and re-run, with the same result. The bot seems to misunderstand the "author=" field. Materialscientist (talk) 13:45, 7 November 2011 (UTC)
Strange: I haven't touched the bot for some time, and it works fine in some cases. Martin(Smith609 – Talk) 17:17, 20 November 2011 (UTC)
The specified link looks to be working fine now. Perhaps this was an intermittent server problem? Marked as resolved. Martin(Smith609 – Talk) 15:42, 21 January 2012 (UTC)
Fixed; thanks. You are welcome to edit the page yourself, by the way! Martin(Smith609 – Talk) 16:05, 21 January 2012 (UTC)
Apparently improper converts to doi.
In trying to pare down the incomplete doi references page, I've run into a few places where the Citation bot has replaced a jstor.org link with an incorrect doi. These are
A step in the right direction...Naraht (talk) 19:42, 16 December 2011 (UTC)
The bug is in the url2template function at line 631 of expandFns.php. The regular expression fails to recognize all types of JSTOR URLs. It could probably be fixed by changing the regex into something like "~jstor\.org/(?!sici).*[/=](\d+)~". However, ideally the bot should retrieve data from the SICI; not sure whether it already has code available that does that. Ucucha (talk) 19:57, 16 December 2011 (UTC)
Should we use the bug template for that?Naraht (talk) 17:14, 22 December 2011 (UTC)
Yes, the bot extracts info where it finds a SICI. I've implemented the suggested regexp modification: Fixed in GitHub Pull 409Martin(Smith609 – Talk) 15:56, 21 January 2012 (UTC) {{resolved}}
This was garbage-in-garbage-out. See the para containing "Fuerst", with its corrupted ref tag. We've previously discussed handling of mismatched markup tags, braces, etc, this looks like another example. LeadSongDogcome howl! 23:15, 19 January 2012 (UTC)
A specific edit to the bot's code is requested below.
Also: this citation should be expanded with a DOI: Kroger, B. (2007). "Early orthoceratoid cephalopods from the Argentine Precordillera". Journal of Paleontology. 81 (6): 1266–1283.
Citation bot evidently created a page in the Template: namespace containing wikitext from a citation that appears in several Wikipedia articles. I'm not sure which article Citation bot was editing at the time.
That's (a large part of) this bot's job. It generates a citation automatically from a database, and stores it on a Wikipedia page, so that multiple articles can easily reuse the same citation. In this case, {{cite doi|10.1093.2Feurpub.2Fckn139}} will produce the citation. Ucucha (talk) 00:30, 2 December 2011 (UTC)
This is from a few months back (29 September). In the [Comparison of regular expression engines] article, the bot deleted the definition text for a ref with note "Combined duplicate references" resulting in a cite error since it was not a duplicate.
The ref was named "unicode optional 1" with spaces. An earlier ref was labeled "unicode_optional" underscores. Perhaps this combination triggered a false hit in the duplicate reference matching logic.
We can't proceed until
Agreement on the best solution
Requested action from maintainer
Investigate and fix please.
The bot was confused by Template:R, which should not be mixed & matched with <ref> tags. Martin(Smith609 – Talk) 17:24, 21 January 2012 (UTC)
Citation deleted due to a stray newline?
I saw this revision on my watchlist. It corrected a dash issue in one citation, but deleted another citation. I don't know much about such things, but I'm guessing the bot was confused by an extra newline in the middle of the second one. Ntsimp (talk) 14:44, 9 January 2012 (UTC)
The presence of newlines in the second one is not an error - newlines are perfectly valid in citation templates, provided that they don't occur partway along either: (i) the template name; (ii) a parameter name; or (iii) the value of a named parameter (named or unnamed). In this case the two newlines occur (i) between a parameter value and the pipe character which separates that parameter from the next, and (ii) between the }} which closes the template and the </ref> which closes the reference. Both of these are legal uses of newlines. --Redrose64 (talk) 18:20, 9 January 2012 (UTC)
I said I didn't know much. So why did the bot delete a whole reference? Ntsimp (talk) 20:33, 9 January 2012 (UTC)
OK, what's the problem? It changed the doi that was outside of a reference to being inside which strikes me as the correct thingto do.Naraht (talk) 16:50, 9 January 2012 (UTC)
Deleterious: Human-input data is deleted or articles are otherwise significantly affected. Many bot edits require undoing and/or Catastrophical: The bot should stop editing immediately
What happens
References are deleted. Not sure if that's an isolated incident or a more general issue.
A specific edit to the bot's code is requested below.
It's the same issue as I reported above. Not an isolated incident, at least. U+003F?
This case looks a bit more subtle. Here it was evidently tripped up by citations that were duplicates except for page numbers. LeadSongDogcome howl! 20:07, 8 December 2011 (UTC)
Not quite; the cases also differed in |url= and |title=. I haven't been able to find the bot code that performs these changes; Citation bot's code is rather unstructured in my opinion. Ucucha (talk) 20:19, 16 December 2011 (UTC)
Can you suggest resources that would help me to improve the structure? I've learned a lot since I started coding the bot but good structural practice is something that's hard to pick up intuitively. Martin(Smith609 – Talk) 15:47, 21 January 2012 (UTC)
A lot of it is probably my personal preference and habit; I've written some similar code for my own use (most of it at [40]), and approached it differently. In your code, many of the files have non-intuitive names (e.g. DOItools.php has stuff that deals with much more than DOIs, expand.php isn't the only place where citations are expanded, and what it has to do with expandFns.php I'm not entirely sure). I think using objects (for example, to represent a single citation) would help, and avoiding some globals. You could also make blocks of code (files, classes, groups of methods, and soon traits) for related functionality (e.g., expanding DOIs, finding DOIs, dealing with fishy parameters, and whatever else the bot does). Ucucha (talk) 18:34, 21 January 2012 (UTC)
The bot stumbles very often on the {{reflist|refs= <ref></ref>}} referencing style, by deleting some references. I can provide at least several examples of that, a recent one here. Materialscientist (talk) 08:53, 22 December 2011 (UTC)
A typo meant that if the stem of a reference was the same as the name of another (e.g. "name=ref" and "name=ref_2"), the bot would spot the stem "ref" and mark the references as duplicates with the same name. This was the case in Germanium, and I've fixed this bug in r413. Do any of the examples you mention not conform to this pattern? (NB references aren't being deleted in Gary Speed in the latest build.)Martin(Smith609 – Talk) 16:50, 21 January 2012 (UTC)
Articles on medical subjects
An editor -- SandyGeorgia -- complains of the edits made by the DOI bot. She says the bot edits do not follow the standard practice of editors of medical articles:
"We don't use capital case in article titles on Wikipedia-- we use sentence case (which means we don't capitalize the words in the article titles). [Confirmed by WP:MOS, yet DOI bot is substituting capital case.] And, most medical articles use the Diberri format, which uses only one author field."
Please advise. Fconaway (talk) 04:36, 29 December 2011 (UTC)
WP:MOS is about articles, not citations. How to write citations is covered in WP:CITE. That guideline mostly says follow the established style in the article, and if the article relies on a particular printed style guide, follow it. This bot (Citation bot AKA DOI bot) only operates on citations that use citation templates. There is no manual of style for the citation templates, so there is no answer to your question. I suggest you either write a style manual for the citation templates or stop using them.
If you are the editor invoking Citation bot, you are responsible for the use. If the established style in an article is to use sentence-case capitalization for the title of citations, and Citation bot defies the established style in the article, stop using Citation bot, or fix the titles manually. Jc3s5h (talk) 16:47, 29 December 2011 (UTC)
I would think the citation bot could assess existing use of |author= versus |last1=, and existing case in titles, and follow the existing conventions if present. I'm sure this would be helpful to and appreciated by wikiproject medicine and though it's not my code I wouldn't think it would be that difficult to implement. Rjwilmsi 20:19, 30 December 2011 (UTC)
It could guess; if existing titles have lots of capitals and the title to be added has few capitals, the bot could refuse to act, and vice versa. But the bot couldn't change the case of a title due to proper names and a few names that retain a lower case letter at the beginning, such as "iPod". Jc3s5h (talk) 20:26, 30 December 2011 (UTC)
It sounds as if the bot is taking the title value from CrossRef rather than from PubMed data. I would suggest that if a PMID is found, the PubMed data should be preferred over the CrossRef date, if only for its openness. That said, I don't think anyone will be terribly confused one way or the other by titlecase/sentencecase choices. Just so long as we don't inflict ALLCAPS TITLES on readers. LeadSongDogcome howl! 23:20, 9 January 2012 (UTC)
Is there any way of converting from Title Case to sentence case, whilst keeping proper nouns capitalized? Martin(Smith609 – Talk) 17:28, 21 January 2012 (UTC)
The error was AdsAbs', but I've incorporated a fix in r414. Martin(Smith609 – Talk) 17:46, 21 January 2012 (UTC)
Handling angle brackets (greater than/less than)
I don't think the Citationbot is handling angles correctly. On Rubidium telluride, the {{cite doi|10.1002/1521-3773(20020802)41:15<2725::AID-ANIE2725>3.0.CO;2-G}} doesn't seem to get handled correctly, and yet, when I enter that doi at dx.doi.org, it doesn't seem to bat an eyelash. I've tried changing the angles to %3C and %3E but that doesn't seem to help. (Similar issues apparently at Jack D. Dunitz Any ideas?Naraht (talk) 11:03, 5 January 2012 (UTC)
If I recall correctly, dot-encoding (.3C and .3E) works, because citation bot anchorencodes the DOIs. Ucucha (talk) 13:12, 5 January 2012 (UTC)
I try those techniques. (But if one of you wants to try first and let me know, go ahead. :) )Naraht (talk) 02:45, 6 January 2012 (UTC)
Looks like both techniques would work. I tried changing one them to the lt; > with amper and that worked in the *preview* and was able to connect to the dot encoded version. Now how do we fix this disconnect and/or document it.Naraht (talk) 11:00, 6 January 2012 (UTC)
The problem is that MediaWiki changed the Anchorencode parserfunction so that it removes anything contained within <>s. I've updated the bot accordingly but will have to work out how to go through the backlog. Martin(Smith609 – Talk) 18:19, 21 January 2012 (UTC)
MD5 problems?
I've attempted to change two cite broken doi entries to cite doi and jump the queue. But I get "Edit may have failed. Retrying: xxx Still no good. One last try: Failed. Error code: BADMD5: The supplied MD5 hash was incorrect." after it seems to have gotten to the end. What's going on with that? (See http://toolserver.org/~verisimilus/Bot/DOI_bot/doibot.php?doi=10.1063%2F1.3633090 as an example) Naraht (talk) 03:20, 7 January 2012 (UTC)
Investigating; and as a note to myself the capitalization of |last1= has gone a bit funny... Martin(Smith609 – Talk) 22:14, 21 January 2012 (UTC)
Blocked
First and foremost - I have great respect to this bot and its operator.
Block summary: the bot was introducing minor changes, some of which were questionable and too many of which (IMHO) were hardly an improvement. The block is to initiate a dialogue with the person behind those edits. (If he/she does not show up, then I propose the bot operator makes the user identifiable).
Difs. Let us go from last edit back.
Surely there are many constructive edits.
Most edits bring infinitesimal changes, i.e. spamming watchlists. This is a minor issue and the proposal here is to see who initiated the edit (and engage into a friendly dialogue with them), and maybe amend the bot code not to save minor edits. Examples [41][42][43][44][45][46][47][48][49][50][51][52][53][54][55][56][57], .. , this can go for pages.
[58][59] In some edits the bot corrects the article title (removes italics from species names) perhaps using some html reference, whereas the actual title uses italics (say, [60]).
More on titles: [61] the first bot-corrected ref is capped in the PMC, but not in the original [62]. The second ref title should contain Interleukin-21, not IL-21 (perhaps taken from erroneous PMC), same thing with the third bot-corrected title which should start with "The" per doi (doi is the primary link, pmc and pmid are derivatives). The issue is not about these changes, but about the bot algorithm of title correction. Materialscientist (talk) 06:48, 10 January 2012 (UTC)
Here too the bot was overwriting user input with data from the PMC database. It should not have been. Fixed in GitHub Pull 417Martin(Smith609 – Talk) 22:01, 21 January 2012 (UTC)
[63][64][65]. These are examples of a need for human intervention, which has not happened (yet), thus my point of identifying the bot user.
Not sure what you mean by 'bot user'; the bot runs automatically. In these cases the bot's edits did not alter what was displayed on the page. Martin(Smith609 – Talk) 22:01, 21 January 2012 (UTC)
A while ago we had a discussion about when ISSNs should be added and when they should be removed. If this information was added to the Cite Journal documentation, it has since been removed. The bot will accord with whatever the consensus is. Martin(Smith609 – Talk) 22:01, 21 January 2012 (UTC)
[68]. This is perhaps my personal issue with the bot: it adds a behemoth '''|postscript=<!-- Bot inserted parameter. Either remove it; or change its value to "." for the cite to end in a ".", as necessary. -->{{inconsistent citations}}''' merely to say that a full stop is missing or excessive. Materialscientist (talk) 05:18, 10 January 2012 (UTC)
Again, this solution was the result of protracted discussion on these pages. Feel free to resurrect this discussion if the solution is not to your taste. Martin(Smith609 – Talk) 22:01, 21 January 2012 (UTC)
Is this block not a bit harsh to initiate a discussion? Maybe I missed it, but did you try to discuss this first with the bot operator whether it would be possible to identify the editor requesting the edits, Materialscientist? And as far as I can see, most edits are minor, not 'null'-type. What do you think of as a 'minor' edit, is that not personal taste? --Dirk BeetstraTC 05:49, 10 January 2012 (UTC)
Yes, admittedly harsh, and admittedly biased (against some implementations in the bot algorithm). The trigger was tweaking the titles, which is apparently a minor bug needing a fix. The bot is very useful, but also makes controversial changes; it often tries to fix semi-broken code where human intervention is required. Given its high rate of edits, I do not believe the bot can operate autonomously at this stage. Materialscientist (talk) 05:59, 10 January 2012 (UTC)
I think that if you check PMID11830664 and PMID17032458 you will see that the removal of the italics in the titles is based on the source, not a problem with the bot itself. You may prefer the CrossRef version to the PubMed version, but that's hardly an unambiguous or irreversible choice. If we at some point think that is the right preference, another bot run past the article would restore the italicised (CrossRef) version. We already ask this bot to do some very subtle things, let's not make it more failure prone by fussing over trivia such as watchlist spamming. Most of us see far less valuable "contributions" from human editors without batting an eyelash. LeadSongDogcome howl! 06:48, 10 January 2012 (UTC)
There are a range of items here, from acceptable minor edits adding a correct |issue=, to errors in ISSN formatting, to a long-winded comment over a full stop (correct, but can the message be shorter?), to errors in title re-formatting. Which editor initiated the edits hardly seems relevant, these are bot code problems or improvements. I think it would be sensible that the bot temporarily disable all changes to casing/italics of titles, and a later discussion establishes what changes, if any, the bot could accurately make to titles. Can we establish a list of 'blocker' problems that need to be addressed for an unblock:
Changes to italics or casing of titles (also Wikiproject medicine concern)
Loss of |id= ISSN xxxxxxxx
use of |last1= etc. when the rest of the article uses |author= (also Wikiproject medicine concern)
would be three for me that would address a lot of the oustanding concerns. Rjwilmsi 08:45, 10 January 2012 (UTC)
Per Rjwilmsi I'll sum up what I am not happy about (I am the blocker, but surely my suggestions need consensus and discussion.
Changing article titles: somewhat critical, not per changes themselves but per algorithm. In the examples above, the bot apparently took secondary source, ignoring the titles which were provided by a human editor and were correct per original source.
Done
Removal of id=ISSN: not crucial to me per se, but puzzling, as in some (very recent) edits the bot actually added ISSNs.
Author field: I am neutral to the use of "last"/"author" but strongly support reversal of the code to the point of not touching author fields truncated with et al. This is how the bot used to operate in a recent past, and too many articles still keep the author lists with et al. (not only medical convention). What happens now is the automatic operation of Citation bot 1 triggers changes to the article even if the only change is to expand et al. lists.
Please explain how the bot modifies the parsed output in these cases (save the addition of metadata). Martin(Smith609 – Talk) 22:07, 21 January 2012 (UTC)
Saving minor changes. The user can not predict the changes the bot will bring to the article, thus I propose the bot does not save changes which are "too minor". Before defining "too minor" I suggest the following: the bot operates in a mode similar to the interactive reflinks mode: i.e. the user manually saves the changes proposed by the bot (and thus takes responsibility :-).
Last point: I've blocked Citation bot 1 (talk·contribs) which was running automatically, at a significantly high rate. Citation bot (talk·contribs) is not blocked, and the crucial difference Citation bot is manually activated. This leaves username and hope that the user checked the bot edits. Materialscientist (talk) 09:23, 10 January 2012 (UTC)
Article titles maybe should be left alone, or stripped of all formatting and then checked whether they are correct (and then still). Only fill in when empty?
ISSN one is indeed a bit puzzling .. but should the ISSN not be in an ISSN-field, and maybe id is deprecated?
I don't mind et al.-lists, but I would prefer that the bot expands them, and IMHO preferably into forms where the names of the authors are (were available) the same as the on-wiki pagenames (or redirects thereto). Even if that is the only change to the article, I prefer the full-credit and since Wikipedia is turning more into a research tool, at some point we may want to be able to have a neat trick so we can search for all articles which have articles by Dr. J. Doe cited (I know, some get excessively wrong, but then the template should impose 'et al.').
There are many users who will also save those minor changes. I think that would be detriment to the function of the bot, is adding only one pagenumber a minor edit .. it may be just adding a '1' to one empty parameter??
Let me get this clear, Citation bot is the one that is user-operated, and Citation bot 1 is doing essentially the same, but 'chooses its own articles'/works from a list?
Still I don't think this is worth blocking over. Was discussion with the bot-owner about these issues tried before blocking? --Dirk BeetstraTC 10:12, 10 January 2012 (UTC)
The two cases of removed ISSN mentioned (above numbered as 36 and 37) were of ISSN1524-7848 and ISSN0820-4217. In both these cases, the serial title found on WorldCat varies from that provided by the human editor. The short titles provided by the human editors were ambiguous as well. Perhaps a better result would have been the application of a tag such as {{dubious}} to attract human attention, but removing the conflicting ISSN seems pretty reasonable to me in that circumstance. LeadSongDogcome howl! 14:20, 10 January 2012 (UTC)
The |id= field is not deprecated, but it is preferred not to use this when another param is designed specifically for the task. {{citation}} and {{cite journal}} both have an |issn= parameter, as do {{cite book}}, {{cite encyclopedia}} and others, although it's not documented on all of them. Where such parameter exists, the use of |issn=0123-4567 is preferable to |id={{ISSN|0123-4567}}. --Redrose64 (talk) 14:26, 10 January 2012 (UTC)
Agreed. The template /doc needs some cleanup on this to clarify all the identifiers that are directly supported vice the ones that need to go via the |id= parameter. I discovered this morning that id={{OL|123456M}} can be replaced by |ol=123456M, and there are likely others not documented.LeadSongDogcome howl! 16:32, 10 January 2012 (UTC)
I have been working this. The Citation Style 1 template documentation pages vary from minor to major points. I created {{Citation Style documentation}} to create the doc pages in consistent chunks. I have more work to do before I want to deploy it, mainly in making it a bit easier to edit. ---— Gadget850 (Ed)talk 18:37, 11 January 2012 (UTC)
With regard to the et al., the bot can hardly be expected to implement a stylistic choice that hasn't been made. If editors care about this, the best approach is to use the |display-authors=3 (or whatever value they like), so the metadata can attribute up to ten authors without the template imposing them all against editors' stylistic choice. LeadSongDogcome howl! 17:08, 10 January 2012 (UTC)
There were voices for and against et al. expansion, but consider this: in the past, the user could control this option by adding or removing et al. Now the bot will "ignore" (in the code, the article display is kept) user input in this field. As a result, some scientific article become really hard to edit because their code is swamped by ref. code. My proposal to the bot is simply to be more flexible. Materialscientist (talk) 23:01, 10 January 2012 (UTC)
Re issn, I believe that this information is almost always redundant and useless for publications for which more accurate identifiers (doi etc) are available. Issn only tells you what journal it is, doi tells you the specific article. But probably actual removal of issns is the sort of editorial decision that should be done by humans not bots. —David Eppstein (talk) 17:12, 10 January 2012 (UTC)
(butting in). Fully agree on redundancy of issn when there is doi/pmid/jstor/or alike, but, there are many legitimate links which don't have those and where the journal is hard to locate, especially good old, yet discontinued journals. Materialscientist (talk) 03:11, 14 January 2012 (UTC)
So how would you suggest the bot should handle the specific mismatches in the examples above, David? Flag them for human action? LeadSongDogcome howl! 17:16, 10 January 2012 (UTC)
You mean like here? Why do you think it's necessary that the bot do anything in this case? —David Eppstein (talk) 18:41, 11 January 2012 (UTC)
That's a good example. I doubt the bot would correctly have processed that malformed input, but it should have gotten attention long ago:
{{ cite journal | author=Zalivako, Anke | title=A Critique of the Preservation of Moscow's Planetarium | url=http://muse.jhu.edu/journals/future_anterior/v005/5.1.zalivako.pdf | journal=Future Anterior | year=2008 | volume=5 | issue=1 | id=ISSN 1549-7915 (print), ISSN 1934-6026 (online) }}
rendered
Zalivako, Anke (2008). "A Critique of the Preservation of Moscow's Planetarium"(PDF). Future Anterior. 5 (1). ISSN 1549-7915 (print), ISSN 1934-6026 (online).
Should probably have been just:
{{ cite journal | author=Zalivako, Anke | title=A Critique of the Preservation of Moscow's Planetarium | url=http://muse.jhu.edu/journals/future_anterior/v005/5.1.zalivako.pdf | journal=Future Anterior | year=2008 | volume=5 | issue=1 | issn=1934-6026 }}
rendered
Zalivako, Anke (2008). "A Critique of the Preservation of Moscow's Planetarium"(PDF). Future Anterior. 5 (1). ISSN1934-6026.
(this is a bit troublesome as the p-ISSN isn't found on worldcat while the e-ISSN is, both are valid, and the template doesn't take an |e-issn= parameter, hence human intervention would clearly have been better) LeadSongDogcome howl! 14:59, 13 January 2012 (UTC)
Action from maintainer
Is there anything outstanding that requires my attention before the bot is resumed? Martin(Smith609 – Talk) 22:10, 21 January 2012 (UTC)
Thank you for replies and fixes. Sorry if some of the issues were discussed/decided before (I understand you follow the developed consensus), but I would rather launch (new) straw polls. Materialscientist (talk) 23:56, 21 January 2012 (UTC)
I note that these ASINs are identical to the ISBN-10s of the books concerned; in the ISBN-10 system, X is a valid character in position 10. So, instead of using |asin=159116401X and |asin=159116415X, use |isbn=159116401X and |isbn=159116415X - by using ISBN, library catalogues may be linked as well as Amazon, rather than directing people to one specific supplier. --Redrose64 (talk) 18:56, 24 January 2012 (UTC)
Redrose, ISBNs should not be used as a substitute for ASINs in references because it is the Amazon.com page being referenced not the book. If you want to continue this discussion do so at Talk:List of Neon Genesis Evangelion chapters and leave a {{Please see}} on my talk page. – Allen4names 05:57, 27 January 2012 (UTC)
The bot has joined the Democrats :-D. Just a hunch, the replication lag is high for the past few days, i.e. some databases are not catching up. Materialscientist (talk) 22:51, 26 January 2012 (UTC)
Looks like two independent issues. The first one (Immelt) looks like a race condition, which we've seen before. The bot should be noting the pre-run version number. After saving a new version, it should check the history. If there has been an intervening version saved by another user, it should self-undo, delay a few minutes, then run again fresh. The second one's simpler. Empty refs should simply be deleted, not combined.LeadSongDogcome howl! 15:27, 27 January 2012 (UTC)
In the two instances shown in the diff, the bot mistakenly interprets the entries as journal articles instead of the books (which they are), and starts modifying the entries erroneously.
I can't blame the bot for not fixing that correctly, but I don't think it should be finding an ISBN there. The check on line 1190 of Citation bot's DOItools.php seems to assume that any 10- or 13-digit number is an ISBN, which is risky. However, I think the bug that led to this problem is in the first part of the regex there, which has (?!< instead of the correct <code(?<! for the lookbehind. Ucucha (talk) 23:41, 14 February 2012 (UTC)
The presumed ISBN isn't even valid. It seems obvious that the bot ought to verify them before adding them. Looks like two distinct problems to be fixed.LeadSongDogcome howl! 02:23, 15 February 2012 (UTC)
Finding duplicate "cite doi" templates in an article
Is there a tool or another way to find duplicate uses of the same "cite doi" reference in an article? Preferably a particular "cite doi" template should only be used once in an article and other references to the same source should use "ref name" and link to only one "cite doi" template. Acadēmica Orientālis (talk) 11:33, 19 February 2012 (UTC)
Poll2
The bot is altered to remove ISSN if a solid link (doi/pmid/pmc/jstor) is present and not remove ISSN otherwise. (Correct me if I missed a "solid link" type.)
SupportRDBrown (talk) 03:29, 24 January 2012 (UTC)
Neutral Additional 'solid' links would include arxiv/bibcode/jfm/mr/osti/ssrn/zbl Headbomb {talk / contribs / physics / books} 15:54, 28 February 2012 (UTC)
Poll3
The bot is altered to leave the author field untouched if it contains "et al". (Just a matter of flexibility - an editor can always remove et al and rerun the bot.)
Support with the exceptions of |last1= and |first1= (or equivalent) which it should certainly fill if empty. Headbomb {talk / contribs / physics / books} 15:52, 28 February 2012 (UTC)
I'm attempting to cite this paper on JSTOR in the article Coherentism. I used {{Cite jstor}} but when I hit "jump the queue", it cannot actually retrieve the metadata from JSTOR.
Well, following that link, the bot says that JSTOR didn't find the doi. The whole question of how those things should be encoded still seems buggy or badly explained. When should it be percent-encoded, when dot-encoded, when just plain? Further, the JSTOR doi should not be a substitute for the original publisher's doi, it should only be used when there is no publisher's doi available. In this case Springer has a munged PII-as-DOI that should be turned up on Crossref: doi:10.1007/s11229-006-9062-8
I didn't think there were supposed to be papers with both a DOI resolving to JSTOR and a publisher DOI going elsewhere (obviously paper with JSTOR ID and publisher DOI fine)? Rjwilmsi 18:35, 28 February 2012 (UTC)
Right. The problem may be that the publisher DOI was not found, so the JSTOR DOI was used. Really though, what is the value of using the JSTOR DOI vice the JSTOR ID and corresponding url? So far as I can see, it just confuses matters. We should just leave the DOI for the publisher. Let the reader decide if xhe wants one or the other. LeadSongDogcome howl! 20:12, 28 February 2012 (UTC)
There are many papers with both a JSTOR doi and a publisher doi, since in many cases JSTOR indexes journals for which recent issues are also available from the publisher. As LeadSongDog says, it makes sense in that case to have the JSTOR id and the publisher doi—the JSTOR doi doesn't add anything. Ucucha (talk) 21:24, 28 February 2012 (UTC)
As a humanities guy, I tend to not worry about DOIs. In philosophy, the vast majority of sources are book sources, and the papers tend to just supplement the books. Given this, I don't really worry too much about the DOI, and just use JSTOR URLs for my private research. I use {{Cite jstor}} because I can simply find the paper on JSTOR, copy the segment of the URL with the JSTOR ID and then paste it in. —Tom Morris (talk) 22:15, 28 February 2012 (UTC)
This is a late report on an edit last October. Is the issue simply the change from spelled-out firstnames to initials? The current version still does that. LeadSongDogcome howl! 04:55, 4 March 2012 (UTC)
Don't really see what's the problem with that link of yours LeadSongDog. All the bot did was remove unused and useless parameters (bunch of empty |author#=). Headbomb {talk / contribs / physics / books} 05:34, 4 March 2012 (UTC)
Perhaps it was this earlier edit by Citationbot3 in 2009 that was the issue? I don't really see it as a problem either. After all, the full names are still available from crossref if they are wanted. LeadSongDogcome howl! 06:04, 4 March 2012 (UTC)
Sorry, maybe I wasn't clear enough - the bot appears to have made a number of successive edits over the past few months, getting the first and last names of five authors mixed up and at one point splitting up the names into 6 sets of first = and last =. Need to review all the change log for all the edits to really see what's going on. See [70], [71], [72], [73] The original list of five authors, manually entered by G716, was changed by the bot to a list of 6, incorrect, authors Decstop (talk) 06:23, 4 March 2012 (UTC)
Now I see. In this edit back in 2009 the bot tried to parse "author = Paul Cowpertwait, Graeme Wake, Robert D. Anderson, Howard Edwards, and Shayle Searle", but got it very wrong, confusing first and last names. The later edits eventually fixed it, so that the current rev displays as "Cowpertwait, P.; Wake, G.; Anderson, R. D.; Edwards, H.; Searle, S."
All the same, this is yet another example of why hiding content in unwatched template subpages is a bad idea. LeadSongDogcome howl! 07:00, 4 March 2012 (UTC)
Poll1
The bot is altered not to add "|postscript=<!-- Bot inserted parameter. Either remove it; or change its value to "." for the cite to end in a ".", as necessary. -->{{inconsistent citations}}" to citation templates. (Better solution is welcome, but as a separate thread).
Oppose unless meaningful reasons to deviate from existing consensus are demonstrated. Martin(Smith609 – Talk) 02:30, 24 January 2012 (UTC)
I'd wish for a less horrible message, but that is no reason to block the bot over. Personally, I would prefer for the bot to determine whether to add or remove that ".", and then add or remove it. Headbomb {talk / contribs / physics / books} 19:18, 8 March 2012 (UTC)
IMO, Citation Bot really should check for the input matching '10.Foobar' before doing anything else. Headbomb {talk / contribs / physics / books} 06:21, 4 March 2012 (UTC)
Cite|PMID templates not being filled in
Hey there, I noticed that a number of {{Cite|PMID}} templates were not being filled in on some gene pages, such as this one, which was created in January. I've pushed a few of them manually (through the 'jump the queue' button), but I was wondering if automatic completion would be restored at some point. Thanks for the awesome utility- it's a huge help! Pleiotrope (talk) 17:42, 13 March 2012 (UTC)
{{cite doi}}, {{cite pmid}} etc are nice ways of adding references and I see the need for a {{cite ISBN}}/{{cite isbn}}. Books are frequently cited in articles and most new books and most books published in the past 20-30 years have an ISBN. Adding book citations by hand is tedious and some of the citation tools are sporadic in operation or are a little "fiddly". I would like to have a {{cite ISBN}} template created and then a bot goes off to find the biblio data to build a reference along the same lines as what happens with {{cite doi}}. There are concerns about availability of ISBN data in the quantities that would be used by WP, however WorldCat allows up to 1000 ISBN queries per day with other options available. That may be sufficient for WP needs. The citation templates would then be placed in a Category:Cite ISBN templates similar to Category:Cite doi templates, Category:Cite pmid templates etc.
In the meanwhile, there's nothing wrong with {{cite book}}, which will get pretty much the same result without creating yet another unwatched template subpage. Initially (before bot improvement populates it) that would look like:
Hi, i just created the template {{cite isbn}} for the purpose of synchronization of citations on different pages. I suppose i could add the openlibrary link to the fill-out form. Although not perfect, this would ease the creation of new references. BR84 (talk) 18:20, 18 March 2012 (UTC)
I've seen this happen millions of time, it always annoyed me. Can't really say why I didn't bother to report it until now. Headbomb {talk / contribs / physics / books} 20:23, 13 March 2012 (UTC)
Similar things seem to happen for the bibcode database. It uses |author1= instead of |last1=, but gets all |firstn= and all other |lastn= correctly. [78]. Headbomb {talk / contribs / physics / books} 19:36, 24 March 2012 (UTC)
This is one of those weird cases where content is in PMC but not Pubmed. A check finds no pmid and presumably returns 0 rather than null. Clearly the bot should not use this as output unless the template expects 0 as a special case.LeadSongDogcome howl! 21:29, 28 March 2012 (UTC)
People seem to now be pasting DOI's like this alot: {{cite journal|doi= doi:10.1017/S175569100900704X}}
We can't proceed until
Agreement on the best solution
Requested action from maintainer
Clean up these DOI's automatically (ie remove the doi:)
Unusual date of publication causes problems
Had to edit Template:Cite pmid/20301430 by hand to correct publication date. PubMed says "1993-2006 Jul 31 [updated 2010 Nov 02]". -- Richiez (talk) 22:39, 9 April 2012 (UTC)
Deleterious: Human-input data is deleted or articles are otherwise significantly affected. Many bot edits require undoing.
What happens
The ketogenic diet article has footnotes (notes) and reference citations. Some of the footnotes have reference citations. These were removed by the bot.
What should happen
It should leave citations within footnotes alone. Since it probably can't tell the difference between a reference footnote and a note footnote, this "feature" should probably be disabled.
Hi, wondering about Special:Contributions/KBrainbridge. Either this are all good faith edits then hopefully the bot could do them or something strange is going on? 21:13, 8 February 2012 (UTC)
Looks like the journal has moved to a new publisher. This demonstrates the advantage of DOIs over URLs. My opinion is that using the bot to edit URLs risks overriding deliberate user intention, so wouldn't be a good idea. Martin(Smith609 – Talk) 13:21, 29 April 2012 (UTC)
Improvement: The bot would be much better if the bot would also check for existing merges like in that edit amde by me after the bot was activated by me.
What happens
missed to combine a ref
Relevant diffs/links
see above
Replication instructions
see above
We can't proceed until
Agreement on the best solution
Requested action from maintainer
check for bare urls and existing refs using citation templates/non linkrot urls and combine them
on pages like http://en.wikipedia.org/wiki/Sleep_deprivation when you click jump the queue on PMC 353042, the bot runs and then complains that the cite PMC is a redirect so it does not edit. But, the redirect is to Template:Cite pmid/ with no number at all. Try it: {{Cite pmc | 353042 }}
We can't proceed until
Agreement on the best solution
Requested action from maintainer
fix it??
I'd suggest avoiding {{Cite pmc}}, it's pretty much an unused orphan. If all you have is a pmc, try {{Cite journal}} with parameter |pmc= instead. Will propose deleting.LeadSongDogcome howl! 18:47, 10 April 2012 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
-->
What happens
The bot seems to be repeatedly creating redirects to nonexistent templates/pages. I tagged several of these for speedy yesterday and it looks like there are now more.
Catastrophical: The bot stopped editing immediately
What happens
The bot does not initialize
Replication instructions
"The page you requested is hosted by the Toolserver user verisimilus, whose account has expired. Toolserver user accounts are automatically expired if the user is inactive for over six months. To prevent stale pages remaining accessible, we automatically block requests to expired content...If you think you are receiving this page in error, or you have a question, please contact the owner of this document: verisimilus [at] toolserver [dot] org. (Please do not contact Toolserver administrators about this problem, as we cannot fix it—only the Toolserver account owner may renew their account.)"
We can't proceed until
Operator: Bot operator's feedback on what is feasible
The account must be renewed before the bot will be operable again. JJJ (talk) 01:05, 19 April 2012 (UTC)
this is a recurring problem. someone should find a way to keep the account from becoming inactive or apply for an exception to this rule. —Chris CapocciaT⁄C 21:24, 19 April 2012 (UTC)
All sorts of things stop working. I understand toolservers policy, but this account is not exactly a "here today, gone tomorrow" account. It would be nice if they emailed the person in charge a week before at least. That would keep us from singing the blues. AManWithNoPlan (talk) 23:20, 19 April 2012 (UTC)
I agree. This type of service should never expire. I have used this bot on thousands of edits. We should contact them about this. Or, does anyone know of another *nix toolsever hosting service? --Thorwald (talk) 05:12, 20 April 2012 (UTC)
The username on the toolserver account is versimilus, but that user now uses a different account on Wikipedia, Smith609. Perhaps because of the 'name-change', the versimilus toolserver account isn't checked as frequently. The Wikipedia user account redirects to the new account, but the toolserver account possibly has no way to 'redirect' to the new account and the name on the account (possibly) cannot be changed, so continuously checking this account registered to the old username could be a hassle. -- JJJ (talk) 07:23, 20 April 2012 (UTC)
Martin is aware of the issue. Realistically, it's simpler for him to just set a calendar reminder to renew rather than to change toolserver account policies.LeadSongDogcome howl! 14:59, 20 April 2012 (UTC)
As of now, the bot is still blocked. What's the estimated time to get this reinstated? --Mblumber (talk) 04:26, 22 April 2012 (UTC)
These things take time. Human time and computer time. Last time it took several days for Mr. Citation to contact them, and for them to act. So, I felt a great disturbance in the Wiki, as if a thousand citations cried out in horror and were suddenly silenced. :-) He knows and it will get done when it gets done. Nagging won't help. AManWithNoPlan (talk) 18:48, 22 April 2012 (UTC)
It was |unused_data=, not |unused_date=. None of the citation templates use positional parameters - they all use named parameters exclusively - so when the bot encounters information in a positional parameter it attempts to put it into the best named parameter. Anything that can't be so assigned ends up in |unused_data= for a manual fix later on. --Redrose64 (talk) 19:21, 4 July 2012 (UTC)
I assume the bot went wrong because the reference's name was either a short or a long version of another reference's name (ARana vs. ARanard and AR), but that wouldn't fully explain its behaviour. For now the bot has stopped removing the reference. Huon (talk) 03:39, 11 June 2012 (UTC)
It's odd to have multiple named references to the same book at different pages. Such cases are what {{sfn}} is for.LeadSongDogcome howl! 05:23, 11 June 2012 (UTC)
If the article evolved with references to different sources, and then one source is used twice with different pages, it would be understandable that the editor who added the twice-cited source didn't want to bother and go through the whole article and convert it to short footnotes. Jc3s5h (talk) 12:02, 11 June 2012 (UTC)
I agree the references could be simplified, but that's no reason for the bot to remove one of them. I expect the user who wrote the draft was not aware of {{sfn}}. Huon (talk) 14:37, 11 June 2012 (UTC)
A specific edit to the bot's code is requested below.
Requested action from maintainer
Citation bot to not use the deleted template any more.
The offending PMC value PMC2726758 has a corresponding PMID, so if possible, the Bot should do a PMID lookup and generate a cite PMID template. Some older PMC values may only have the PDF, without a matching PMID entry. RDBrown (talk) 08:31, 5 July 2012 (UTC)
Note that {{cite pmc}} template was deprecated for a long time and is now deleted: Wikipedia:Templates_for_discussion/Log/2012_May_21#Template:Cite_pmc. See also earlier entries on this page. Most PMC articles have a DOI and are also listed on PubMed, so you are able to use {{cite doi}} and {{cite pmid}} templates. kashmiri 09:46, 16 August 2012 (UTC)
The template seems to have been deleted by someone who didn't understand how it worked. I've restored it. Martin(Smith609 – Talk) 07:35, 18 August 2012 (UTC)
Examples: Attention: This template ({{cite jstor}}) is deprecated. To cite the publication identified by jstor:1305363, please use {{cite journal}} with |jstor=1305363 instead., Attention: This template ({{cite jstor}}) is deprecated. To cite the publication identified by jstor:1303024, please use {{cite journal}} with |jstor=1303024 instead.
We can't proceed until
Bot operator's feedback on what is feasible
Requested action from maintainer
fix it, or explain the the bug is with jstor, and not the bot itslef
jstor refs also don't get expanded if the doi is used instead of the jstor parameter. —Chris CapocciaT⁄C 13:48, 18 June 2012 (UTC)
Must be a problem with the JSTOR API. I'll investigate. Martin(Smith609 – Talk) 08:19, 17 August 2012 (UTC)
JSTOR's API seems to be down. I'm following this up with JSTOR. Martin(Smith609 – Talk) 07:46, 18 August 2012 (UTC)
Here's the response from JSTOR:
Thanks for your message. The JSTOR API has been taken down for the time being. There are currently no plans for when this interface may be available again. Apologies for any inconvenience caused here.
This seems like the place to bring up this question, but if not I'd appreciate someone directing me to where would be a good spot. A {{cite isbn}} template was created not that long ago but it seems to lack the automated process that {{cite doi}} has (I think that is this bot). Is there a way to get this template added to the automated process? I'd be willing to help work on it, but I have no experience with bots so would at the very least greatly appreciate guidance. Zfeinst (talk) 16:24, 30 August 2012 (UTC)
Please don't go there. The template {{cite isbn|0836221192}} (which renders: {{cite isbn|0836221192}} ) doesn't add anything to {{cite book|isbn=0836221192}} which renders as . ISBN0836221192. {{cite book}}: Missing or empty |title= (help), and introduces all the same problems that {{cite doi}} et al. bring. The subpages, isolated from the articles which transclude them, do not adapt to the citation format in the article. LeadSongDogcome howl! 16:52, 30 August 2012 (UTC)
Zfeinst:I don't know why yoou added this to the bot-filled templates, as it is not bot-filled.
LeadSongDog: The template documentation is wrong and the template is rather borked. When you use {{cite isbn|0836221192}}, it shows as {{cite isbn|0836221192}} with the edit link because the subtemplate does not exist until you edit it. If you do use an existing subtemplate, then {{cite isbn|978067144133}} shows as {{cite isbn|978067144133}} because the template requires a /, thus {{cite isbn/978067144133}} shows as {{cite isbn/978067144133}}. But if you use a slash in the non-existing template {{cite isbn/0836221192}}, it shows as {{cite isbn/0836221192}}. ---— Gadget850 (Ed)talk 17:30, 30 August 2012 (UTC)
At the time I added it I thought it was bot-filled, have just undid my edit there. My bad, thanks for pointing that out.Zfeinst (talk) 17:36, 30 August 2012 (UTC)
I have made minimal changes to enhance the usefulness of this: optional page, pages and ref parameters are now passed to the subpage. The first two existing subpages and the preload definition are updated to support these parameters. Please see the new testcases, and sandbox for testing any further updates.
We will also need a year suffix parameter for 1996b etc in articles. I am thinking of |ys= to define the "b" etc, since the value will nearly always be a single letter there doesn't seem much point in making the name long. Thinking further: the user has to know the year anyway to decide on the suffix, it will make the article source clearer to define the year-plus-suffix together, thus just passing through the year parameter will be better. Usage in the subpages would be, for example, |year={{{year|1996}}} . (updated Mirokado (talk) 09:40, 31 August 2012 (UTC))
The first subpage Template:Cite isbn/978006039436 illustrates some of the confusion which could arise if care is not taken with these templates: the subpage name has one isbn but the cite book definition uses a different one. We need to decide how to correct this: which isbn did the book being referred to have?
I think we should be providing hyphenated 13-digit isbns in the citation definition for MOS conformance (at least the 13-digits) and general readability. --Mirokado (talk) 23:32, 30 August 2012 (UTC)
I have finished "messing about" for a while now... The testcases page is working as expected and I'm happy with the support for four article-specific parameters:
|ref= and |year= needed particularly for articles using Harvard style (supporting year tidily required rather more extensive changes than I had at first expected)
|page= and |pages= for single citations with a page range.
Comments of course welcome... --Mirokado (talk) 17:40, 31 August 2012 (UTC)
While ISBN-13 should always be preferred to ISBN-10, it can always be computed from the ISBN-10, so it doesn't matter too much, so long as what is given is correct. The hyphenation is not based on a fixed pattern. Different ISBNs are hyphenated differently, so the templates and the bot should not fidget with hyphenation unless it is looked up from an authoritative reference.
I hate to harp on this, but again, it is better not to use these subpage templates at all. They cause no end of problems, even if they work as intended. They can't adapt to {{usedmy}} or {{usemdy}}. Indeed, they don't adapt to any article-specific formatting conventions. The bot should replace any transclusion instances in articlespace with equivalent {{citation}} or {{cite book}} (as appropriate) once the data has been populated. Then article editors can tweak format as necessary without it spilling over into other articles' formats. LeadSongDogcome howl! 19:11, 31 August 2012 (UTC)
Thanks for the comments. Certainly there will be articles for which the subpage templates are not suitable. The ones generating {{cite journal}} work quite well (for articles using cs1 style) as they rarely contain dmy/mdy/iso dates which would not suit all articles, a doi or pmid identifies a unique object, and it is at least arguable that a particular name, initials format for journal article authors can coexist with something else for web pages, books etc. The book citations I have seen also overwhelmingly contain just year so date formats should not be an insuperable problem.
More problematic is the fact that an isbn does not uniquely identify an object and its page ranges. I am thinking about that issue, which of course almost certainly affects some articles for which different editors have referred to different editions of a book without adding separate citation definitions. This also means that, I agree, a bot is not really suitable except perhaps for post facto tidying up.
I would quite like to see how far we can get with this template but I will not be using it at all on real articles until the user interface is satisfactory and reasonably stable. --Mirokado (talk) 22:53, 31 August 2012 (UTC)
Issues which need to be sorted out include:
isbns do not determine a unique object
the #ifexists function is "expensive", I don't think we can have it used once the subpage has been created
Something that always worries me about subpage citations is that they're very unlikely to be on anybody's watchlists, hence more susceptible to vandalism than usual. —David Eppstein (talk) 01:08, 1 September 2012 (UTC)
I'm not sure what you mean by "isbns do not determine a unique object". Do you have examples? Perhaps you are thinking of books in a series, which have both the series isbn and the volume isbn, but in such cases each isbn has its own "object".
I agree that for books the dateformat is less significant, but perhaps I wasn't clear enough about the real problem. The editors of all the different articles that transclude the templates are entitled to choose how reformat them. Imagine having an article going through FAR, only to have it fail because suddenly the rendered citations change without any record in the article history, when editors on an unrelated stub wanted to use a different format. The only ways to prevent that are to not use the subpage templates, or to make them unique to each article using them (which defeats the purpose). LeadSongDogcome howl! 01:26, 1 September 2012 (UTC)
Reuse? Wouldn't that defeat the purpose? It certainly violates the standard, not to mention making life difficult for librarians (the people who really matter). LeadSongDogcome howl! 02:58, 5 September 2012 (UTC)
Reuse is actually encouraged by some organizations to save money (we have to deal with what people do, not what they should do). Also, different editions often have the same ISBN and different printings with the exact same information end up with different ISBN. It is not a single unique identifier. For example see warning from : http://www.fictiondb.com/faq.htm Hardcover vs. Softcover. It is best to just use cite book. AManWithNoPlan (talk) 03:45, 7 September 2012 (UTC)
Just to emphasize the risks noted above, there is no guarantee that numerous ISBN numbers have not been reused for newer books or a specific book have other edition numbers not matching the {cite_isbn|<ISBN>} on file, confusing people who find a different title/number. Meanwhile, we want to leave each Cite_isbn/xx subpage unprotected to allow others to expand detail and fix typos. However, the use of {Cite_isbn} with unprotected subpages would quickly enter the frequently read pop-culture articles, which do cite from books, and famous books would more likely use {Cite_isbn} entries. Note those articles rarely cite from DOI numbers, which are more common in medical or other journal-related articles. Plus, I can confirm the recent horrors of unprotected templates, even outside of pop-culture article vandalism, where some well-meaning admins unprotected many string-handling templates a year ago, and within months, almost all were vandalized/hacked by IP-address users, then re-protected. It took a while to find where the vandalism was hidden, and with {Cite_isbn}, there would be strong temptation to rename any book as "Celebrity Xxx drug abuse with underage students". The reason vandalism remains for weeks or months is because it is 50x times easier to vandalize than to find and correct. I dred the protected subtemplates of Template:Taxobox or others when attempting improvements, but they must be protected due to extreme risk of hacking, and if {Cite_isbn} subpages were to get automatic protection, then that would thwart the promise of "fix one place" for improvement everywhere. Instead, the reality is most likely, "vandalize one place for embarrassing insults" everywhere. Hence, it is better to repeat a {cite_book} in 20 celebrity articles, rather than risk hideous {cite_isbn} vandalism to attack 20 semi-protected pages, where 99.99% of readers, for days/weeks, would search and not know if or where vandalism had occurred, with drugs for "underage students" (re: History of Facebook). Since there is a critical need to protect string-utility templates from registered editors, then imagine the risk to {cite_isbn} entries in pop-culture articles. Hence, for those reasons, I agree with the fears expressed above, and advise the new {cite_isbn} should be kept limited in usage. -Wikid77 (talk) 15:38, 10 September 2012 (UTC)
The bot (or person) should convert {{cite isbn}} to {{cite book}} and then delete the template. {{cite isbn}} must be destroyed AManWithNoPlan (talk) 01:27, 11 September 2012 (UTC)
Add parameters sep and fmter to select formatter
To solve the problem of incompatible citation format, perhaps alongside some hand-coded citations, then the Template:Cite_isbn and bot (if any) should handle new parameters "fmter=cite book" and "sep=," to optionally put commas as the separator between phrases in the citations. With parameter "fmter=cite book" then the actual stored format would be "{{ {{{fmter|cite book}}} |isbn=1112223334}}" where the name of the cite-template (fmter) could be passed as some other template than {cite_book}. The most obvious small format difference is likely to be separator "," rather than dot "." and the short name "sep" avoids the spelling glitch "seperator" with syllable "-er". Although unusual, the unique parameter name "fmter" would also allow wikisearch to track the use within all prior articles as usage expands (I had imagined calling fmter as "citer" but that word is very common from French sources, already matching 95,000 articles). -Wikid77 (talk) 13:34, 10 September 2012 (UTC)
Stop the bot from reverting good edits
Status
{{wontfix}}; problem is with the inflexible cite doi template
It is not clear this was an intended edit. It looks more like a race condition. Attempting to reproduce, it seems even stranger. Perhaps the comment is confusing the bot? LeadSongDogcome howl! 15:35, 18 October 2012 (UTC)
I keep reverting its reverts such as this latest one. Fix this or I'll have to block the bot. Urhixidur (talk) 16:00, 18 October 2012 (UTC)
This one on the same article is a better example. Clearly not a race as it happens nearly an hour after the previous edit. —David Eppstein (talk) 19:56, 18 October 2012 (UTC)
Weird bug. Seems to be triggered by an edit of a {{Cite pmid}} or {{Cite doi}} sub-template, giving the contributor the impression of being stalked by the bot. Once reverted, it seems to keep quiet. Urhixidur (talk) 21:00, 18 October 2012 (UTC)
Note that the bot will edit citations so they comply with Template:Cite_doi#Formatting. Doesn't explain the comment handling but does explain the abbreviation of author forenames. Martin(Smith609 – Talk) 08:42, 19 October 2012 (UTC)
Template:Cite_doi#Formatting is incredibly vague as far as author formatting goes. The Manual of Style does not mandate abbreviations, and that entry in the template's documentation seems pointless: simply require doi and pmd entries to specify the first/last parameters of the citation templates, and let them handle it. Urhixidur (talk) 13:11, 19 October 2012 (UTC)
this edit by the bot shows that it does not force abbreviations onto citations. Not systematically, anyway. Urhixidur (talk) 13:15, 19 October 2012 (UTC)
I am having a similar problem with my edits being reverted by the bot. See for example this. Boghog (talk) 21:27, 19 October 2012 (UTC)
A cite doi template is designed to be used on multiple pages. It is important to some editors that references within an article are formatted consistently. For this reason it is necessary to be able to predict how a cite doi template will be formatted. If it is important to you that an author's initial is not followed by a period, then I suggest that you either generate consensus for this change to be made on all pages that use a cite doi template, or do not use the cite doi template on articles where the additional period is deleterious. Martin(Smith609 – Talk) 08:25, 20 October 2012 (UTC)
I agree that it is important to have consistently formatted references within an article. There are cases where an individual {{cite doi}} template is transcluded into a set of articles that are otherwise consistently formatted in a different style. Where is the consensus that all cite doi templates must use the same style? This would seem to directly contradict WP:CITEVAR. This proposal might be a solution, but unfortunately it has not yet been implemented. In the mean time, I have replaced the cite doi template in the specific example mentioned above with in-line templates in the articles that tranclude this template. Boghog (talk) 09:12, 20 October 2012 (UTC)
The {{cite doi}} family of templates is pretty clear about initials only. Some fields insist on initials so that you don't know if the author is a girl or a boy (and to a lesser extend race, ethnicity, etc.). Of course one could counter that is it well documented that ones sex effects ones interpretation of data. If one really wants to control the formatting of a reference, don't use {{cite doi}}, use {{cite journal}}. AManWithNoPlan (talk) 19:58, 20 October 2012 (UTC)
The articles that I was referring to use the Vancouver system (initials without periods). The cite doi template documentation states that "this template need[s] to be formatted consistently so they match the reference formatting of all the articles they are used in"". What if all the articles that use a specific cite doi template are otherwise consistently formatted using another style? Of course, I have in the past and I will continue in the future substituting these templates where appropriate. Boghog (talk) 21:25, 20 October 2012 (UTC)
Just to be clear, I have never myself inserted a {{cite doi}} into an article. I am editing articles where Vancouver system has previously been established and someone else has inserted a cite doi template. Substituting templates, especially if the same template has been inserted into several articles becomes tedious. Boghog (talk) 06:14, 21 October 2012 (UTC)
Search for {{cite doi| and replace with {{cite journal|doi=, then let bot do the rest. Not too bad. Simple fact is you are cleaning up someone else's mess. AManWithNoPlan (talk) 14:01, 21 October 2012 (UTC)
Thanks for the suggestion. However I want to avoid introducing verbose "first1, last1, first2, last2, ..." parameters and instead use a single author parameter for the author list. Using the single author parameter reduces clutter and page load times. Unfortunately the bot will not do that but User:Diberri's Wikipedia template filling tool will. Boghog (talk) 15:16, 21 October 2012 (UTC)
If load times really matter to you, then you should not use {{cite doi}} because that adds an extra template between your page and the {{cite journal}} template. On a personal side note, I hate {{cite doi}},{{cite pmc}},{{cite isbn}},{{cite pmid}}, etc. because of this. AManWithNoPlan (talk) 15:32, 21 October 2012 (UTC)
The reasons I dislike these templates are different: (1) they force {{cite}} formatting, making them inconsistent whenever they're used in articles that have different formatting such as that of {{citation}}; (2) they are unlikely to be watchlisted, making it likely that vandalism to them will take much longer to detect. —David Eppstein (talk) 21:32, 21 October 2012 (UTC)
But {{cite}} is merely a redirect to {{citation}} so it doesn't matter which is used, the output is the same. --Redrose64 (talk) 22:22, 21 October 2012 (UTC)
I meant the "cite x" series of templates. Citation Style 1, to give them their proper name. I prefer Citation Style 2, and splitting off the citations into separate little templates gets in the way of that. —David Eppstein (talk) 00:04, 22 October 2012 (UTC)
It is pretty clear by now that the problem is in formatting. The cite doi templates should be used as data repositories (and hence be as complete as possible) but should not force a formatting over another. Hence their invocations should be accompanied by a format= parameter that would allow the invoker to choose his citation format (initials vs. full names, semi-colons vs. periods, etc.). This change is not a minor one, however. Urhixidur (talk) 13:08, 22 October 2012 (UTC)
We've known for a long time that cite doi and kin are fundamentaly flawed concepts, but their convenience means that editors can't resist using them. Data should not be in template space. Format choices should be local to the transcluding page. But this is not template talk:cite doi. Let's stick to discussions of the bot. LeadSongDogcome howl! 14:10, 22 October 2012 (UTC)
So, can we agree that there is no bug in the bot, and that the solution is to avoid the evil templates designed for lazy editor? :-) AManWithNoPlan (talk) 16:51, 22 October 2012 (UTC)
Well, except for the "evil" bit. They were definitely well-intentioned, but afaisi misconceived.LeadSongDogcome howl! 22:19, 22 October 2012 (UTC)
No, I do not agree that the bot is bug-less. Its description talks repeatedly of adding data (which it does, mostly), but the one thing I've seen it do repeatedly is remove data by stripping first names down to initials. I think it simply needs to be improved to recognize initials vs. full names for what they are. There also seems to be a problem with its extraction of ending page numbers, but that's clearly a different bug. Urhixidur (talk) 11:09, 23 October 2012 (UTC)
If you want to make a case for the templates needing distinct handling of initials vs forenames, that's a valid perspective, but this is not the place for pursuing it. What the bot is doing, though, is in keeping with its assigned task and with the documentation at template:cite doi/doc#Formatting. You always have the option of not using cite doi. If the subpage is only transcluded in one place, you could apply bot exclusion to the subpage, though I would not really want to encourage that: as it freezes the data, you might as well just subst the template into the article.LeadSongDogcome howl! 15:26, 23 October 2012 (UTC)
Removing wrong data is well within the rights of the bot. And full names (not just initials) is wrong data in the {{cite doi}} templates. Like we tell little kids "You get what you get, and you don't throw a fit". Use {{cite doi}} and all you get is initials. AManWithNoPlan (talk) 16:44, 24 October 2012 (UTC)
The problem is not with the data, but an inflexible template that does not conform to WP:CITEVAR. One can argue not use it or to substitute it, but because the template is so easy to use, it ends up being widely misused. Boghog (talk) 17:23, 24 October 2012 (UTC)
The problem isn't with what the data is, but where the data is. It does not belong in templatespace, but that is what has happened. Separation of code from data is a basic principle of robust software systems design. LeadSongDogcome howl! 17:39, 24 October 2012 (UTC)
I agree in theory but nevertheless there are situations where storing data in template space may be desirable. There have been a number of templates whose data has been entirely moved to template space. For example the data in ~10,000 {{GNF_Protein_box}} templates has smoothly been moved to template space without controversy. Because of random vandalism, there has also been an effort to validate "immutable" data in main space (see for example Chembox validation). Citation data should also be "immutable". Perhaps what is necessary is a new namespace reserved for immutable data. Boghog (talk) 18:56, 24 October 2012 (UTC)
Another potential solution may be Wikidata. Boghog (talk) 19:22, 24 October 2012 (UTC)
Well, meta:WikiScholar is part of meta:Wikicite, which in turn depends on Wikidata. These should eventually help, but it won't be anytime very soon. Citations just aren't sexy enough to get the needed priority. Infoboxes, though, they make the first round :-) LeadSongDogcome howl! 20:21, 24 October 2012 (UTC)
In the mean time, I have added a cautionary hatnote to the template documentation that will hopefully reduce the misuse of this template. The only caution previously in the documentation was buried in the middle of the text where many editors (including myself) probably would not notice it. Boghog (talk) 06:42, 25 October 2012 (UTC)
I have it selected for as a gadget. When that didnt respond, I have tried to use it by demoepedia and toolserver, and both failed multiple tries. It worked wonders for my edits of Matthew Lyon, but has seemed to not respond to my edits on the page for 1922 Vanderbilt football team.
What should happen
I would just like it to make it so one reference can have multiple instances in the page (such as when there are superscript a and b on the reference list) instead of 45 Ibids
We can't proceed until
Agreement on the best solution
Requested action from maintainer
I would just like to know what I am doing wrong if I am doing something wrong, or if not fixing whatever problems are on the other side. Thanks
I suggest you read Help:Footnotes and {{cite}}. This is a bug report page for the Citation bot. kashmiri 20:52, 22 October 2012 (UTC)
Citation bot deleting stuff at Comparison of birth control methods
Currently, nothing happens at all... the page is not edited, a bare link to a book inserted here remained unchanged even after running the 'bot using this link.
What should happen
The 'bot used to be able to find date and ISBN for a Google book reference and insert them into the reference on-page when activated by a user.
Could you spell out the specific references that you hope that the bot will expand? Thanks, Martin(Smith609 – Talk) 07:55, 18 August 2012 (UTC)
This is is working for me, and no updates in months. Tagging as resolved. Plus that page had many messed up references that I fixed, and the bot then worked on them. AManWithNoPlan (talk) 23:24, 27 October 2012 (UTC)
When the title of a journal is in caps (such as CFSK-DT) and is linked to the appropriate article, the bot undoes the caps, thereby breaking the link to the article about the journal.
What should happen
The bot should be configured such that it leaves linked journal titles alone if the target article has a title that is in all-caps.
Changing the code to prevent the bot from making similar edits in the future
You need to add the journal acronym to the list of capitalization exclusions; see the bot's user page for instructions. Martin(Smith609 – Talk) 07:57, 18 August 2012 (UTC)
Marking as resolved. I added it to the exclusions list. Interesting that person who reported the "bug" didn't fix it. AManWithNoPlan (talk) 23:27, 27 October 2012 (UTC)
This functionality cannot be added without consensus, reflected at Template:Cite isbn/doc, that this formatting choice should be enforced on all cite isbn templates (see example at cite doi). Even then, I'm not sure that I'll have time to code support for the Cite ISBN family of templates. Martin(Smith609 – Talk) 14:41, 27 October 2012 (UTC)
Improvement: The bot would be much better if ...
-->
What happens
I don't know of/think there are any bots that convert bare URL refs to cite web refs
What should happen
<ref>http://elvey.com/insecure</ref> should become <ref>{{cite title=blah, author = blah, date=fee, url="http://elvey.com/insecure", archiveurl="<the bot could trigger the Internet Archive and/or WebCite to archive the URL...>"... }}.
Why not have a bot do this? Has it been attempted before? I'd guess it has... I'm wondering why it isn't happening. --Elvey (talk) 01:08, 5 October 2012 (UTC)
Try using Reflinks for a while, and you'll see that there are quite a few parameters that don't get populated properly. GoingBatty (talk) 01:25, 5 October 2012 (UTC)
A fine idea but not one I have the time or ability to implement. Sorry! Martin(Smith609 – Talk) 08:27, 20 October 2012 (UTC)
The article, "I-696: Three Pedestrian Plazas Over Freeway" in the MDOT Context Sensitive Solutions Case Study: Metro Region by the Michigan Department of Transportation appears on page B1-17. The study uses chapter-based pagination, so that's the 17th page of chapter B1, or page B1-17. It isn't a page range, but twice now I've reverted the bot. I'm not sure what can be done to avoid false positives like that. One revert is understandable, but twice (yes, I know, user-requested and no one's looking for the reverts in the page history first) means that the bot has now been blocked from editing the page. Imzadi 1979→ 08:25, 25 October 2012 (UTC)
Easy to fix. use this as your dash character &#x 2012; (but with the space removed). AManWithNoPlan (talk) 03:03, 26 October 2012 (UTC)
Except that the page number doesn't, and shouldn't, have a dash. Blindly converting hyphens in page numbers to dashes just doesn't work though, because they aren't all wrong. Imzadi 1979→ 03:16, 26 October 2012 (UTC)
Then use the hyphen HTML code & #8209; in the rare cases that a hyphen is needed. 99% of the time a dash is right and a hyphen is wrong. AManWithNoPlan (talk) 13:30, 26 October 2012 (UTC)
No, no! The thread is not about "how to circumvent the bot" – but how to improve the bot. Based on the example above, I would suggest that if a page number consists of a letter followed by a number, then any hyphen between them should not be converted to a dash. That is, I am proposing a new rule for the bot. Any objections or examples to the contrary? kashmiri 15:03, 26 October 2012 (UTC)
Therefore, B1-B5 should be considered normal page number, but B1-17 should be considered a dash. AManWithNoPlan (talk) 23:18, 26 October 2012 (UTC)
A dash would be misleading as dashes normally denote page range, not a single page. My suggestion stands. kashmiri 14:15, 27 October 2012 (UTC)
What about cases where an editor provides "E310-15" to denote the range "E310-E315"? In this case, a hyphen would be incorrect. I would suggest that the use of &-codes as suggested above is the best way to make it clear that an editor's choice of dash glyph is intentional. Martin(Smith609 – Talk) 14:43, 27 October 2012 (UTC)
Bot added this tag to article: 'This article does not cite any references or sources' but there were plenty of references
The bot did not add the {{references|date=November 2012}} - it was already present in your first draft, so you must have put it there yourself.
In the version immediately prior to the bot edit, the {{references|date=November 2012}} is clearly present. This version has two instances of the following construct:
<ref name="BCC Musgrave Park">[http://www.belfastcity.gov.uk/parksandopenspaces/parks/musgravepark.asp Belfast City Council - Musgrave Park]</ref>
One of the first group may be simplified to <ref name="BCC Musgrave Park" />, similarly, three of the second group may be simplified to <ref name="Keep Britain Tidy" />, and that is what the bot did. It added no maintenance templates. --Redrose64 (talk) 11:10, 19 November 2012 (UTC)
Can you suggest an algorithm by which the bot can tell whether the data specified as an author is a person's name or not? I can't do anything otherwise. Martin(Smith609 – Talk) 07:37, 18 August 2012 (UTC)
If the bot wants to add |first1=Inc, it's probably not a person's name. GoingBatty (talk) 04:04, 22 August 2012 (UTC)
The bot runs to completion, but there is nothing in the article that it needs to change. That's not a bug.
LeadSongDogcome howl! 05:23, 21 December 2012 (UTC)
Cite DOI doesn't create template for doi:10.2307/461317
Place {{cite doi|10.2307/461317}} on a page. Then click on the "jump the queue" link.
We can't proceed until
Agreement on the best solution
JSTOR DOI's don't work right anymore (note that {{cite jstor}} is just a wrapper for {{cite doi}}). There is nothing the bot can do. AManWithNoPlan (talk) 17:15, 24 December 2012 (UTC)
I made a suggestion a while back about making the template display an error message informing the editor that the JSTOR API is down. --— Gadget850 (Ed)talk 10:42, 25 December 2012 (UTC)
I'm not sure the bot fails for all JSTOR articles. I tried it with doi:10.1086/591861, and it worked fine.–Temporal User (Talk) 01:13, 26 December 2012 (UTC)
That is not a JSTOR DOI, that is a DOI owned by the Journal. So, {{cite jstor}} does not work for that article. AManWithNoPlan (talk) 02:17, 26 December 2012 (UTC)
Now that I read this again, I don't see why JSTOR came up. --— Gadget850 (Ed)talk 02:55, 26 December 2012 (UTC)
{{cite jstor}} is just a wrapper for {{cite doi}}, and the DOI that the complaint is about is a JSTOR DOI. It does not matter if you use {{cite jstor|461317}} or {{cite doi|10.2307/461317}} the result is the same (other than {{cite jstor}} avoids tell you that the bot will fill it in for you). AManWithNoPlan (talk) 04:38, 26 December 2012 (UTC)
How do you distinguish between a JSTOR DOI and a DOI owned by the journal? –Temporal User (Talk) 08:58, 28 December 2012 (UTC)
How do I determine if any DOI, not just 10.2307, is JSTOR or not? (Also, I don't think 10.2307 is JSTOR. doi:10.2307/2998600 links to a non-JSTOR page.) –Temporal User (Talk) 05:51, 29 December 2012 (UTC)
10.2307/2998600 is a JSTOR DOI, but the journal seems to have taken it over from JSTOR. Off course, it still does not exist in the databases that the bot uses because JSTOR doesn't do what it should do. Of course they aren't the only offenders http://en.wikipedia.org/wiki/Category:Pages_with_inactive_DOIs 15:11, 29 December 2012 (UTC)
The bot needs to remove the special code that accesses the JSTOR databases and switch to the Cross-Ref code (unless it already does that). AManWithNoPlan (talk) 17:54, 31 December 2012 (UTC)
The code does this:
} else {
echo "not found in JSTOR?";
}
} else {
// Not a JSTOR doi, use CrossRef
$crossRef = $crossRef?$crossRef:crossRefData(urlencode(trim($p["doi"][0])));
}
Note that after the not found in JSTOR complaint, the code does not even try CrossRef AManWithNoPlan (talk) 17:57, 31 December 2012 (UTC)
This page's citations are indeed an ugly jumble. So perhaps the bot is easily confused. Perhaps this page should be fixed by hand? Choor monster (talk) 17:19, 4 January 2013 (UTC)
I assume that you mean this edit. There were four lines modified there.
In the first two, the bot was trying to make sense of the positional parameter | 2:38 pm|. None of the citation templates recognise positional parameters. The "p. m." visible in the References section is due to the |page=m - presumably the bot has decided that pm is short for p.m. and interpreted it as a page identifier. The nonexistent parameter |unused_data= is used by Citation bot to hold anything that it can't place in a recognised parameter, hence |unused_data= 2:38
In the third, this was a conversion of an absolute |url=http://www.amazon.com/dp/0060925000 to the relative |asin=0060925000 which is essentially the same thing, but is future-proof: if Amazon change their URL format, we only need to modify one or two templates, not thousands of pages. The |accessdate=2012-10-31 was removed because access dates are only meaningful when there is a URL.
In the fourth, this was a substitution of the European guillemet with the English double quote.
This is a good example of GIGO at work. --Redrose64 (talk) 19:24, 4 January 2013 (UTC)
OK, thanks. This makes sense. I've been slowly replacing some wikilinks with cite templates, getting the hang of it on other pages. But this page is just one giant mess.
The 2:38 pm doesn't make any sense to me, so I've just finished cleaning up after the bot here. I get the "asin". Is there supposed to be something similar for Google Books? I discovered at some point I was editing down the Google links to just the ID--who cares what the search criterion was that I used to get there? Choor monster (talk) 20:11, 4 January 2013 (UTC)
Well, there's {{Google books|7ydCAAAAIAAJ|History of the Western Insurrection|page=42}} This renders as:
Access dates are only useful for online sources. This ref has no URL, hence it's not online, therefore an access date is meaningless. --Redrose64 (talk) 23:08, 6 January 2013 (UTC)
It seems that for this particular source, someone forgot to put in the URL. But in principle, there are on-line sources that are not accessible with a URL. For example, computers in a few government offices that are open to the public, but which must be accessed in person at the government office. Jc3s5h (talk) 23:54, 6 January 2013 (UTC)
It looks like it removed the access dates in this edit but this time it had an archived URL.--Astros4477 (talk) 04:16, 7 January 2013 (UTC)
An |archiveurl= is not the same as a |url=. Although not explicitly stated at Template:Cite web#URL, the indentation there implies that the use of either |accessdate= or |archiveurl= also requires a |url= to be provided. --Redrose64 (talk) 10:23, 7 January 2013 (UTC)
There is no implication, it is explicit: see [[Template:Cite web#Syntax "and is ignored if parent is not used". |archiveurl= is a child of |url=.--— Gadget850 (Ed)talk 11:05, 7 January 2013 (UTC)
So, I think we all agree that the accessdate removal is not a bug but a feature when there is no url AManWithNoPlan (talk) 03:57, 12 January 2013 (UTC)
Input was a malformed wikitext to begin, with |id=ISSN 0134-3084 rather than |id={{ISSN|0134-3084}} . This is a plausibly common class of error that either this or another bot could address. LeadSongDogcome howl! 22:09, 27 February 2013 (UTC)
Not a bug. When parsing the named parameters of templates, the MediaWiki parser doesn't care how many spaces there are between the equals sign and the parameter value. You can put one space or two - or none, or any other number. It doesn't matter. --Redrose64 (talk) 15:09, 11 April 2013 (UTC)
I added names to a cite doi page (subpage). The bot deletes all names from "last10=" and higher. Note that this is a legal option since CS1 is upgraded into Lua.
If the template is intended to be substituted, please feel free to remove the speedy deletion tag and please consider putting a note on the template's page indicating that it must be substituted so as to avoid any future mistakes (<noinclude>{{substituted}}</noinclude>).
If you think this page should not be deleted for this reason, you may contest the nomination by visiting the page's talk page, where you can explain why you believe the page should not be deleted. However, be aware that once a page is tagged for speedy deletion, it may be removed without delay. Please do not remove the speedy deletion tag from the page yourself, but do not hesitate to add information in line with Wikipedia's policies and guidelines. If the page is deleted, and you wish to retrieve the deleted material for future reference or improvement, you can place a request here. Illia Connell (talk) 03:36, 29 April 2013 (UTC)
{{resolved}}
be certain bot chooses |author= with letter string and not number string
In relation to the above, this edit was also necessary as a result of the same bot edit. --Redrose64 (talk) 15:22, 28 May 2013 (UTC)
Unfortunately I no longer have time to handle bugs that arise due to unpredictable human input error. Other users are welcome to suggest code alterations that would resolve this bug. Martin(Smith609 – Talk) 14:24, 2 June 2013 (UTC)
enter this reference and see whether it expands correctly after clicking "jump the queue": Attention: This template ({{cite doi}}) is deprecated. To cite the publication identified by doi:10.1002/1531-8249(199911)46:5<770::AID-ANA13>3.0.CO;2-U, please use {{cite journal}} (if it was published in a bona fide academic journal, otherwise {{cite report}} with |doi=10.1002/1531-8249(199911)46:5<770::AID-ANA13>3.0.CO;2-U instead.
We can't proceed until
Agreement on the best solution
It looks like that's a bad DOI, not a bot problem. Worked around at the article.LeadSongDogcome howl! 18:07, 3 July 2012 (UTC)
DOI is correct (just click the link above to see), the problem is with the bot. Still, thanks for manually entering the reference in the article. kashmiri 21:49, 3 July 2012 (UTC)
It may be my firewall had trouble with that ridiculously long url. In any case, it's temporary. Once it passes, we'll see if things work again.LeadSongDogcome howl! 05:14, 4 July 2012 (UTC)
DOI.org resolver returns correct address http://onlinelibrary.wiley.com/doi/10.1002/1531-8249(199911)46:5<770::AID-ANA13>3.0.CO;2-U; Wiley Online Library further processess this URL, redirecting it to /abstract and, until 7 July, adding a maintenance message as a query string. Let me stress again that the problem is with the DOI bot which in all probability is unable to correctly resolve two subsequent colons and/or less/greater-than signs. kashmiri 13:02, 4 July 2012 (UTC)
Are there other examples where the double colon presents a problem? LeadSongDogcome howl! 13:22, 4 July 2012 (UTC)
Using the anchorencode magicword gives a non-functional result...
*Attention: This template ({{cite doi}}) is deprecated. To cite the publication identified by doi: 10.1002/1531-8249(199911)46:53.0.CO;2-U , please use {{cite journal}} (if it was published in a bona fide academic journal, otherwise {{cite report}} with |doi= 10.1002/1531-8249(199911)46:53.0.CO;2-U instead.
I do not really feel like investigating whether DOI bot works with anchorencode or not. The DOI is not a very typical one, albeit used regularly by that journal. What I want to point out that both the DOI bot and the Wiki software have problems passing/rendering this string correctly (see my entry above where I had to use nowiki tags around the URL as a workaround). The problem lies not with your firewall or Wiley pages; it is with the bot and Wiki software (just have a look at how the DOI got misformed by the bot). kashmiri 17:23, 4 July 2012 (UTC)
Ok, I think I'm getting somewhere. You're right, it wasn't my firewall, it rather was my local cache that was puking when Wiley forced the redirect. A manual cache purge and reload gets me past that. The cite doi subpage that I (manually) created was at the wrong path, so I've redirected it from
I'm not sure why cite doi names the subpage with anchorencode but generates the url with urlencode, but that's what it seems to do. (While anchorencode uses .3A urlencode uses %3A for the same character.)
Looks like the encoding is all working fine; I suspect that there was a temporary problem with a remote server. Unless there are further reports of similar errors I'll mark this as resolved. Martin(Smith609 – Talk) 07:54, 18 August 2012 (UTC)
The problem is still there with DOIs that contain < and > characters, I had it a few days back, just hold on a day or two pls and I will give more details here. kashmiri 10:39, 18 August 2012 (UTC)
OK, got it. Try the bot to expand this DOI: 10.1002/1098-2264(2000)9999:9999<::AID-GCC1018>3.0.CO;2-E. kashmiri 10:56, 18 August 2012 (UTC)
That woud be doi:10.1002/1098-2264(2000)9999:9999<::AID-GCC1018>3.0.CO;2-E. Notice the round parentheses are urlencoded to %28 and %29, and the angle brackets urlencode to %3C and %3E. When I click on that link, I initially get a "malformed request" failure, but on refreshing my cache it comes through as http://onlinelibrary.wiley.com/doi/10.1002/1098-2264%282000%299999:9999%3C::AID-GCC1018%3E3.0.CO;2-E/abstract, as expected. LeadSongDogcome howl! 17:30, 30 August 2012 (UTC)
I know what it would be and I definitely know how to encode it manually. But try to get DOI bot generate proper listing for this DOI using the Cite doi template. It does not work for me, and has nothing to do with cache. Try to experiment in your sandbox. kashmiri 22:52, 30 August 2012 (UTC)
Other DOIs from the same issue seem to work, so I can't replicate this. Martin(Smith609 – Talk) 12:23, 17 August 2013 (UTC) {{resolved}}
A specific edit to the bot's code is requested below.
Requested action from maintainer
Fix algorithm
I don't think this is a bug. It's a combination of two things: (1) you wrote the author name in the same format that would be used for a single author whose last name is "Sandhu BK" and whose first names are "Brueton MJ"; how is the bot to tell that's not what you meant? The correct author name format for cite pmid is very strict, and demands a semicolon in place of the comma you used. (2) Because of the strict author formatting in cite pmid, the bot reformats names to be in the "Last, F. M." format rather than the "Last FM" format you used. This is usually a good thing (if you had used a semicolon properly, it would have worked) but interacts badly with part (1). —David Eppstein (talk) 02:29, 23 November 2012 (UTC)
The problem in this case is that the {{cite pmid}} should not have been transcluded into Johanson–Blizzard syndrome article in the first place since {{cite pmid}} (and citation bot) enforces a style that is incompatible with the predominate citation style used in that article. As I have stated before, the {{cite pmid}} template is so easy to use, it becomes widely misused. Boghog (talk) 08:13, 23 November 2012 (UTC)
Definitely a bug since Citation bot created the template in the first place. Urhixidur (talk) 11:52, 23 November 2012 (UTC)
The reason the template was created because someone inserted it into the article. The fundamental problem is not this particular instance of this template, but the template itself since it is inflexible and enforces a particular citation format in contradiction to WP:CITEVAR. Boghog (talk) 23:41, 23 November 2012 (UTC)
The real problem is not with Citation bot, but rather the inflexible {{cite pmid}} template. It appears that help is on the way, but it may be sometime before it is implemented. Boghog (talk) 00:04, 24 November 2012 (UTC)
I concur with Boghog. The kind of name mangling reported here has occurred in several other occasions, always because of commas. In this case there was but one, but I've seen others where there was a list of 3+ names, thus 2+ commas. That should have tipped off the bot. Also see the "Jr." bug below (applicable to "Sr.", "IIIrd", and so forth). Urhixidur (talk) 14:28, 24 November 2012 (UTC)
Maintainer: A specific edit to the bot's code is requested below.
-->
Requested action from maintainer
Fix the bug
There seems to be some problem with Springer's data validation before export to crossref. If you go to that abstract on Springer's site, then click on "export citation" and choose plain text, you get:
Reference Type: Journal Article
Author: Haubrich, C. Haubrich
Author: Krings, T. Krings
Author: Senderek, J. Senderek
Author: Züchner, S. Züchner
Author: Schröder, J. Schröder
Author: Noth, J. Noth
Author: Töpper, R. Töpper
Primary Title: Hypertrophic nerve roots in a case of Roussy-Lévy syndrome
Journal Name: Neuroradiology
Cover Date: 2002-11-01
Publisher: Springer Berlin / Heidelberg
Issn: 0028-3940
Subject: Medicine
Start Page: 933
End Page: 937
Volume: 44
Issue: 11
Url: http://dx.doi.org/10.1007/s00234-002-0847-2
Doi: 10.1007/s00234-002-0847-2
Clearly the author names have been mangled, unless one expects "Bond, J. Bond" as the normal form. Still, the handling was incorrect. It is not clear if the cause was the umlauts in the names, or something else, . (As an aside, please note that this paper is a primary source, and so not a wp:MEDRS.) It might be helpful to state the dois for whichever other articles show similar problems. LeadSongDogcome howl! 20:54, 23 November 2012 (UTC)
Interesting observation. It would indicate then that Springer uploads bibliographical data to DOI databases using an automated system and that that system is buggy or misconfigured. As to other DOIs, as I mentioned, others from http://link.springer.com/journal/234/44/11/page/1 do not seem to work. I tried one or two others but don't want to spoil bot tests by getting all the templates created right away. kashmiri 22:14, 23 November 2012 (UTC)
I'm still not sure where to look for the instances of cite doi that you find to be incorrectly handled. The only one I could find by searching was the single example above. I'll create one without umlats, e.g. Attention: This template ({{cite doi}}) is deprecated. To cite the publication identified by doi:10.1007/s00234-002-0860-5, please use {{cite journal}} (if it was published in a bona fide academic journal, otherwise {{cite report}} with |doi=10.1007/s00234-002-0860-5 instead. and another with, e.g. Attention: This template ({{cite doi}}) is deprecated. To cite the publication identified by doi:10.1007/s00234-002-0829-4, please use {{cite journal}} (if it was published in a bona fide academic journal, otherwise {{cite report}} with |doi=10.1007/s00234-002-0829-4 instead. to see what happens.LeadSongDogcome howl! 20:13, 28 November 2012 (UTC)
Evidently, it's not about the umlauts. GIGO as suspected.LeadSongDogcome howl! 22:24, 28 November 2012 (UTC)
E.g.,
{{cite doi|10.1007/s00234-002-0860-5}}
gives:
Attention: This template ({{cite doi}}) is deprecated. To cite the publication identified by doi:10.1007/s00234-002-0860-5, please use {{cite journal}} (if it was published in a bona fide academic journal, otherwise {{cite report}} with |doi=10.1007/s00234-002-0860-5 instead.
– even when downloaded references (both RIS and TXT) are absolutely correct. Just go ahead and try any other article from the table of content that I linked above. kashmiri 00:55, 29 November 2012 (UTC)
This may be a knockon effect of Springer's recent platform changes. In the mean time, I'd suggest you just don't use cite doi for Springer. LeadSongDogcome howl! 23:04, 30 November 2012 (UTC)
The Information is now correct, BUT the first and last names are reversed. AManWithNoPlan (talk) 18:12, 17 December 2012 (UTC)
Looks like the bot edited the page three times in as many minutes. The first edit was pretty normal for Citation bot (tidying a Google Books URL and altering a hyphen to an en-dash per MOS:ENDASH). The next edit by NinaGreen (talk·contribs) did not alter anything of what Citation bot had done; yet Citation bot immediately reverted NinaGreen's edit. Two minutes later, NinaGreen made a different edit, again not altering any of the bot's first amendment - but this was also reverted by the bot. --Redrose64 (talk) 11:26, 5 January 2013 (UTC)
If you look closely at the diff, you'll see mismatched braces on {{harvnb}}. This seems to be what tripped up the bot. Once I fixed the braces, the bot left the wikitext as it was (it made no edit). It seems obvious that catching such mismatches is a basic syntax-check that ought to be happening, perhaps with every edit commit. LeadSongDogcome howl! 05:23, 6 January 2013 (UTC)
Mismatched braces might explain the first revert, but not the second. --Redrose64 (talk) 15:50, 6 January 2013 (UTC)
Looks like an edit conflict error. Don't know whether this can be fixed. Martin(Smith609 – Talk) 16:56, 17 August 2013 (UTC)
Dumb question
I am trying to see what effect the bot has in improving citation for "bare-urls" at Camp Dubois. They seem like bare urls to me but no change is made by the bot. So, I guess I don't understand what a bare-url is, or I don't understand the bot, probably both. Any help in understanding - to make me a better more efficient editor appreciated. Alanscottwalker (talk) 14:49, 12 January 2013 (UTC)
I think the url= is munging the link recognition. If CBot does not work, then User:Dispenser/Reflinks should after you remove those snippets. --— Gadget850 (Ed)talk 15:27, 12 January 2013 (UTC)
Thanks, I added the url= after the bot did not make any change to see if that would cause the bot to do something, which it did not. I will let you know how it goes. Alanscottwalker (talk) 15:42, 12 January 2013 (UTC)
User input of additional author / editor names is being removed. As part of the transition to Lua based citation templates, handled by Module:Citation/CS1, it is now possible to add an arbitrarily large number of author and editor fields to many of the standard templates rather than the historical 9 author and 4 editor limit. The bot needs to be updated to recognize this. In the example link below, someone added the additional fields and the bot reverted them. In addition, the bot should be modified to import additional author / editor information when possible.
Fix the bot to handle the additional author / editor fields correctly.
Another example. The paper definitely has 12 authors, and authors 10/11/12 on the paper exactly matched those which were in the template prior to the bot edit. --Redrose64 (talk) 09:36, 3 May 2013 (UTC)
Another. There were two cases here, in the same paragraph: to find them in the diff, search for |last10= --Redrose64 (talk) 10:01, 3 May 2013 (UTC)
Fixed in GitHub Pull 428 I've fixed this bug. Disruption to service could have been avoided if I'd been notified in advance of the planned change to Template:Citation. If there is a similarly urgent need to contact me in future, please feel free to e-mail me through my user page. Martin(Smith609 – Talk) 14:15, 2 June 2013 (UTC)
Should be able to just use {{cite journal}} with |mr=0102537, then let the bot expand it. Or doesn't that work? LeadSongDogcome howl! 16:18, 11 April 2013 (UTC)
I have software that is capable of doing this. For your example, it produces
{{citation
| last = Grothendieck | first = Alexander
| journal = The Tohoku Mathematical Journal
| mr = 0102537
| pages = 119–221
| series = Second Series
| title = Sur quelques points d'algèbre homologique
| volume = 9
| year = 1957}}
which renders as
Grothendieck, Alexander (1957), "Sur quelques points d'algèbre homologique", The Tohoku Mathematical Journal, Second Series, 9: 119–221, MR0102537.
It can also go the other way, from Wikipedia citation or cite templates to BibTeX. However, it currently only runs on OS X. If you're interested in trying it out, drop me an email. —David Eppstein (talk) 16:51, 3 May 2013 (UTC)
{{resolved}}
Using "Expand citations" link in 'p-tb' (toolbox in sidebar) does not open in new tab or return user to page they were on.
We can't proceed until
Agreement on the best solution
Not exactly a bug, but if you'd like to suggest a way to implement this, feel free; the related scripts are all available. Martin(Smith609 – Talk) 14:45, 2 June 2013 (UTC)
template:Cite doi/10.1021.2Fed044p475: after I manually corrected cs1 error (author name repeated), the bot reversed my edit, back into the error situation.
The detection of invalid parameters and the move to 'unused_data' is now obsolete for the Lua versions of the citation templates. These templates now detect invalid parameter names and fields without parameters and immediately show an error and add a category. --Gadget850talk 12:37, 25 April 2013 (UTC)
Fixed in GitHub Pull 437 {{resolved}}
I added manually param "displayauthors=9" (new CS1 citation option). The bot changes it into "display-authors=9". This is not an existing parameter name, so the ref shows an error.
What should happen
Leave it alone? It should not write an illegal name in the first place.
What error? Both |displayauthors= and |display-authors= should be supported as synonyms of each other. There is no reason for the bot to be converting one into the other, but it doesn't seem harmful or generate an error as far as I can see. Dragons flight (talk) 23:50, 25 April 2013 (UTC)
You are right. This was not one of the reappearing errors, and I missed the synonym thing. Changed level of severity for the useless edit. Or withdraw? -DePiep (talk) 00:02, 26 April 2013 (UTC)
For doi 10.1056/NEJMoa1304617 Citation Bot wants to add an incorrect (and implausibly large) page value of "130424140638006". This bizarre page value is actually in the source DOI metadata (shame on you New England Journal of Medicine!)
It would be a good idea if the bot had some basic sanity checking to avoid generating obviously invalid values (i.e. 100 trillion page value could have been rejected by sanity checking).
Feel free to suggest a sanity check in PHP; I'd be happy to add this to the source code. Martin(Smith609 – Talk) 14:35, 2 June 2013 (UTC)
On the assumption that such a sanity check is non-trivial (or at least beyond the ability of the submitter), I'm closing this bug. Martin(Smith609 – Talk) 11:35, 17 August 2013 (UTC)
A pipe correctly used as a parameter separator within {{cite web}} i.e. |url=http://benoitpoirierdambreville.com/about/|accessed 30 April 2013 was altered to %7C which made it and the following word become part of the preceding URL, i.e. http://benoitpoirierdambreville.com/about/%7Caccessed
What should happen
Regarding the incorrect construct |accessed 30 April 2013:
first preference: alter this to the correct construct |accessdate=30 April 2013
second preference: simply insert the missing equals giving |accessed=30 April 2013 which would have thrown the error Unknown parameter |accessed= ignored (|accessdate= suggested) (help) which would have placed the page into Category:Pages with citations using unsupported parameters, and so can be fixed up by other processes later
In most cases that match this syntax, the pipe is part of the URL. I can't see a way for the bot to always get this right; if you can think of an algorithm to improve its guessing, feel free to suggest a change to the code. Martin(Smith609 – Talk) 14:33, 2 June 2013 (UTC)
Pipe is part of the url syntax? This is {{cite web}}, with still the |accessed=-error, but untouched by the bot:
a. {{cite web|url=http://benoitpoirierdambreville.com/about/|accessed 30 April 2013|title=Some title}} →
The bot need not recognise the pipe as part of the url. My algorithm: if left alone, (for the cite template to handle), it ends fine. -DePiep (talk) 17:16, 2 June 2013 (UTC)
I've moved this back from User talk:Citation bot/Archive1 because it was not resolved. The literal pipe character is not valid in a URL (see rfc:3986), although it may be represented as %7C. There are no templates - not even {{cite web}} - where the literal pipe character is anything other than a separator between parameters. Therefore, the pipe character cannot form part of a URL when that URL is a parameter's value. --Redrose64 (talk) 23:15, 4 June 2013 (UTC)
Good you retrieved it.
Maybe the bot error cause here is, that the sequence is: url=...|something-not-a-parameter-because-no-=-sign. (which is a rare sequence!). -DePiep (talk) 23:26, 4 June 2013 (UTC)
I can't see any legitimate reason for a trailing pipe in a URL. Another example here of the bot changing it to % encoded form. There's no good reason why the sequence [pipe][whitespace][pipe] or [pipe][pipe] should be left at all in the parameter list, they can always be safely changed to [whitespace][pipe]. LeadSongDogcome howl! 18:45, 12 June 2013 (UTC)
Not just about a trailing pipe. The OP example url above ends up with the pipe, escaped, midstring (the first space is end-of-url separator). citation bot should not be commanded by extra whitespace requirements in a template parameter list. The general question is: is there a situation possible that we have a non-escaped pipe in an url that should stay there (in a {{citation}}; escaped or not)? If yes: does CS1 itself handle it correctly? If no: citation bot should not touch the pipe as url. -DePiep (talk) 20:18, 12 June 2013 (UTC)
The MediaWiki template parser treats any bare pipe within pairs of double braces as a parameter separator, without exception. The only way that a pipe within pairs of double braces is not so treated is when it is not a bare pipe - it has been further enclosed by some other Wiki markup such as a piped link [[foo|bar]] (including image syntax), a pair of triple braces (as used in template coding as e.g. {{{url|}}}), etc.
To force the MediaWiki parser to treat a pipe within pairs of double braces (but outside of other markup) as a literal character and not as a param separator, it needs to be intentionally encoded in some way: if it is part of a URL, then %7C will do it; elsewhere (such as in a webpage title) the syntax | works.
Many people put templates into pages in such a manner that all superfluous whitespace is removed, so a pipe used as a param separator will often be preceded by characters which are not whitespace; it's not unusual for such characters to be the value of a |url= --Redrose64 (talk) 22:47, 12 June 2013 (UTC)
So if I understand correctly, there is no reason to treat a pipe in any URL as anything other than a parameter separator. Right? LeadSongDogcome howl! 01:23, 13 June 2013 (UTC)
Well, yes... but the pipe isn't in the URL, it separates the URL from whatever comes next. Some users put the pipe directly after the URL; some use a space and then the pipe; some (like this guy use a space and two pipes. But in all cases, the pipe is taken by the MediaWiki parser as a parameter separator, never as part of a parameter's value. --Redrose64 (talk) 06:48, 13 June 2013 (UTC)
So it is only citation bot that gets this wrong. The bot did not cut off url-before-a-pipe in these two examples. Module:Citation/CS1 and the three of us talking here, agree. One more thing, about the three options Redrose64 wrote in the original bug report. If the bot corrects internally this behaviour (a straight pipe is always a parameter separator), the first two options are not needed (they are editor/bot edits anyway, not bot or CS1 logic). -DePiep (talk) 07:31, 13 June 2013 (UTC)
It would be very useful if the bot added a single space before each pipe within any cite code that it edits. The code becomes more readable when that is done. Likewise, spaces before or after "equals" signs are rarely necesaary. 212.139.111.187 (talk) 19:19, 7 July 2013 (UTC)
re Not just off-topic, it is contrary to the topic! (I do accept your good faith contribution). It is not about our reading, it is about the bots reading ability. -DePiep (talk) 19:56, 7 July 2013 (UTC)
If the bot ensured there is a single space before every pipe within each cite that it touches, adding the space where missing, then it wouldn't misread the URL or any other part of the cite code. 212.139.111.83 (talk) 18:15, 8 July 2013 (UTC)
You propose to change the template formatting to serve & solve this bot's problem. A. won't happen, and B. why not change the bot to read a template correctly? -DePiep (talk) 18:48, 8 July 2013 (UTC)
Most code examples seemingly show a space before each pipe. Most humans seem to add a space before each pipe when manually creating a reference. Some people also add a space after each pipe. Sure, change the bot so that it recognises that quote, space, pipe or right-curly-brace signifies the end of a parameter value, but please also change the bot so that it adds a space before each pipe when editing a reference. 212.139.111.83 (talk) 20:47, 8 July 2013 (UTC)
This is getting funny. For the bot to add a space before a (parameter ending) pipe, it would have to recognise a pipe as such. That is what we both want. EOT. -DePiep (talk) 21:01, 8 July 2013 (UTC)
Some URLs contain a pipe. I don't have an example to hand, but say journal.com/?page=auo|ahi&aht. As such, the bot has to guess whether the intended behaviour is a URL with a pipe in it, or a separate parameter. the bot uses an equals sign as a clue: parameters always contain an equals sign. If there's no equals, it guesses that the pipe's part of the URL. A properly-formatted citation (with or without spaces) will never throw a problem. Hence, marking resolved. Martin(Smith609 – Talk) 11:31, 17 August 2013 (UTC) {{resolved}}
Deleterious: Human-input data is deleted or articles are otherwise significantly affected. Many bot edits require undoing.
What happens
|page=693 was marked as |pages=693. Fixed. Bot added |pages=693–697; discussion 697–700. Rm'd single |page=693. Bot readded |page=693, causing error, x3. Stopped after 4th removal of old |page=.
Hmmm, all those edits were inside a very short time window, it's not that urgent to correct the errors, why not hang back a few minutes instead of edit warring with a bot? LeadSongDogcome howl! 03:55, 11 May 2013 (UTC)
It's just really frustrating. Seems to have stopped doing it after I rolled it back on another page instead of just editing.--Aurictalk 10:43, 11 May 2013 (UTC)
You might have an idea there: if the bot ignores an undo but respects a rollback as an instruction not to do it again, that might explain why this sequence ceased. --Redrose64 (talk) 15:43, 11 May 2013 (UTC)
Hello, I'm BracketBot. I have automatically detected that your edit to Kakkonto may have broken the syntax by modifying 1 "[]"s. If you have, don't worry, just edit the page again to fix it. If I misunderstood what happened, or if you have any questions, you can leave a message on my operator's talk page.
List of unpaired brackets remaining on the page
Determination of puerarin, daidzein, paeoniflorin, cinnamic acid, glycyrrhizin, ephedrine, and \6]-gingerol in Ge-gen-tang by micellar electrokinetic chromatography |year=1997 |last1=Huang |first1=
There is no response from bot maintenance to prevent these errors. See Module talk:Citation/CS1. -DePiep (talk) 21:51, 28 May 2013 (UTC)
FYI Bot introducing an error in the same topic |displayauthors= (though not exactly the same): [89]. -DePiep (talk) 14:53, 30 May 2013 (UTC)
{{cite journal}} throws |displayauthors= suggested when there are exactly nine authors, and |displayauthors= has not been set. It's a hint that you may either add further authors beyond nine, since we now support many more, or to add an explicit |displayauthors=9 if the paper credits exactly nine authors. --Redrose64 (talk) 16:13, 30 May 2013 (UTC)
re Redrose64. This is not just {{cite journal}}, it is basically Module:Citation/CS1 (see it's talk). I say: Citation bot should not introduce (or: re-introduce) this error. (same link again) shows it did (currect ref #7). -DePiep (talk) 19:57, 31 May 2013 (UTC)
I do agree that the bot shouldn't make these errors; however, I am unable to fix it - it's not my bot - so I blocked it (after notifying WP:BON) because the number of bug reports on this page that are simply being ignored was far too great. --Redrose64 (talk) 20:06, 31 May 2013 (UTC)
Of course I am aware of your knowledge. I only noted my point, here, to wake up and help the bot master. Also I did not raise a new full bug issue here. Why bother? -DePiep (talk) 21:02, 31 May 2013 (UTC)
This bug report is unintelligible. I have no idea what "Lua CS1 conforming" means, nor how to find out. Please list concrete requests that I am able to action. Martin(Smith609 – Talk) 14:22, 2 June 2013 (UTC)
The bot edit (the actual edit, mentioned in the bot report) introduces two redtext errors in the reference list: [90]. The cause is explained in the CS1 help page, I won't repeat here. As Redrose64 explained another way using the red error text, this is because of the new Module:Citation/CS1. Strange that you claim to have solved the errors, but do not have a clue when someone writes "CS1". -DePiep (talk) 16:55, 2 June 2013 (UTC)
I couldn't see any red-text errors, so am marking this report as resolved. Please file a new bug report if there's anything outstanding, with a detailed explaination of the required action. Martin(Smith609 – Talk) 10:21, 17 August 2013 (UTC)
Valid accessdate removed
[91]: a valid |accessdate= removed by the bot. -DePiep (talk) 18:26, 2 June 2013 (UTC)
Actually, no, that was not a valid accessdate. Accessdates should be used for sources that are changeable, which is not the case for a print source. The bot's action was correct. Check the hidden cats of this page, after your incorrect reversal of the bot, it is now in Category:Pages using citations with accessdate and no URL, awaiting cleanup... --Randykitty (talk) 18:30, 2 June 2013 (UTC)
The reference uses {{cite web}}, so not print. Or, how would the bot have concluded it were print? -DePiep (talk) 18:42, 2 June 2013 (UTC)
Like the cat says: accessdate but no URL. --Randykitty (talk) 18:49, 2 June 2013 (UTC)
Then possible conclusions: 1. It was print → change {{cite web}} into a more apropriate template. 2. url can be recovered and added: accessdate can stay. 3. No url available: rm accessdate, rm |url=. The bot concluded one, while it would require user-intervention to get the right conclusion & solution. -DePiep (talk) 19:03, 2 June 2013 (UTC)
I don't see how a mere bot could ever decide whether this should be a cite journal, cite news, or any other citation template instead of cite web. I don't know whether the bot tries to recover URLs through Google, but it does add URLs (if needed) if it has something to go by (like a PMID; in the present case, even a human editor may not be able to readily see what the editor who added the reference actually intended: to me it seems like a journal article, but without Googling it and finding something online, I can't be certain). As it is, the citation template itself signals an error (by adding the hidden cat) and the simplest fix is to remove the accessdate (which indeed doesn't make sense in the absence of an URL). If that is the incorrect action, then it is the responsibility of the editor triggering the bot to check and correct where necessary. It's a tool and, unfortunately, at this point an imperfect one, but it still is extremely helpful to complete citations that miss info (like dois, PMIDs, title, etc, etc). It's an enormous timesaver if you can simply insert {{cite journal |PMID=1234567}} and then have the bot fill in the nitty-gritty details. --Randykitty (talk) 08:33, 3 June 2013 (UTC)
Automated botreporter reinstalled. It was not removed by the bot maintainer. -DePiep (talk) 08:07, 3 June 2013 (UTC)
Randykitty, if the bot can't decide it should leave the citation alone. No mishandling the error situation. End of story. There was a extensive and inconclusive discussion on how to handle exactly these situations at VP/P. It clearly did not conclude one can delete such a lone accessdate without checks. So the bot should follow. And the invoking editor is not responsible for a bots automated action. -DePiep (talk) 08:56, 3 June 2013 (UTC)
This citation has a date, thus the access date is not required. --Gadget850talk 23:09, 3 June 2013 (UTC)
Another solution for the same problem: if there are more authors than listed (say there are 12 authors and 4 are listed in the reference here), one should always write |displayauthors=3 (4 - 1 = 3) which produces "et al." correctly. This same is valid for situation: "9 listed, unknown number --> |displayauthor=8". -DePiep (talk) 09:12, 3 June 2013 (UTC)
The simple, general answer, which should always work, is to set the default to |displayauthors=1, then let editors change it if they prefer something else. LeadSongDogcome howl! 21:35, 3 June 2013 (UTC)
I am such an editor, and I would not like to have a bot say "1" when I took time to descibe more authors. -DePiep (talk) 23:21, 3 June 2013 (UTC)
So set it to the value you prefer. It's trivial to do. Or perhaps you simply want every article to default to your preference to show all authors? That won't wash, it has been discussed many times. LeadSongDogcome howl! 04:55, 5 June 2013 (UTC)
The behavior of the CS1 templates is to show all authors that someone has taken the time to enter, unless the editor filling in the template also chooses to add a |displayauthors= setting in order to force truncation. If 33 authors are entered then 33 are shown. One can think of this as a default setting of infinite for displayauthors, though that is still restrained by the willingness of editors to actually type in a long author list. As far as I know there is no requirement anywhere that an editor has to enter every author. I also don't know of any broad policy that specifies any particular level of truncation as required or preferred. Which suggests that the only controlling issue is WP:CITEVAR, i.e. that citations should generally be consistent within individual articles.
The only exception to the above rule is that if an editor enters exactly 9 authors, then we presently treat that as if |displayauthors=8 were set because that matches the historical behavior of the templates prior to Lua, i.e. entering exactly nine authors resulted in eight authors being display followed by an "et al.". Dragons flight (talk) 05:57, 5 June 2013 (UTC)
It is a good idea to fill in as many authors as are credited on the paper, even if it is desired to display fewer - this is so that the COinS metadata is populated properly.
Good catch. So unless we find consensus on changing the behaviour, the post-Lua templates should continue to treat display-authors=8 as the implicit default value. Otherwise we substantially change a huge number of articles without the consent of their editors. If they want some different behaviour they can say so by specifying an explicit value for display-authors. LeadSongDogcome howl! 16:50, 6 June 2013 (UTC)
Preferred CS1 behaviour is not exactly the topic here (this page). The primary issue is that when the bot adds exactly 9 authors, CS1 produces the error message (see the pre-Lua behaviour noted above). A bot introducing an error message is seriously at fault. This specific situation should be covered (prevented) by the bot.
Now the bot can arrive at 9 authors for 2 reasons:
For historical reason, the bot never looks further than 9. If so, the bot should add |displayauthosrs=8 (produces et al., correct).
There are exactly 9 authors in the source. If the bot knows this, it should add |displayauthors=9. (No et al. will be added, correct).
Bot action should not require manual action afterwards to remove the error message. -DePiep (talk) 17:21, 6 June 2013 (UTC)
I'm marking this resolved; please file a bug report using a link at the top if there is any outstanding action required of me. Martin(Smith609 – Talk) 10:17, 17 August 2013 (UTC) {{resolved}}
The bot is trying to add an f word with every edit
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
The bot adds letter f in the beginning of the article, at least when called manually using the "citation" button.
Replication instructions
Try any article
We can't proceed until
A quick fix
Requested action from maintainer
delete the spurious "f" in the first line of the bot code
I've looked at several of the user-activated edits that the bot made in the last few hours and I don't see a single instance of this. Can you give an example? --Randykitty (talk) 15:20, 3 June 2013 (UTC)
Hit the "Citation" button (with me, it is in the row with the ""Save page", "Show preview" buttons: below the editbox)
After bot processing (maybe some seconds), very first character now is "f".
Clearly, no much editors did save this. -DePiep (talk) 15:33, 3 June 2013 (UTC)
I see it now, it only occurs if you run the bot while editing. I have a link in my toolbox, which processes an article in a sort of batch mode. Just tried that on The New England Journal of Medicine and then it doesn't happen. My guess is that this is a simple typo introduced by the modifications made to the bot yesterday. Correcting it should be fairly trivial for somebody who knows bots. Unfortunately, I can program in PASCAL, but not much more (and to show my age: I can read FORTRAN and have heard of COBOL... :-). --Randykitty (talk) 15:50, 3 June 2013 (UTC)
So, a bug it is. -DePiep (talk) 20:11, 3 June 2013 (UTC)
Difflinks please. No need to send others on hunting expeditions through your contribs looking for edits.LeadSongDogcome howl! 21:59, 3 June 2013 (UTC)
and btw, the bot master is not interested. Why who how spend time? -DePiep (talk) 22:07, 3 June 2013 (UTC)
also, I get the smell that LeadSongDog is related to the bot owner. -DePiep (talk) 22:15, 3 June 2013 (UTC)
In the sense that I've long made use of this bot on WP, yes. In the sense that I've tried to help others use it and improve it, yes. On a couple of occasions I've given him an email prod when things seemed to be deteriorating here. But I wouldn't know him if I passed him on the street. Is there some reason for hurling accusations around? Have you been having an unpleasant day? In any case, I am quite sure you've failed to read the history here. The bot owner has for many years undertaken the support of an unending stream of whimsical changes in the stylistic requirements the bot needs to support, largely without complaint. I see no reason you should heap vitriol upon him in this fashion. Now if you had volunteered to undertake support of the bot and been rebuffed, that might be a different matter... LeadSongDogcome howl! 22:33, 3 June 2013 (UTC)
hurling accusations around I do not. I point to failings in the bot. If the bot is this important, we should not be dependent on a personal behaviour. And I maitain that the recent (June 2nd) edits by botowner do not show a grasping of the changes. Looking for CS1 would prevent ignorance. -DePiep (talk) 23:02, 3 June 2013 (UTC)
Your "smell" implication that I was somehow inappropriately connected to Smith609 was indeed a wp:SOCK accusation, and unfounded to boot. "Looking for CS1 would prevent ignorance" is a statement free of useful content. You'll need to actually say what you mean if you wish to have it understood and acted upon. LeadSongDogcome howl! 04:57, 4 June 2013 (UTC)
This is clearly a simple typo in rev428 to the bot. I'm sure that it'll be readily remedied.LeadSongDogcome howl! 21:22, 4 June 2013 (UTC)
I am seeing the same problem. In fact, I added a report below (not knowing this discussion was underway. (It's nice to know that I'm not the only editor experiencing the problem.) So it has not been fixed. Thanks. – S. Rich (talk) 02:27, 5 June 2013 (UTC)
Citation bot introduced an error. -DePiep (talk) 22:56, 3 June 2013 (UTC)
Look closer. The before version had many different values for |last2=. Garbage in, garbage out. If you look at the recommended citation on the e-journal it reads "Ginter, T.N.; Gregorich, K.E.; Loveland, W.; Lee, D.M.; Kirbach, U.W.; Sudowe, R.; et al.(2002). Confirmation of production of element 110 by the (208)Pb(64-Ni,n)reaction. Lawrence Berkeley National Laboratory: Lawrence Berkeley National Laboratory. Retrieved from: http://escholarship.org/uc/item/8pr676mb" which is to say, only six authors displayed.LeadSongDogcome howl! 05:07, 4 June 2013 (UTC)
many different values for |last2=??? Which input error do you point at? What is the garbage in? If there were multiple entrances for the same |last2= or its synonyms like |author2=, there would have been an error message. The before-input had no errors and was in line with the documentation. -DePiep (talk) 09:54, 4 June 2013 (UTC)
Before code: {{cite journal|doi=10.1103/PhysRevC.67.064609|title=Confirmation of production of element 110 by the <sup>208</sup>Pb(<sup>64</sup>Ni,n) reaction|year=2003|author=Ginter, T. N.|journal=Physical Review C|volume=67|page=064609 |last2=Gregorich|first2=K.|last3=Loveland|first3=W.|last4=Lee|first4=D.|last5=Kirbach|first5=U.|last6=Sudowe|first6=R.|last7=Folden|first7=C.|last8=Patin|first8=J.|last9=Seward|first9=N.|first10=P. |last10=Wilk|first11=P. |last11=Zielinski|first12=K. |last12=Aleklett|first13=R. |last13=Eichler|first14=H. |last14=Nitsche|first15=D. |last15=Hoffman |bibcode = 2003PhRvC..67f4609G|issue=6 }} →
Ginter, T. N.; Gregorich, K.; Loveland, W.; Lee, D.; Kirbach, U.; Sudowe, R.; Folden, C.; Patin, J.; Seward, N.; Wilk, P.; Zielinski, P.; Aleklett, K.; Eichler, R.; Nitsche, H.; Hoffman, D. (2003). "Confirmation of production of element 110 by the 208Pb(64Ni,n) reaction". Physical Review C. 67 (6): 064609. Bibcode:2003PhRvC..67f4609G. doi:10.1103/PhysRevC.67.064609.
Again: what is the error? And, of course, why did the bot not remove that "error"? Instead, it introduced one. -DePiep (talk) 10:04, 4 June 2013 (UTC)
Sorry, my previous comment pointed at the wrong changed para: You had this for ref name="267Ds"
{{cite journal|doi=10.1103/PhysRevC.51.R2293|title=Evidence for the possible synthesis of element 110 produced by the <sup>59</sup>Co+<sup>209</sup>Bi reaction|year=1995|author=Ghiorso, A.|journal=Physical Review C|volume=51|pages=R2293|last2=Lee|first2=D.|last3=Somerville|first3=L.|last4=Loveland|first4=W.|last5=Nitschke|first5=J.|last6=Ghiorso|first6=W.|last7=Seaborg|first7=G.|last8=Wilmarth|first8=P.|last9=Leres|first9=R.|first10=A. |last10=Wydler|first11=M. |last11=Nurmia|first12=K. |last12=Gregorich|first13=K. |last13=Czerwinski|first14=R. |last14=Gaylord|first15=T. |last15=Hamilton|first16=N. J. |last16=Hannink|first17=D. C. |last17=Hoffman|first18=C. |last18=Jarzynski|first19=C. |last19=Kacher|first20=B. |last2=Kadkhodayan|first21=S. |last2=Kreek|first22=M. |last2=Lane|first23=A. |last2=Lyon|first24=M. A. |last2=McMahan|first25=M. |last2=Neu|first26=T. |last2=Sikkeland|first27=W. J. |last2=Swiatecki|first28=A. |last2=Türler|first29=J. T. |last2=Walton|first30=S. |last2=Yashita|bibcode = 1995PhRvC..51.2293G|issue=5 }}
which renders as {{quote|Ghiorso, A.; Somerville, L.; Loveland, W.; Nitschke, J.; Ghiorso, W.; Seaborg, G.; Wilmarth, P.; Leres, R.; Wydler, A.; Nurmia, M.; Gregorich, K.; Czerwinski, K.; Gaylord, R.; Hamilton, T.; Hannink, N. J.; Hoffman, D. C.; Jarzynski, C.; Kacher, C. (1995). "Evidence for the possible synthesis of element 110 produced by the 59Co+209Bi reaction". Physical Review C. 51: R2293. doi:10.1103/PhysRevC.51.R2293. {{cite journal}}: |first20= missing |last20= (help); |first21= missing |last21= (help); |first22= missing |last22= (help); |first23= missing |last23= (help); |first24= missing |last24= (help); |first25= missing |last25= (help); |first26= missing |last26= (help); |first27= missing |last27= (help); |first28= missing |last28= (help); |first29= missing |last29= (help); |first2= missing |last2= (help); Cite has empty unknown parameter: |action required from maintainer= (help)
Howdy. I noticed that the bot did this again on June 27 here. I fixed it. Is there any kind of update when this bug might be fixed?--Rockfang (talk) 22:53, 22 July 2013 (UTC)
In this case the bot saw an instance of {{cite journal}} without |journal= but with |work=Behav Ecol. What it should have done is change para work to para journal while keeping the value, to have |journal=Behav Ecol. Instead it simply added |journal=Behav Ecol. The old template tolerated this redundancy, but the new one (post-Lua) does not, creating a problem. LeadSongDogcome howl! 18:02, 23 July 2013 (UTC)
"Inconvenience: Humans must occasionally make immediate edits to clean up after the bot" (I would call this worse than "inconvenient", because if a human doesn't notice then it's stuck wrong until it's fixed.)
What happens
"A et al" is changed to "A et al, B, C, D", and "A, B, C" (in one author field" is changed to "A, B, C, B, C"
It seems to be impossible to guard against this type of problem, because the root cause is that the citation template parameters were not filled in properly. But this shouldn't be dismissed; if citations were always filled out perfectly then this bot would never be needed! I'm not sure if this was an automated change since the end of the edit summary seems to be cut off. If it was manually triggered by an editor than obviously that's that editor's fault for not checking the preview properly. In that case please let us know how to find out who the user is so that they can be informed. If it was an automatic change then I think the bot should be disabled for all such changes, especially if the maintainer is no longer active. Quietbritishjim (talk) 20:53, 14 June 2013 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
Erroneous parameters are reinstated. DOI query evidently returning the wrong values, so I edited manually. Bot is reinstating erroneous " last2 = Atiyah
What should happen
No edit was required, the bot should not reinstate erroneous data.
I've now blanked the parameters and hope this will stop a further automated edit. I wasn't expecting the citation bot to revisit the page, perhaps the bot should not re-edit pages it has already touched once? Gareth Jones (talk) 15:31, 18 June 2013 (UTC)
Hmmm, OCLC4668945893 and Bibcode:1961PCPS...57..902K both show that second author, which metadata the ADS attributes to CROSSREF. If they are in error, the bot should ignore |last2=<!-- leave blank to exclude bad data --> I'm not sure about leaving the parameter undefined. LeadSongDogcome howl! 21:15, 19 June 2013 (UTC)
Then OCLC and Bibcode both have bad data. The article is by Kingman (only), "Communicated by M. F. Atiyah" meaning I believe that Atiyah is acting as journal editor. The pdf that one gets from the publisher says "How to cite this article: J. F. C. Kingman (1961). The single server queue in heavy traffic. Mathematical Proceedings of the Cambridge Philosophical Society, 57, pp 902904 doi:10.1017/S0305004100036094". So I believe this is not a Citation bot bug. But is there some way to tell the bot not to touch this field, other than locking it out of the citation altogether? —David Eppstein (talk) 00:09, 25 June 2013 (UTC)
A correction request has been submitted to the ADS, we'll see if they correct their record. Meanwhile, something's odd about what happened with this, but it should work with this. I'll see if I can determine why they're treated differently.LeadSongDogcome howl! 06:25, 29 June 2013 (UTC)
{{resolved}}
Inconvenience: The bot does not create a template, from a valid DOI
What happens
Trys, but can't find anything
What should happen
Provider: John Wiley & Sons, Ltd
Content:text/plain; charset="UTF-8"
TY - JOUR
AU - Milius, Susan
TI - Botany under the mistletoe: Twisters, spitters, and other flowery thoughts for romantic moments
JO - Science News
JA - Sci News
VL - 158
IS - 26-27
PB - Society for Science & the Public
SN - 1943-0930
UR - http://dx.doi.org/10.2307/4018592
DO - 10.2307/4018592
SP - 411
EP - 413
PY - 2000
ER -
This is a database error. JSTOR - to where this DOI links - sometimes provides duplicate dois. Try using the DOi from Wiley & Son. Martin(Smith609 – Talk) 16:06, 17 August 2013 (UTC)
Look, I'm sure I'll get an earful about why the hell I even bothered to add {{thinsp}} ( thinsp is a "thin space"). But whether or not that was a good idea, the bot certainly shouldn't do what it did -- no doubt this wanders into an unanticipated corner case in some regex.
Interesting. The xml from pubmed for that cited source shows:
<ArticleTitle>"No longer Gage": frontal lobe dysfunction and emotional changes.</ArticleTitle>
That should give |title="No longer Gage": frontal lobe dysfunction and emotional changes. as the value to be automagically inserted by the bot. Your version substituted &squot; for the leading ", presumably to improve legibility, but breaking the automatic check against published metadata. I note that of the published works citing that paper, many mangle the quotation marks in even stranger ways. This may be a special-enough case to be worth avoiding cite journal entirely, manually writing the cite doi subpage instead. The problem only arises when the title properly begins with a double quotation mark, invoking the substitution of the single... LeadSongDogcome howl! 22:03, 6 August 2013 (UTC)
I don't think you're understanding the problem -- nothing to do with the title field. I put {{thinsp}} in 4 fields [102] -- last1, last2, last3, title; also, in title I changed double quotes to single quotes. The bot came in after me and, er, adjusted [103]. In last1, 2, 3 it turned {{thinsp}} into {{T.} (with only one brace on the right!), but it left the {{thinsp}} in the title field alone.
Later [104] I tried using   instead of {{thinsp}} and the bot left that alone. So there's something about the {{thinsp}} which is confusing it -- except in the title field!
But your comments raise some interesting points:
1. Surely there are plenty of times that a title contains double-quotes; since the bot knows that the title itself will be enclosed in double-quotes by the cite template, it ought to be made smart enough to change the doubles to singles, and even (as I did manually) add a thinsp at left or right or both as needed. (Whether you want to detect nested quotes in the title and handle them, I leave to you. This well illustrates why the British system of single-inner, double-outer really is better -- under it, to embed a quote within a quote doesn't require "toggling" -- or 'toggling' -- all the inside quotes.)
2. You mentioned "manually writing the cite doi subpage instead". Sure that's fine for some situations, but how to keep the bot from coming in after and fooling with what the human's done? I expected there to be some magic bot_hands_off parameter I could add to the doi cite, but there doesn't seem to be such a thing. (Perhaps in the presence of such a keyword the bot might add, as a <!- -> comment, what it would have put in that field, so a human can judge for himself.)
3. Or, just the fact that a human edited at all might act as an implicit instruction to the bot to keep hands off. In this case, again, the bot might still come in and add a <!- -> comment at the end of fields it "disagrees" with.
4. I realize that (3) above may raise some issues when the source databases the bot is working from update their records i.e. I can imagine that, right now, if an external database makes a spelling correction in a bibliographic record, the bot might propagate that to the doi cite -- if any human intervention blocks the bot permanently, then that cite would never get updated. One way to handle this is the following: any time there's a human edit, the bot comes in and adds a <!- comment to the effect, "Live value is human-supplied. I would have supplied value X." Later, if there's a database update changing one of the field values, the bot looks at that field in the doi cite. If there's no <!- such as I just described, the bot can assume the field was bot-supplied, and therefore it updates the field with the new value from the external database. But if there's a <!- comment, then bot keeps its hands off. Of course, a not-uncommon special case might be that the new value from the external database matches the human-supplied value, because the human knew early that the value was wrong and corrected it. This might actually be worth detecting so the comment can be changed in that case to "Database value finally reflects human-supplied value".)
OK, enough commentary! EEng (talk) 23:24, 6 August 2013 (UTC)
No, I understood that problem in the forenames, I just couldn't identify its source. I've looked at the code, but I can't claim to have much understanding of it. It seems to be buried several layers down in unicode conversions, and PHP isn't my bag. With regards the toggling issue, it strikes me that the proper place for that functionality is in the templates. If that can't be done, then such failing should be documented so that all editors, human or bot, know that they have to pre-toggle the title quotation marks and in what fashion to do so in order to match the template's expectations.
To have the bot keep "hands off", see User:Citation_bot#Stopping the bot from editing. It's pretty simple. However, there's some merit in saying "for cite doi subpages, leave completed templates untouched if they have been human-edited following the bot's initial filling-out". My personal perspective is rather more drastic: eliminate cite doi template subpages entirely. Just subst them. They are a cute-but-bad idea. That seems, however, not to be a widely-held view and I do not really wish to beat a wp:DEADHORSE. LeadSongDogcome howl! 14:53, 7 August 2013 (UTC)
If the bot is erroneously adding a DOI, author, etc to a citation, and you want to stop it adding the data again, you need to put a comment in place of the appropriate parameter – because the bot will not overwrite existing data.
Well, clearly the bot will overwrite existing data -- maybe it should read will not overwrite a value that contains a comment??? I'm gonna try that experiment.
For those of us not steeped in this corner of WP: what is the significance of that? What is this metadata fed to? Is it different if a template e.g. { { thinsp } } is used instead? EEng (talk) 22:45, 7 August 2013 (UTC)
{{thinsp}} is inconsistent with the Cite Doi formatting requirements. If you wish to use this template, then unfortunatly you won't be able to use Cite Doi - using Cite Journal in an article will be fine, and the bot won't alter the forename punctuation. Martin(Smith609 – Talk) 17:34, 17 August 2013 (UTC)
This edit request has been answered. Set the |answered= or |ans= parameter to no to reactivate your request.
It's not easy for a novice to find the relevant documentation. It could be better integrated but I lack the time and skill. As a stopgap I suggest that something like the following be added as hatnotes (or whatever) to Template:Cite doi/subpage
citation bot added " |last=Schwartz" when the correct " |last13=Schwartz" already existed
in Cancer and nausea, citation bot incorrectly added " |last=Schwartz" (diff). This creates the somewhat confusing error message "More than one of |last1= and |last= specified". —Chris CapocciaT⁄C 01:18, 18 September 2013 (UTC)
Manually corrected. The entries for authors 13 and 14. Not sure what was confusing about the error message, it seems to have fingered the problem.LeadSongDogcome howl! 13:35, 18 September 2013 (UTC)
manual correction is always possible, but the fact is that authors from bibcodes are still being incorrectly listed with the last name of the first author as author1 and the last name of the last author as last. then citation bot must be giving up and not trying to do any of the rest of the authors. expand any citation from just the bibcode, and you should get the same result. —Chris CapocciaT⁄C 04:22, 30 September 2013 (UTC)
I encountered a similar issue, where 'last' entries were being added when numbered last names (last1, last2 etc.) were already in use. diff. I had to manually revert. Modest Geniustalk 19:40, 30 November 2013 (UTC)
I've found a couple of pages that have a {{cite doi|10.2307/number}} that the citebot doesn't seem to be able to handle. 10.2307 is JSTOR. OTOH, if I change that to {{cite jstor|number}} it seems to be able to find the appropriate information and create the template. Any ideas? At least as of right now, Vetigastropoda is an example. I could change it to {{cite jstor|1306561}}, but I'm going to leave it there as an example for the moment.Naraht (talk) 13:40, 18 September 2013 (UTC)
Seems that doi is not resolved by Crossref, and searching JSTOR for the number doesn't find the article either, even though it does find it as the first hit when searched for "Vetagastropoda". Perhaps still in the process of registration? LeadSongDogcome howl! 13:24, 19 September 2013 (UTC)
I agree it isn't caught by Crossref, how did you search JSTOR? The search in google that comes up if you search for JSTOR does find for 1306561. This article is from 1997, so it doesn't seem that it is too recent. Is it proper to change it to a cite jstor?Naraht (talk) 21:49, 22 September 2013 (UTC)
I plugged in "Vetagastropoda" to the search box at jstor.org, and it found the article. LeadSongDogcome howl! 03:40, 23 September 2013 (UTC)
Your bug report didn't say where you saw this problem, making it difficult for others to duplicate and diagnose. LeadSongDogcome howl! 04:27, 9 October 2013 (UTC)
I see it everywhere. Every page of Wikipedia. I have tried Firefox and Chrome so it is not my browser, and I have tried two different computers on different IP addresses, so it isn't my computer. (And there is no field for "location of problem" so how am I supposed to know that's needed?) There should be no problem duplicating it. Just click that "cite" button and try it.-- Brainy J ~✿~ (talk) 13:12, 9 October 2013 (UTC)
I think you may be misunderstanding how Citation Bot works. Citation bot operates on existing citations, as far as I know. For example, if you create a template like {{cite doi|10.4035/foo.bar.20130915}} in an article and then Preview or Save your article, Citation Bot will come along in a short while and fill in that template for you using the citation information that exists at the web page of that doi. Citation Bot can sometimes also expand existing citations in articles using the "Expand Citations" Gadget, which you can add to your left-side Toolbox by enabling it in Preferences. I have not had much luck with expanding citations using this Gadget on articles, but it works very well for me on cite doi templates. Citation Bot does not just swoop in and fill in partial citations in articles after you have edited them.
Citation Bot is intended to work as described in User:Citation_bot#Function, though I think that some of those functions do not work right (JSTOR doi values, for example, are problematic because of a problem that appears to reside on the JSTOR end of things). I don't think Citation Bot can expand citations based only on a URL, as you hope it will. I've never tried it with an ISBN.
And one last note: where is this "Cite" button that you are clicking? Editors work in many different ways. I don't see or use a "Cite" button in my editing interface. Sorry to be obtuse and blind, but I just add citations by hand. – Jonesey95 (talk) 22:09, 9 October 2013 (UTC)
I'm guessing BrainyJ's using Wikipedia:RefToolbar/2.0, since the throbber is from File:RefToolbar spinning throbber.gif. It's enabled for all editors by default, and does have a "Cite" button, but that doesn't use this bot. Perhaps BJ's has javascript blocked? LeadSongDogcome howl! 03:30, 10 October 2013 (UTC)
Nope, but thanks for linking that since I now know that it's Wikipedia:RefToolbar/1.0 that is bugging out on me. It's not easy to figure out what the names of all these things are. -- Brainy J ~✿~ (talk) 13:22, 11 October 2013 (UTC)
Please send your bot over there. I created a small problem. Thanks. 7&6=thirteen (☎) 00:32, 10 November 2013 (UTC)
Fixed. In the future, the best thing to do is to copy and paste the DOI from the article or its web page. – Jonesey95 (talk) 00:50, 10 November 2013 (UTC)
I think that part of the problem is that 88-85957 - 36 is not a valid ISBN no matter how you hyphenate or space it: there are only nine digits. Even making it up to 10 or 13 by putting 0- or 978-0- at the start does not make it valid (ISBN0-88-85957-36 Parameter error in {{ISBN}}: checksumISBN978-0-88-85957-36 Parameter error in {{ISBN}}: checksum); but if we assume that it's the last digit that's missing, the only possibility that would make a valid ISBN is a 6, i.e. ISBN88-85957-36-6 --Redrose64 (talk) 23:24, 3 December 2013 (UTC)
That looks correct, as per OCLC43082665. I note that it's also available in an English edition as OCLC51769082. However, it is clearly wrong for the bot to insert an invalid isbn, even if it is just reformatting another invalid isbn. If it can't verify it, it should flag it as invalid. LeadSongDogcome howl! 23:46, 3 December 2013 (UTC)
There will soon be code in the CS1 module that checks ISBNs for characters other than 0-9, hyphen, and X. ISBNs with extraneous characters in CS1 citations will be flagged as erroneous. Citation bot might be able to fix some of these errors, but simply stripping out spaces and other characters may not result in a valid ISBN; a common invalid ISBN entered by editors lists both the 10- and 13-digit ISBNs, like this: Title. ISBN0-521-45834-X, 9780521458344. {{cite book}}: Check |isbn= value: invalid character (help)
The bot will need to be clever. If the bot can't create a valid ISBN, it should leave it alone (it will already be flagged as an error by the CS1 module). – Jonesey95 (talk) 00:31, 4 December 2013 (UTC)
Extraneous spaces or misplaced hyphens in ISBNs do not render them invalid: for example, as demonstrated above, ISBN88-85957-36-6 is valid - but so are ISBN8885957366ISBN8-8-8-5-9-5-7-3-6-6ISBN88-85957 36-6ISBN88 85957-36 6ISBN88 85957 36 6 which are all parsed as exactly the same valid ISBN - it is only the digits themselves which are significant to an ISBN parser. However, ISBN 88-85957 - 36-6 is technically valid (because the ten digits taken together are valid), but the MediaWiki parser chokes on spaces occurring adjacent to hyphens and so won't mark it up as an ISBN. --Redrose64 (talk) 11:00, 4 December 2013 (UTC)
The good news, I suppose, is that the bot drew human attention to an error which had been in place for six years, since this edit initially added the reference to the article. LeadSongDogcome howl! 14:46, 4 December 2013 (UTC)
Sounds like this is resolved; unmark if not. Martin(Smith609 – Talk) 18:35, 18 January 2014 (UTC)
The ADS database record seems rather bizarre at Bibcode:2005JChPh.123c4103G, showing "pp. 034103-034103-7" for a 7 page article. Crossref just shows it as page 034103.LeadSongDogcome howl! 18:48, 14 December 2013 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
the bot fails to create the citation template for doi 10.1161.2FCIR.0b013e318214914f
What should happen
it is pulling the correct citation, but for some reason fails to make the template
Relevant diffs/links
{{cite doi|10.1161.2FCIR.0b013e318214914f}}
Replication instructions
try to generate citation for 10.1161.2FCIR.0b013e318214914f
We can't proceed until
Agreement on the best solution
Requested action from maintainer
I have used this bot hundreds of times without seeing this problem before. I wish that someone more knowledgeable than me could explain to me why the bot would fail to operate properly in this instance, and advise me about what I should do. I could manually make the citation but there may be a bot bug here.
This is user error. The DOI includes bogus hidden characters right before the / Try this:
Attention: This template ({{cite doi}}) is deprecated. To cite the publication identified by doi:10.1161/CIR.0b013e318214914f, please use {{cite journal}} (if it was published in a bona fide academic journal, otherwise {{cite report}} with |doi=10.1161/CIR.0b013e318214914f instead.
Ah, thank you. I see what I did incorrectly. Blue Rasberry (talk) 12:43, 15 December 2013 (UTC)
I first remove vandalism from this page, then when I clicked expand citations in my tols meny, the bot "undid my edit. (my first edit, the bot's edit and my second edit)
We can't proceed until
Agreement on the best solution
Looks like a race condition (a.k.a. wp:edit conflict). All four edits were inside a 4 minute window, and that is a large, extensively referenced article which takes quite a while to process. The bot didn't exactly undo your revert, it simply started from the same (vandalized) version as you did, and by the time it committed, you had jumped in with your quick revert. It's rare but it does happen. Unfortunately, it happens disproportionately often on these massive articles.LeadSongDogcome howl! 13:54, 20 December 2013 (UTC)
I suppose the bug is arguably with the citation template?--Elvey (talk) 01:20, 22 December 2013 (UTC)
I don't know much about templates, but it appears that the underlying {{PMC}} template does not like it when spaces are included in the parameter value. In the above example, spaces were added before and after the PMC ID, resulting in something that looks like item C below:
If the bot inserted the spaces in the curly brackets, it should probably stop doing so. I suppose someone with template programming knowledge could also change {{PMC}} so that it was more tolerant of white space. – Jonesey95 (talk) 15:45, 22 December 2013 (UTC)
Thanks, Jonesey95 and Redrose64 !--Elvey (talk) 02:46, 24 December 2013 (UTC)
Link to blocked editor
I have a question about the edit summary of this edit by the bot: [106]. There is absolutely nothing wrong with the edit itself, and I'm not objecting to it in any way. However, the edit summary refers to an editor who is currently blocked, and so I have a question as to what the link to the username indicates. Does this mean that someone who is blocked can, nonetheless, activate the bot to make an edit? --Tryptofish (talk) 17:04, 8 September 2013 (UTC)
The bot currently allows trolls to spoof usernames and make abusive edits via edit summary vandalism. It was just used to fake my username as if I activated it on Niggers in the White House.
Secure the ability of who can launch the bot to prevent username spoofing.
Another possible change, if the maintainer is unwilling to add this feature: Add a note like "(username not verified)" when the bot is not run by a logged-in user. – Jonesey95 (talk) 03:59, 17 September 2013 (UTC)
The 'bot has no means to know if the user is logged into Wikipedia. For all we know, this guy could be editing the white house article just to disrupt Wikipedia. K7L (talk) 15:15, 17 September 2013 (UTC)
Following the discussion that I started above at #Link to blocked editor, I find it significant that other editors are now raising what has ended up being a very closely related issue. I really wish that there were more interest amongst bot developers in finding solutions to it, because I think that it is an issue that will not go away. It's probably the case that most bot developers just don't believe that it's a significant enough problem to devote time and effort to, but I believe that I am not alone in the community in wishing that a solution would be found. --Tryptofish (talk) 18:43, 17 September 2013 (UTC)
Let's reach consensus: What's better: to allow users to associate themselves with a bot's edit, so that genuine questions can be addressed to them, or to force the user that triggers a bot run to remain anonymous, so that the only way of leaving unusual edit summaries is to make an edit one's self? Martin(Smith609 – Talk) 07:31, 27 September 2013 (UTC)
There are issues both ways. On one side, the requesting name can be falsified, which could give the appearance that banned editors are editing articles by bot-proxy. Or the falsified name could be used to tie an editor to an article they've never edited for embarrassment reasons. On the other hand, by not linking any name, we have no idea who's requesting the bot edits. The way to fix this is simple: the request page needs to be on a Wikimedia site that uses the Single User Login system so that the request can be tied to an account or an IP for public listing in the edit summary. That can't be done at the moment though. Imzadi 1979→ 07:41, 27 September 2013 (UTC)
To the question of which is better, I see a contradiction. On the one hand, there is the argument that bots, by definition, only make constructive edits, never disruptive ones, so a bot edit cannot do harm, no matter who triggers it. On the other hand, the question here indicates that other editors might have questions about a bot edit, where they would want to ask something to the editor who triggered the bot. Why would anyone want to question an edit that cannot, by definition, ever do any harm? You can't have it both ways. If it doesn't matter who triggered the bot, why do we need to identify who did it? If we need to identify who did it, why does it not matter if we identify them correctly? --Tryptofish (talk) 18:39, 27 September 2013 (UTC)
While the bot can't determine if a user is logged in, it could conceivably check that his latest contribution was recent (one day, perhaps). It also could advise users that it has made edits in their name so that they can repudiate spoofed ones. A spoofed user could respond reasonably, as by excluding the bot from that target page. It is also conceivable to provide a way for the user to tell the bot not to edit on his behalf (e.g. in prefs or on the userpage).LeadSongDogcome howl! 22:04, 27 September 2013 (UTC)
Those are very good ideas – for any bot. The more I think about, the more I think it's just a matter of time until, following some contentious arbitration case, an editor who was topic-banned by the arbs will be accused of making a bot edit in the area of the topic ban, with drama to follow. --Tryptofish (talk) 22:28, 27 September 2013 (UTC)
"Expand citations" gadget in toolbox linked to old version?
Clicking on Toolbox -> Expand citations (the "Citation expander gadget" in Preferences -> Gadgets) results in edits by version 442 of the bot, even though version 458 is current.
Create a cite doi template, let the bot auto-fill it (it uses v458), then remove some parameters and re-fill it using the Citation Expander gadget in your left-side toolbox.
We can't proceed until
A specific edit to the bot's code is requested below.
Requested action from maintainer
Fix the gadget so that it uses the current version.
Operator: Bot operator's feedback on what is feasible
Requested action from maintainer
Correction of error
The author parameter was not specified according to the template documentation. Garbage in, garbage out. It'd be nice to handle this error more gracefully, but this isn't something I can do at the moment. Martin(Smith609 – Talk) 21:30, 29 January 2014 (UTC)
I have fixed the above citations and a few others in the same article. Randykitty: for future reference, wikilinks should not be used in author parameters. – Jonesey95 (talk) 03:53, 30 January 2014 (UTC)
Thanks for that, Jonesey95! I know that we have the parameter "authorlink" for these cases. However, I suspect that there will be many, many cases where editors have used wikilinks in author parameters. That means that all those references will be "garbage out" after citationbot cleans an article up. The syntax may have been wrong, but at least the references were correct, so the "garbage in" part was, IMHO, less serious than the "garbage out" part, rendering correct syntax but incorrect references. Smith609, I immensely appreciate the good work you are doing with this bot, but I think your dismissal of this error is a bit too fast. I don't know zilch about bot programming myself, so I have no clue how much work this is, but perhaps the bot could leave citations that have wikilinks in the authorfields alone and tag them for human attention? Or just simply remove the wikilinks first? We would then lose those links, but I think that would be preferable to introducing incorrect references. Thanks! --Randykitty (talk) 09:47, 30 January 2014 (UTC)
Randykitty, for what it's worth, I have edited many thousands of {{cite doi}} and {{cite journal}} templates with various errors in them, and I have seen astoundingly few author parameters with multiple authors and wikilinks in the same parameter, which is the situation required for this bug to manifest itself. There is an amazing variety of ways in which WP editors have constructed these templates, some of which "work" but are malformed or generate error messages, and others of which do not work. Among this wondrous diversity, I do not recall seeing this particular variety of error before. I'm guess I'm saying "don't worry too much about it."
That said, I will put in a feature request that wikilinks in author parameters generate a hidden category in these citations. That way, editors might be able to fix the problems before they are whacked by Citation bot. – Jonesey95 (talk) 15:38, 30 January 2014 (UTC)
Excellent, glad to hear that this is a rare error, guess I underestimated the "median WP editor", they do an even better job than I thought :-) Thanks. --Randykitty (talk) 15:48, 30 January 2014 (UTC)
A separate authorlink parameter is not an adequate substitute for wikilinks in the author field when only part of the author field should be linked (e.g. when it contains names of multiple authors). And separating everything out into firstn lastn authorn-link, while usually the right choice, is also not always an adequate substitute, for instance when some of the authors are institutions rather than people or when the preferred formatting of author names is different than what can be achieved by separating them out. So I object to the claim that having wikilinks within the author field should be disallowed or treated as garbage; it seems a perfectly acceptable use case (if rare) to me. —David Eppstein (talk) 08:31, 31 January 2014 (UTC)
But the 'author' field should not contain multiple authors; |authors= would be much more suitable in these cases. Martin(Smith609 – Talk) 10:00, 31 January 2014 (UTC)
Nevertheless, |author= is an alias of |authors= and the underlying template code treats them identically. Like it or not, there are large numbers of citations that contain a single author parameter that were generated for example using Diberri's template filler tool. Other editors have then imbedded wiki links into the author lists. Not everyone likes to use verbose "first1, last1, ..." parameters. Using a single author parameter to store the author list is much more compact and less cluttered. Boghog (talk) 11:00, 31 January 2014 (UTC)
You don't need to split every author into lastn/firstn pairs; corporate authors can go into authorn parameters quite happily. It is perfectly legitimate to put e.g. |last1=Doe|first1=John|authorlink1=John Doe (biologist)|author2=United Nations Secretariat|author3=Ban Ki-moon|authorlink3=Ban Ki-moon. See also Template talk:Citation#Lastname, Firstname. --Redrose64 (talk) 11:12, 31 January 2014 (UTC)
This edit was performed (but not saved) by opening an edit window and clicking on the "Citations" button.
We can't proceed until
Agreement on the best solution
Requested action from maintainer
Spank bot, please.
Diannaa, your link merely opens the edit window for the article. Looking back through the page history, I see two recent citation bot edits - this and this. Which one is in error; or is it both, or neither? --Redrose64 (talk) 15:07, 6 May 2012 (UTC)
Sorry for not replying sooner. It looks like the first of the two edits is where most of the damage occurred; the bot removed a bunch of citations outright. But both were bad; the second edit totally removed the citation named "notes". User:Pumpkin Sky undid the two edits to restore the lost citations. -- Dianna (talk) 18:46, 14 June 2012 (UTC)
Better algorithm for duplicate references or stop removing duplicates.
The bot went wrong by removing the second of two references. The article uses a list of named references in the reference section between reference and /reference. The reference in the text can be referred to by name, but the full reference needs to be in the reference list. StarryGrandma (talk) 00:08, 16 June 2012 (UTC)
Two different references with the same name seems to confuse the botAManWithNoPlan (talk) 04:25, 20 June 2012 (UTC)
Remove or alter the code to not change «/», or possibly make exceptions for cases as shown in example.
It changes them in Russian too. If an article title has « in I would hope that the bot would keep it like that. The English translation doesn't have the offending character in, but the bot shouldn't alter the Russian. Secretlondon (talk) 01:20, 1 August 2012 (UTC)
I'm pretty sure that this behaviour is in order to comply with style guidelines that recommend the use of straight quotes. Could you consult the WP:MOS and confirm how it suggests that quotes are handled? Thanks. 131.111.184.106 (talk) 07:32, 18 August 2012 (UTC)
Fix the bot, or perhaps remove support for {{cite pmc}}
The Citation bot does all sorts of horrible things with {{cite pmc}}, but we closed those bugs when {{cite pmc}} was deleted. Some one recreated {{cite pmc}} for no good reason. AManWithNoPlan (talk) 02:02, 25 September 2012 (UTC)
Can you point to some articles that are affected by this behaviour, so that I can investigate? Thanks. Martin(Smith609 – Talk) 08:34, 20 October 2012 (UTC)
That given example is only transcluded one place, in a userspace subpage: Special:WhatLinksHere/Template:Cite_pmc/1293471 Perhaps the bot shouldn't be operating on pages outside articlespace (except for the specific areas of templatespace needed)? LeadSongDogcome howl! 22:13, 22 October 2012 (UTC)
There are several instances of wikilinks to these templates (not transclusions) in user:NTox/CSD log that may be connected. LeadSongDogcome howl! 04:51, 23 October 2012 (UTC)
Seems to be 17 instances over the past year or so. Some of these AMWNP has corrected, only to be reverted by the bot! We could just pipetrick the uses of cite pmc with equivalent uses of cite pmid, or (better) of cite journal. It isn't that commonly used anyhow. LeadSongDogcome howl! 20:04, 21 August 2013 (UTC)
A very special case. The vast majority of cites to books.google.com will not be to pages about that service. Even then, few will be naked URLs. Fixed the wikitext.LeadSongDogcome howl! 22:01, 27 February 2013 (UTC)
I agree that it's very much an edge case, but it may be worth excluding the bot from the Google Books article (or the topic area), if this gets more annoying. — daranz [ t ] 22:04, 27 February 2013 (UTC)
Adds last2 but not first2, and with pre-exisiting co-authors still present, the last2 name is then shown twice.
Existing entry has last1=... , first1=... and coauthors=... . Bot adds last2=... , but not first2=... . The pre-existing co-authors=... part remains, so duplicate surname shows for the "last2" name. Requires manual addition of the first2=... part and manual deletion of the coauthors=... part.
This seems like an ongoing theme. Articles have inconsistent mix of uses between |authors=, |author1=/|author2=..., |last1=/|last2=, and |coauthors=. These mixes do not seem to be consistently resolved, either with the bot or manually. There should be a discussion on this question in a broader forum. LeadSongDogcome howl! 14:13, 1 March 2013 (UTC)
I remember seeing "coauthors (deprecated)" somewhere on my travels. There needs to be a simple way to convert existing coauthors data to whatever is the new way of managing extra names, as well as some sort of error message shown when editors try to use the coauthors element within new citations. -- 212.139.104.161 (talk) 14:59, 1 March 2013 (UTC)
Yes, but many editors still use it. Trying to convert multiple names in the coauthors field automatically is not going to be a simple task. --— Gadget850 (Ed)talk 15:03, 1 March 2013 (UTC)
An easy first step would be getting rid of all the unpopulated instances of |coauthors=| (and variations for whitespace). Then we'd at least know how many substantive instances we have to deal with.LeadSongDogcome howl! 17:49, 1 March 2013 (UTC)
RefToolbar and maybe ProveIT support coauthors. Probably other tools as well. --— Gadget850 (Ed)talk 19:31, 1 March 2013 (UTC)
The bot, when processing this article, appears to have removed re-used named citations that were embedded in tables, as well as remove content from those tables. I don't see any immediate logic for why this was the case.
A specific edit to the bot's code is requested below.
Requested action from maintainer
Review of what happened here.
(talk page stalker) - It appears that the bot incorrectly deleted two instances of <ref name="foobar" />...</ref> - GoingBatty (talk) 22:39, 14 March 2013 (UTC)
Thanks for your interest, but grouping the references is a bit premature as the article is only just starting. Please don't group the references in any of the other articles which I am currently in the process of creating because they are only drafts, at this stage. --Rskp (talk) 06:41, 30 March 2013 (UTC)
This refers to this edit and this one. The main question is: why is the bot making edits in user space at all? --Redrose64 (talk) 14:19, 30 March 2013 (UTC)
Have undone them, so no probs. But definitely weird. Do you know an editor who can fix the bot? --Rskp (talk) 00:34, 31 March 2013 (UTC)
It's easy to stop the bot editing a given page using a template (I think ("{Bots| deny=Citation Bot"} will do it). Martin(Smith609 – Talk) 14:51, 2 June 2013 (UTC)
copy and paste a DOI from the Journal of experimental biology website. The bot claim the DOI won't work even though it is correct to the human eye. When editing, there seems to be an extra character between the / and jeb, or some kind of unseen markup. Deleting this character and the / then putting the / back in on the second link, and the bot worked, I have left the first doi broken.
We can't proceed until
Operator: Bot operator's feedback on what is feasible
Wikipedia talk:Articles for creation/Benny Award had been moved to mainspace as Benny Award. The bot then overwrote the redirect that had been created as part of that move. This meant that there were two similar pages (although not identical) where there should only have been one.
Looks like a race condition. Not sure how I can resolve? — Preceding unsigned comment added by Smith609 (talk • contribs)
Would it not be possible to have the bot check the target article history after saving a revision? If edited within the runtime window (plus perhaps 2 minutes) there is a substantial risk of a race, so self-revert and restart.LeadSongDogcome howl! 19:39, 19 August 2013 (UTC)
Great idea! Thanks! I'll implement this in the next update. Martin(Smith609 – Talk) 06:59, 20 August 2013 (UTC)
Another case. I gave up looking for these some months ago, but it's clearly still happening - here's another. --Redrose64 (talk) 15:27, 7 December 2013 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
One of the {{cite news}} templates had two of its parameters placed between valid HTML comment markers, i.e. <!--| archiveurl=http://www.webcitation.org/5nospLAte | archivedate=February 25, 2010--> the bot has taken the open-comment marker <!-- and decided it was "unused data", moved it after the close-comment marker
Does this systematically occur when going from, say, two digits to three? Or could it be that the data the bot is using is from a pre-publication version that has different pagination? Martin(Smith609 – Talk) 14:30, 2 June 2013 (UTC)
I really don't know. The point is that before the bot edit, the |pages= parameter held a meaningful value, even if, as a single page number, it should actually have been in the |page= parameter; but after the bot edit, it is no longer meaningful, because the number representing the end of the page range is lower than the number representing the start of the range. I have found out that the second example was probably supposed to be |pages=3–11 which suggests that it has been truncated; but using the same method on the first, I came up with |pages=87–96 which cannot become |pages=97–10 by simple truncation. --Redrose64 (talk) 22:56, 4 June 2013 (UTC)
Hi there. This bot is blocked and inactive, as a result its bot flag will soon be removed to try and tidy up our list of accounts with bot flags. If you have any problems get in touch with me on my talk page! ·Add§hore·Talk To Me! 10:42, 28 May 2013 (UTC)
Why is the bot blocked? I can't work out how to find this out. Martin(Smith609 – Talk) 14:25, 2 June 2013 (UTC)
I miss this bot. It helped me every day. Blue Rasberry (talk) 19:09, 31 May 2013 (UTC)
Same here :-( Meodipt (talk) 07:05, 1 June 2013 (UTC)
What the heck?? This is one of the most useful tools here. Please have this fixed soonest. Not having this tool creates a HUGE amount of extra work. --Randykitty (talk) 09:50, 1 June 2013 (UTC)
well we've got an editor that wants to list 33 authors in a citation because the new system lets them. meanwhile, other editors use diberri's style which shows 3–5 authors. i remember the diberri crowd getting upset a while back because authors 6–9 were being added. i can't imagine how they'll respond when the size of their articles doubles for all the tertiary authors. —Chris CapocciaT⁄C 10:44, 1 June 2013 (UTC)
The essential problem is that the maintainer of this bot hasn't apparently fixed any bugs since August, and hasn't even commented on any bug reports since October. In that time about 40 bug reports have been filed. One of the most frequently recurring is the incompatibility with expanded Lua author lists, where the bot is destructively reverting the efforts of Wikipedians, though that's certainly not the only issue. Essentially, it appears that this bot has been abandoned by its author. If someone else wants to take up the mantle of fixing and maintaining this bot, it appears that the source code is here. Dragons flight (talk) 17:52, 1 June 2013 (UTC)
Went ahead and updated the userpage to reflect the inactivity. RIP Citation bot! --Nathan2055talk - contribs 18:50, 1 June 2013 (UTC)
I just emailed the bot owner and to my bafflement, nobody had tried that yet, so he was blissfully unaware of these issues (apparently having been less active on WP). Unfortunately, he's leaving for a vacation, so he won't be able to help anytime soon. I must say that I find the cavalier way in which this important tool has been blocked and, by blanking the user page, people are apparently simply putting this bot aside, highly disconcerting. I mean, really, this guy is currently less active and people just keep posting notices here and nobody even tries to email him directly? If this were some minor tool, OK, but this is a hugely important bot and its non-functioning is causing broken templates ("will be completed in the next few minutes") all over (I think it also does the cite pmid template, not just the doi one). I realize that people editing articles on manga, computer games, or sports figures will not be inconvenienced, but anybody editing in the sciences will be severely handicapped by the absence of this bot. --Randykitty (talk) 21:03, 1 June 2013 (UTC)
This user's unblock request has been reviewed by an administrator, who accepted the request.
I believe this issue to be resolved. As such, it should be safe to unblock the bot. Please let me know if any further changes are required. Martin(Smith609 – Talk) 14:18, 2 June 2013 (UTC)
Are any of the bugs that were causing the bot to be blocked (getting into edit wars with users and removing correct information from citations) fixed? Because "owner is going on vacation and still won't look at anything for a while" sounds like the opposite of resolved to me. —David Eppstein (talk) 16:27, 2 June 2013 (UTC)
Bug again[112]: not resolved (note that maintainer does not have a clue about "CS1", while the discussion has multiple links).
Bug [[#A {{cite doi}} reversed into error]] [113] was not a section (misformed when entered). It now is. Still open, and CS1 related: the bot re-introduces a CS1 error.
Clearly the maintainer has not grasped the CS1 effects. They are mentioned multiple times on this page, including the module and help links. Also I find it strange that we are supposed to mail, while this page is automated for bug reports. -DePiep (talk) 17:35, 2 June 2013 (UTC)
Bug #Bot assumes that pipe is part of URL declared "won't fix" after bad reasoning. (probably introduces errors since CS1 is introduced; at least in the example. -DePiep (talk) 17:37, 2 June 2013 (UTC)
Bug #unused_data: no response. It introduces a CS1 error. -DePiep (talk) 17:43, 2 June 2013 (UTC)
A temporary fix is to add: "{{bots|deny=Citation bot}}" to the article. I strongly agree that full CS1 support needs to be implemented, but at the same time, blocking the bot is causing more problems than it solves. Boghog (talk) 19:26, 2 June 2013 (UTC)
Doesn't sound "temporal" to me. Some CS1-issues (introducing errors) are ignored while declared "all fine". The automated bug reporter was removed (by Nathan2055) and not restored [114]. Bot owner now asks to mail for urgent fixes, and asks pre-announcements for lager exterior changes (as was going over to CS1 I assume). Strangely enough, such a "please adapt the bot to the new situation" are many on this page, but were not read as such (even if they were made post-change with diffs). Meanwhile, yesterday some 12 bugreports have been archived [115]. Some have been changed bugstatus yesterday, but without elaborating the discussion with post-CS1 effects. -DePiep (talk) 07:28, 3 June 2013 (UTC)
I would like to note that this bot is user-activated (even when it fills in cite doi templates, because that us triggered by editor an editor adding such a template). I always use the "history" link that is displayed after the bot has run and look at the changes. If I use a bot like this (reflinks is another example), it is my responsibility to check that the bot did not introduce errors. Of course, ultimately the bot should run flawlessly, but at this point, not having it is worse. --Randykitty (talk) 07:55, 3 June 2013 (UTC)
There are examples in this buglist that the bot came back and re-introduced errors after I corrected them. What you are saying is more like WP:AWB, where you have to click personally to enter a proposed edit. This bot is not like that. You cannot make me responsible for what the bot does automatically, sometimes days after my edit. Also this remark contradicts the advice to enter {{bots|deny=citation bot}}. -DePiep (talk) 08:29, 3 June 2013 (UTC)
Automated botreporter reinstalled. It was not removed by the bot maintainer. [116] -DePiep (talk) 08:43, 3 June 2013 (UTC)
Quite simply: since the bot relies on {{citation}} templates (basically, that was {{citation/core}}, now Module:citation/CS1) it should comply with that one. It should not introduce errors because of this omission. Putting the load and responsibility on the invoking editor for this is incorrect attitude -- that is not why is is a bot. -DePiep (talk) 08:29, 3 June 2013 (UTC)
This user's unblock request has been reviewed by an administrator, who accepted the request.
Please also unblock User:Citation bot 1, which was missed when the above unblock request was actioned. This block is stopping me from fixing bugs! Thanks. Martin(Smith609 – Talk) 09:23, 17 August 2013 (UTC)
Accept reason:
this one unblocked as well Graeme Bartlett (talk) 21:54, 22 August 2013 (UTC)
remove line |author1=C.A.R.. Wait for bot to restore it.
We can't proceed until
A specific edit to the bot's code is requested below.
Requested action from maintainer
Fix or stop
I added {{nobots}} to that citation so that at least we can have a correct citation while this bug is open. Once the bug is fixed, bot access to the cite should be restored. —David Eppstein (talk) 07:04, 24 June 2013 (UTC)
Deleterious: Human-input data is deleted or articles are otherwise significantly affected. Many bot edits may require undoing.
What happens
{{Cite doi/10.1021.2Fed044p475}} has had a problem from the start. It places the author's name in both the author1 parameter and the last1/first1 parameters, which some of us catch with a code in our .css file.
Please find an adjustment for this bot that will cause it to respond appropriately to author parameters in Cite (whatever) templates (especially the Cite doi ones). Note... if you place the following code in your vector.css file (or whatever skin you use to edit), then you will see the error code for this problem at the link showing what happens specified above.
The bot incorrectly adds a reference of a book review to a citation of a book. In the instance cited below, it added a reference to a review in Nature to a citation of De Lacy O'Leary's How Greek Science Passed to the Arabs. As is not uncommon for reviews, the review had the same title as the book under review.
I've just noticed that this error happened to a citation, not to a cite book template. Apparently (or so it seems from the documentation) Citation Bot will not make such changes to cite book templates.
However, in the same article Citation Bot changed a bunch of cite book templates to citation templates. This opens the way for a two step corruption process in which on round one a cite book template is changed to a citation template, and in round two the citation template is corrupted by adding a review. SteveMcCluskey (talk) 17:26, 9 July 2013 (UTC)
Looks like a
This is a database error. The review is listed on the AdsAbs database with the author given for the book. As such, there's no way to discriminate a book review from a journal. Can you think of any way to solve this? Martin(Smith609 – Talk) 16:53, 17 August 2013 (UTC)
Look, I'm sure I'll get an earful about why the hell I even bothered to add {{thinsp}} ( thinsp is a "thin space"). But whether or not that was a good idea, the bot certainly shouldn't do what it did -- no doubt this wanders into an unanticipated corner case in some regex.
Interesting. The xml from pubmed for that cited source shows:
<ArticleTitle>"No longer Gage": frontal lobe dysfunction and emotional changes.</ArticleTitle>
That should give |title="No longer Gage": frontal lobe dysfunction and emotional changes. as the value to be automagically inserted by the bot. Your version substituted &squot; for the leading ", presumably to improve legibility, but breaking the automatic check against published metadata. I note that of the published works citing that paper, many mangle the quotation marks in even stranger ways. This may be a special-enough case to be worth avoiding cite journal entirely, manually writing the cite doi subpage instead. The problem only arises when the title properly begins with a double quotation mark, invoking the substitution of the single... LeadSongDogcome howl! 22:03, 6 August 2013 (UTC)
I don't think you're understanding the problem -- nothing to do with the title field. I put {{thinsp}} in 4 fields [121] -- last1, last2, last3, title; also, in title I changed double quotes to single quotes. The bot came in after me and, er, adjusted [122]. In last1, 2, 3 it turned {{thinsp}} into {{T.} (with only one brace on the right!), but it left the {{thinsp}} in the title field alone.
Later [123] I tried using   instead of {{thinsp}} and the bot left that alone. So there's something about the {{thinsp}} which is confusing it -- except in the title field!
But your comments raise some interesting points:
1. Surely there are plenty of times that a title contains double-quotes; since the bot knows that the title itself will be enclosed in double-quotes by the cite template, it ought to be made smart enough to change the doubles to singles, and even (as I did manually) add a thinsp at left or right or both as needed. (Whether you want to detect nested quotes in the title and handle them, I leave to you. This well illustrates why the British system of single-inner, double-outer really is better -- under it, to embed a quote within a quote doesn't require "toggling" -- or 'toggling' -- all the inside quotes.)
2. You mentioned "manually writing the cite doi subpage instead". Sure that's fine for some situations, but how to keep the bot from coming in after and fooling with what the human's done? I expected there to be some magic bot_hands_off parameter I could add to the doi cite, but there doesn't seem to be such a thing. (Perhaps in the presence of such a keyword the bot might add, as a <!- -> comment, what it would have put in that field, so a human can judge for himself.)
3. Or, just the fact that a human edited at all might act as an implicit instruction to the bot to keep hands off. In this case, again, the bot might still come in and add a <!- -> comment at the end of fields it "disagrees" with.
4. I realize that (3) above may raise some issues when the source databases the bot is working from update their records i.e. I can imagine that, right now, if an external database makes a spelling correction in a bibliographic record, the bot might propagate that to the doi cite -- if any human intervention blocks the bot permanently, then that cite would never get updated. One way to handle this is the following: any time there's a human edit, the bot comes in and adds a <!- comment to the effect, "Live value is human-supplied. I would have supplied value X." Later, if there's a database update changing one of the field values, the bot looks at that field in the doi cite. If there's no <!- such as I just described, the bot can assume the field was bot-supplied, and therefore it updates the field with the new value from the external database. But if there's a <!- comment, then bot keeps its hands off. Of course, a not-uncommon special case might be that the new value from the external database matches the human-supplied value, because the human knew early that the value was wrong and corrected it. This might actually be worth detecting so the comment can be changed in that case to "Database value finally reflects human-supplied value".)
OK, enough commentary! EEng (talk) 23:24, 6 August 2013 (UTC)
No, I understood that problem in the forenames, I just couldn't identify its source. I've looked at the code, but I can't claim to have much understanding of it. It seems to be buried several layers down in unicode conversions, and PHP isn't my bag. With regards the toggling issue, it strikes me that the proper place for that functionality is in the templates. If that can't be done, then such failing should be documented so that all editors, human or bot, know that they have to pre-toggle the title quotation marks and in what fashion to do so in order to match the template's expectations.
To have the bot keep "hands off", see User:Citation_bot#Stopping the bot from editing. It's pretty simple. However, there's some merit in saying "for cite doi subpages, leave completed templates untouched if they have been human-edited following the bot's initial filling-out". My personal perspective is rather more drastic: eliminate cite doi template subpages entirely. Just subst them. They are a cute-but-bad idea. That seems, however, not to be a widely-held view and I do not really wish to beat a wp:DEADHORSE. LeadSongDogcome howl! 14:53, 7 August 2013 (UTC)
If the bot is erroneously adding a DOI, author, etc to a citation, and you want to stop it adding the data again, you need to put a comment in place of the appropriate parameter – because the bot will not overwrite existing data.
Well, clearly the bot will overwrite existing data -- maybe it should read will not overwrite a value that contains a comment??? I'm gonna try that experiment.
For those of us not steeped in this corner of WP: what is the significance of that? What is this metadata fed to? Is it different if a template e.g. { { thinsp } } is used instead? EEng (talk) 22:45, 7 August 2013 (UTC)
{{thinsp}} is inconsistent with the Cite Doi formatting requirements. If you wish to use this template, then unfortunatly you won't be able to use Cite Doi - using Cite Journal in an article will be fine, and the bot won't alter the forename punctuation. Martin(Smith609 – Talk) 17:34, 17 August 2013 (UTC)
Why should the bot ever edit more than once?
Excuse me, but can you point to the "Cite doi formatting requirements" you're talking about? And assuming, for the sake of argument, that I introduced something to the cite that isn't normal or acceptable for some obscure reason, does that really justify the bot's completely screwing up the cite by causing the braces to become unbalanced? I have taken the liberty of removing the "invalid" tag. And please don't archive discussion threads until others have a chance to comment on the "resolution". EEng (talk) 03:33, 18 August 2013 (UTC)
These formatting "requirements" appear to be rules meant to lend uniformity to otherwise arbitrary choices (e.g. "Smith, John Albert" vs. "Smith, J.A." vs. "Smith, J. A."), which is really something that the editors concerned can work out among themselves. And even assuming that there's some good reason for cites to be restricted to this format, that doesn't explain why the bot goes haywire and messes up the cite completely should a cite, for whatever reason, not conform to it.
Look, what I don't understand is why the bot needs to ever touch a template more than once? A template appears that's just a bare doi (or pmid, or whatever). The bot fills it in. Why can't that be the end of it, as far as the bot is concerned? If, after that, editors have their reasons for adjusting the template, why should the bot keep jumping back in?
If there's something I'm missing in offering the above reasoning, please explain. EEng (talk) 00:45, 23 August 2013 (UTC)
1) To update the citation 2) To enforce formatting standards 3) To fix mistakes 4) etc... Headbomb {talk / contribs / physics / books} 01:22, 23 August 2013 (UTC)
1) Can you give an example of a citation being updated? I'm not certain I understand what you mean (though I can guess some possibilities).
2) As to "enforcing formatting standards", I addressed that above. Why should the bot override the judgment of human editors?
3) How does the bot recognize a "mistake"? See 2) above.
1) I don't feel like digging through thousands of edits to find an example. Examples of updates are adding arxiv/Bibcodes/DOI/JSTOR/... identifiers as they become available. (See [124] for an example from Bibcode Bot.) 2) Because "the judgment of human editors" in this case very often means breaking citations. If for example, you want your citation to be in "J. Smith (2009), bla bla bla.", this will clash with every other citations that are in "Smith, J. (2009). bla bla bla." format. The cite doi framework needs to be in that format, otherwise citations cannot be relied upon to produce the expected output. If you don't like this, then simply don't use {{cite doi}} templates and instead use {{cite journal}}. 3) If you change Smith, J. to Jason Smith, this is a mistake. If you put the journal title in the volume parameter, this is a mistake. Headbomb {talk / contribs / physics / books} 02:58, 23 August 2013 (UTC)
Saying, "if you don't like the way the bot works then don't use it" isn't really helpful -- that would be an excuse for the bot doing all kinds of silly things right and left. (Remember, this thread began because the bot does something it should never do, under any circumstances, which is to leave the source text with unbalanced braces.) The discussion here should be over: What bot behavior is going to be most helpful to the most editors?
I don't see why humans should be any less trustworthy in this context than in any other. No one's saying people should be able to change "Smith, J." to "Jason Smith", but as seen elsewhere on this page ([125], and see also the talk about quotes near the beginning of this thread [126]) there are certainly places where humans know better than the robot. Here [127] we're told that "the bot will not overwrite existing data" -- is that in fact true, and if so what's the exact logic of that? EEng (talk) 06:46, 26 August 2013 (UTC)
<bump> I would really appreciate an answer per above. What I'm looking for is the precise test by which the bot decides whether or not it should overwrite existing data. Or is the code somewhere I can look at it myself? EEng (talk) 13:51, 29 August 2013 (UTC)
Here's the Source code (link listed on page User:Citation_bot). The function you are probably interested in is if_null_set(). Do note that I'm currently re-building the bot from the ground up, and that the next major release may (or may not) resolve these problems. Martin(Smith609 – Talk) 09:29, 5 September 2013 (UTC)
To save me setting up svn and so on, can you point me to the appropriate module? EEng (talk) 04:57, 6 September 2013 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
Bot added a "last=" parameter when "last1=" was already present, and the value inserted into "last=" is not correct. Also, the bot failed to enter the remaining authors.
Hmmm, perhaps the article PMID10638459 is being conflated with the subsequent Erratum, which is mentioned in the pubmed record. I won't bother commenting on Springer's idea that readers should pay to read an erratum on something they themselves published in error. LeadSongDogcome howl! 04:17, 24 October 2013 (UTC)
Deleterious: Human-input data is deleted or articles are otherwise significantly affected. Many bot edits require undoing.
-->
What happens
Article has some commented-out access dates, in various forms and with all different dates. Bot changes them all to "<!-- accessdate=2007-08-19 -->", thus losing the original information such as it is.
What should happen
If something is commented out, just leave it alone.
It looks like all comments in references were replaced with the same text. I reverted the whole edit.
Perhaps related to above bug report. In each case, the inserted text was copied from hidden text found as the value of a parameter in a citation template. It was copied over similar hidden texts found as the value of a parameter in a citation template. LeadSongDogcome howl! 03:54, 30 October 2013 (UTC)
Template:Cite doi/10.1038.2Fnchembio.189 was created by, and has only ever been edited by, the Citation bot family. The present version uses {{Cite journal}} which has been given several mutually-exclusive parameters: |first1=L.|last2=Rawat|author2=Rawat|first2=M.|author3=La Clair|last4=Jothivasan|last3=La Clair|last5=Budiarto|first3=J.|author4=Jothivasan|first4=K.|last7=Claiborne|author5=Budiarto|last8=Helmann|first5=T.|last1=Newton|last9=Fahey|first6=J.|author7=Claiborne|last6=Hamilton|author6=Hamilton|first7=A.|author8=Helmann|first8=D.|author9=Fahey|first9=C.
We can't proceed until
Agreement on the best solution
Can you spell out the problem and desired outcome? At a cursory glance the above string doesn't seem to cause an error. Martin(Smith609 – Talk) 14:37, 2 June 2013 (UTC)
{{cite journal}} now uses the new Module:Citation/CS1 core. The new core has more error checking, such as duplicate parameters. See Help:CS1 errors. The older version of this citation was:
Newton, L.; Rawat, M.; La Clair, J.; Jothivasan, K.; Budiarto, T.; Hamilton, J.; Claiborne, A.; Helmann, D.; Fahey, C. (2009). "Bacillithiol is an antioxidant thiol produced in Bacilli". Nature chemical biology. 5 (9): 625–627. doi:10.1038/nchembio.189. ISSN1552-4450. PMC3510479. PMID19578333. {{cite journal}}: More than one of |author2= and |last2= specified (help); More than one of |author3= and |last3= specified (help); More than one of |author4= and |last4= specified (help); More than one of |author5= and |last5= specified (help); More than one of |author6= and |last6= specified (help); More than one of |author7= and |last7= specified (help); More than one of |author8= and |last8= specified (help); More than one of |author9= and |last9= specified (help); Unknown parameter |month= ignored (help)
If you check the history, you will see that the bot added these duplicate parameters. --Gadget850talk 15:32, 2 June 2013 (UTC)
Here's a similar example from a different cite doi template. The template, for some reason, had "author1" and "first1" for the authors' names, which is incorrect but was working. I ran the citation expander bot on it, and it blew up with duplicate parameters. The bot added "last" params for each author, leaving the "author" parameters in place. I don't expect the bot to deal well with a poorly-formatted edge case like this one, but if it can do so without too much trouble, go for it. (BTW, I love the new feature that cleans up the obnoxious and previously time-consuming "displayauthors" Lua error. Great work.) – Jonesey95 (talk) 01:45, 22 August 2013 (UTC)
I have been encountering lots of this problem as well, as can be seen in this diff showing the result of Citation bot's work. This is not an incorrect edge case where "author" and "first" were originally being wrongly used simultaneously. In this case, only "author" was being used, and Citation bot decided to add "last" parameters, causing errors to be displayed on the processed page. This has happened to me many times, although I usually catch it in preview when using the bot to assist with new citations. I think the desired behavior would be for the bot to refrain from adding "last" or "first" to templates with existing "author" fields, would it not? Azaghal of Belegost (talk) 18:24, 14 January 2014 (UTC)
Should be fixed in r520, for good this time! Martin(Smith609 – Talk) 17:24, 2 February 2014 (UTC)
Hi. The "class" parameter of Template:Cite arXiv should only be inserted when it cannot be read off from the "eprint" parameter, but the bot seems to supply one indiscriminantly. It's not fatal, but hurts the readability a bit. Compare the first and the second examples in Template:Cite_arxiv#Examples, for when it's necessary and not. Thanks in advance.
Can you provide a logic that will resolve this? For example, "If the arXiv id begins with a letter, unset the class parameter"? I will be able to integrate the requested function with this information, but am not familiar enough with the arXiv format to try to guess it myself. Thanks, Martin(Smith609 – Talk) 08:48, 17 August 2013 (UTC)
If |class=foobar matches |eprint=foobar/barboo, then remove |class=foobar. That would probably cover 99% of cases. Headbomb {talk / contribs / physics / books} 18:32, 17 August 2013 (UTC)
@Martin: Your idea is probably correct, but I think checking the existence of a slash is surer. Let me elaborate:
Arxiv.org uses two formats, the old one, and the new (i.e. current). In the old, eprint=<class>/<numbers>, and in the new, eprint=<numbers>, or more precisely, <YYMM.numbers>. Since it's two-digit YY, it'll sometime change, and that's the subtlety I indicated above.
In the old format (which had been used for papers submitted by some 2007), you shouldn't add the "class", or if one is already present, class should be deleted. -- Teika kazura (talk) 02:04, 19 August 2013 (UTC)
Great, thanks, I'll implement this in the next major update – hopefully in the next few weeks. Martin(Smith609 – Talk) 07:02, 19 August 2013 (UTC)
Think this is fixed as of r527. Martin(Smith609 – Talk) 07:35, 3 February 2014 (UTC)
Automatic addition of authors beyond original nine puts doi param in a funny place in the source code
A specific edit to the bot's code is requested below.
Requested action from maintainer
place added author/first/last params after the existing nine author params but before the doi param
Here's a diff that shows what I mean. The original citation, created by an older version of the bot, limited the authors to nine (I understand why this happened). With the revised bot's wonderful ability to add more authors, using the citation expander tool to add more authors adds "last10" before the doi, then "first10" and the rest of the author params after the doi. It looks strange when you edit the template, though of course it renders just fine. – Jonesey95 (talk) 01:56, 22 August 2013 (UTC)
revert the diff and try to expand again using citation expander gadget
We can't proceed until
Bot operator's feedback on what is feasible
Requested action from maintainer
Test it to see if it's a citation bot bug or a problem with the XML in PubMed or something else.
I used citation expander on about ten existing cite pmid templates with just nine authors, expecting to gain the remaining authors, and this is the only one that this happened to. The other ones expanded just fine. – Jonesey95 (talk) 02:13, 22 August 2013 (UTC)
Another example, but different: This one added only first names. Is this cite pmid specific? I did maybe 100 cite doi expansions and did not notice this problem. – Jonesey95 (talk) 02:15, 22 August 2013 (UTC)
This seems to happen when there is an "author" parameter. When this happens and I remove the "author" parameter, replacing it with a blank "last1 =" and "first1 =" and run the citation expander again, it usually works fine. – Jonesey95 (talk) 05:47, 23 August 2013 (UTC)
Here's an example of this or a similar problem still happening. I thought that 449 was in production, but it looks like I'm still using 443. If you look at the subsequent history, you can see the manual changes I made to clean it up so that citation bot could fill it in cleanly. I don't know why this workaround works, but it does. It seems to happen only when something is unusual in the first author's parameter(s), like when "author" exists or "last" and "first" exist instead of "last1" and "first1". – Jonesey95 (talk) 23:14, 26 August 2013 (UTC)
The output page from the bot shows "Processing page Template:Cite doi/XXXX" at the top, and it looks like it should be a link to the template, but it's a link to the WP main page.
What should happen
The link should go to the template. The link at the bottom, which is produced only if there are changes (could you display the link at the bottom even for unchanged cite templates as a workaround?), works fine.
Replication instructions
Expand any cite doi template.
We can't proceed until
Operator:
Requested action from maintainer
Make the link connect to the expected page.
This has been a bug as long as I have been using citation bot (3-4 months). The link at the top of the output page goes to the WP main page instead of to the template, and if there are no changes to the citation, there is no link to the template at the bottom of the page. The output in question appears to be produced in line 25 of expand.php. – Jonesey95 (talk) 13:37, 22 August 2013 (UTC)
My question is simple: when was this bot approved for this behavior? Specifically, where was it discussed to have the bot run on new FAC nominations? With the US 23 example linked above, nothing visible was changed, but some extra spaces were removed. Is that not a violation of bot policy to only make changes if they'll be visible? Imzadi 1979→ 02:57, 26 August 2013 (UTC)
Odd that there's no visible change - not sure what's caused that. You mention that your watchlist is clogged – which implies that this is a widespread problem. Could you give a list of a dozen or so links, please, so that I can work out how to replicate the bug? Note also that you can choose to hide bot edits by clicking the link at the top of your watchlist. Thanks, Martin(Smith609 – Talk) 08:29, 31 August 2013 (UTC)
This missed part of my point: the edit says it was "user activated", but by whom? Apparently the submission of an article to FAC is enough to activate this bot, so I ask where was that behavior approved? Each of my FAC submissions this year has been followed by an edit to the nominated article within the first 24 hours or so, prompting me to check to see if it's a worthless edit like this, or if I need to revert and clean up automated changes made to the citations. All were under the guise of "user activated", yet I did no such thing!Imzadi 1979→ 16:12, 11 September 2013 (UTC)
There's no way of knowing this. Sorry. Martin(Smith609 – Talk) 07:36, 3 February 2014 (UTC) {{resolved}}
Bot does not display "et al." for articles with more than 30 authors
expand a citation (like the one linked above) for an article with more than 30 authors.
We can't proceed until
A specific edit to the bot's code is requested below.
Example article containing more than 30 authors. Cite doi template for this article is listed above. I have not looked at the XML at the doi's source to see if 30 authors is a limitation imposed by the article's XML itself, but it seems like a citation bot limit, since it happens with journals all over the place.
One note: This situation has the potential to put us in the same situation we ended up in with the "displayauthors" Lua error that appears on 11,000+ articles that have 9 authors listed in the cite journal template, so the enhancement, if implemented, should be done in a way that avoids splashing a bunch of red text all over the place. Ideally. I think. Or maybe that's what we want, once there is a fix, so that we can run the citation expander on those articles again and get the problem fixed. – Jonesey95 (talk) 12:57, 26 August 2013 (UTC)
This is a bot error and should be easy to fix. I'll get onto it. Martin(Smith609 – Talk) 07:39, 31 August 2013 (UTC)
It looks like this fix results in |displayauthors=30 being set. We might want this to be 29 instead, so that the article shows 29 authors and "et al." instead of exactly 30 authors. For an article with 40 authors, I would want to see some indication in the citation that the article has more than the 30 authors listed in the citation. Am I explaining this clearly? – Jonesey95 (talk) 15:16, 5 September 2013 (UTC)
Gotcha. Will implement anon. Martin(Smith609 – Talk) 08:14, 9 October 2013 (UTC)
Use the citation expander gadget. Notice the shorter-than-usual output.
We can't proceed until
Bot operator's feedback on what is feasible
I've tried to use the citation expander gadget to add remaining authors to this citation, but the bot appears to stop processing and generate a portion of the normal output without adding the authors. It looks like it is finding a graceful way to exit out of the code for some reason, but I don't think it should. I haven't seen this one before. I did note that there are curly braces in the title parameter. I don't know if that causes any problems. – Jonesey95 (talk) 23:55, 26 August 2013 (UTC)
Expand any cite doi template using the citation expander gadget
We can't proceed until
Operator: Bot operator's feedback on what is feasible
Requested action from maintainer
Try to replicate, undo whatever changed in the last hour or so.
Here's the output I get:
Establishing connection to Wikipedia servers ... Using account Citation bot. Fetching parameter list ... done.
Welcome to Citation Bot
Please wait while the Citation bot processes the page you requested.
Activated by Jonesey95
Expanding 'Template:Cite_doi/10.1073.2Fpnas.0401080101'; will commit edits.
Revision #443
[00:00:00] Processing page 'Template:Cite_doi/10.1073.2Fpnas.0401080101' — edit—history
* Looking for bare references...
* Tidying reference tags...
- No references found.
- No references found.
----
Blank page. Perhaps it's been deleted?
** Blank page retrieved.
# # #
End of output
# # #
I don't think this is a bug of this bot. When I use Reflinks, I get the usual output page, but all changes are lost when I then go back to the WP edit window. --Randykitty (talk) 14:19, 29 August 2013 (UTC)
I think it has indeed to do with the change to https. However, I changed my preferences and then tried to login again and then switch to good old http, to try whether the bot would now work, but somehow I stay stuck in https. Darned inconvenient... --Randykitty (talk) 14:26, 29 August 2013 (UTC)
I've changed the preferences for the bot account (so that the bot doesn't use SSL) and this seems to have fixed the bug that was originally reported. RandyKitty, could you confirm whether it's fixed the bug that you were encountering too? Cheers, Martin(Smith609 – Talk) 07:37, 31 August 2013 (UTC)
I ran into that bug a few times over the last few days, possibly up to a frequency of "always". Now I'm not running into it anymore. So it looks fixed to me. Headbomb {talk / contribs / physics / books} 12:38, 31 August 2013 (UTC)
(PS this bot will get to work on that category soon. You are welcome to make a start, but it might be that the bot is faster!) Martin(Smith609 – Talk) 14:35, 5 September 2013 (UTC)
That is good news. The category had over 12,000 entries a couple of weeks ago. By focusing on expanding the cite doi templates, I have reduced it to a little under 8,000. The remaining articles will fit into a few general types:
About 500 {{cite pmid}} templates and articles that incorporate them. See above for a bug that adds first names but not last names and appears to be limited to cite pmid templates. That should be fixed before a bot processes the cite pmid templates.
About 800 {{cite doi}} templates for articles with exactly 9 authors, or for articles that will not expand beyond 9 authors for some reason. I recommend adding |displayauthors=29 to these citations so that they will show all authors and remove the error message. I recommend a value of 29 instead of 9 or some other value because the 9-author articles should show all authors, and the unexpanded templates may eventually be able to expand to more than 9 authors, up to the current apparent limit of 30.
Articles that are displaying the error but that are not using {{cite doi}}. Many of these articles do not use the same author citation format as the citation bot, and they should not be "corrected" to do so. For example, if authors are shown as "Smith M" or "Smith, M" as opposed to the citation bot's "Smith, M.", the bot should not necessarily change that formatting. Adding the displayauthors param to these citations might be a good fix, but it could also be problematic if the article has a consistent citation format that shows a certain number of authors.
Articles that incorporate newly-error-free {{cite doi}} templates but that need a null edit to be removed from the category. I don't know how many of these there might be.
Edge cases. Lots of edge cases. I'll be happy to help clean those up.
I welcome your thoughts. – Jonesey95 (talk) 15:12, 5 September 2013 (UTC)
Understood! I hope to have a draft version of the new script together within the next fortnight and would welcome your input when I have made some trial edits. I'll keep you posted. Martin(Smith609 – Talk) 16:05, 5 September 2013 (UTC)
When the bot posts the report, it starts something like
Please wait while the Citation bot processes the page you requested. Activated by Headbomb Expanding 'OPERA_experiment'; will commit edits. Revision #442 [00:00:00] Processing page '[https://en.wikipedia.org/wiki/ OPERA_experiment]' — [https://en.wikipedia.org/w/index.php?title=OPERA_experiment&action=edit edit]—[https://en.wikipedia.org/w/index.php?title=OPERA_experiment&action=history history]
That link looks like it's going to the Beta version running r442 (where the fix hasn't propogated yet). Try the non-beta link (substitute citation-bot for DOI_bot in the URL). The beta link'll be fixed in a while. Martin(Smith609 – Talk) 14:32, 5 September 2013 (UTC)
I think clicking the "jump the queue" link when you create a brand new cite doi template results in this error. – Jonesey95 (talk) 13:57, 5 September 2013 (UTC)
Some google book entries have both first/last name and author fields populated.
What should happen
This should be either/or - don't populate both as "More than one of author1= and last= specified" generates a warning in the cite book / cite journal template.
[132] with the page in whatever state it was in before the aforementioned edit.
We can't proceed until
Agreement on the best solution
I mistakenly posted my experience of this bug in another section, but this diff demonstrates the problem again. I'm not sure what consensus needs to occur for the solution. If "author" is specified, shouldn't the bot just refrain from adding "last" or "first" parameters? Azaghal of Belegost (talk) 18:39, 14 January 2014 (UTC)
A specific edit to the bot's code is requested below.
Can you suggest a protocol by which the bot can determine whether the pipe is genuinely part of the title paramete (so should be escaped as you suggest), or whether a user has mis-typed the following parameter — for example title = All about things | journal - Journal of Science or title = All about things | <!-- url = http://not.working.com -->
Perhaps the OP should have said "A pipe encountered in a datum retrieved from an external source must be escaped with {{!}} before adding it to the template as a parameter value." Seen this way there's no ambiguity. EEng (talk) 17:13, 9 September 2013 (UTC)
Thanks, now I understand the problem. Will do! Martin(Smith609 – Talk) 07:35, 27 September 2013 (UTC)
Institutional author formatted as "lastname, firstname" ("Association, American Bar" instead of "American Bar Association")
Citations for journal articles with exactly nine authors display a Lua error message. Citations for journal articles with thirty or more authors are created with exactly thirty authors but do not display "et al.", even if the original article has more than thirty authors. Both situations are undesirable.
We can't proceed until
Maintainer: A specific edit to the bot's code is requested above.
Requested action from maintainer
Both types of citations should have "|displayauthors=29" added so that the 9-author citations will show all authors with no errors, and 30-author citations will show 29 authors and then "et al." to indicate that there are more than 29 authors.
I think I've thought this one all the way through. There may be some nuance that I'm missing. – Jonesey95 (talk) 21:17, 20 September 2013 (UTC)
Citation Bot 2 is sweeping through and creating new citations with this same error, requiring more human intervention, even though this bug was reported four months ago. Is this any easy fix? – Jonesey95 (talk) 19:04, 19 January 2014 (UTC)
In rev 503, I have made it so that any addition of 'last9' or 'author9' will add 'display-authors = 29'. Not the most elegant solution, but will this patch the hole? Martin(Smith609 – Talk) 21:47, 19 January 2014 (UTC)
Thanks! That should work fine. Is there a link where I can test a relatively stable beta version on existing cite doi templates? – Jonesey95 (talk) 21:42, 20 January 2014 (UTC)
Thanks. I tried to run a page through the bot (r506) at that page, but I received an error: Fatal error: Call to undefined function expand() in /data/project/citations-dev/public_html/doibot.php on line 105. – Jonesey95 (talk) 00:59, 21 January 2014 (UTC)
I've not updated doibot.php yet (so I guess the link I gave you isn't much help - sorry!). I'm checking that the most important outstanding bugs are resolved, then will work on getting the user-facing parts of the bot operational. Shouldn't be long now (next few weeks?). Martin(Smith609 – Talk) 07:46, 3 February 2014 (UTC)
Oh, that was easier than I expected. The link should now work. Do please check the output carefully and let me know of any errors. Martin(Smith609 – Talk) 08:35, 3 February 2014 (UTC)
Not fixed yet. This edit expanded a 9-author article to 30 authors and then added "display-authors=8" instead of "display-authors=29". It should have added "display-authors=29". – Jonesey95 (talk) 05:38, 3 February 2014 (UTC)
In my experience, the golden rule of bot operating is 'preserve the format that was already there'. Under the old template, the listing of 9 authors would cause the first eight only to display, followed by 'et al' (I think). The bot's trying not to change this in case this behaviour was desired by the original editor. But I can modify the behaviour if you think it necessary. Martin(Smith609 – Talk) 07:43, 3 February 2014 (UTC)
Bracketbot
Hello, I'm BracketBot. I have automatically detected that your edit to Mesoionic may have broken the syntax by modifying 1 "[]"s. If you have, don't worry: just edit the page again to fix it. If I misunderstood what happened, or if you have any questions, you can leave a message on my operator's talk page.
List of unpaired brackets remaining on the page:
|title=Structure of the mesoionic compound N-\1-methyl-3-(p-tolyl)-4-(1,2,3-triazolio)]acetamidate (MMTAT), C12H14N4O |year=1983 |last1=Papageorgiou |first1=M. |last2=Kokkou |first2=S. C.
Thanks, BracketBot (talk) 12:56, 26 September 2013 (UTC)
"Oh, hi, I'm Citationbot. Wow, thanks. I've been looking everywhere for that other bracket! So you're that big strong Bracketbot I've heard so much about. Gosh, you look like you must be 64-bit -- such big quads! -- a complex instruction set with hardcoded ABS. Why don't you come into my domain? That's not my usual protocol, but a girl feels so secure when a guy has so much onboard cache. I wasn't expecting to host, so pardon my open proxy -- a bit RISCé, perhaps, but just something I wear around the server farm. Virtual mammary memory? -- oh no, these dual cores are absolutely real! All native configuration -- no upgrades at all. Should I slip into a more user-friendly interface -- something GUI, perhaps? Oh, you prefer command-line? -- kinky! ..."
Later: "Oh, Bracketbot! I can't believe your high refresh rate. Lucky I'm in protected mode so you won't have to worry about any viruses! My husband has a really short cycle time and his puny little floppy drive is subject to frequent hardware failures, so sometimes I have to use manual simulation! And I've never had 10 gigabytes of hard drive before! Let's FTP! ... Oh god! I'm downloading ..."
You are an evil, evil man... :)Naraht (talk) 15:54, 26 September 2013 (UTC)
Personally I think "wicked" has a more lofty ring to it, but I'll take "evil". Thanks! EEng (talk) 18:22, 26 September 2013 (UTC)
Brilliant! You will feature prominently in my Oscar acceptance speech. EEng (talk) 00:20, 1 October 2013 (UTC)
Done. While you two bots were flirting, I fixed the bracket problem. Carry on. Don't let me interrupt. – Jonesey95 (talk) 20:04, 26 September 2013 (UTC)
That's OK. They're both configured for quick reboot. EEng (talk) 20:27, 26 September 2013 (UTC)
Hard as this may be to believe, someone seems to have completely missed this point of all this. [134]EEng (talk) 20:48, 26 September 2013 (UTC)
Could someone clarify whether this is a bug and if so what action is required? Martin(Smith609 – Talk) 09:49, 19 January 2014 (UTC)
I don't know whether it's a bug or GIGO. The bot inserted the journal article title as "... N-\ ..." instead of "... N-[ ...". The journal article in question is [135]. – Jonesey95 (talk) 16:34, 19 January 2014 (UTC)
See this edit. The bot seems to have forgotten to add |lastn= parameters to match all of the |firstn= parameters that it added. The bot added |author1= and |last= which are aliases of each other and as such, when present simultaneously in a citation, cause an error. I have reverted.
A similar bug was reported above and a fix was attempted. It appears to have failed to fix the problem, which appears to manifest itself when |author is already present. – Jonesey95 (talk) 04:04, 9 October 2013 (UTC)
There are other things wrong. The robot uses |pages= when it should use |page= and occasionally adds extraneous text: |journal=Astronomical Journal (ISSN 0004-6256). The ISSN belongs in |issn=. The title of the journal is The Astronomical Journal. And too much white space. I have fixed one of the mangled citations.
Not a bug but I wonder if long long citations like those in Neil deGrasse Tyson shouldn't be named and moved into {{reflist}} as part of |refs=. It gets them out of the article text so makes reading in the edit window easier. I've moved the citation that I fixed as an example.
Get "lucky" with well-timed page-moves and bot-executions
We can't proceed until
Agreement on the best solution
Requested action from maintainer
Decide how to handle it then make necessary code changes
Are you activating the bot yourself, or is the bot jumping in and editing your article after you save it?
I believe that this is a page caching problem that exists for all quick-succession edits, not just edits by Citation Bot. I have experienced it when doing a manual edit of a cite doi template, followed a few seconds later by running Citation Bot using the "Expand citations" gadget in my Toolbox. The bot edits the previous version, ignoring my manual edits. Let's see if I can find it.... Maybe bugzilla:46014, and it looks like a patch has been developed for the MediaWiki software, presumably to be deployed at some future date. – Jonesey95 (talk) 04:08, 10 October 2013 (UTC)
The edit summary in the diff above says "user activated" so I assume that's the case. Even of the bot can't detect a race condition as it is happening, it can go back and detect that a race condition MIGHT have occurred and either undo its edit or, if that fails (perhaps due to another intervening edit) notify the editors possibly-conflicting edits so they can straighten things out by hand. davidwr/(talk)/(contribs) 04:23, 10 October 2013 (UTC)
I understand race conditions to be fixedin r500+ so have marked thsi | status = {{resolved}} Martin(Smith609 – Talk) 07:55, 3 February 2014 (UTC)
Removes AfC submission template appearing after reflist template
Bot deletes "AFC submission" template occurring after "reflist" template in pages under Wikipedia talk:Articles for creation/ ... this "unsubmits" a draft article, potentially leaving the article submitter (very likely a new user) unaware of what has happened or what they should do about it - they are likely to assume their submission is still waiting for review or has just been ignored
This bug is similar to one reported above. The citation bot sometimes has trouble with citations when "author" parameters are present. – Jonesey95 (talk) 15:14, 17 October 2013 (UTC)
Fourth author added, using |author4=Van Den Hoek Ostende LW - but it was already present, in slightly different form, as |last4=Hoek Ostende|first4=L. W.
What should happen
Modify either of the existing parameters, don't add one that is an alias of |last4=
Could you please, please, please tell the bot to stop making changes to comments within cite templates. Here is a case where in three different places it replaces a simple comment with a large amount of unrelated text. It is also doing meaningless whitespace churn which makes the diffs even harder to read. Here is another edit where it took away a valuable comment about an alternate source that I had. If the bot sees a comment, just leave it alone!! I and others have complained about this comment botching above, but nothing seems to be happening.
We can't proceed until
Agreement on the best solution
Assumed to be fixed in r505. Martin(Smith609 – Talk) 16:44, 2 February 2014 (UTC)
Comment in parameter not blocking citation bot from filling it
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
The documentation page says "If the bot is erroneously adding a DOI, author, etc to a citation, and you want to stop it adding the data again, you need to put a comment in place of the appropriate parameter – because the bot will not overwrite existing data." I did so, and the bot removed my comment and filled the parameter.
What should happen
The bot should comply with its documentation, declining to fill a parameter with a comment in it.
Remove my edits and run the bot on the above cite doi template.
We can't proceed until
Operator: Bot operator's feedback on what is feasible
Ok, so the bot is now respecting the exclusion protocol's "deny", but the other problem here is that the bot would (if not denied) insert {{{last|Voelker}}} when according to both pubmed and crossref it should be satisfied with {{{last1|Bachert}}} and {{{last5|Voelker}}}. Fortunately in this particular case it doesn't matter, the template is never transcluded. LeadSongDogcome howl! 20:37, 22 November 2013 (UTC)
Deleterious: Human-input data is deleted or articles are otherwise significantly affected. Many bot edits require undoing.
What happens
error More than one of |author2= and |last2= specified because author and last are the same person, redundant. | author2 = von Bohlen und Halbach O
<br/>| last2 = Von Bohlen Und Halbach | first2 = O.
When you remove the redundant author| author2 = von Bohlen und Halbach O
Sorry. Duplicate of "Author was already present" and "Organochloride bad edit: spurious lastn parameters added" above. I agree: stop the bot now. It is causing a lot of work to be undone. Quebec99 (talk) 16:22, 14 December 2013 (UTC)
Looking at the transcript, I see
1: Tidy citation and try to expand
- Checking CrossRef for more details [DOItools.php/expand_from_crossref] + last2: Von Bohlen Und Halbach (ok) - Searching PubMed... no results. 1 result found; updating citation + author2: von Bohlen und Halbach O 2: DOI already present 3: PMID already present - Checking PMID12845521 for more details [DOItools.php/get_data_from_pubmed] + author2: von Bohlen und Halbach O 4: Expand citation - Checking CrossRef for more details [DOItools.php/expand_from_crossref] + last2: Von Bohlen Und Halbach (ok) - No changes since last PubMed search. 5: Formatting and other tweaks First: Cite Doi formatting [expand.php/expand_text] + first2: O.
Clearly the difference in capitalization between Crossref and PubMed is confusing the bot, as it doesn't have a problem with the first author. Will exclude the bot from this record for now.LeadSongDogcome howl! 18:24, 14 December 2013 (UTC)
The surname components 'Von ' and 'Und ' are now [r492] always placed in lowercase. Are there any other components that should receive the same treatment? Martin(Smith609 – Talk) 18:32, 18 January 2014 (UTC)
Improvement: The bot would be much better if ...
Cosmetic: The bot modifies appearance (e.g. capitalization) but not content
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
Deleterious: Human-input data is deleted or articles are otherwise significantly affected. Many bot edits require undoing.
Catastrophical: The bot should stop editing immediately
-->
What happens
Bot creates DOI pages with <noinclude>{{template doc|Template:cite_pmid/subpage}}</noinclude>
What should happen
should use {{Documentation|etc.</noinclude>
We can't proceed until
A specific edit to the bot's code is requested below.
Requested action from maintainer
Can be fixed in lines 70, 86 and 91. Don't know if the same code is used in other pages, but I had a look around and didn't see any other occurrences.
Thanks for the detailed report; fixed. Martin(Smith609 – Talk) 18:21, 18 January 2014 (UTC)
It seems the bot can recognize "displayauthors", so whatever mechanism does that, to replace it with "displayauthors" OR "display-authors".
I'm pretty sure it's better, just not fixed. See this recent diff. Under line 64, it fails to recognize display-authors=3, and goes on to add displayauthors=30.
A tip: I could be totally wrong about this, but the r490 fix message log lists it in uppercase ("Displayauthors", not "displayauthors"). Case-sensitivity perhaps? meteor_sandwich_yum (talk) 22:38, 1 February 2014 (UTC)
Seems like the sensible thing for the bot to do is replace the alias displayauthors with display-authors and be done with it. The alias should be deprecated to avoid creating more such. LeadSongDogcome howl! 02:34, 2 February 2014 (UTC)
Citation bot 487 created new template with multiple errors
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
Bot created a template with |last= and |last1=, which it should not do. It also created a template with exactly four editors without adding |displayeditors=29, which it should do (see similar bug report above with more details). The bot also did not format the editors' names in the same way that it formats author names. The bot also left HTML markup in |title=. It also failed to pick up all of the authors' last names.
What should happen
Bot should not create redundant parameters. Bot should not create new templates that result in displayauthors or displayeditors errors. It should also display names consistently and remove HTML markup from parameter values.
Bot altered the name of a reference from <ref name="ReferenceA"> to <ref name="RailMagazine2007">, which caused an error Cite error: The named reference ReferenceA was invoked but never defined (see the help page). which was not previously present.
The citation is expanded correctly, however the following text appears, twice, at the top of the page:
Warning: session_start() [<a href='function.session-start'>function.session-start</a>]: Cannot send session cookie - headers already sent by (output started at /data/project/citations/public_html/text.php:9) in /data/project/citations/public_html/expandFns.php on line 3
I simply typed in: ref name="Ma1999" cite journal pmid=10508203 /ref, with the appropriate brackets and clicked the citation button at the bottom of the editing pane.
We can't proceed until
Agreement on the best solution
Requested action from maintainer
Look at Line 3 for problems?
Random warning message appears, twice at top of page being edited every time I use the Citation bot. --Iamozy (talk) 20:00, 22 January 2014 (UTC)
Will be resolved when the new version rolls out later this year. Martin(Smith609 – Talk) 19:32, 27 January 2014 (UTC)
Bot recreated deleted templates that have no apparent incoming links
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
The bot created a handful of templates that have no apparent incoming links. I CSD'd them, and the bot recreated them. The bot should not autonomously recreate deleted templates, especially if they have no incoming links.
Replication instructions
Investigating: see which pages are triggering template creation...
We can't proceed until
Input from editors
Citation bot 2 created the following templates. I can't figure out why they were created in the first place, and I can't figure out why they were recreated after I CSD'd them. Time stamps are UTC-8.
14:34, 2014 January 28 (diff | hist) . . (+145) . . N Template:Cite doi/10.2307.2F10.1086.2F663750 ([cw512]New page from {{Cite doi}} template) (current)
14:34, 2014 January 28 (diff | hist) . . (+103) . . N Template:Cite pmid/25104635 ([cw512]New page from {{Cite pmid}} template) (current)
14:34, 2014 January 28 (diff | hist) . . (+103) . . N Template:Cite pmid/25104016 ([cw512]New page from {{Cite pmid}} template) (current)
14:34, 2014 January 28 (diff | hist) . . (+103) . . N Template:Cite pmid/40794784 ([cw512]New page from {{Cite pmid}} template) (current)
14:34, 2014 January 28 (diff | hist) . . (+103) . . N Template:Cite pmid/25108119 ([cw512]New page from {{Cite pmid}} template) (current)
14:34, 2014 January 28 (diff | hist) . . (+103) . . N Template:Cite pmid/25100523 ([cw512]New page from {{Cite pmid}} template) (current)
14:34, 2014 January 28 (diff | hist) . . (+103) . . N Template:Cite pmid/25108201 ([cw512]New page from {{Cite pmid}} template) (current)
14:34, 2014 January 28 (diff | hist) . . (+103) . . N Template:Cite pmid/25107223 ([cw512]New page from {{Cite pmid}} template) (current)
14:34, 2014 January 28 (diff | hist) . . (+103) . . N Template:Cite pmid/40936072 ([cw512]New page from {{Cite pmid}} template) (current)
14:34, 2014 January 28 (diff | hist) . . (+103) . . N Template:Cite pmid/25207853 ([cw512]New page from {{Cite pmid}} template) (current)
14:34, 2014 January 28 (diff | hist) . . (+109) . . N Template:Cite pmid/10.1086/663750 ([cw512]New page from {{Cite pmid}} template) (current)
14:34, 2014 January 28 (diff | hist) . . (+103) . . N Template:Cite pmid/27532624 ([cw512]New page from {{Cite pmid}} template) (current)
14:34, 2014 January 28 (diff | hist) . . (+103) . . N Template:Cite pmid/25107749 ([cw512]New page from {{Cite pmid}} template) (current)
14:34, 2014 January 28 (diff | hist) . . (+103) . . N Template:Cite pmid/25107510 ([cw512]New page from {{Cite pmid}} template) (current)
14:34, 2014 January 28 (diff | hist) . . (+103) . . N Template:Cite pmid/30064609 ([cw512]New page from {{Cite pmid}} template) (current)
14:34, 2014 January 28 (diff | hist) . . (+103) . . N Template:Cite pmid/25130323 ([cw512]New page from {{Cite pmid}} template) (current)
Can you explain what happened here and how it might be prevented? – Jonesey95 (talk) 23:06, 28 January 2014 (UTC)
When the bot creates these templates, it now mentions the page that includes them in the edit summary. This should help to understand (and if necessary debug) why the templates are being created. Do let me know if the problem persists. Martin(Smith609 – Talk) 08:07, 31 January 2014 (UTC)
No further reports, so I'll queue this for archiving; do re-activate the report if the bug recurs. Martin(Smith609 – Talk) 08:02, 3 February 2014 (UTC)
Inserting some weird source code at the top of pages.
From what I have seen, this happens every time I run the script by clicking the citations button at the bottom of the edit pane. It MAY be browser specific or something like that? If you are unable to reproduce, please let me know and I will certainly help.
We can't proceed until
Agreement on the best solution
Requested action from maintainer
Not sure what to say here other than it should be fixed?
Will be fixed when the new version is rolled out. Want to be thorough in checking it for bugs before it goes live! Martin(Smith609 – Talk) 22:04, 2 February 2014 (UTC)
THANKS! --Zackmann08 (talk) 22:05, 2 February 2014 (UTC)
doi_inactivedate/doi_brokendate requires user editing
It looks like there is no change in the rendered output or even the content of the citation. – Jonesey95 (talk) 06:09, 23 December 2013 (UTC)
Indeed. The change is cosmetic and restricted to the source code. You have to admit that it is not clear what she bot sees to make such useless changes, detracting from the easy overview of the source code. Debresser (talk) 09:41, 23 December 2013 (UTC)
Yes, since these are named parameters, leading and trailing whitespace is insignificant - newlines and spaces are both stripped out, so must function identically, so altering one to the other has a cosmetic effect on the page source but no effect on the rendered page. If this bot were using AWB to make such edits, it would have been directed to WP:AWB#Rules of use item 4 a long time ago; probably had AWB access revoked for repeated violation; possibly even blocked for disruptive editing. But the last time that I blocked this bot, there was an outcry. --Redrose64 (talk) 22:47, 23 December 2013 (UTC)
This edit only occurs in Cite Doi / Cite PMID namespace, and occurs for legibility reasons and to comply with formatting guidelines. Martin(Smith609 – Talk) 18:22, 18 January 2014 (UTC)
What formatting guidelines? --Redrose64 (talk) 19:14, 18 January 2014 (UTC)
What formatting guidelines? How does this improve legibility? There are thousands of Cite Doi / Cite PMID pages, so that is no reason not to take care of this. Debresser (talk) 17:28, 19 January 2014 (UTC)
A fresh example (from R487) at this edit, where for PMID22258985 the value of last1 is incorrectly added as editor1-last and first2 is incorrectly added as editor2-first. LeadSongDogcome howl! 17:26, 13 January 2014 (UTC)
As of r491Editor is only added if CrossRef reports no Journal Title. Martin(Smith609 – Talk) 18:19, 18 January 2014 (UTC)
It's been a while that I couldn't run the bot in the "edit" mode, i.e., after clicking "edit" on an article and a radio button "citations" at the bottom. This mode is most important for the bot, as it allows to preview changes before they are automatically saved (with potential errors, as done when clicking "expand citations"). Please restore this mode, or tell me if I miss something obvious. The error messages when trying to use this mode look like
Warning: session_start() [<a href='function.session-start'>function.session-start</a>]: Cannot send session cache limiter - headers already sent (output started at /data/project/citations/public_html/text.php:9) in /data/project/citations/public_html/expandFns.php on line 3 Materialscientist (talk) 00:44, 17 January 2014 (UTC)
Looks like a victim of the move to the new servers. I'll investigate ASAP. Martin(Smith609 – Talk) 17:23, 18 January 2014 (UTC)
Just came here to report that I am experiencing the same problem, and to express my appreciation for the work you do with citation bot :) AioftheStorm (talk) 21:15, 18 January 2014 (UTC)
It works now, but attempts to add the above error message to the top of the edited article. Materialscientist (talk) 13:30, 25 January 2014 (UTC)
It works now without the error. AManWithNoPlan (talk) 19:55, 6 February 2014 (UTC)
Can't run the bot from the toolbox on the left side
It is possible that this bug has been fixed in a later revision of the bot. – Jonesey95 (talk) 00:34, 5 February 2014 (UTC)
Fix in newer versions. AManWithNoPlan (talk) 21:53, 6 February 2014 (UTC)
Latest update
Please make the Written to Example. red agin, it is easier to see if the article has been edited that way. (t) Josve05a (c) 10:04, 4 February 2014 (UTC)
The bot is changing {{cite web|url=http://www.ibo.org/school/search/index.cfm?programmes=&country=®ion=&find_schools=Find|url=http://www.ibo.org/school/search/index.cfm?programmes=&country=US®ion=&find_schools=Find|title=Find an IB World School|accessdate=June 18, 2013}} to {{cite web|DUPLICATE_url=http://www.ibo.org/school/search/index.cfm?programmes=&country=®ion=&find_schools=Find|url=http://www.ibo.org/school/search/index.cfm?programmes=&country=US®ion=&find_schools=Find|title=Find an IB World School|accessdate=June 18, 2013}} (chaning url to DUPLICATE_url)
It really can't delete the URL, since they are not the same and the bot has no way of knowing which one is better. Putting the DUPLICATE_url in the wiki makes it clear that there is a problem that requires human intervention. AManWithNoPlan (talk) 21:50, 6 February 2014 (UTC)
Improvement (restoring functionality that used to exist)
What happens
"jump the queue" failed to create a cite doi template that had not previously existed
What should happen
"jump the queue" should create a new template.
Replication instructions
Edit a sandbox page and add a new cite doi template (one that does not exist). Click Preview, then click "jump the queue" in the Preview section of the page. Error message is below.
We can't proceed until
Operator: Bot operator's feedback on what is feasible
I added {{cite doi|10.1086/663750}} to my Sandbox page, clicked Preview, and then clicked "jump the queue" in the Preview mode. I like to do this so that I can create a cite doi template before saving an article with a new cite doi template. I received the following response from Citation Bot:
Welcome to Citation Bot
Please wait while the Citation bot processes the page you requested.
Revision #532
Establishing connection to Wikipedia servers ...
Using account Citation bot.
Fetching parameter list ... done.
Expanding 'Template:Cite doi/10.1086.2F663750'; will commit edits.
Page 'Template:Cite doi/10.1086.2F663750' not found.
End of output
# # #
I then created the cite doi template by hand, populated the doi parameter, and Citation Bot filled in the rest with no trouble. Something about the "jump the queue" link didn't work for me. This used to work just fine. – Jonesey95 (talk) 22:25, 4 February 2014 (UTC)
I have the same problem since about a week. -- Daniel Mietchen (talk) 09:12, 7 February 2014 (UTC)
I've noticed this problem as well. Looking through the bot's contribs, it hasn't created any new {{cite doi}} templates in the last 500 edits - it is not just the 'jump the queue' option that isn't working. I tried creating a blank template as the error says the bot couldn't find the template and running it again but it still didn't work. SmartSE (talk) 14:59, 9 February 2014 (UTC)
Looks like the queue is not being processed regularly; seems I have still to get my head around the WMFlabs scheduling services. As a temporary workaround, accessing http://tools.wmflabs.org/citations/citewatch.php should cause the queue to be processed. (It may look like nothing's happening, but whilst the page is open User:Citation bot 2 will be processing the queue.) Martin(Smith609 – Talk) 08:35, 11 February 2014 (UTC)
At line 2890 (and following) there's a repeat of ten lines with 80 through 89 vice 90 through 99
I'm not at all sure, but this line looks to me like it has reversed the cases. That is vice if (!($authors = $this->get('author'))) $authors = $this->get('authors'); should have if (!($authors = $this->get('authors'))) $authors = $this->get('author');
Deleterious: Human-input data is deleted or articles are otherwise significantly affected. Many bot edits require undoing.
What happens
Bot changes the number parameters to issue parameters in cite templates.
What should happen
This should not be done. Semantic information is lost. Some journals use number instead of issue and an editor may choose one over the over on purpose. Using the proper naming aids people in finding the articles. Using improper naming is adds unnecessary confusion. The number parameter is a proper alias of issue and is not deprecated. There is no mandate to change these in a systematic manner.
comment out the code causing this change. It's ruining human input.
I understand number to be a synonym of issue. Please link to a case where the different parameter name produces different output. Martin(Smith609 – Talk) 18:37, 18 January 2014 (UTC)
I have just noticed that the bot is blocked based on this report. This seems like a heavy-handed approach. Could the bot be unblocked until community consensus is reached? Martin(Smith609 – Talk) 19:52, 20 January 2014 (UTC)
AFAIK the bot isn't blocked, and hasn't been since 2 June 2013. Here's the log; the last blocker was me, on 28 May 2013; and that block was lifted by somebody else five days later. However, it's clear from the subsequent reports on this page that the reasons that I blocked the bot have not all been resolved. --Redrose64 (talk) 20:33, 20 January 2014 (UTC)
Hi. Yes, the output is identical but some journals use "number" and some use "issue" and a user has presumably chosen one over the other for a reason. (As I had with the example URL I gave.) If "issue" were always the best thing to use, there'd be no point in having "number" as an alias. Having issue vs number correct can be a boon to focusing on the correct journal, especially with medical journals where they often have very similar titles. Having it wrong introduces an air of uncertainty. So I think the case for doing nothing, clearly outweighs the case for changing the parameter for uniformity purposes.
As for my blocking of Citation bot 1, although I haven't used any of the citation bots, I'm wondering if all versions of this it ought to be blocked because the documentation and functionality have major issues. Take a look at the edit summary in the example link I gave in the bug report. 1) there's a dangling pipe character, 2) there's a confusing bot link that even more confusingly points to a confusing shortcut that leads to a use page, 3) the edit summary doesn't even match the work that's been done. That's a very bad edit summary. I don't know if those things have been fixed in newer versions or how much of the edit summary is due to user sloppiness, but any bot leading to edit summaries like that needs re-examination in my book. If all of it cannot be explained away by user sloppiness, it's the bot's problem. WORSE, the documentation, especially for bot shutdown, for this collection of bots is atrocious! It's completely unclear WHICH bot to shutdown in the case of problems. At least in some (if not all) cases, there's no way to determine which bot is responsible for which edits. I followed the instructions as close as possible but I have no idea if I shutdown the correct bot in the example link case. In my opinion, this entire system of bots needs he documentation completely revamped with a focus on a clear and precise organization and wording. It's totally unclear to me how the documentation for this group of bots is organized and how User:Citation bot relates to User:Citation bot 1 through User:Citation bot 9 (which doesn't even have a user page). Heck, the shutdown page only lists User:Citation bot and bots 1 through 4. What is going on here? This whole situation is a absolute confusing mess to decipher. Users and admins should not have to waste days just to figure out what the story is with a bot.
Lastly, from the other messages on this bot's talk pages, it's clear that I am not the only one rather annoyed with these bots. Martin, I'm worried that there's a bit of a problem with your view of bots. Given the great damage a malfunctioning bot can do in a short period of time, our approach to bots is very conservative. The whole bot approval procedure is designed such that there's community consensus needed FOR the edits a bot makes. Users should be encouraged to shutdown bots if there may be a genuine problem; asking for community consensus to stop a bot from making questionable edits is sort of backwards. Where is the consensus to make number to issue changes in the first place? I quick perusal of (all 9!) bot approval requests did not find this as a stated goal bot task. (It kind of rankles me the wrong way that so much onus is on the user to learn exactly what the bots are approved to do. Each bot's user page should have a clear list of concise tasks it is allowed to do.) Assuming I am not overlooking this as an approved functionality (link please), it means that the bot (bots?) are in violation WP:Bot policy. As I recall I quickly searched for this feature approval back when I blocked the bot too (because it would have strongly opposed this task for the bot). All in all, and I'm sorry for sounding harsh, I just feel like this bot not being maintained as well as it should.
PS Again, please forgive any false assumptions I have made about the bots, as I haven't used them myself and don't know how they work. Point out any misunderstanding I have but also try to be charitable in understanding my general point of view (for instance over the quality of the documentation) rather than dismissing my whole view on the basis of a minor false premise. Jason Quinn (talk) 01:41, 21 January 2014 (UTC)
'number' is not a documented parameter in Template:Cite Journal. It is difficult to argue that the bot is 'breaking' anything by replacing an undocumented parameter (which is retained to make life easier for editors) with a documented parameter.
Perhaps there is a case to introduce a 'number' parameter to Cite Journal, in which case I would recommend that you put forward the case there. You may wish to address the following questions: If the distinction between 'number' and 'issue' is useful, then why is it not made in the formatted output? Can you give a (real life) example of journals being distinguished in this way, perhaps in the reference section of a journal article, or in a library catalogue? If two journals have similar names, isn't the ISSN much more useful to distinguish between them than what they call their parts?
Incidentally, if you can give a list of journals whose names are confusing, it would be easy to modify the bot to specify an ISSN for these journals – which in my view could provide a more prominent and useful distinction than something readers can only discover if they edit the page and find and decipher the original reference's code.
The best solution here is to gain consensus before blocking a very widely used tool, particularly if you are unsure of the consequences. In this process, it will be possible to gather community consensus as to whether a particular problem (such as an unclear link!?) provides grounds to block a bot, and if so, which account it is necessary to block.
The bot's currently undergoing a fundamental re-write.
See User:Citation_bot/block for what each bot account does, and a page's edit history to see which account made an edit. See User:Citation_bot#Function for a list of the bot's functions. For what it's worth, the mis-typing of 'issue' as 'number' has in the past been a common mistake (function 2); I suspect that it's only recently that the template was modified to gracefully accommodate this mistake. Martin(Smith609 – Talk) 08:26, 21 January 2014 (UTC)
Hi, Martin. I know you did it to be helpful but please do not chop comments up into pieces to reply to each point. This is completely non-standard for Wikipedia talk pages and ultimately makes it more difficult for later readers to know who wrote what. If you like you can use bullet points to signal each new topic to which you are replying. In order to try to recover the threading, I have reassembled my message and grouped your replies, whose context is still mostly clear with no wording changes.
As for number, it is in the {{cite journal}} documentation. You are correct it does not currently have a parameter description but it once did (it was introduced with this edit back on 8 September 2010 and stayed around until 6 April 2012 when it was removed but not for any good reason that I know). My argument above does not concern itself with the rendered display of the template so questions indicate you are still missing the point regarding the usefulness of number being used in the template.
As for blocking the bot, as per WP:BOTISSUE, blocking a bot that is making changes not explicitly approved is a perfectly valid reason to block a bot. It's possible I overlooked the specific approval this bot had to make those "number" to "issue" changes (I did look). Unless that is found, this bot may be in violation of bot policy. Further, any user who suspects a bot is malfunctioning or noticed a new potential bug should be encouraged to block a bot ASAP. I find resistance to this idea, and thinking community consensus is necessary, rather obtuse. If this were a specifically approved bot activity, then it would be a different story but as far as I've found, it's not. For bots community consensus is needed to allow the bot do changes, not to stop it.
As for the bot functions I have seen most of that material. The problem is the documentation is a pretty disorganized, blob-like, and imprecise. For instance, even the bot functions you linked to above states: "Citation bot only amends the parameters of Citation templates." Well, that's not true. Here is a case where it's changing the parameters not amending them. The bot documentation needs to reflect the precision and clarity of a detail-orientated person. Then, as I pointed out for User:Citation_bot/block, it doesn't even list all the bots. It only lists 5 out of 10. Or maybe it does? I don't know. It's just totally confusing how these bots are related to each other, which is my point. If the bot is being re-written, I suggest it be considered a new bot and these bots retired rather than amending the bot to the ever-growing list here.
Quite frankly I don't understand the mindset at work here. It feels like real problems are being downplayed. This not what I would expect from the author or maintainer of a bot. To quote the Spiderman cliché "with great power comes great responsibility". This is the attitude I prefer to see. Jason Quinn (talk) 22:27, 22 January 2014 (UTC)
The bot needs to support the parameter aliases. Unfortunately the cite template documentation is incomplete so some of these problems/concerns are only discovered during use. Whether we gain from editor use of issue versus number doesn't really matter here as the bot & other cite tools have to support the parameters and aliases either way. Rjwilmsi 09:30, 21 January 2014 (UTC)
The CS1 source lists number as just an alias to issue (as people have already said), but here is the url https://en.wikipedia.org/wiki/Module:Citation/CS1/Configuration The distinction of issue and number is only present in the wiki source, not the rendered html, so I think that using that distinction to transfer information is a poor assumption, particularly given that most cite examples already are probably randomly set to issue or number based on author preference and not what the journal does. If their is possible confusion about what is meant, then use pmc, doi, pmid, url, issn, etc. to point the reader to the correct journal. AManWithNoPlan (talk) 02:31, 24 January 2014 (UTC)
Please update the documentation at Cite Journal / whereever else so that the correct use of the 'number' and 'issue' parameter is unambiguous. This will make the distinction clear to future editors and flag up the usage to the communities that use the templates. I will then update the bot to follow the documented usage. Martin(Smith609 – Talk) 19:38, 27 January 2014 (UTC)
The |number= parameter is now listed as deprecated. Thus the bot's current behaviour – changing |number= to |issue= – is correct. Having resolved this issue, I am requesting that User:Citation bot 1 is unblocked. Martin(Smith609 – Talk) 10:47, 1 February 2014 (UTC)
Wait a minute, where and when did |number= get deprecated? As far as Module:Citation/CS1 is concerned, |number= is an active and valid alias for |issue= and will remain so for the foreseeable future.
There having been no discussion to arrive at a determination to deprecate |number= as an alias of |issue= in CS1 citations, I have undone the change made by Editor Smith609 to Help:Citation_Style_1#Edition_identifiers with this edit.
This user's unblock request has been reviewed by an administrator, who declined the request. Other administrators may also review this block, but should not override the decision without good reason (see the blocking policy).
Please unblock User:Citation bot 1 (which shares this talk page). The bot was blocked because it was changing 'number' to 'issue'. It was unclear whether this was the correct behaviour. It has now been clarified that 'number' is a deprecated parameter and should be replaced with 'issue.
Decline reason:
Procedural decline: I hate to sound bureaucratic, but for unblock request tracking purposes, the unblock request MUST be made on that bot's individual usertalk page (been there, done that myself). Once the unblock is accepted, the redirect can be reinstated DP 14:11, 9 February 2014 (UTC)
If you want to make any further unblock requests, please read the guide to appealing blocks first, then use the {{unblock}} template again. If you make too many unconvincing or disruptive unblock requests, you may be prevented from editing this page until your block has expired. Do not remove this unblock review while you are blocked.
Now that this discussion has been resolved, the bot will treat |number= as a valid alias as 'issue' and will not rename it. Martin(Smith609 – Talk) 08:07, 3 February 2014 (UTC)
A barnstar for you!
The Original Barnstar
Thank you, you have been very helpful to me as a new user and contributor. Tonythetiger89 (talk) 16:29, 15 August 2013 (UTC)
Cite an article with a first name containing a dash.
We can't proceed until
A specific edit to the bot's code is requested below.
Requested action from maintainer
Update author formatting on Cite Doi templates
At least in Germany, "H.-P." would be the canonical shortening. No whitespace is inserted, as this is a single first name, and there was no whitespace in it originally either.
Worse, the bot even undid my manual fix in this edit, probably triggered by adding the issue, a similar change happened in one where it added the ISBN. --Chire (talk) 10:56, 3 July 2013 (UTC)
The bot undoing a manual fix is a recurring issue. I am wondering why a bot is getting so much leeway. "done so many other good things" will wear out. -DePiep (talk) 20:13, 7 July 2013 (UTC)
The bot is exclusion compliant. If there's a problem, just tell it to stay out. In those two cases, though, the entirely wrong template was in use. The sources cited were conference proceedings, not journals. In general, cite doi is only reliable for journal citations.LeadSongDogcome howl! 21:56, 9 July 2013 (UTC)
Really? This limitation is mentioned nowhere in the cite doi documentation — it says that it uses cite journal syntax, but not that it is only for journal articles. Many conference proceedings papers have dois, and the template documentation says that it can be used whenever a source has a doi. If so, it should be stated much more prominently in the docs, and this is yet another reason to ditch cite doi. —David Eppstein (talk) 22:35, 9 July 2013 (UTC)
re LeadSongDog: The bot is exclusion compliant. If there's a problem, just tell it to stay out. You do not get the concept of a bot. That "problem" is not mine to solve. It is a bug.
Frankly, LeadSongDog, your contributions on this page are predominantly along that line: "nothing wrong with the bot, it's you who has to act/think". Sure. Though your remarks look simple, they are employing a trick of deviation again and again. Now, since you play down bugs this way repeatedly, what is your association with the bot master? -DePiep (talk) 00:07, 10 July 2013 (UTC)
@DePiep: If you actually read all of what I've said here, I think you'll see your opinion was hasty. My only associations with the bot master are attempting to help resolve issues when editors find problems with a very useful but (like all others) imperfect bot. When this happens, which is obviously more than we would like, I try to clarify what the use-case is to facilitate debugging. In some simple cases, such as [[./#The bot is trying to add an f word with every edit]] (above), I've even been able to identify parts of the bot code that need to be fixed, but I don't understand it well enough to take on its operation myself. On occasion when the bot owner is not on-wiki I have used his enabled wiki-email to advise him there is an issue. That is what is known as constructive engagement. I suggest you try it yourself, rather than repeatedly engaging in groundless innuendo about fellow editors. LeadSongDogcome howl! 03:27, 10 July 2013 (UTC)
How exactly does "journal vs. proceedings" play a role here? (Plus, the citation bot should make the decision based on the DOI, instead of me having to use a different template!) Are journal authors not allowed to have dashes in their name? I bet I can find a "true" journal article where the same thing happens!. E.g. here: https://en.wikipedia.org/w/index.php?title=Template:Cite_doi/10.1007.2Fs10618-012-0300-z&oldid=562662774 - while it at least did not undo my manual changes, it failed to give the first name correctly: it lost the dash this time completely.
Just to recall: I'm more concerned about the handling of the dash, than of the bot undoing my changes. The process of preventing the bot from further edits is well documented. But I'd prefer if it gets the dashes right, without manual editing... --Chire (talk) 22:17, 13 July 2013 (UTC)
A proceedings should be cited with {{cite conference}}, which uses different formatting than {{cite journal}}. {{cite doi}} uses journal formatting, so it should not be used for conference papers. As for the example you show, the correct abbreviation for the third author is "Kriegel, H.-P.", so I agree that citation bot's removal of the dash is incorrect. It's hard to tell whether the problem is a bug in how the bot is coded, or a bug in the data source that the bot is using for this citation, though. —David Eppstein (talk) 22:30, 13 July 2013 (UTC)
I'm not seeing a data source for the abbreviation "H.-P." vice the spelled-out forename "Hans-Peter" given at crossref. Is there one? LeadSongDogcome howl! 06:12, 15 July 2013 (UTC)
Please ensure that the formatting instructions at Template:Cite doi/doc are updated to reflect the consensus reached here. Martin(Smith609 – Talk) 13:58, 17 August 2013 (UTC)
Bot limits editors to four first names and four last names.
What should happen
Bot should retrieve all editors and add "displayeditors=29" parameter if there are exactly four editors.
Replication instructions
Run the citation expander on a citation that has four editors listed but more than four editors in the original work. Here's one example: Template:Cite_doi/10.1007.2F978-0-387-78705-3 (revert the citation to four editors and then run the bot on it).
We can't proceed until
Bot operator's feedback on what is feasible
Requested action from maintainer
Remove four-editor limit from bot code and add "displayeditors=29" to citations with exactly four authors.
The bot should add "displayeditors=29" if there are exactly four editors to avoid the Lua error described for exactly 9 authors above. – Jonesey95 (talk) 23:49, 21 September 2013 (UTC)
Citation bot 458 fails to copy accented characters
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
Bot should copy a name like "Röser" correctly, but instead it writes "r�Ser". I have created many citations with the bot and have seen it copy accented characters correctly, so I don't know what's different in this case.
See the link above. Try to create a cite doi template from the above doi, or maybe one from the same source.
We can't proceed until
Operator: Bot operator's feedback on what is feasible
I'm wondering if this might be related to this edit [139] recently where the bot de-accented the characters from some names (I don't know if this is correct) but also completely lost the vowel in Guzmán, changing it to Guzmn. Metadox (talk) 04:46, 27 October 2013 (UTC)
Please would you actually fix some of these issues, or must I block the bot again? --Redrose64 (talk) 14:43, 21 November 2013 (UTC)
Note that there's a major bot overhaul underway; some volunteers are re-writing it in perl. I hope that this overhaul will address all bugs reported here. In light of this, I'm reluctant to spend too much of my currently limited time fixing bugs in the current PHP implementation. Sorry that this means a period of slight inconvenience. Martin(Smith609 – Talk) 08:46, 26 November 2013 (UTC)
That wasn't actually {{para|date||, it read |year=2009-03-17. The correct behavior would have been to change it to |date=2009-03-17. I do not, though, understand why rev 442 is still running after all the subsequent code revisions that @Smith609: has made. LeadSongDogcome howl! 15:17, 21 November 2013 (UTC)
I'll use my data dump to look for other journal cites with this |date= in |year= problem. That aside, the citationbot could at least ignore an ISO date for endash conversion. Rjwilmsi 17:22, 21 November 2013 (UTC)
That would be useful, especially if there's a ready way to fix them by bot or AWB. I'm not sure there's a clean way for a bot to know that |year=2009-03-17 was intended by the human editor as an ISO date, rather than a (malformed) year range, except by comparison with records in citation databases. Such a date range would, of course, call for the endash. LeadSongDogcome howl! 18:48, 21 November 2013 (UTC)
Further to the above: the bot also removed diacritics (|last11=Füssel became |last11=Fussel), inserted undesirable spaces (|first11=H.-M. and |first15=J.-P. became |first11=H. -M. and |first15=J. -P.), and capitalised a word which is not normally capitalised (|last15=van Ypersele became |last15=Van Ypersele. There was also no need to remove the linebreaks: that just makes comparison more difficult. --Redrose64 (talk) 18:59, 21 November 2013 (UTC)
(Please see also above at #Abbreviation of first names with dashes.) PNAS calls them Hans-Martin Füssel and Jean-Pascal van Ypersele. Pubmed also has those forenames in its XML, but gives the abbreviation as HM Füssel (neither hyphen nor endash) and JP van Ypersele (ditto). Crossref gives them H.-M. Fussel (no umlat) and J.-P. van Ypersele. So none of the database sources gives the expected abbreviated form H.-M. Füssel we'd want. Should the bot trust Crossref which mangles the surname or Pubmed which just mangles the initials? LeadSongDogcome howl! 19:55, 21 November 2013 (UTC)
I further note at Template:Cite_doi/doc the instruction about date values (to not provide a day), which would seem to entirely rule out yyyy-mm-dd date values. The two articles which transclude the template, Global_warming and United_Nations_Framework_Convention_on_Climate_Change, both use a mix of "d Month yyyy" and "Month yyyy", so the proper value for the parameter should be |date=March 2009, which would be consistent with other citations in those articles. LeadSongDogcome howl! 20:38, 21 November 2013 (UTC)
It's odd to rule out a day - the way I read "Do not specify a day." is "do not use the |day= parameter" (it's been deprecated for years). --Redrose64 (talk) 21:55, 21 November 2013 (UTC)
I think the documentation means "in |date= don't give a full date including the day, just give month & year. Or just give year in |year=". Reason being that print journal publications came out weekly/monthly or less often, so issue number & month/year mattered, exact day did not. |day= is long gone so it's not referring to that. However, now we have online-only or online first publishing, I'm particularly thinking of PLoS as a good example, the full date of publication is given on the website and it seems most sensible to give that. So while I agree with where the documentation was coming from, I'm not sure it's fully up to date for online/online first journal publications. Either way a full date or month & year lives in |date=, not |year= and the citation bot needs a bit more validation, year ranges could be checked to match nnnn-nnnn format before endashes applied. Now that we have the CS1 date errors category some cleanup will happen over time. Rjwilmsi 22:37, 21 November 2013 (UTC)
I tweaked the doc to clarify the intention refers to the day and month parameters, not the day and month parts of the |date= value. LeadSongDogcome howl! 14:20, 20 December 2013 (UTC)
Other than a block per WP:3RR, how can I stop the bot doing this? --Redrose64 (talk) 21:39, 21 November 2013 (UTC)
Put {{nobots}} on the page should do it. Rjwilmsi 22:22, 21 November 2013 (UTC)
Or {{bots|deny=Citation bot}} more specifically. Debresser (talk) 05:39, 23 December 2013 (UTC)
Not a bug. The bot should changes dashes in a |year= (which could only be e.g. 2000-2001), but should not change dashes in a |date= (which could be formatted in YY-MM-DD). Martin(Smith609 – Talk) 08:00, 3 February 2014 (UTC)
Please note that the ToolServer, on which the bot is hosted, is scheduled to be decommissioned on Jan 6th. I am in the process of transferring the bot to WikiMedia servers at http://tools.wmflabs.org/citations/. If you can help, please let me know. Otherwise I will endeavour to complete the transition as soon as possible. Martin(Smith609 – Talk) 17:32, 3 January 2014 (UTC)
User:Mattsenate has started the process of porting to Labs - see here and here, and we are planning to help with the transition in any way we can. We are also thinking about how to handle references more systematically across the English Wikipedia, other languages and other Wikimedia projects. More via email. -- Daniel Mietchen (talk) 22:28, 3 January 2014 (UTC)
! Function updateBacklog() in expandFns.php requires mysql support, not yet available on WMFlabs servers. Backlog not updated. (t) Josve05a (c) 16:23, 12 January 2014 (UTC)
This error is because I've not yet managed to implement MySQL support on the WMF servers; it just means that the bot won't have a record of which pages it has edited recently. As the bot is only running on WP page 'on demand' at the moment, the lack of this record-keeping is not problematic. Martin(Smith609 – Talk) 11:45, 13 January 2014 (UTC)
Have any of the above bugs been addressed? Just asking so that I know whether to re-report them / continue to work around them.
Bugs not marked as 'resolved' have not been addressed: this is the same instance of the bot, just running on the new host. (The new revision ID reflects changes to the source necessary for the re-homing.) I and some others are currently planning the future of Citation Bot, which will entail a fundamental re-write – so bug reports are useful in that they will help us to ensure that the bugs do not afflict the re-written script, but unfortunately will not be acted upon for a while. Martin(Smith609 – Talk) 11:45, 13 January 2014 (UTC)
The "Expand citations" link in the left side Tools list still points to the old tool (for me, at least). – Jonesey95 (talk) 16:30, 12 January 2014 (UTC)
@Jonesey95: not for me. Have you purged/cleared your cache? (t) Josve05a (c) 16:37, 12 January 2014 (UTC)
I have purged and cleared my cache, but if it's working for you, that's fine. I'm sure mine will catch up.
Thanks, I've updated this link. Martin(Smith609 – Talk) 11:45, 13 January 2014 (UTC)
I think that the left-side Tools bar still has the old link, despite what is written above. I got this link by selecting the Citation Expander Gadget in my Preferences. I believe that MediaWiki:Gadget-citations.js loads this Tool, and that page appears to have the old URL.
I also fixed the cite pmid documentation template a few hours ago. It had the old link. There appear to be other links to one of the bot's old addresses in various places, like Template:ISBN. Please do a search for them, as I do not know all of the old addresses. Some of them point to toolserver.org, others point to toolserver.de or something like that. Template:Update_children might also need updating, even though it's not a citation bot link. – Jonesey95 (talk) 23:40, 17 January 2014 (UTC)
Citation bot 4 "Cat528-dev" made many unexplainable and inappropriate changes to Memristor
I can't explain what happened here. Valid CS1 cite templates like "cite journal" and "cite press release" (and many others) were turned into "citation" templates. "author3=and others" was added to some citations. Blank |postscript= was added to many citations.
Is that approval still reasonable, given the change from Citation/core to Module/CS1? I don't think so, but Trappist the monk may be able to speak to the details.
Notwithstanding that approval, the bot put "author3=and others" into the citation using doi=10.1063/1.126902. The third author for that citation should have been "last3=Gerber | first3= C." It did the same for doi=10.1109/MC.2011.48 and doi=10.1088/0957-4484/20/34/345201
It also failed to change the spacing in "first2 = M.P.", for what it's worth.
New version live
Just to let you know that the new version of the bot has now gone live at http://tools.wmflabs.org/citations/. Users should not notice any difference, other than the resolution of many long-standing bugs. Please do be extra vigilant when initiating the bot (in case there's anything I've missed) and continue to report bugs here as usual. Cheers, Martin(Smith609 – Talk) 07:47, 4 February 2014 (UTC)
This version is a substantial rewrite and has bugs that the old version did not have (see below). It's not just a bug fix version. Have you considered whether it might be appropriate to file a new BRFA in order to get a wider set of (willing) QA testers who can put the new code through its paces? I worry that you might be asked to undo large batches of Citation Bot changes that cause errors like the ones filed below, or that the bot might be stopped by a short-fused admin. – Jonesey95 (talk) 22:29, 4 February 2014 (UTC)
Since the old version also had multiple outstanding bugs and was blocked by an admin, I don't think we're any worse off under the new system! The bot's currently only running at the request of individual users, who are hopefully checking the bot's output and reporting bugs as they arise. This should allow us to iron out any remaining bugs. Martin(Smith609 – Talk) 08:37, 12 February 2014 (UTC)
Not a bug. The page incorrectly listed multiple authors in the 'author' parameter. It should have either used the 'authors' parameter, or listed authors separately. Martin(Smith609 – Talk) 08:39, 12 February 2014 (UTC)
Clearly a bug. Author is an alias of authors and are treated identically by the {{cite journal}} template code. Boghog (talk) 14:24, 12 February 2014 (UTC)
At Laura Oldfield Ford the "display-authors" parameter was added so that all nine listed authors showed in the reference and "et al." did not. This would be desirable if there were only nine authors; however there are lots more (however the citation template has or had a limit of line).
What should happen
"Et al." should remain if there are to be eight or nine authors listed.
That book review blog has the byline "staff blogger". I've removed the long list of authors that didn't belong. LeadSongDogcome howl! 14:25, 4 February 2014 (UTC)
Bot cat515-dev converted "trans_quote" to "trans_chapter", which does not make sense and creates a CS1 error
Bot replaced |trans_quote= with |trans_chapter=. |trans_quote= is not valid, but |trans_chapter= doesn't make sense as a replacement. |trans_chapter= without |chapter= creates a CS1 error, "Pages with citations using translated terms without the original".
Not the right solution, I believe. |trans_quote= is not a valid parameter. It properly produces an error. Changing it to |trans_chapter= is not a good fix, since that does not match the original editor's intent. It should be left alone in the "unsupported parameter" category for a human editor to sort out. Thanks. – Jonesey95 (talk) 02:56, 13 February 2014 (UTC)
The bot will leave anything alone that is in its list of supported parameters. Anything that's not there will be considered a typo and the bot will do its best to fix it. Martin(Smith609 – Talk) 08:15, 13 February 2014 (UTC)
I don't have the energy to clarify much more, but it does not make sense to me to put invalid parameters in the list of supported parameters. – Jonesey95 (talk) 15:04, 13 February 2014 (UTC)
Bot 4 r531 replaced "cite doi" with "citation", creating an empty citation
A specific edit to the bot's code is requested below.
Requested action from maintainer
Stop replacing cite doi with citation. It produces empty citations.
The code for Citation Bot 4 needs to be checked to ensure that, at a minimum, it is replacing only Module/CS1 templates with "citation". Please stop Bot 4 until this is done. – Jonesey95 (talk) 15:13, 4 February 2014 (UTC)
Gosh, I didn't spot this. Bot 4 only runs under my direct supervision; I won't run it again until I've fixed this. Thanks for the spot. Martin(Smith609 – Talk) 08:16, 7 February 2014 (UTC)
Citation bot dev532 did not convert first names to initials in cite doi template
Bot failed to add "displayauthors=29" when 30 authors were added to the cite doi template. Rationale for this necessary addition is in multiple bug reports above (in short, there are more than 30 authors in the original article), and this bug was reported as fixed in a previous version.
What should happen
Bot should add "displayauthors=29" to any cite doi template where the number of authors is changed, there is no existing displayauthors or display-authors parameter, and there are exactly 9 or exactly 30 authors in the template.
A specific edit to the bot's code is requested below.
Requested action from maintainer
Bot should add "displayauthors=29" to any cite doi template where the number of authors is changed, there is no existing displayauthors or display-authors parameter, and there are exactly 9 or exactly 30 authors in the template after the bot has completed its addition of authors.
Citation bot 2 cw530 created Cite PMID template based on new Cite Jstor template
Citation Bot should ignore Cite Jstor unless the JSTOR API is working. It certainly should not create a Cite PMID template based on a Cite JSTOR template, using the JSTOR value.
Get well soon Citation Bot. The bot also seems unable to process dois. Blue Rasberry (talk) 17:41, 10 February 2014 (UTC)
Please give an example of how to replicate this bug, so I can fix it. Martin(Smith609 – Talk) 08:38, 11 February 2014 (UTC)
Works for me; unable to replicate. I created the nonexistent {{cite pmid|1980765}} in my Sandbox, hit Preview, clicked "jump the queue", and the bot created the template just fine. I still can't click "jump the queue" to create a {{cite doi}} template; I reported this error above. – Jonesey95 (talk) 08:31, 12 February 2014 (UTC)
Super, thanks for the examples; I'll investigate. Martin(Smith609 – Talk) 08:06, 13 February 2014 (UTC)
I don't know if this is the same bug, but when I click the Expand Citations gadget (or the Click here link in the template doc) in {{Cite_doi/10.1111.2Fresp.12102}}, for example, the bot gives me minimal feedback and does not expand the template. Here's the output:
Revision #532
Establishing connection to Wikipedia servers ...
Using account Citation bot.
Fetching parameter list ... done.Activated by Jonesey95
Expanding ''; will commit edits.
[21:22:16] Processing page '' — edit—history
* 0 {{Citation}} templates and 0 {{Cite XXX}} templates identified. Using dominant template {{Cite XXX}}.
# No changes required.
# # # -[#TODO unhandled DB request]-
End of output
# # #
It doesn't do anything. – Jonesey95 (talk) 02:00, 21 February 2014 (UTC)
It is Expanding ''; will commit edits. . For some reason it does not get the page name. AManWithNoPlan (talk) 01:56, 21 February 2014 (UTC)
Bot dev532 added |eprint=, which does not appear to be valid
Bot replaced |url= with |eprint=, which created two errors and one undesirable result in one citation. First, |format= requires |url=, so when |url= went away, an error was displayed. Second, |eprint= results in an "unsupported parameter" error. Third, the title link went away when |url= was removed.
What should happen
Bot should not create |eprint= and should not remove |url= when it is present.
A specific edit to the bot's code is requested below.
Requested action from maintainer
Bot should not create |eprint= and should not remove |url= when it is present.
I disagree with your assertion that the bot should preserve the url. The correct parameter to have used was |arxiv=, not |eprint=, but in this case the arxiv parameter was already set correctly (and the url was completely redundant). The correct thing to do in this case was to convert url to arxiv, notice that arxiv was already present and matching, just delete the url, and also delete the no-longer-relevant |format=. —David Eppstein (talk) 08:06, 17 February 2014 (UTC)
|url= is useful because it links |title=. If url is removed, title won't have a link. It has been stated elsewhere that readers are more likely to understand a linked title than a doi, pmid, pmc, or arxiv link. That seems a reasonable accommodation to me. I would delete the url only if it were likely to lead the reader somewhere unhelpful; that will take human judgment. – Jonesey95 (talk) 15:40, 17 February 2014 (UTC)
Deleting Cite PMID templates
Users of Citation bot might be interested in this discussion to delete the Cite PMID template family. Martin(Smith609 – Talk) 08:49, 12 February 2014 (UTC)
Bot tries to edit the article "wiki" in interactive mode
The bot tries to edit the "wiki" article instead of the specified page in interactive mode, both when using the "expand citations" gadget and when using the interactive interface.
The bot fails to load an article. I get the same result, no matter on wich article I try to load. E.g. I tried to run Atlántida (opera) and the bot gave this result: [140]. When I tried to run the bot on TheBus (Honolulu) I got this result: [141]. (The same reslut).
We can't proceed until
Bot operator's feedback on what is feasible
I get the same error/non-response when trying to run the Expand Citations gadget on {{Cite doi/10.1097.2FAOG.0b013e318193c677}}. I believe that it may be related to DOIs that are not linked on dx.doi.org. The PMID is valid in my template, but the bot does not seem to try to use it to fill in the template. – Jonesey95 (talk) 15:01, 13 February 2014 (UTC)
Please would you actually fix some of these issues, or must I block the bot again? --Redrose64 (talk) 14:43, 21 November 2013 (UTC)
Note that there's a major bot overhaul underway; some volunteers are re-writing it in perl. I hope that this overhaul will address all bugs reported here. In light of this, I'm reluctant to spend too much of my currently limited time fixing bugs in the current PHP implementation. Sorry that this means a period of slight inconvenience. Martin(Smith609 – Talk) 08:46, 26 November 2013 (UTC)
That wasn't actually {{para|date||, it read |year=2009-03-17. The correct behavior would have been to change it to |date=2009-03-17. I do not, though, understand why rev 442 is still running after all the subsequent code revisions that @Smith609: has made. LeadSongDogcome howl! 15:17, 21 November 2013 (UTC)
I'll use my data dump to look for other journal cites with this |date= in |year= problem. That aside, the citationbot could at least ignore an ISO date for endash conversion. Rjwilmsi 17:22, 21 November 2013 (UTC)
That would be useful, especially if there's a ready way to fix them by bot or AWB. I'm not sure there's a clean way for a bot to know that |year=2009-03-17 was intended by the human editor as an ISO date, rather than a (malformed) year range, except by comparison with records in citation databases. Such a date range would, of course, call for the endash. LeadSongDogcome howl! 18:48, 21 November 2013 (UTC)
Further to the above: the bot also removed diacritics (|last11=Füssel became |last11=Fussel), inserted undesirable spaces (|first11=H.-M. and |first15=J.-P. became |first11=H. -M. and |first15=J. -P.), and capitalised a word which is not normally capitalised (|last15=van Ypersele became |last15=Van Ypersele. There was also no need to remove the linebreaks: that just makes comparison more difficult. --Redrose64 (talk) 18:59, 21 November 2013 (UTC)
(Please see also above at #Abbreviation of first names with dashes.) PNAS calls them Hans-Martin Füssel and Jean-Pascal van Ypersele. Pubmed also has those forenames in its XML, but gives the abbreviation as HM Füssel (neither hyphen nor endash) and JP van Ypersele (ditto). Crossref gives them H.-M. Fussel (no umlat) and J.-P. van Ypersele. So none of the database sources gives the expected abbreviated form H.-M. Füssel we'd want. Should the bot trust Crossref which mangles the surname or Pubmed which just mangles the initials? LeadSongDogcome howl! 19:55, 21 November 2013 (UTC)
I further note at Template:Cite_doi/doc the instruction about date values (to not provide a day), which would seem to entirely rule out yyyy-mm-dd date values. The two articles which transclude the template, Global_warming and United_Nations_Framework_Convention_on_Climate_Change, both use a mix of "d Month yyyy" and "Month yyyy", so the proper value for the parameter should be |date=March 2009, which would be consistent with other citations in those articles. LeadSongDogcome howl! 20:38, 21 November 2013 (UTC)
It's odd to rule out a day - the way I read "Do not specify a day." is "do not use the |day= parameter" (it's been deprecated for years). --Redrose64 (talk) 21:55, 21 November 2013 (UTC)
I think the documentation means "in |date= don't give a full date including the day, just give month & year. Or just give year in |year=". Reason being that print journal publications came out weekly/monthly or less often, so issue number & month/year mattered, exact day did not. |day= is long gone so it's not referring to that. However, now we have online-only or online first publishing, I'm particularly thinking of PLoS as a good example, the full date of publication is given on the website and it seems most sensible to give that. So while I agree with where the documentation was coming from, I'm not sure it's fully up to date for online/online first journal publications. Either way a full date or month & year lives in |date=, not |year= and the citation bot needs a bit more validation, year ranges could be checked to match nnnn-nnnn format before endashes applied. Now that we have the CS1 date errors category some cleanup will happen over time. Rjwilmsi 22:37, 21 November 2013 (UTC)
I tweaked the doc to clarify the intention refers to the day and month parameters, not the day and month parts of the |date= value. LeadSongDogcome howl! 14:20, 20 December 2013 (UTC)
Other than a block per WP:3RR, how can I stop the bot doing this? --Redrose64 (talk) 21:39, 21 November 2013 (UTC)
Put {{nobots}} on the page should do it. Rjwilmsi 22:22, 21 November 2013 (UTC)
Or {{bots|deny=Citation bot}} more specifically. Debresser (talk) 05:39, 23 December 2013 (UTC)
Not a bug. The bot should changes dashes in a |year= (which could only be e.g. 2000-2001), but should not change dashes in a |date= (which could be formatted in YY-MM-DD). Martin(Smith609 – Talk) 08:00, 3 February 2014 (UTC)
Please note that the ToolServer, on which the bot is hosted, is scheduled to be decommissioned on Jan 6th. I am in the process of transferring the bot to WikiMedia servers at http://tools.wmflabs.org/citations/. If you can help, please let me know. Otherwise I will endeavour to complete the transition as soon as possible. Martin(Smith609 – Talk) 17:32, 3 January 2014 (UTC)
User:Mattsenate has started the process of porting to Labs - see here and here, and we are planning to help with the transition in any way we can. We are also thinking about how to handle references more systematically across the English Wikipedia, other languages and other Wikimedia projects. More via email. -- Daniel Mietchen (talk) 22:28, 3 January 2014 (UTC)
! Function updateBacklog() in expandFns.php requires mysql support, not yet available on WMFlabs servers. Backlog not updated. (t) Josve05a (c) 16:23, 12 January 2014 (UTC)
This error is because I've not yet managed to implement MySQL support on the WMF servers; it just means that the bot won't have a record of which pages it has edited recently. As the bot is only running on WP page 'on demand' at the moment, the lack of this record-keeping is not problematic. Martin(Smith609 – Talk) 11:45, 13 January 2014 (UTC)
Have any of the above bugs been addressed? Just asking so that I know whether to re-report them / continue to work around them.
Bugs not marked as 'resolved' have not been addressed: this is the same instance of the bot, just running on the new host. (The new revision ID reflects changes to the source necessary for the re-homing.) I and some others are currently planning the future of Citation Bot, which will entail a fundamental re-write – so bug reports are useful in that they will help us to ensure that the bugs do not afflict the re-written script, but unfortunately will not be acted upon for a while. Martin(Smith609 – Talk) 11:45, 13 January 2014 (UTC)
The "Expand citations" link in the left side Tools list still points to the old tool (for me, at least). – Jonesey95 (talk) 16:30, 12 January 2014 (UTC)
@Jonesey95: not for me. Have you purged/cleared your cache? (t) Josve05a (c) 16:37, 12 January 2014 (UTC)
I have purged and cleared my cache, but if it's working for you, that's fine. I'm sure mine will catch up.
Thanks, I've updated this link. Martin(Smith609 – Talk) 11:45, 13 January 2014 (UTC)
I think that the left-side Tools bar still has the old link, despite what is written above. I got this link by selecting the Citation Expander Gadget in my Preferences. I believe that MediaWiki:Gadget-citations.js loads this Tool, and that page appears to have the old URL.
I also fixed the cite pmid documentation template a few hours ago. It had the old link. There appear to be other links to one of the bot's old addresses in various places, like Template:ISBN. Please do a search for them, as I do not know all of the old addresses. Some of them point to toolserver.org, others point to toolserver.de or something like that. Template:Update_children might also need updating, even though it's not a citation bot link. – Jonesey95 (talk) 23:40, 17 January 2014 (UTC)
At Laura Oldfield Ford the "display-authors" parameter was added so that all nine listed authors showed in the reference and "et al." did not. This would be desirable if there were only nine authors; however there are lots more (however the citation template has or had a limit of line).
What should happen
"Et al." should remain if there are to be eight or nine authors listed.
That book review blog has the byline "staff blogger". I've removed the long list of authors that didn't belong. LeadSongDogcome howl! 14:25, 4 February 2014 (UTC)
Cite an article with a first name containing a dash.
We can't proceed until
A specific edit to the bot's code is requested below.
Requested action from maintainer
Update author formatting on Cite Doi templates
At least in Germany, "H.-P." would be the canonical shortening. No whitespace is inserted, as this is a single first name, and there was no whitespace in it originally either.
Worse, the bot even undid my manual fix in this edit, probably triggered by adding the issue, a similar change happened in one where it added the ISBN. --Chire (talk) 10:56, 3 July 2013 (UTC)
The bot undoing a manual fix is a recurring issue. I am wondering why a bot is getting so much leeway. "done so many other good things" will wear out. -DePiep (talk) 20:13, 7 July 2013 (UTC)
The bot is exclusion compliant. If there's a problem, just tell it to stay out. In those two cases, though, the entirely wrong template was in use. The sources cited were conference proceedings, not journals. In general, cite doi is only reliable for journal citations.LeadSongDogcome howl! 21:56, 9 July 2013 (UTC)
Really? This limitation is mentioned nowhere in the cite doi documentation — it says that it uses cite journal syntax, but not that it is only for journal articles. Many conference proceedings papers have dois, and the template documentation says that it can be used whenever a source has a doi. If so, it should be stated much more prominently in the docs, and this is yet another reason to ditch cite doi. —David Eppstein (talk) 22:35, 9 July 2013 (UTC)
re LeadSongDog: The bot is exclusion compliant. If there's a problem, just tell it to stay out. You do not get the concept of a bot. That "problem" is not mine to solve. It is a bug.
Frankly, LeadSongDog, your contributions on this page are predominantly along that line: "nothing wrong with the bot, it's you who has to act/think". Sure. Though your remarks look simple, they are employing a trick of deviation again and again. Now, since you play down bugs this way repeatedly, what is your association with the bot master? -DePiep (talk) 00:07, 10 July 2013 (UTC)
@DePiep: If you actually read all of what I've said here, I think you'll see your opinion was hasty. My only associations with the bot master are attempting to help resolve issues when editors find problems with a very useful but (like all others) imperfect bot. When this happens, which is obviously more than we would like, I try to clarify what the use-case is to facilitate debugging. In some simple cases, such as [[./#The bot is trying to add an f word with every edit]] (above), I've even been able to identify parts of the bot code that need to be fixed, but I don't understand it well enough to take on its operation myself. On occasion when the bot owner is not on-wiki I have used his enabled wiki-email to advise him there is an issue. That is what is known as constructive engagement. I suggest you try it yourself, rather than repeatedly engaging in groundless innuendo about fellow editors. LeadSongDogcome howl! 03:27, 10 July 2013 (UTC)
How exactly does "journal vs. proceedings" play a role here? (Plus, the citation bot should make the decision based on the DOI, instead of me having to use a different template!) Are journal authors not allowed to have dashes in their name? I bet I can find a "true" journal article where the same thing happens!. E.g. here: https://en.wikipedia.org/w/index.php?title=Template:Cite_doi/10.1007.2Fs10618-012-0300-z&oldid=562662774 - while it at least did not undo my manual changes, it failed to give the first name correctly: it lost the dash this time completely.
Just to recall: I'm more concerned about the handling of the dash, than of the bot undoing my changes. The process of preventing the bot from further edits is well documented. But I'd prefer if it gets the dashes right, without manual editing... --Chire (talk) 22:17, 13 July 2013 (UTC)
A proceedings should be cited with {{cite conference}}, which uses different formatting than {{cite journal}}. {{cite doi}} uses journal formatting, so it should not be used for conference papers. As for the example you show, the correct abbreviation for the third author is "Kriegel, H.-P.", so I agree that citation bot's removal of the dash is incorrect. It's hard to tell whether the problem is a bug in how the bot is coded, or a bug in the data source that the bot is using for this citation, though. —David Eppstein (talk) 22:30, 13 July 2013 (UTC)
I'm not seeing a data source for the abbreviation "H.-P." vice the spelled-out forename "Hans-Peter" given at crossref. Is there one? LeadSongDogcome howl! 06:12, 15 July 2013 (UTC)
Please ensure that the formatting instructions at Template:Cite doi/doc are updated to reflect the consensus reached here. Martin(Smith609 – Talk) 13:58, 17 August 2013 (UTC)
I can't explain what happened here. Valid CS1 cite templates like "cite journal" and "cite press release" (and many others) were turned into "citation" templates. "author3=and others" was added to some citations. Blank |postscript= was added to many citations.
Is that approval still reasonable, given the change from Citation/core to Module/CS1? I don't think so, but Trappist the monk may be able to speak to the details.
Notwithstanding that approval, the bot put "author3=and others" into the citation using doi=10.1063/1.126902. The third author for that citation should have been "last3=Gerber | first3= C." It did the same for doi=10.1109/MC.2011.48 and doi=10.1088/0957-4484/20/34/345201
It also failed to change the spacing in "first2 = M.P.", for what it's worth.
Bot 4 r531 replaced "cite doi" with "citation", creating an empty citation
A specific edit to the bot's code is requested below.
Requested action from maintainer
Stop replacing cite doi with citation. It produces empty citations.
The code for Citation Bot 4 needs to be checked to ensure that, at a minimum, it is replacing only Module/CS1 templates with "citation". Please stop Bot 4 until this is done. – Jonesey95 (talk) 15:13, 4 February 2014 (UTC)
Gosh, I didn't spot this. Bot 4 only runs under my direct supervision; I won't run it again until I've fixed this. Thanks for the spot. Martin(Smith609 – Talk) 08:16, 7 February 2014 (UTC)
Bot failed to add "displayauthors=29" when 30 authors were added to the cite doi template. Rationale for this necessary addition is in multiple bug reports above (in short, there are more than 30 authors in the original article), and this bug was reported as fixed in a previous version.
What should happen
Bot should add "displayauthors=29" to any cite doi template where the number of authors is changed, there is no existing displayauthors or display-authors parameter, and there are exactly 9 or exactly 30 authors in the template.
A specific edit to the bot's code is requested below.
Requested action from maintainer
Bot should add "displayauthors=29" to any cite doi template where the number of authors is changed, there is no existing displayauthors or display-authors parameter, and there are exactly 9 or exactly 30 authors in the template after the bot has completed its addition of authors.
Get well soon Citation Bot. The bot also seems unable to process dois. Blue Rasberry (talk) 17:41, 10 February 2014 (UTC)
Please give an example of how to replicate this bug, so I can fix it. Martin(Smith609 – Talk) 08:38, 11 February 2014 (UTC)
Works for me; unable to replicate. I created the nonexistent {{cite pmid|1980765}} in my Sandbox, hit Preview, clicked "jump the queue", and the bot created the template just fine. I still can't click "jump the queue" to create a {{cite doi}} template; I reported this error above. – Jonesey95 (talk) 08:31, 12 February 2014 (UTC)
Super, thanks for the examples; I'll investigate. Martin(Smith609 – Talk) 08:06, 13 February 2014 (UTC)
I don't know if this is the same bug, but when I click the Expand Citations gadget (or the Click here link in the template doc) in {{Cite_doi/10.1111.2Fresp.12102}}, for example, the bot gives me minimal feedback and does not expand the template. Here's the output:
Revision #532
Establishing connection to Wikipedia servers ...
Using account Citation bot.
Fetching parameter list ... done.Activated by Jonesey95
Expanding ''; will commit edits.
[21:22:16] Processing page '' — edit—history
* 0 {{Citation}} templates and 0 {{Cite XXX}} templates identified. Using dominant template {{Cite XXX}}.
# No changes required.
# # # -[#TODO unhandled DB request]-
End of output
# # #
It doesn't do anything. – Jonesey95 (talk) 02:00, 21 February 2014 (UTC)
It is Expanding ''; will commit edits. . For some reason it does not get the page name. AManWithNoPlan (talk) 01:56, 21 February 2014 (UTC)
Could you check to see whether this is now working? Let me know the outcome. Thanks. Martin(Smith609 – Talk) 18:13, 5 March 2014 (UTC)
Working for me. I tried it on twenty or so cite doi templates and cite pmid templates, and it worked every time. – Jonesey95 (talk) 18:53, 5 March 2014 (UTC)
The bot tries to edit the "wiki" article instead of the specified page in interactive mode, both when using the "expand citations" gadget and when using the interactive interface.
We can't proceed until
User
-->
Could you confirm whether this has been fixed? If not, could you help by providing a more detailed description of how to replicate the bug? Thanks. Martin(Smith609 – Talk) 19:42, 3 March 2014 (UTC)
It's been fixed; thanks! Graham87 14:35, 9 March 2014 (UTC)
The bot fails to load an article. I get the same result, no matter on wich article I try to load. E.g. I tried to run Atlántida (opera) and the bot gave this result: [142]. When I tried to run the bot on TheBus (Honolulu) I got this result: [143]. (The same reslut).
We can't proceed until
Bot operator's feedback on what is feasible
I get the same error/non-response when trying to run the Expand Citations gadget on {{Cite doi/10.1097.2FAOG.0b013e318193c677}}. I believe that it may be related to DOIs that are not linked on dx.doi.org. The PMID is valid in my template, but the bot does not seem to try to use it to fill in the template. – Jonesey95 (talk) 15:01, 13 February 2014 (UTC)
Still not running, e.g. at:
Revision #532
Establishing connection to Wikipedia servers ...
Using account Citation bot.
Fetching parameter list ... done.Activated by LeadSongDog
Expanding 'Amyotrophic_lateral_sclerosis'; will commit edits.
[01:28:06] Processing page 'Amyotrophic lateral sclerosis' — edit—history
* 0 {{citation}}: Empty citation (help) templates and 0 {{Cite XXX}} templates identified. Using dominant template {{Cite XXX}}.
# No changes required.
# # # -[#TODO unhandled DB request]-
End of output
# # #
This despite presence of cite journal templates. LeadSongDogcome howl! 01:40, 6 March 2014 (UTC)
The Bot, in my experience (I used the "jump the queue" button if that's relevant), repeatedly trascribes only the first page in the DOI URL's stated page range in the pages field.
What should happen
It should copy across the whole page range, probably replacing the hyphen with an N-dash (see MOS).
Try creating a hitherto non-existent {{cite doi}} and press "jump the queue", as a bit of a guess?
We can't proceed until
Input from editors
It seems the bot has not processed an existing page since 11 Feb except for a few page-range hyphenation edits. Something went amiss with rev 535?LeadSongDogcome howl! 08:07, 16 February 2014 (UTC)
Reporter, could this be because the database doesn't store details of the final page in a range? Can you confirm that the bot has a way to know what the final page in the range is, and let me know what that is? Martin(Smith609 – Talk) 19:35, 3 March 2014 (UTC)
Potential edit wars: the bot violates WP:RETAIN and adds extraneous code where not asked to
What happens
The bot converts authorlists from
What should happen
author=Norman, Michael J.; Andrew, James E.; Bett, Thomas H.; Clifford, Roger K.; England, John E.; Hopps, Nicholas W.; Parker, Kenneth W.; Porter, Kenneth; Stevenson, Mark to |first1=..|last1=Norman|first2=.., etc
A fix (when consensus obtained and reflected in doc)
This edit also did not space author initials. It should and always has, per WP:INITS. – Jonesey95 (talk) 06:13, 5 February 2014 (UTC)
WP:INITS applies to text and not necessarily to citations. Boghog (talk) 12:39, 15 February 2014 (UTC)
At first glance: The |authors= parameter exists for multiple authors. Placing multiple authors in the |author= parameter suggests a user input error. The bot preserves output formatting whilst making metadata available to services that use it. Martin(Smith609 – Talk) 08:12, 7 February 2014 (UTC)
Please reconsider this position. Consider that Diberri's citation template filler tool creates|author= populated with a comma-separated list of author names. This tool has been very widely used.LeadSongDogcome howl! 22:56, 7 February 2014 (UTC)
Indeed, we have several popular tools that semi-automatically fill in reference templates. Citation Bot shouldn't discriminate and change their formats (which are all valid). Materialscientist (talk) 00:41, 8 February 2014 (UTC)
There are two ways forwards here:
Start a discussion at Template talk:Cite journal to discuss changing its documentation, which clearly states that both 'author' and 'authors' are aliases of 'last' and should only be used for a single author. The template's documentation should then be updated based on the outcome of this discussion;
Request that the maintainers of these tools modify their software so that it is consistent with the documented behaviour of the templates that they are producing.
I vaguely recall the original discussion in favour of using separate author parameters; I think that the points in favour were that this made it easier to ensure that authors were consistently formatted in a page, and that it produced better metadata for users that rely on metadata to locate references in libraries that they have access to (or users that use metadata in other ways). Martin(Smith609 – Talk) 08:13, 13 February 2014 (UTC)
relevant:[151] and [152] . Seems to have happened when the great doc consolidation was going on. I don't find any related discussion. Any way you look at it, there is a lot of wiki code in place using each alias last, last1, author, or authors with a list of names, frequently with irregular separators such as , ; & and with (or even none before et al.). The bot should not be assuming wiki text is correct in all respects, that's what it's for... LeadSongDogcome howl! 15:04, 13 February 2014 (UTC)
Prior to this edit, the documentation very clearly listed two options: "Using last name, first name" and "Using author". Hence "author" and not "authors" was the recommended parameter name for storing multiple authors in a single field. Furthermore there is nothing in the current or previous documentation to suggest that if multiple authors are stored in a single parameter, that parameter should be named "authors" in preference to "author". The only discussion that I can find for the above mentioned change in the documentation is this. Hence it appears that this change in the documentation was made without consensus. Boghog (talk) 11:52, 15 February 2014 (UTC)
The current documentation is ambiguous. The documentation states that "last1, author, authors, author1" are all aliases of "last". If "author" is an alias of "last" and "authors" is an alias of "last" then "author" is an alias of "authors". Q.E.D. Also in this edit to the documentation that was made by an IP with no consensus for the change on the talk page:
For a very long time before that, a single author parameter was (and still is) the accepted way to store a multiple author list in the Vancouver system style citation. In the interest of full discloser, I partially undid the above change in this series of edits.
The advantage of the Vancouver system is that it is very compact and avoids the clutter of "first1, last1, first2, last2, ..." parameters. The number of Wikipedia editors far exceeds the number of meta data users. IMHO, having more compact citation templates imbedded in the raw text of Wikipedia articles is a distinct advantage. The Vancouver system is quite often used when there is a pmid. In these cases, the only meta data one needs is the pmid. The rest of the data can be downloaded from PubMed which is a more reliable source of bibliographic information than Wikipedia. Boghog (talk) 19:59, 13 February 2014 (UTC)
Boghog, does this report even have anything to do with your Vancouver preference? The names are already expanded with full first names. I'm not understanding why the original report is even a problem. Final article formatting is identical. If you don't want the bot to manage formatting in a way that improves availability of metadata while preserving final output formatting, don't use the bot. It can very easily be blocked on a per article or per citation basis. —Chris CapocciaT⁄C 14:43, 27 February 2014 (UTC)
@Chris Capoccia: You have completely misunderstood the above discussion. First of all, I did not originally file this bug report. Second, I did not trigger the bot on the above mentioned edits. Others did. This is a legitimate problem that requires resolution. Boghog (talk) 19:54, 3 March 2014 (UTC)
@Chris Capoccia:Final article formatting is identical. – False Without being asked, the bot converted first names to first initials. Boghog (talk) 22:07, 3 March 2014 (UTC)
@Chris Capoccia:does this report even have anything to do with your Vancouver preference? – Not in the original report, but definitely relevant to the bot maintainers response to the report. The argument advanced above that this bug report isn't a bug is based on {{cite journal}} documentation. However the current documentation includes Vancouver system examples where the recommended author parameter is "author" and not "authors". Hence storing multiple authors in the "author" parameter is consistent with the present documentation, and in particular the Vancouver system examples. Boghog (talk) 02:12, 4 March 2014 (UTC)
"the current documentation includes Vancouver system examples where the recommended author parameter is "author" and not "authors". Hence storing multiple authors in the "author" parameter is consistent with the present documentation, and in particular the Vancouver system examples."
The bot is editing in accordance with the long-standing documentation on {{cite journal}} that was modified by a single editor a few months ago. – Jonesey95 (talk) 02:41, 4 March 2014 (UTC)
@Jonesey95: Please get your facts straight. As I already explained above, I restored, not added the documentation concerning Vancouver system to the cite journal documentation. Huge difference. I would appreciate if you would careful re-read the above discussion (search for "In the interest of full discloser") and modify your comments accordingly. Boghog (talk) 04:33, 4 March 2014 (UTC)
Boghog is correct. The single |author= parameter was in the Vancouver section of the cite journal documentation for quite a while before the edits cited above. I retract my statements. – Jonesey95 (talk) 05:17, 4 March 2014 (UTC)
@Jonesey95: Thanks! I appreciate that. Boghog (talk) 05:35, 4 March 2014 (UTC)
It is important to note that "using author" was part of the original cite journal documentation that was written in 12 September 2006 which is more than year before I made my first edit on 12 October 2006 to Wikipedia. Boghog (talk) 04:56, 4 March 2014 (UTC)
The Vancouver system was explicitly added to the cite journal documentation on 26 September 2009, modified 18 October 2013 to replace "author" with "first1, last1, ...", and restored on 22 November 2013 by me. Boghog (talk) 05:04, 4 March 2014 (UTC)
Just to remind everyone what this bug report was originally about, it is the insistence by the bot, contrary to long standing and widespread usage, that if multiple authors are placed in a single data field, the parameter name "authors" instead of "author" must be used. One consequence of this was the error mentioned above. As previously reported, another type of error is this in which author names are duplicated both in the citation template and in the rendered citation. One can argue if the type behavior in the link of this bug report is acceptable or not (IMHO it is not), but clearly the behavior linked in the previous bug report is not acceptable. Boghog (talk) 01:14, 4 March 2014 (UTC)
It sounds like the resolution of this discussion is that the current behaviour is acceptable. If so, I'll mark this thread as {{resolved}}. Martin(Smith609 – Talk) 19:38, 3 March 2014 (UTC)
This issue is not resolved (see above). Boghog (talk) 19:54, 3 March 2014 (UTC)
Can't reproduce: can you give an example of where the bot changes the English variety (WP:RETAIN, quoted in report)? Martin(Smith609 – Talk) 19:36, 3 March 2014 (UTC)
OK, @Smith609: I think I now understand your comment. WP:RETAIN was linked in the original bug report. I believe @Materialscientist: meant it in a broad sense, and not narrow sense that is implied by that specific link. "Retaining the existing variety" not only applies to language, but also to dates (as the hatnote in WP:RETAIN that redirects to WP:DATERET implies), and to citation formats as per WP:CITEVAR. Boghog (talk) 20:56, 3 March 2014 (UTC)
Please help us out and use <sarcasm><sarcasm/> tags when appropriate ;-) Boghog (talk) 01:28, 4 March 2014 (UTC)
Sorry - had been skimming reports quickly, am with you now. Do ping me when you feel there's consensus on the above discussion, and this consensus is reflected by the appropriate documentation; I'll then modify the bot's code accordingly. Martin(Smith609 – Talk) 08:07, 4 March 2014 (UTC)
Meanwhile, I've modified the bot code (as of r540) so that it doesn't modify author lists as readily. Martin(Smith609 – Talk) 20:10, 8 March 2014 (UTC)
Maybe not the same "bug", but the bot keeps duplicating authors by adding last= when the authors are already listed as author=. Examples are numerous, a fresh one is [153]. Materialscientist (talk) 22:38, 5 March 2014 (UTC)
It seems the bot can recognize "displayauthors", so whatever mechanism does that, to replace it with "displayauthors" OR "display-authors".
I'm pretty sure it's better, just not fixed. See this recent diff. Under line 64, it fails to recognize display-authors=3, and goes on to add displayauthors=30.
A tip: I could be totally wrong about this, but the r490 fix message log lists it in uppercase ("Displayauthors", not "displayauthors"). Case-sensitivity perhaps? meteor_sandwich_yum (talk) 22:38, 1 February 2014 (UTC)
Seems like the sensible thing for the bot to do is replace the alias displayauthors with display-authors and be done with it. The alias should be deprecated to avoid creating more such. LeadSongDogcome howl! 02:34, 2 February 2014 (UTC)
Return of this bug
I am copying this section from the archive, because this bug is back in dev532. Please fix again. Thanks. – Jonesey95 (talk) 05:39, 7 March 2014 (UTC)
Thanks for the bump; fixed again in 539. Martin(Smith609 – Talk) 19:25, 8 March 2014 (UTC)
Proxy Error
The proxy server received an invalid response from an upstream server.
The proxy server could not handle the request GET /citations/doibot.php.
Reason: Error reading from remote server
You may also see a faster response if you don't run the bot in Slow Mode (remove 'slow=1' from the URL); the AdsAbs server seems to be unresponsive or otherwise slowing down the bot, and this server is only consulted in 'slow mode'.
This bug appears to be fixed or vanished??? AManWithNoPlan (talk) 00:53, 20 March 2014 (UTC)
Probably a WMFlabs issue then. I'll mark as resolved. Martin(Smith609 – Talk) 21:42, 22 March 2014 (UTC) {{resolved}}
New version live
Just to let you know that the new version of the bot has now gone live at http://tools.wmflabs.org/citations/. Users should not notice any difference, other than the resolution of many long-standing bugs. Please do be extra vigilant when initiating the bot (in case there's anything I've missed) and continue to report bugs here as usual. Cheers, Martin(Smith609 – Talk) 07:47, 4 February 2014 (UTC)
This version is a substantial rewrite and has bugs that the old version did not have (see below). It's not just a bug fix version. Have you considered whether it might be appropriate to file a new BRFA in order to get a wider set of (willing) QA testers who can put the new code through its paces? I worry that you might be asked to undo large batches of Citation Bot changes that cause errors like the ones filed below, or that the bot might be stopped by a short-fused admin. – Jonesey95 (talk) 22:29, 4 February 2014 (UTC)
Since the old version also had multiple outstanding bugs and was blocked by an admin, I don't think we're any worse off under the new system! The bot's currently only running at the request of individual users, who are hopefully checking the bot's output and reporting bugs as they arise. This should allow us to iron out any remaining bugs. Martin(Smith609 – Talk) 08:37, 12 February 2014 (UTC)
Endorsing as [155] Refs 2 and 3 "Bio link" should be combined but aren't. If this is not fixed in a week, I'm going to start a proposal that the Articles for Creation project de-list Citation bot from it's help templates as the bot appears to not be doing anything useful for us. It's been over 10 days since this complaint was opened and per WP:BOTCOM bot operators are supposed to respond promptly on complaints. Hasteur (talk) 13:20, 22 April 2014 (UTC)
This function has been discontinued, as it was difficult to accommodate Template:Reflist; when time permits I plan to re-enable it for pages that don't use this template. Martin(Smith609 – Talk) 15:21, 2 May 2014 (UTC)
Bot dev543 added |issue= when |number= was already present
Faced with this typo | last=Russell | first=MJ | last2=Hall | first2=AJ | last3=Cairns-Smith | first3=AG | last4=Braterman | first=PS | where the final item should be | first4=PS |, citation bot does this | last=Russell | DUPLICATE_first=MJ | last2=Hall | first2=AJ | last3=Cairns-Smith | first3=AG | last4=Braterman | first4=P. S. | first=PS |
What should happen
However, I think it would be far more logical and useful for it to declare the second occurrence as being the duplicate like this: | last=Russell | first=MJ | last2=Hall | first2=AJ | last3=Cairns-Smith | first3=AG | last4=Braterman | first4=P. S. | DUPLICATE_first=PS |
Type the wrong number on one of the author parameters creating a duplicate then run citation bot over the page
We can't proceed until
Agreement on the best solution
Requested action from maintainer
Mark the second occurrence as the duplicate.
The value that the template sees is the last value listed. Your solution would thus change the appearance of the citation, which should be avoided unless necessary. Martin(Smith609 – Talk) 13:17, 28 June 2014 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
Bot removes accessdate from citations. There is no consensus to make this removal. The bot, which, as you know, I love dearly, is exceeding its mandate with this type of edit.
A specific edit to the bot's code is requested below.
Requested action from maintainer
Disable the portion of the code that removes accessdates until there is consensus to do so.
This one could get you blocked, which would be unfortunate. Please desist ASAP. – Jonesey95 (talk) 04:36, 17 June 2014 (UTC)
Since when did we start recording the date someone looked at material printed on dead trees? An edition of a book once printed is unchanging. Surely, the dates were redundant? -- 79.67.242.63 (talk) 16:20, 17 June 2014 (UTC)
I agree. Accessdate without url is a clear error and the bot is doing us a favor by cleaning it up. —David Eppstein (talk) 16:29, 17 June 2014 (UTC)
There have been multiple discussions about |accessdate= at the various template/module/CS1 talk pages. At least:
No complete consensus has been reached. However, there appeared to be a general leaning toward: A. Adding an |accessdate= is not required for unchanging source material. B. Having an |accessdate= on a citation to an "unchanging" source is not an actual error (i.e. it is not something that negatively affects the ability of the citation to perform its primary function of providing information pointing to the reference which supports the article text). C. There is certainly no consensus that removing an |accessdate= with a valid date is appropriate at any time.
Without a clear consensus that |accessdate= should be removed, it is not something a bot should be doing. Bots are supposed to only do things for which there is a clear consensus. There is not such a consensus for removing |accessdate=. Only performing actions for which there is consensus is one of the prime requirements for a bot.
What we're talking about here is this line of code in the bot. The functionality approved five years ago was listed as "remove redundant parameters", which accessdate without URL clearly was. If much more recent discussions have created some doubt about that, the template /doc has not kept pace: Template:Cite_book/doc#URL says that accessdate requires url, which seems pretty explicit to me. LeadSongDogcome howl! 20:40, 17 June 2014 (UTC)
|accessdate= without a valid |url= is not necessarily a clear error. Editors do crazy stuff, like putting URLs in |title= or putting url: instead of url=, or adding an archiveurl without a url, or adding a URL but forgetting the "url=". If a bot had removed |accessdate= from any of those citations, it would have done so in error. You can quibble with my examples, but those are just edits I have performed in the last 24 hours. There are plenty more like them.
Among other reasons, the edits above show why there is no consensus for a bot to remove |accessdate= from citations. I am not against a bot doing so after a consensus is reached, followed by a well-thought-out BRFA. None of that has happened yet, so the bot should not perform such edits. – Jonesey95 (talk) 21:00, 17 June 2014 (UTC)
Those are all instances of you editing {{cite web}}, which is not what we are discussing. Your original complaint was about the bot removing the accessdate parameter from {{cite book}}. It is a different thing. LeadSongDogcome howl! 21:29, 17 June 2014 (UTC)
Accessdates to fixed media seem rather absurd, and in many cases I would remove them. But other cases might be debatable, and I don't trust this bot to make an appropriate distinction. Until there is consensus on when to remove accessdates, and confidence in the bot doing so appropriately, I think it should desist from doing so. ~ J. Johnson (JJ) (talk) 21:32, 18 June 2014 (UTC)
That's just the point. No one has pointed to any instances of the bot removing accessdate from non-fixed media. LeadSongDogcome howl! 22:08, 18 June 2014 (UTC)
OK, here's one, where |website=, an alias for |work=, was used instead of |url=, a mistake by an editor that needs human correction. I have made that correction.
Here's a second one like the first one. Citation bot also removed |accessdate= from citations that had URLs without "|url=", removing useful information.
Here's a third one, where the invalid |nurl= was populated with a valid URL. Again, a mistake by an editor that needs human correction.
Each of these edits shows why a bot has not yet been approved for this action. Editors do crazy stuff, and a bot, if it were approved to remove accessdates from citations for fixed media, would need to be more careful and more limited in action than Citation bot is being. – Jonesey95 (talk) 04:35, 19 June 2014 (UTC)
Thank you. Those are useful examples, even though they represent, as you say "crazy stuff". Each of those is an instance of {{cite web}}, with no url provided. Odd that the template code didn't throw up an error message, or did it? LeadSongDogcome howl! 06:30, 19 June 2014 (UTC)
Yes, each of those citations produced an error. Follow the links above and click on the previous version in the history, and you can see all sorts of errors.
I am not personally against an approved bot performing edits to remove accessdate on cite journal and cite book references, but the bot would need some basic coding (like looking for "http" somewhere in the citation, which could have meant avoiding all of the above edits) to avoid false positives. Some editors may object even to this removal, however, which is why consensus is needed before a bot can be approved. That's just basic WP process. – Jonesey95 (talk) 06:47, 19 June 2014 (UTC)
So do we have concurrence that there are instances of questionable removal of accessdate? ~ J. Johnson (JJ) (talk) 20:25, 20 June 2014 (UTC)
Yes, the bot acted incorrectly, but not on input which was correct to begin with. It changed bad wikitext to less-bad wikitext, where it could in some cases have fixed it. It looks like some of the issues could be addressed here in the code. This would be the place to fix the "website" ==> "url" parameter name substitution. I'm not sure if the "nurl" ==> "url" could also be addressed here, but possibly. The third example, where "url=" is missing but the url itself is present might have to be handled elsewhere, I can't find that in the bot code. (Also in the same table is the evident source of the error in handling "coauthor": it clearly is not correct to replace it with "author2" unless there is only one "author" listed. This requires a more nuanced approach than that part of the code can provide.) Anyhow, I'm sure these bugs are fairly simply fixed when the operator resurfaces. LeadSongDogcome howl! 21:45, 20 June 2014 (UTC)
All the bot has to do is look for the presence of "http://" or "https://" somewhere in the citation, at which point it should skip that citation entirely. Or the bot could stop removing accessdates until there is consensus that removing accessdates is the right way to address this citation error message. – Jonesey95 (talk) 01:30, 21 June 2014 (UTC)
The Cite Journal documentation specifies that the "accessdate" parameter requires that the "url" parameter is set. See Template:Cite_journal#URL. The bot is behaving in accordance of this; if the documentation needs updating, please discuss at Tempalte talk:Cite Journal and notify me when changes are made so that I can modify bot behaviour. Martin(Smith609 – Talk) 13:13, 28 June 2014 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
The bot changed the "coauthors" parameter to "authors" while leaving the separate "author" parameter untouched; this causes the primary author to be hidden.
Relevant diffs/links
Diff here - the bottom change (Heizer et al) is the problem edit.
Replication instructions
Create a cite with "author" and "coauthors" parameters, and run the bot...
We can't proceed until
Agreement on the best solution
Requested action from maintainer
Fix this, so that the author is merged into authors as the first author, or something similar, or an exception so the bot doesn't break the ref.
No idea how that mess got to be FA-class. It was improperly using template:Aut as the value for |author= and |coauthor= in CS1 templates. Cleaned up now. I hope there are not many more like that? LeadSongDogcome howl! 19:35, 27 June 2014 (UTC)
Bot completely broke
this edit was completely malformed- lots of "# # # Citation Bot : Template Placeholder 19 # # #". --PresN 20:13, 27 June 2014 (UTC)
Garbage in, garbage out. Who created that crap? LeadSongDogcome howl! 20:42, 27 June 2014 (UTC)
Endorsing as [156] Refs 2 and 3 "Bio link" should be combined but aren't. If this is not fixed in a week, I'm going to start a proposal that the Articles for Creation project de-list Citation bot from it's help templates as the bot appears to not be doing anything useful for us. It's been over 10 days since this complaint was opened and per WP:BOTCOM bot operators are supposed to respond promptly on complaints. Hasteur (talk) 13:20, 22 April 2014 (UTC)
This function has been discontinued, as it was difficult to accommodate Template:Reflist; when time permits I plan to re-enable it for pages that don't use this template. Martin(Smith609 – Talk) 15:21, 2 May 2014 (UTC)
Bot dev543 added |issue= when |number= was already present
Faced with this typo | last=Russell | first=MJ | last2=Hall | first2=AJ | last3=Cairns-Smith | first3=AG | last4=Braterman | first=PS | where the final item should be | first4=PS |, citation bot does this | last=Russell | DUPLICATE_first=MJ | last2=Hall | first2=AJ | last3=Cairns-Smith | first3=AG | last4=Braterman | first4=P. S. | first=PS |
What should happen
However, I think it would be far more logical and useful for it to declare the second occurrence as being the duplicate like this: | last=Russell | first=MJ | last2=Hall | first2=AJ | last3=Cairns-Smith | first3=AG | last4=Braterman | first4=P. S. | DUPLICATE_first=PS |
Type the wrong number on one of the author parameters creating a duplicate then run citation bot over the page
We can't proceed until
Agreement on the best solution
Requested action from maintainer
Mark the second occurrence as the duplicate.
The value that the template sees is the last value listed. Your solution would thus change the appearance of the citation, which should be avoided unless necessary. Martin(Smith609 – Talk) 13:17, 28 June 2014 (UTC)
Inconvenience: Humans must make immediate edits to clean up after the bot
What happens
The bot replaced |coauthors= with |authors= in four places. In all cases, since there was already data in either |author= or |first= and |last=, this created a duplicate parameter error. In the third edit, the bot properly parsed all the additional author names into |first#= and |last#=, but left them in |authors= as well. I'm mystified as to why it only parsed the names in this one correction instead of in all of them.
What should happen
Ideally, the bot would perform as it did in the third instance, except without replacing |coauthors= with |authors=, but instead deleting |coauthors= and what it contains, since its contents will have been properly parsed.
This edit shows the same error. |coauthors= should be replaced with |authors= only if there are no authors already listed. There are currently no articles in that error category. – Jonesey95 (talk) 23:24, 1 June 2014 (UTC)
Might I suggest that instead of just reverting the bot's mistake to the previous use of |coauthors= that you change the bot's |authors= to |author2=? That way the page gets fixed and doesn't have to be revisited. Just a thought.—D'Ranged 1VTalk 06:34, 8 June 2014 (UTC)
It seems to me there are really two kinds of problems here. First is the stuffing of multiple authors ("co" or otherwise) into a single parameter. In such cases merely removing the "co" does nothing about the problem, is entirely trivial. Second is stuffing both last and first names into a single |author=. Due to the difficulty of reliably parsing such data it would be better to leave it in a form that essentially flags it as having a problem. Searching for "author" or even "author2" could (and presumably does) include many valid instances, whereas "coauthors" effectively says "here be a problem!" As the bot can't fix these kinds of errors, I think it should not change the parameter, but leave it as a flag for a human editor. ~ J. Johnson (JJ) (talk) 22:27, 8 June 2014 (UTC)
I am fixing one or two of these errant edits every day. Please comment out this code. Monkbot is programmed to do this replacement in a way that has passed a careful set of BRFAs. Thanks. – Jonesey95 (talk) 04:23, 23 June 2014 (UTC)
Yes. "|authors" has no benefit over coauthors, and should be deprecated itself. At the very least a bot should not be mindlessly making such a pointless change. ~ J. Johnson (JJ) (talk) 20:22, 23 June 2014 (UTC)
Smith609, I just fixed 19 of these errors that were created in the last 24 hours. I have posted here. I have notified the bot's owner on the owner's Talk page. I have e-mailed the bot's owner. I don't know what else to do to get this bug fixed. I don't want to stop this bot, because it is extremely useful, but this is getting old, and it should be easy to comment out this code.
The only reason I am seeing this bug is that the articles are being placed into Category:Pages with citations having redundant parameters when there is already a |last= and |authors= is added. If |authors= is already present and |coauthors= is replaced with |authors=, information (the first author/authors in the citation) will be silently removed from the rendered article. I have no way of finding that kind of damage. – Jonesey95 (talk) 14:49, 25 June 2014 (UTC)
For a sense of scale, in the last 500 contribs, going back to 19 June, there are 56 matches for dev543 and "authors. Removed redundant parameters". Some of those are simply changing empty parameter coauthors to authors, but it is still creating an error, with both last and authors present. LeadSongDogcome howl! 23:12, 25 June 2014 (UTC)
Can someone who is familiar with PHP take a look at the code to see which lines might be commented out to turn off this |coauthors= renaming feature? The main code is here. Thanks. – Jonesey95 (talk) 04:45, 26 June 2014 (UTC)
Fixed in GitHub Pull 552. The problem arose when the deprecated parameter |coauthors= was removed from the bot's list of valid parameters. It will now be replaced with a supported synonym, |author2=. Martin(Smith609 – Talk) 15:06, 28 June 2014 (UTC)
Unless you're using logic like that programmed into Monkbot, which is really a collection of five different programs to parse |coauthors=, I suggest leaving |coauthors= alone. Will Citation Bot change |coauthors= to |author2= if |last2= is already present? It should not, because it will create redundant parameter errors. Will Citation Bot ignore |coauthors= within {{citation}} or when ref = harv is present? It should, because it will break links from {{sfn}} references. This is why BRFA and test edits are a good thing. – Jonesey95 (talk) 15:59, 28 June 2014 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
Bot removes accessdate from citations. There is no consensus to make this removal. The bot, which, as you know, I love dearly, is exceeding its mandate with this type of edit.
A specific edit to the bot's code is requested below.
Requested action from maintainer
Disable the portion of the code that removes accessdates until there is consensus to do so.
This one could get you blocked, which would be unfortunate. Please desist ASAP. – Jonesey95 (talk) 04:36, 17 June 2014 (UTC)
Since when did we start recording the date someone looked at material printed on dead trees? An edition of a book once printed is unchanging. Surely, the dates were redundant? -- 79.67.242.63 (talk) 16:20, 17 June 2014 (UTC)
I agree. Accessdate without url is a clear error and the bot is doing us a favor by cleaning it up. —David Eppstein (talk) 16:29, 17 June 2014 (UTC)
There have been multiple discussions about |accessdate= at the various template/module/CS1 talk pages. At least:
No complete consensus has been reached. However, there appeared to be a general leaning toward: A. Adding an |accessdate= is not required for unchanging source material. B. Having an |accessdate= on a citation to an "unchanging" source is not an actual error (i.e. it is not something that negatively affects the ability of the citation to perform its primary function of providing information pointing to the reference which supports the article text). C. There is certainly no consensus that removing an |accessdate= with a valid date is appropriate at any time.
Without a clear consensus that |accessdate= should be removed, it is not something a bot should be doing. Bots are supposed to only do things for which there is a clear consensus. There is not such a consensus for removing |accessdate=. Only performing actions for which there is consensus is one of the prime requirements for a bot.
What we're talking about here is this line of code in the bot. The functionality approved five years ago was listed as "remove redundant parameters", which accessdate without URL clearly was. If much more recent discussions have created some doubt about that, the template /doc has not kept pace: Template:Cite_book/doc#URL says that accessdate requires url, which seems pretty explicit to me. LeadSongDogcome howl! 20:40, 17 June 2014 (UTC)
|accessdate= without a valid |url= is not necessarily a clear error. Editors do crazy stuff, like putting URLs in |title= or putting url: instead of url=, or adding an archiveurl without a url, or adding a URL but forgetting the "url=". If a bot had removed |accessdate= from any of those citations, it would have done so in error. You can quibble with my examples, but those are just edits I have performed in the last 24 hours. There are plenty more like them.
Among other reasons, the edits above show why there is no consensus for a bot to remove |accessdate= from citations. I am not against a bot doing so after a consensus is reached, followed by a well-thought-out BRFA. None of that has happened yet, so the bot should not perform such edits. – Jonesey95 (talk) 21:00, 17 June 2014 (UTC)
Those are all instances of you editing {{cite web}}, which is not what we are discussing. Your original complaint was about the bot removing the accessdate parameter from {{cite book}}. It is a different thing. LeadSongDogcome howl! 21:29, 17 June 2014 (UTC)
Accessdates to fixed media seem rather absurd, and in many cases I would remove them. But other cases might be debatable, and I don't trust this bot to make an appropriate distinction. Until there is consensus on when to remove accessdates, and confidence in the bot doing so appropriately, I think it should desist from doing so. ~ J. Johnson (JJ) (talk) 21:32, 18 June 2014 (UTC)
That's just the point. No one has pointed to any instances of the bot removing accessdate from non-fixed media. LeadSongDogcome howl! 22:08, 18 June 2014 (UTC)
OK, here's one, where |website=, an alias for |work=, was used instead of |url=, a mistake by an editor that needs human correction. I have made that correction.
Here's a second one like the first one. Citation bot also removed |accessdate= from citations that had URLs without "|url=", removing useful information.
Here's a third one, where the invalid |nurl= was populated with a valid URL. Again, a mistake by an editor that needs human correction.
Each of these edits shows why a bot has not yet been approved for this action. Editors do crazy stuff, and a bot, if it were approved to remove accessdates from citations for fixed media, would need to be more careful and more limited in action than Citation bot is being. – Jonesey95 (talk) 04:35, 19 June 2014 (UTC)
Thank you. Those are useful examples, even though they represent, as you say "crazy stuff". Each of those is an instance of {{cite web}}, with no url provided. Odd that the template code didn't throw up an error message, or did it? LeadSongDogcome howl! 06:30, 19 June 2014 (UTC)
Yes, each of those citations produced an error. Follow the links above and click on the previous version in the history, and you can see all sorts of errors.
I am not personally against an approved bot performing edits to remove accessdate on cite journal and cite book references, but the bot would need some basic coding (like looking for "http" somewhere in the citation, which could have meant avoiding all of the above edits) to avoid false positives. Some editors may object even to this removal, however, which is why consensus is needed before a bot can be approved. That's just basic WP process. – Jonesey95 (talk) 06:47, 19 June 2014 (UTC)
So do we have concurrence that there are instances of questionable removal of accessdate? ~ J. Johnson (JJ) (talk) 20:25, 20 June 2014 (UTC)
Yes, the bot acted incorrectly, but not on input which was correct to begin with. It changed bad wikitext to less-bad wikitext, where it could in some cases have fixed it. It looks like some of the issues could be addressed here in the code. This would be the place to fix the "website" ==> "url" parameter name substitution. I'm not sure if the "nurl" ==> "url" could also be addressed here, but possibly. The third example, where "url=" is missing but the url itself is present might have to be handled elsewhere, I can't find that in the bot code. (Also in the same table is the evident source of the error in handling "coauthor": it clearly is not correct to replace it with "author2" unless there is only one "author" listed. This requires a more nuanced approach than that part of the code can provide.) Anyhow, I'm sure these bugs are fairly simply fixed when the operator resurfaces. LeadSongDogcome howl! 21:45, 20 June 2014 (UTC)
All the bot has to do is look for the presence of "http://" or "https://" somewhere in the citation, at which point it should skip that citation entirely. Or the bot could stop removing accessdates until there is consensus that removing accessdates is the right way to address this citation error message. – Jonesey95 (talk) 01:30, 21 June 2014 (UTC)
Looking for 'http' is a good idea, but what if the user mistyped this too? It sounds like the community has a low tolerance for 'Garbage in, Garbage out' bot actions, and would prefer a 'Garbage in, Garbage left as garbage' approach, so I'll aim to adopt this where I can. As such I've updated the bot (in the latest release, soon to go live) so that it doesn't touch the accessdate. Martin(Smith609 – Talk) 15:02, 28 June 2014 (UTC)
Thank you. I would characterize the community as having a low tolerance for erroneous edits by bots that could have been avoided. There is a lot of garbage out there; that's just reality. Bots need to live in the real world and avoid making errors to the extent they can. – Jonesey95 (talk) 16:07, 28 June 2014 (UTC)
And perhaps the authors/coauthors issues could also be dealt with? ~ J. Johnson (JJ) (talk) 20:32, 28 June 2014 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
The bot changed the "coauthors" parameter to "authors" while leaving the separate "author" parameter untouched; this causes the primary author to be hidden.
Relevant diffs/links
Diff here - the bottom change (Heizer et al) is the problem edit.
Replication instructions
Create a cite with "author" and "coauthors" parameters, and run the bot...
We can't proceed until
Agreement on the best solution
Requested action from maintainer
Fix this, so that the author is merged into authors as the first author, or something similar, or an exception so the bot doesn't break the ref.
No idea how that mess got to be FA-class. It was improperly using template:Aut as the value for |author= and |coauthor= in CS1 templates. Cleaned up now. I hope there are not many more like that? LeadSongDogcome howl! 19:35, 27 June 2014 (UTC)
Bot completely broke
this edit was completely malformed- lots of "# # # Citation Bot : Template Placeholder 19 # # #". --PresN 20:13, 27 June 2014 (UTC)
Garbage in, garbage out. Who created that crap? LeadSongDogcome howl! 20:42, 27 June 2014 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
Bot should copy a name like "Röser" correctly, but instead it writes "r�Ser". I have created many citations with the bot and have seen it copy accented characters correctly, so I don't know what's different in this case.
See the link above. Try to create a cite doi template from the above doi, or maybe one from the same source.
We can't proceed until
Operator: Bot operator's feedback on what is feasible
I'm wondering if this might be related to this edit [157] recently where the bot de-accented the characters from some names (I don't know if this is correct) but also completely lost the vowel in Guzmán, changing it to Guzmn. Metadox (talk) 04:46, 27 October 2013 (UTC)
This edit may be similar. It created last17=Schr|[Ouml]|Der.
The accents are not stored correctly in CrossRef, so this is a case of GIGO. Martin(Smith609 – Talk) 13:33, 12 July 2014 (UTC)
Bot cat515-dev converted "trans_quote" to "trans_chapter", which does not make sense and creates a CS1 error
Bot replaced |trans_quote= with |trans_chapter=. |trans_quote= is not valid, but |trans_chapter= doesn't make sense as a replacement. |trans_chapter= without |chapter= creates a CS1 error, "Pages with citations using translated terms without the original".
Not the right solution, I believe. |trans_quote= is not a valid parameter. It properly produces an error. Changing it to |trans_chapter= is not a good fix, since that does not match the original editor's intent. It should be left alone in the "unsupported parameter" category for a human editor to sort out. Thanks. – Jonesey95 (talk) 02:56, 13 February 2014 (UTC)
The bot will leave anything alone that is in its list of supported parameters. Anything that's not there will be considered a typo and the bot will do its best to fix it. Martin(Smith609 – Talk) 08:15, 13 February 2014 (UTC)
I don't have the energy to clarify much more, but it does not make sense to me to put invalid parameters in the list of supported parameters. – Jonesey95 (talk) 15:04, 13 February 2014 (UTC)
Can you suggest a scalable fix? i.e. how should the bot know which parameters to 'leave alone', and which are misspellings that should be corrected? If trans_quote is used often, then perhaps we should add a tracking category to template:Cite journal. Martin(Smith609 – Talk) 09:23, 9 March 2014 (UTC)
How would this fix ensure that misspelt parameters (e.g. juornal) were replaced with their correct spellings? Martin(Smith609 – Talk) 18:07, 10 March 2014 (UTC)
Also that category seems to list all citations with unsupported parameters, not just those using trans_quote. Martin(Smith609 – Talk) 18:07, 10 March 2014 (UTC)
Bot replaced |url= with |eprint=, which created two errors and one undesirable result in one citation. First, |format= requires |url=, so when |url= went away, an error was displayed. Second, |eprint= results in an "unsupported parameter" error. Third, the title link went away when |url= was removed.
What should happen
Bot should not create |eprint= and should not remove |url= when it is present.
A specific edit to the bot's code is requested below.
Requested action from maintainer
Bot should not create |eprint= and should not remove |url= when it is present.
I disagree with your assertion that the bot should preserve the url. The correct parameter to have used was |arxiv=, not |eprint=, but in this case the arxiv parameter was already set correctly (and the url was completely redundant). The correct thing to do in this case was to convert url to arxiv, notice that arxiv was already present and matching, just delete the url, and also delete the no-longer-relevant |format=. —David Eppstein (talk) 08:06, 17 February 2014 (UTC)
|url= is useful because it links |title=. If url is removed, title won't have a link. It has been stated elsewhere that readers are more likely to understand a linked title than a doi, pmid, pmc, or arxiv link. That seems a reasonable accommodation to me. I would delete the url only if it were likely to lead the reader somewhere unhelpful; that will take human judgment. – Jonesey95 (talk) 15:40, 17 February 2014 (UTC)
This bug is still present in dev543. See this edit. – Jonesey95 (talk) 16:42, 13 March 2014 (UTC)
|url= is also useful when the DOI, PMC, PMID, or other identifier is incorrect. I have recently fixed dozens of citations with incorrect DOI, PMID, and PMC values. Removing the URL removes a reader's second chance to find the correct cited source. Please do not remove the URL. – Jonesey95 (talk) 02:05, 5 April 2014 (UTC)
I have had similar experiences correcting DOIs. I think it would be best to preserve the urls. —PC-XT+ 06:53, 7 April 2014 (UTC)
I've implemented these suggestions in r572. Martin(Smith609 – Talk) 13:04, 12 July 2014 (UTC)
Deleting Cite PMID templates
Users of Citation bot might be interested in this discussion to delete the Cite PMID template family. Martin(Smith609 – Talk) 08:49, 12 February 2014 (UTC)
{{resolved}}
Bot adds "first1, last1, ..." parameters to citations where authors have already been specified with the "authors" parameter
This part of the bot code should be modified so that it respects WP:CITEVAR
It was suggested (but not agreed to) in a previous bug report, that if multiple authors are placed in a single field, the parameter name should be "authors" instead of "author". The above diff demonstrates the following errors occur when the bot processes a citation where the full list of authors is stored in a single "authors" parameter:
No matter how many authors are present, the "authors" parameter is renamed to author4.
By adding "first1, last1, ..." parameters, the author format changes from the Vancouver system (authors separated by commas, no punctuation between last names and first/middle initials, and no periods after initials) to a different system (authors separated by semicolons, comma between last and first/ initial, and periods after initials). Furthermore the new citations are not even consistently formatted (some contain spelled out first names and others contain first name initials.) Both behaviors are a clear violation of WP:CITEVAR. At the very least, if additional "first1, last1, ..." parameters are added to a citation that only contains a single "authors" parameter where the authors are formatted in Vancouver style, the following parameters should also be added: "| authorformat = vanc | author-separator = , | author-name-separator =   ". A much better solution is not to add the "first1, last1, ..." parameters in the first place (see below).
Adding "first1, last1, ..." parameters to an article that has not used them previously is a violation of WP:CITEVAR. Please note there is no consensus that "first1, last1, ..." or "author1, author2, ..." parameters must be used to stored multiple authors. Per consensus and long established usage and consistent with the current {{cite journal}} documentation, full author lists can be stored in a single field called either "authors" or "author" without need for additional numbered author parameters. Boghog (talk) 21:18, 7 March 2014 (UTC)
Please point me to the page that documents the correct use of the 'authors' parameter so that I cam update the bot's behaviour accordingly. (I've now added the parameter to the list of valid parameters, so it won't be changed to author4.) Martin(Smith609 – Talk) 18:53, 8 March 2014 (UTC)
Great, thanks. So, the "author" and "authors" parameters should only be used to give the last name of the author, and should never contain wikilinks? Sounds like much of the recent discussion here has taken a different stance. Martin(Smith609 – Talk) 08:28, 9 March 2014 (UTC)
I have always worked under the assumption that |last=|first= are for a person and |author= is for a committee, department or organisation. I have therefore avoided using |authors=, |coauthor= and |coauthors=. -- 79.67.241.76 (talk) 00:32, 26 March 2014 (UTC)
If it is desired to wikilink the author(s), the |authorlink= parameters are available for that. |authorlink=John Doe|last=Doe|first=J. or |authorlink=John Doe|author=Doe, J. will both yield output that is visually identical to |author=[[John Doe|Doe, J.]] but emit clean metadata. It's all covered at Template:Cite journal#csdoc_author. --Redrose64 (talk) 10:54, 26 March 2014 (UTC)
Whether the citation uses author, or authors, or last, it's less confusing to humans to stick with that same alias for the corporate authors too, e.g. |last=International Council of Medical Journal Editors
Please point me to the page that documents the correct use of the 'authors' parameter so that I can update the bot's behaviour accordingly. – Template:Cite_journal#Usage (See Vancouver system citation of a scientific journal - Year ). The bot should not touch the 'author(s)' parameter or add additional "author1, author2, ..." or "first1, last1, first2, last2, ..." parameters if not asked to. Boghog (talk) 14:07, 29 March 2014 (UTC)
Citation bot has been doing this sort of thing for more than three years, turning this <ref name="iauc5731">
{{cite journal | author=J. Ripero, F. Garcia, D. Rodriguez, P. Pujol, A. V. Filippenko, R. R. Treffers, Y. Paik, M. Davis, D. Schlegel, F. D. A. Hartwick, D. D. Balam, D. Zurek, R. M. Robb, P. Garnavich, B. A. Hong | year=1993 | title=Supernova 1993J in NGC 3031 | journal=[[IAU Circular]] | volume=5731 | pages=1 | bibcode=1993IAUC.5731....1R }}</ref> into this <ref name="iauc5731">{{cite journal | author=J. Ripero, F. Garcia, D. Rodriguez, P. Pujol, A. V. Filippenko, R. R. Treffers, Y. Paik, M. Davis, D. Schlegel, F. D. A. Hartwick, D. D. Balam, D. Zurek, R. M. Robb, P. Garnavich, B. A. Hong | year=1993 | title=Supernova 1993J in NGC 3031 | journal=[[IAU Circular]] | volume=5731 | pages=1 | bibcode=1993IAUC.5731....1R | last2=Garcia | last3=Rodriguez | last4=Pujol | last5=Filippenko | last6=Treffers | last7=Paik | last8=Davis | last9=Schlegel }}</ref> and I dread to think how many thousands of articles now have corrupted author lists. -- 79.67.241.229 (talk) 23:59, 3 April 2014 (UTC)
An edit from a few hours ago.... Citation bot adds |last2=|first2=, etc, but one of those numbered entries contains a duplicate of the author name already previously specified in the |author= parameter. -- 79.67.241.229 (talk) 01:21, 4 April 2014 (UTC)
yes, it's still making a mess, such as this edit. When a citation had a preexisting |author=Smith AB the bot is adding |last2=Jones |first2=C. D. (etc.) creating a mix of styles even after last month's changes to the code. See the top of this page. LeadSongDogcome howl! 04:23, 4 April 2014 (UTC)
The lack of a space before the pipes for the added author names makes them incredibly hard to read. Please add a space before each pipe when adding parameters to horizontal-format references. -- 79.67.241.229 (talk) 09:15, 4 April 2014 (UTC)
When Citation Bot comes across something like this:
In both cases the amended references would be complete. The now redundant original parameter could be manually removed later. -- 79.67.241.229 (talk) 13:08, 7 April 2014 (UTC)
The proposed solution immediately above is not acceptable as it is a clear violation of WP:CITEVAR. It is much better to leave the "authors" parameter untouched and not introduce "first1, last1, first2, last2, ..." parameter bloat. Boghog (talk) 14:04, 7 April 2014 (UTC)
I don't think WP:CITEVAR was ever intended to apply to minutiae such as this. We've recently seen |coauthor= and |coauthors= deprecated. Looking forward to seeing |authors= and |editors= deprecated, and |author= clearly defined as being intended for a single entry, not for lists of names. -- 79.67.241.229 (talk) 14:13, 7 April 2014 (UTC)
WP:CITEVAR was ever intended to apply to minutiae such as this – false First of all, replacing a single parameter with multiple parameters is not minutiae, especially if the citation has a large number of authors. Second
there have been multiple discussions on whether [CITEVAR] applies to these trivial aspects of internal formatting, and the discussions always end with a rough consensus that it does apply to basically anything and everything. — User:WhatamIdoing — diff
Also please note that the change is not only to the internal format of the citation, but can also affect how the citation is rendered. Boghog (talk) 09:22, 13 April 2014 (UTC)
|author= clearly defined as being intended for a single entry, not for lists of names – false. That change in the documentation was made fairly recently without wide consensus. Furthermore the current cite journal documentation (see Template:Cite_journal#Usage, see Vancouver system citation of a scientific journal - Year ) clearly documents placing a list of names in a single author parameter. There is no reason to deprecate the author parameter. Boghog (talk) 15:32, 7 April 2014 (UTC)
Many of the references in these articles were generated using Diberri's template filling tool which generates both compact templates and also more compact output in the rendered citations by adopting Vancouver system author formatting. One alternative possibility is to use something like the {{Vcite2 journal}} template which enforces Vancouver system output no matter how the authors are input. (Note this template uses the same Module:Citation/CS1 that {{cite journal}} uses.) This template could also be modified to parse the single author parameter to produce proper author metadata. This would allow both compact templates and full metadata making everyone happy. Boghog (talk) 15:44, 7 April 2014 (UTC)
There are distinctly separate issues here of 1) whether "author(s)" should be split into first/last parameters, and 2) whether multiple authors (as implied by "authors") should go into a single parameter. I say the latter is such an auful abuse of the data it ought to be deprecated (like "coauthors") regardless of an prior usage. ~ J. Johnson (JJ) (talk) 20:21, 7 April 2014 (UTC)
"first1, last1, first2, last2, ..." is an awful abuse of the data. This parameter bloat unnecessarily inflates the size of citation templates which makes the templates themselves harder to read. In addition, this parameter bloat overwhelms the surrounding wikitext to generate metadata that no one uses. Conciseness is important. The purpose of citations is support the surrounding wikitext, not the other way around. Hence what should be deprecated is "first1, last1, first2, last2, ..." and not "author(s)". Boghog (talk) 20:46, 7 April 2014 (UTC)
The principal of "first, do no harm" should also apply to bots. Regardless of the current documentation, there are a huge number of citations where multiple authors are stored in a single parameter. Furthermore this usage has not been deprecated. Unless (1) the usage is formally deprecated and (2) there is a bot run to systematically convert all these citations, then Citation Bot should not touch the author parameters of these citations, unless the bot can be modified so that it (1) doesn't introduce redundant parameters and (2) doesn't mangle wiki links that are contained within this author lists. To simply state that there is an input error is incorrect. It ignores the reality of how many, many citations are currently formatted. Boghog (talk) 09:22, 13 April 2014 (UTC)
I believe this suggestion to represent the state of play at r570. If this is not the case, please file a new bug report detailing specific violations. Martin(Smith609 – Talk) 12:55, 12 July 2014 (UTC)
A specific edit to the bot's code is requested below.
Thanks for the report; will look into it anon. Martin(Smith609 – Talk) 18:46, 9 March 2014 (UTC)
Everytime I find something the bot does wrong I revert (/undo) the edit. That means that the bot-account will get a notification. Is there someone checking these notifications on a regular basis? (t) Josve05a (c) 19:07, 9 March 2014 (UTC)
No, I wasn't aware of this function. Sounds like a useful feature. Martin(Smith609 – Talk) 19:02, 10 March 2014 (UTC)
Still adds these "so-called" templates. (t) Josve05a (c) 19:57, 7 July 2014 (UTC)
Citation bot 1 [Cat547-dev] changed |isbn2= to |issn=
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
Bot replaced unsupported parameter |isbn2= with |issn=, creating an ISSN error.
What should happen
Bot should ignore |isbn2=, which already causes an "unsupported parameter" error and will be dealt with by a human editor. Changing |isbn2= to |issn= just confuses things.
How many pages contain the "isbn2" parameter? I wonder whether a manual fix might be possible here. Martin(Smith609 – Talk) 22:52, 16 March 2014 (UTC)
I don't know, but a manual fix is the right answer. The bot should leave this parameter alone for a human to fix. – Jonesey95 (talk) 00:03, 17 March 2014 (UTC)
By manual fix, I meant that the typo is probably rare enough that a human could manually replace all the occurrences with the correct parameter name. If it is too common for this approach, could you please provide a list of all common-but-wrong parameters that the bot should leave alone, and detail how humans should be made aware that they need correcting? Martin(Smith609 – Talk) 09:33, 17 March 2014 (UTC)
I agree that the typo is probably rare, and that a human should do these edits, hence my bug report and suggested action above.
The "list" of wrong parameters is all of the parameters that are not parameters or aliases allowable in CS1 cite templates. Editors are already made aware of the need for correction by a red error message that is applied to all citations categorized in Category:Pages with citations using unsupported parameters. At this writing, there are about 6,000 articles in that category. – Jonesey95 (talk) 16:15, 17 March 2014 (UTC)
What about "juornal", which can trivially be corrected to "journal"? Martin(Smith609 – Talk) 21:40, 22 March 2014 (UTC)
No webservice
The URI you have requested, http://tools.wmflabs.org/?503, is not currently serviced.
If you have reached this page from somewhere else...
This URI is not currently part of any tool.
If you're pretty sure this shouldn't be an error, you may wish to notify the project administrators about the error and how you ended up here.
What is worng? (t) Josve05a (c) 16:32, 17 March 2014 (UTC)
No webservice
The URI you have requested, http://tools.wmflabs.org/citations-dev/doibot.php?edit=toolbar&slow=1&user=Josve05a&page=H_II_region, is not currently serviced.
If you have reached this page from somewhere else...
This URI is part of the citations-dev tool, maintained by Smith609.
That tool might not have a web interface, or it may currently be disabled.
If you're pretty sure this shouldn't be an error, you may wish to notify the tool's maintainers (above) about the error and how you ended up here.
If you maintain this tool
You have not enabled a web service for your tool, or it has stopped working because of a fatal error. You may wish to check your logs or common causes for errors in the help documentation.
I'm not available at the moment, but could this be a WMFlabs issue? Is this error message occurring intermittently or consistently? I'm not aware of what action I need to take (and don't have time to investigate at present). Martin(Smith609 – Talk) 21:39, 22 March 2014 (UTC)
It happens all the times I try to access the tool. It has also happened on other tools, so it is possibly a WMFLabs-problem. (t) Josve05a (c) 21:41, 22 March 2014 (UTC)
I don't know if this is a bug in the bot or something on the far end. Clicking on the DOI in this template results in an article, but running the bot on it using the "Expand citations" gadget results in this:
Revision #543
Establishing connection to Wikipedia servers ...
Using account Citation bot.
Fetching parameter list ... done.Activated by Jonesey95
Expanding 'Template:Cite_doi/10.3233.2FSW-2012-0088'; will commit edits.
[04:28:31] Processing page 'Template:Cite doi/10.3233.2FSW-2012-0088' — edit—history
* 0 {{Citation}} templates and 1 {{Cite XXX}} templates identified. Using dominant template {{Cite XXX}}.
* Expand citation: [..> process> rifydoi]
. Checking that DOI 10.3233/SW-2012-0088 is operational...
! Broken doi: 10.3233/SW-2012-0088
- Checking AdsAbs database [..> process> yadsabs]: no record retrieved. [..> process> indpmid]
- Searching PubMed... no results. nothing found.
# No changes required.
# # # -[#TODO unhandled DB request]-
End of output
# # #
Looks to me like that DOI was not yet in the CrossRef database. I'm marking this as fixed as it is not occurring elsewhere. Let me know if this isn't the case. Martin(Smith609 – Talk) 12:45, 12 July 2014 (UTC)
The bot inserted two endashes into |year=, replacing hyphens. I presume that the bot has code looking for "YYYY-YY" in the year field and changes it to "YYYY–YY". That makes sense, but if the bot finds two dashes in the same field, it should either ignore the whole thing or (possibly) change "year" to "date", since the value was intended as YYYY-MM-DD, which needs to go in |date=. The latter "solution" could end up with some buggy changes as well. It's probably better to just ignore this garbage, since it's already flagged with a CS1 date error. A human will get to it eventually. – Jonesey95 (talk) 21:54, 21 March 2014 (UTC)
Physics articles are well known for having many co-authors, sometimes upto 100. Its standard practices just to list the first with et. al. It really does not help to list all authors as in the above diff.--Salix alba (talk): 18:57, 24 March 2014 (UTC)
Salix alba, just add |displayauthors=1 to the citation, which will display it as you desire. Unless it has a bug, Citation bot will respect that parameter. Adding authors to citations is a feature of Citation bot. If you want to keep the bot out of the article entirely, you can do so. Instructions are at User:Citation_bot#Stopping_the_bot_from_editing. – Jonesey95 (talk) 00:41, 25 March 2014 (UTC)
The other problem is that its not picking up the et al. in the first authors name
N. Agafonova et al. (OPERA Collaboration); Aleksandrov; Altinok; Ambrosio; Anokhina; Aoki; Ariga; Ariga; Autiero; Badertscher; Bagulya; Bendhabi; Bertolin; Besnier; Bick; Boyarkin; Bozza; Brugière; Brugnera; Brunet; Brunetti; Buontempo; Cazes; Chaussard; Chernyavsky; Chiarella; Chon-Sen; Chukanov; Ciesielski; Corso (2010). "Observation of a first ντ candidate event in the OPERA experiment in the CNGS beam". Physics Letters B. 691 (3): 138–145. arXiv:1006.1623. Bibcode:2010PhLB..691..138A. doi:10.1016/j.physletb.2010.06.022. {{cite journal}}: Unknown parameter |displayauthors= ignored (|display-authors= suggested) (help)
In the first I've left it in, in the second I've removed it. Neither look ideal as the et al. should be before the name of the collaboration. Related to this is it seems have duplicated the second author, perhaps because it could not understand the first author.--Salix alba (talk): 13:30, 25 March 2014 (UTC)
Thanks for the explanation. I don't think this is a bug, but I'll let the bot's operator weigh in. Since you're doing something non-standard with these author fields, you'll want to use the bot exclusion template and then format the article as you see fit. – Jonesey95 (talk) 14:54, 25 March 2014 (UTC)
I think this is resolved in 566. Let me know if you spot any side-effects. Martin(Smith609 – Talk) 11:34, 12 July 2014 (UTC)
Bot finds inconsistent journal titles, does nothing
For an example, the bot occasionally produces text like this during a run:
* Expand citation: Towards sustainable solutions for fly ash through mechanical Activation [..> process> rifydoi]
. Checking that DOI 10.1016/j.resconrec.2007.06.007 is operational... DOI ok.
- Checking AdsAbs database [..> process> yadsabs]
Match for pagination but database journal "Physics in Medicine and Biology" didn't match "journal = Resources Conservation and Recycling". [..> process> indpmid]
- Searching PubMed... no results. nothing found.
Basically it appears that the bot is looking at the pmid and consulting a database to determine the journal title. Then does nothing.
Run bot on a page with cites containing either DOIs or PMIDs
We can't proceed until
Bot operator's feedback on what is feasible
Requested action from maintainer
explanation
The bot doesn't add information if the information may relate to a different article in a different journal. Martin(Smith609 – Talk) 11:25, 12 July 2014 (UTC)
Please check line 897 of DOItools.php and line 690 of expandFns.php for apparent typos. LeadSongDogcome howl! 04:25, 26 March 2014 (UTC)
{{resolved}} good catch, thanks Martin(Smith609 – Talk) 11:06, 12 July 2014 (UTC)
At Laura Oldfield Ford the bot has twice added "|display-authors=9" to an instance of {{cite web}} with nine listed authors. This has the effect of removing "et al" from the end of the list of authors; however, because the cited article has more than nine authors, "et al" should be present so as not to give the impression that the listed authors are the only authors. I've pointed this out here before. There was a discussion last month at Talk:Laura Oldfield Ford as to whether listing the authors was advisable, which ended inconclusively. If there were a consensus there that listing the authors wasn't advisable, this would still be a problem for any/all other articles using {{cite web}} to cite articles with more than nine authors.
The bot did this edit in order to remove a red CS1 citation error that displays when there are exactly nine authors and no display-authors parameter. I do not know how the bot determined that "9" was the correct value; sometimes it inserts a value of "8" instead. I have added the parameter with a value of "8" in order to mimic the previous display without the error message. – Jonesey95 (talk) 01:50, 27 March 2014 (UTC)
Run bot on a page with a {{cite encyclopedia}} cite.
We can't proceed until
Bot operator's feedback on what is feasible
Which is preferred? the |title= or |encyclopedia=? The {{cite encyclopedia}} template documentation implies they are the same, and yet they display differently. Obviously the bot is getting the title from the ur. AManWithNoPlan (talk) 14:25, 8 April 2014 (UTC)
Is it necessary to do anything in this case? Here is what Module:Citation/CS1 does with |article=, |encyclopedia=, and |title= (and their aliases) (essentially a copy of the comments in the CS1 code):
|encyclopedia= and |title= then map |title= to |article= and |encyclopedia= to |title=
|encyclopedia= and |article= then map |encyclopedia= to |title=
|encyclopedia= then map |encyclopedia= to |title=
|trans_title= maps to |trans_chapter= when |title= is re-mapped
All other combinations of |encyclopedia=, |title=, and |article= are not modified.
Because this code exists, perhaps it is best that Citation bot do nothing when it encounters {{cite enyclopedia}} without |title= but with |article= and |encyclopedia=.
The bot now draws its list of valid parameters from Module:Citation/CS1/Whitelist. I hope that this will resolve the problem. Could you confirm? Thanks. Martin(Smith609 – Talk) 10:48, 12 July 2014 (UTC)
The bot is still making this error. It leaves an auto-generated "too many author paramaters" warning in red in the references, (example). I consider this a disruption to articles, an error is generated where there was none before, consequently I am blocking this bot. No idea how many of these errors the bot has generated, the bot's operator needs to sift through the bot's history and fix them all. SpinningSpark 18:30, 30 June 2014 (UTC)
Spinningspark: As you can read above, the bot's owner has fixed this problem in the source code. He has been having trouble getting the new code uploaded to the tools server. There are currently about 150 instances of this bot-caused error in articles; they are tracked in Category:Pages with citations having redundant parameters and I will be fixing them once the new bot code is running. – Jonesey95 (talk) 02:59, 1 July 2014 (UTC)
There is quite a lot of "above", you might want to be more specific. If this bot is still using the toolserver it is in trouble anyway, the toolserver is about to close down. In fact, I note that several toolserver tools have died this morning and are being redirected to non-existent wmlabs pages. Whatever the cause, the bot needs to stop making these errors. SpinningSpark 05:41, 1 July 2014 (UTC)
Fine, the bot can be unblocked when that goes live. In the meantime, I don't see the benefit of littering the encyclopaedia with red error messages for the sake of removing the deprecated coauthors parameter when this is not even visible to readers. SpinningSpark 15:19, 1 July 2014 (UTC)
Unless I'm mistaken, r552 will not fix this problem when enabled. It appears to assume that the coauthors list should become author2, but that is only the case absent an existing author2, author3, etc. It might be better to just flag such cases for human attention, if the bot can't be made to get it right. LeadSongDogcome howl! 23:31, 10 July 2014 (UTC)
Revision 560 checks whether last2 or author2 is specified; if not, coauthor is renamed 'author2'; if so, 'coauthor' is left alone. Martin(Smith609 – Talk) 10:19, 12 July 2014 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
In this edit to Francisco Vásquez de Coronado, an editor created {{Cite jstor|30246725}}. Citation Bot 2 appears to have created {{Cite pmid/30246725}} based on that edit.
What should happen
Citation Bot should ignore Cite Jstor unless the JSTOR API is working. It certainly should not create a Cite PMID template based on a Cite JSTOR template, using the JSTOR value.
{{resolved}} This is fixed--It now runs, poorly -- see other bugs
What happens
While viewing a page, I click Expand citations. The page redirects and appears to start the processes and posts
Processing page 'Football records in Spain' — edit—history
* 0 {{Citation}} templates and 82 {{Cite XXX}} templates identified. Using dominant template {{Cite XXX}}.
# No changes required.
# # # -[#TODO unhandled DB request]-
End of output
# # #
This has been the case for several weeks but I thought that it would be reported by someone else and fixed by now.
What should happen
The page should process the article and remove duplicates, etc.
Relevant diffs/links
N/A
Replication instructions
100% reproducible by me. Since I don't see others complaining about this, it could be that it's user-specific, but the steps should be simple enough to repro.
We can't proceed until
Agreement on the best solution
Requested action from maintainer
Revert to a version that is working or report to an admin the specific problems with my settings that caused this to stop working.
I noticed this too. It was one of the best things about this tool.--Aurictalk 23:50, 5 April 2014 (UTC)
Is anyone watching this page and trying to determine the root cause of the problem and fix it? Walter Görlitz (talk) 14:31, 9 April 2014 (UTC)
See the top of this talkpage for your answer. LeadSongDogcome howl! 20:25, 9 April 2014 (UTC)
So the person who maintains the script can go weeks without logging in to Wikipedia and can no longer devote extensive time to bot maintenance? Then perhaps the source should be made available to others so that it can be investigated.
It's still dying for me. Walter Görlitz (talk) 01:22, 12 April 2014 (UTC)
Could you point out which citations you are expecting the bot to modify, but it is not modifying? All the templates I saw were Cite webs, which the bot doesn't do much with. Martin(Smith609 – Talk) 13:30, 28 June 2014 (UTC)
I have only just noticed I have edited the article a few times previously (ie replicated the bug) when I clear out the category!
We can't proceed until
Agreement on the best solution
I tried changing the Persondata template to be uppercase parameters as is normal for this template and Citation bot has left the article alone this time. However, it did previously change the uppercase version [158]. No idea why Citationbot is changing the Persondata template anyway? Periglio (talk) 12:51, 6 April 2014 (UTC)
Click Expand citations in toolbar leads to page that gets stuck and doesn't do anything
1. I navigated to this article: https://en.wikipedia.org/wiki/R%C3%B6nninge/Salem_Fotboll . 2. I clicked "Expand citations" under the Tools menu. 3. My browser displays a new page with the headers "Welcome to Citation Bot; Please wait while the Citation bot processes the page you requested." The page looks like it is doing work in the background. 4. I waited 5 minutes. The page's appearance doesn't change. And I don't get taken back to the article.
A specific edit to the bot's code is requested below.
Requested action from maintainer
learn to notice multiple authors and continue the count of authors.
Fixed error for New France article. Bots are great until they mess up the article!
ODNB
{{Resolved}}
Getting "user blocked error" when trying to run doi:10.1093/ref:odnb/48144 using template:cite doi (did it manually)
As an aside - when getting from Oxford Dictionary of National Biography the returned data lacks most fields and has to be manually entered. I assume this is the fault of the meta data being supplied from ODNB.
Does this thing even work anymore? FunkMonk (talk) 04:37, 7 July 2014 (UTC)
Citation bot is currently blocked, pending a bug fix. – Jonesey95 (talk) 04:59, 7 July 2014 (UTC)
Any idea when it'll be back? FunkMonk (talk) 19:26, 7 July 2014 (UTC)
The code has been updated in the last week or two, but the bot operator was having trouble getting it uploaded so that the new code becomes the deployed version. I don't know more than that. – Jonesey95 (talk) 04:19, 8 July 2014 (UTC)
Sorry about the delay. New code's now uploaded and now we're just waiting for an admin to see the unblock request. Martin(Smith609 – Talk) 10:10, 12 July 2014 (UTC)
Bot dev543 converted |title= to |DUPLICATE_title= in the absence of a second title parameter
This may have happened because |title= existed in {{cite web}} in the absence of |url=. This citation is already flagged with a red error message, "|archiveurl= requires |url=", so Citation Bot does not need to take any action unless it can find a working URL for the article and insert it into the citation. – Jonesey95 (talk) 00:09, 31 March 2014 (UTC)
The bot converted |title= to |DUPLICATE_title=, then |url= to |eprint=. Neither 'DUPLICATE_title' nor 'eprint' are supported by the CS1 templates; unknown parameters have generated an error for quite some time. If the bot can't properly repair a citation, perhaps it should just tag it for manual processing. --Gadget850talk 22:55, 31 March 2014 (UTC)
The bot converted |archivedate= to |DUPLICATE_archivedate= in the absence of a duplicate archivedate, causing an error message, in this edit. Please fix. – Jonesey95 (talk) 06:13, 8 June 2014 (UTC)
Here's another one. |title= was converted to |DUPLICATE_title= in the absence of a second |title=. The word "title" did appear in a comment in the citation, FWIW. – Jonesey95 (talk) 05:23, 26 June 2014 (UTC)
List of parameters cleanup
{{resolved}}
I've been working on getting the Citation parameters cleaned up for AWB and someone pointed me to your bot as well. I made an initial edit removing all invalid parameters that were on the list. Then I took the Module:Citation/CS1/Whitelist and created a fresh list of parameters, with 99 authors and 45 editors, with all their attendant numbered options. I lowered the number of "given" and "surname" to 20; as I doubt they would ever be used 99 times in one citation; however, I've arranged the parameters in a way which I hope makes it easy to see what's there and what can be added. I started with the parameters that are all capital letters, followed by parameters using mixed case, followed by numbered parameters using mixed case, followed by lower-case parameters, and ending with numbered lower-case parameters. I hope you find this helpful!—D'Ranged 1VTalk 04:06, 29 May 2014 (UTC)
As of revision 362, the bot now draws all parameters specified in Module:Citation/CS1/Whitelist with the format "['parameter_name'] = true", and treats these as valid spellings. Martin(Smith609 – Talk) 10:44, 12 July 2014 (UTC)
I have now updated the bot so that it runs the latest version of the software, which fixed the outstanding bugs. Martin(Smith609 – Talk) 10:04, 12 July 2014 (UTC)
Accept reason:
Unblock accepted, but could you please run a small number of edits first to confirm all is OK and if not, disable to bot or request it be re-blocked. Nick (talk) 10:14, 12 July 2014 (UTC)
Thanks for the unblock. I've just run a test edit to confirm that there is no disastrous behaviour. Could other users please also confirm that the bot is operating correctly in cases that have previously been problematic, but are marked as fixed? This would be much appreciated; this is my only bot maintenance window for a fair while, so the more that I can fix today the better! Thanks, Martin(Smith609 – Talk) 10:47, 12 July 2014 (UTC)
Nice work. I re-ran a half-dozen |coauthors=-related edits that I had previously undone, and the bot replaced |coauthors= with |author2= correctly in each case. I did notice that sometimes the "last edit" link in the Citation Bot output page linked to the edit before Citation Bot's edit instead of linking to Citation Bot's edit. A minor bug that happened about three out of six times. – Jonesey95 (talk) 14:32, 12 July 2014 (UTC)
I noticed that the edit summary for the latest edits report that they are being done by version "dev561", although some of the fixes were reportedly deployed in version "r572". – Jonesey95 (talk) 15:07, 12 July 2014 (UTC)
r561 hangs and does not complete when clicking "jump the queue" from Cuban macaw. – Jonesey95 (talk) 15:25, 12 July 2014 (UTC)
Seems to hang some of the time and sometimes works. Not sure if it is bot or server problem. AManWithNoPlan (talk) 23:08, 12 July 2014 (UTC)
Doesn't work on quagga either. FunkMonk (talk) 01:27, 13 July 2014 (UTC)
Worked on ones with one author, but not more than one AManWithNoPlan (talk) 03:19, 13 July 2014 (UTC)
Please note that month, coauthor and coauthors are deprecated. -- 79.67.241.255 (talk) 23:32, 21 March 2014 (UTC)
Can you suggest how they should be replaced? I'd be happy to enact this change. Martin(Smith609 – Talk) 21:37, 22 March 2014 (UTC)
coauthor, coauthors
This is a general summary of my observations from fixing these. |coauthor= and |coauthors= are hard to replace, due to the various list formats used. If the bot could detect which one, it could fix it. Otherwise, a human should do it. Common formats are these:
Last, First; Last2, First
Last, FM; Last, FM
Last, F.M.; Last, F.M.
Last, FM, Last, FM
Last, F.M., Last, First M. & Last, First
First Last, First Last, and First Last
First Last; First Last; First Last
The last two are hard for a bot to fix, as the name may be an organization or have a different order for family names or titles, so they should probably be replaced with |author= series rather than |first= and |last= series. There are also other hybrid formats, some using slashes or other symbols to separate authors, and organizations that don't have a "first" name so care is needed. Some citations only use |coauthors= without listing any authors separately in other parameters, and others already have |author= or |first= and |last= or even |author2= or |first2= and |last2= or more, so the next author number would need detecting. —PC-XT+ 05:52, 25 March 2014 (UTC) 05:59, 25 March 2014 (UTC)
As said above, this is a tricky one. I believe |lastn=|firstn= should be used for persons and |authorn= for committees, departments and organisations.
Beware of edits like this where |first= or |first1= contains the whole name, there is no |last= or alias of, and the rest of the names are jammed into |coauthors= in a long list.
When citationbot encounters |first=John Doe and tries to fix it, the end result is two parameters, but with the surname in both: |first=John Doe and |last1=Doe. Note they are widely separated in the template as can be seen in this recent example. Separately look for each of these four parameters: | last1 = Parro | last1 = Keating | last1 = McKay | last1 = Schuerger |. These particular examples have since been manually fixed, but there are many others out there. -- 79.67.241.76 (talk) 00:04, 26 March 2014 (UTC)
Use of |first= without |last= followed by |coauthors= is indeed somewhat common. Apparently, some editors think "first" means the same as |author1=, perhaps due to the 1. Also, I have seen the surname repeated in both |first=, after the given name, and |last=, and doubted a human would do that so often, but didn't know which tool did it. It may be better for documents with doi or other identifiers to retrieve the author list and redo it from scratch. I often do that, myself, then compare changes before saving. Sometimes later authors are listed in a different order, but I don't think it would really matter to change to the doi order. —PC-XT+ 05:26, 27 March 2014 (UTC)
month
|month= can simply be deleted if it is empty or already in |date=. If it is not in |date=, it could be added, using |year= if |date= is empty. It sometimes is a number, 1 to 12, which should be converted to the month name for comparison or adding to |date=. |day=, which is also deprecated, can often be added to |date= in a similar manner, if it is a number from 1 to 31. —PC-XT+ 05:52, 25 March 2014 (UTC) 05:59, 25 March 2014 (UTC)
There's already a bot going around fixing citations found with the |month= parameter within. -- 79.67.241.76 (talk) 00:56, 26 March 2014 (UTC)
Copy the previous version of the above article into a sandbox page and run the bot on it.
We can't proceed until
Bot operator's feedback on what is feasible
Just a wild guess: Citation bot didn't like the comment it found in |doi=. – Jonesey95 (talk) 02:53, 24 March 2014 (UTC)
Citation bot dev543 replaced a comment with garbage in this edit. The documentation says that the bot is supposed to ignore comments: "If the bot is erroneously adding a DOI, author, etc to a citation, and you want to stop it adding the data again, you need to put a comment in place of the appropriate parameter – because the bot will not overwrite existing data." – Jonesey95 (talk) 22:00, 31 March 2014 (UTC)
it looks like rev570 might have been meant to fix that, but I'm not sure it really checks for doi=<!-- this sort of comment --> which seems to have been the problem. LeadSongDogcome howl! 00:05, 23 July 2014 (UTC)
rev570 has not been released yet, as far as I can tell. I had a query about the released version; that query is part of a section that has been archived. – Jonesey95 (talk) 05:25, 23 July 2014 (UTC)
I was just commenting on reading the code, which is published on code.google.com as usual. LeadSongDogcome howl! 06:38, 23 July 2014 (UTC)
This particular report is now fixed, thanks for your comments here. I need to return at some point to save the comment being treated as a DOI, and its spaces being turned into plusses, but at least this is a non-fatal error. Martin(Smith609 – Talk) 05:35, 29 July 2014 (UTC)
Bot 4 r531 replaced "cite doi" with "citation", creating an empty citation
A specific edit to the bot's code is requested below.
Requested action from maintainer
Stop replacing cite doi with citation. It produces empty citations.
The code for Citation Bot 4 needs to be checked to ensure that, at a minimum, it is replacing only Module/CS1 templates with "citation". Please stop Bot 4 until this is done. – Jonesey95 (talk) 15:13, 4 February 2014 (UTC)
Gosh, I didn't spot this. Bot 4 only runs under my direct supervision; I won't run it again until I've fixed this. Thanks for the spot. Martin(Smith609 – Talk) 08:16, 7 February 2014 (UTC)
I have moved this bug report out of Archive1 because it is still outstanding and it is present in Cat543dev. See this edit.
The bot also replaced a cite pmid template with citation, creating an empty citation error. See this edit. – Jonesey95 (talk) 16:42, 26 March 2014 (UTC)
More precisely, it replaced {{cite doi|NUMBER}} with {{citation|NUMBER}} when it should have used {{citation|doi=NUMBER}} instead. LeadSongDogcome howl! 16:51, 26 March 2014 (UTC)
Please do not do this. That proposed edit would remove content from the citation displayed in the article and display a CS1 "missing or empty title" error. Here's the difference:
{{cite doi|10.1021/ja00521a034}} displays:
Attention: This template ({{cite doi}}) is deprecated. To cite the publication identified by doi:10.1021/ja00521a034, please use {{cite journal}} (if it was published in a bona fide academic journal, otherwise {{cite report}} with |doi=10.1021/ja00521a034 instead.
Actually, it's supposed to unify citation types based on the one which is predominant in the article, which it did in the Feline immunodeficiency virus edit, it just did it incorrectly. I've tidied it up manually. LeadSongDogcome howl! 20:10, 26 March 2014 (UTC)
r561 can't fill in cite doi templates when there is more than one author
It may be that these templates can't be filled when there is more than one author in the cited journal article. There may also be something else going on here. This did not happen with the previous version (r543?) – Jonesey95 (talk) 06:12, 13 July 2014 (UTC)
Seems to be the same case when using the side bar expand option too AManWithNoPlan (talk) 04:28, 15 July 2014 (UTC)
I'm seeing the same thing with PMID citations. It seems to get hung up after the first author. --Tryptofish (talk) 15:00, 27 July 2014 (UTC)
^Samuel Lara-Avila, Alexei Kalaboukhov, Sara Paolillo, Mikael Syväjärvi, Rositza Yakimova, Vladimir Fal'ko, Alexander Tzalenchuk, Sergey Kubatkin; Lara-Avila; Kalaboukhov; Paolillo; Syväjärvi; Yakimova; Kazakova; Janssen; Fal'Ko; Kubatkin (2009). "SiC Graphene Suitable For Quantum Hall Resistance Metrology". Nature Nanotechnology. 5 (3): 186–9. arXiv:0909.1193. Bibcode:2010NatNa...5..186T. doi:10.1038/nnano.2009.474. PMID20081845.{{cite journal}}: CS1 maint: multiple names: authors list (link)
Notice the number of last names has increased to ten, there are no first names, and the original eight authors continue to be listed in the "author" parameter.
I have manually fixed the reference, amending back to 8 authors, using "last" and " first" for each and Revision 601513133 contains
The primary reason for the number of authors changing from 8 to 10 is that the orginal author list and article title in the citation were derived from the arXiv data but citation bot has subsequently used the author data from either the Bibcode, doi or PMID data when "correcting" the citation. A problem arises because the latter three identifiers point to a different article!
I have subsequently split the reference into two separate items and Revision 605142634 contains
<ref name="0909.1220">{{cite journal |last=Tzalenchuk |first=Alexander |last2=Lara-Avila |first2=Samuel |last3=Kalaboukhov |first3=Alexei |last4=Paolillo |first4=Sara |last5=Syväjärvi |first5=Mikael |last6=Yakimova |first6=Rositza |last7=Kazakova |first7=Olga |last8=Janssen |first8=T. J. B. M. |last9=Fal'Ko |first9=Vladimir |last10=Kubatkin |first10=Sergey |year=2010 |title=Towards a quantum resistance standard based on epitaxial graphene |journal=Nature Nanotechnology |volume=5 |issue=3 |pages=186–9 |arxiv=0909.1220 |doi=10.1038/nnano.2009.474 |bibcode=2010NatNa...5..186T |pmid=20081845 }}</ref>[example4 2]
^Lara-Avila, Samuel; Kalaboukhov, Alexei; Paolillo, Sara; Syväjärvi, Mikael; Yakimova, Rositza; Fal'ko, Vladimir; Tzalenchuk, Alexander; Kubatkin, Sergey (Submitted 7 July 2009). "SiC Graphene Suitable For Quantum Hall Resistance Metrology". Science Brevia. arXiv:0909.1193. {{cite journal}}: Check date values in: |date= (help)CS1 maint: date and year (link)
^Tzalenchuk, Alexander; Lara-Avila, Samuel; Kalaboukhov, Alexei; Paolillo, Sara; Syväjärvi, Mikael; Yakimova, Rositza; Kazakova, Olga; Janssen, T. J. B. M.; Fal'Ko, Vladimir; Kubatkin, Sergey (2010). "Towards a quantum resistance standard based on epitaxial graphene". Nature Nanotechnology. 5 (3): 186–9. arXiv:0909.1220. Bibcode:2010NatNa...5..186T. doi:10.1038/nnano.2009.474. PMID20081845.
I am wary that citation bot may come back and attempt to add incorrect doi, Bibcode and PMID data to the first of these two references. Will using a commented <!-- none --> for the value prevent that?
Does citation bot ever check that arXiv, doi, Bibcode, PMID and other identifiers actually all point to the same article? I am sure this isn't an isolated case. -- 79.67.241.210 (talk) 12:55, 21 April 2014 (UTC)
Another variation at this edit which changed an incorrect parameter name "author-link4" to an even more incorrect "author-link", so that the name of author1 was linked to the article on author4. It should have changed it to "author4-link", which I have now done manually. LeadSongDogcome howl! 20:32, 15 May 2014 (UTC)
|author-link4= is a legitimate variant and works just fine. There was nothing incorrect about it.
Compare template:citation/doc to template:cite journal/doc closely and you'll see that the aliases are not done consistently. In some places we'd expect author2-link, author2link, authorlink2, author-link2 to all get the same treatment, but in others they do not. This might be a contributing factor, but I don't know. LeadSongDogcome howl! 22:44, 15 May 2014 (UTC)
The CS1 templates and {{citation}} all evolved out of the muck in different ways, created by different groups of editors. It is therefore not surprising that there are differences. As all of these citations have further evolved they are becoming more and more alike but now we have an overabundance of parameters that are aliases of each other. All of these are fully synonymous: |authorn-link=, |author-linkn=, |authornlink=, |authorlinkn=, |subjectlinkn=. Do we really need so many? When the list of pages with deprecated parameters is down to a more manageable level, then perhaps we can consider paring the list down to one or two aliases. My vote would be for |authorlinkn= and it's hyphenated pal |author-linkn=.
|authorn-link= and |authornlink= seem the most logical. These parameters point to the single link for authorn. -- 91.84.87.119 (talk) 15:12, 22 July 2014 (UTC)
The authorlink issue should be resolved now that the bot draws its list of acceptable parameter names directly from the template. And verifying that references are identical is non-trivial; as this is a marginal case and one of GIGO, I'm afraid that I won't be able to work on a fix (although I'd be happy to implement a contributed patch). Martin(Smith609 – Talk) 05:16, 29 July 2014 (UTC)
Is the bot going to come back online?
It's been disabled for over a week now. JMP EAX (talk) 13:53, 27 July 2014 (UTC)
Until I can locate and fix the bug, you can try the dev version; replace "citations" with "citations-dev" in the URL. Martin(Smith609 – Talk) 12:56, 28 July 2014 (UTC)
Thanks, works for me. Maybe the two functions should be split off into different bots? It's a bit annoying that we can't use the cite DOI function whenever something else screws up. FunkMonk (talk) 13:21, 28 July 2014 (UTC)
In the last couple of weeks it's been very slow and most of the time doesn't finish a job or gives a timeout. Is that still a consequence of moving to wmflabs? Reflinks (and some other tools) run again on what appears to be a clone of the late toolserver. Perhaps something for this invaluable tool, too? --Randykitty (talk) 15:24, 28 July 2014 (UTC)
As a temporary workaround, the "citations-dev" fix does work for me, at least for now. If it's unclear, the way to do it is to try to activate the bot (for example, by clicking on "jump the queue" or on the sidebar gadget if you have it), then, when the screen with the bot information appears, change the URL field to read "citations-dev" instead of just "citations", then reload the web page. It works for me. --Tryptofish (talk) 15:50, 28 July 2014 (UTC)
Indeed. I tried it on that page, and the same thing happened to me that happened to you. Perhaps we are actually dealing with two separate bugs here. Let me say, more specifically then, that when it worked for me (at Striatum), it worked in the following manner. There were some sources that I had previously added using the "cite PMID" template. One of them had a single author. When I clicked "jump the queue" for that one, the bot worked. The other two citations had multiple authors. To get them to work, I clicked "jump the queue" successively for each one, and, when that brought up a new tab in my browser, I changed the URL as above, at which point they, too, worked. At the EBSCO page, there are no queues to jump, so I tried the expand citations gadget that I have enabled at the left side of the screen under "tools". As you said, it just seemed to hang, and I gave up after waiting for what seemed like enough time. I guess one possibility is that it's the same bug, but when there are multiple citations simultaneously, one never gets the opportunity to edit the URL before there's a hang. Is there an alternative way to get to "citations-dev" that I don't know about? --Tryptofish (talk) 18:54, 28 July 2014 (UTC)
I waited a long time and then the screen changed from the WP page to the bot page. That's when I could change the URL to the dev one, but that one stalled, too... — Preceding unsigned comment added by Randykitty (talk • contribs) 19:40, July 28, 2014
Then I apparently did not wait long enough. And, in that case, it would indeed appear to be a second bug, because changing to "dev" did not fix it. --Tryptofish (talk) 20:00, 28 July 2014 (UTC)
I've fixed the latter bug in r573. Gosh, WMFlabs does take a long time to load, eh? Martin(Smith609 – Talk) 05:11, 29 July 2014 (UTC)
Yes it takes forever to use it manually and sometimes I get blank output even after waiting/wasting my time. Are the automated/queued runs disabled though? There are pages where I've added {{cite doi}} weeks ago and they still haven't been updated; for example Structured program theorem has some. JMP EAX (talk) 00:55, 30 July 2014 (UTC)
If the bot does not come by to fill in your template, you can go to [159] and let it run for a while (you may not see any output on your browser tab). If you want to watch its progress, you can visit Category:Pages with incomplete DOI references and reload the page periodically. – Jonesey95 (talk) 03:41, 30 July 2014 (UTC)
It seems that eventually it got to some of my work, but there are problems with titles and everything; the templates are created blank basically, with just the doi in them. See Interval (graph theory) for example. JMP EAX (talk) 03:46, 30 July 2014 (UTC)
See bug below. I think this specific discussion is basically over, since the bot is back online and all the problems being discussed are in other bug specific discussions. AManWithNoPlan (talk) 19:06, 2 August 2014 (UTC)
Marking as {{resolved}}, since the bot is back online, and the only ongoing discussion is now moved to bug report below. AManWithNoPlan (talk) 13:59, 5 August 2014 (UTC)
This is a bad way to source the fact that a book was published on a particular date, so it doesn't trouble me much that the bot makes it worse. Can't you find the same information somewhere more reliable and more suited to the purpose, such as worldcat? —David Eppstein (talk) 03:18, 4 August 2014 (UTC)
"Citation" to "Cite" conversions are screwed up, as well as questionable
Quite aside from the questionability of blindly converting "Citation" templates to "Cite", the conversion is f****** things up (I'm just a little pissed about this) and introducing garbage. Stop it! ~ J. Johnson (JJ) (talk) 21:26, 30 March 2014 (UTC)
Specifically, a postscript of "{inconsistent citations}" was added to some of the citations. That is the only visible difference I see in the rendered article before and after Citation Bot made its edits. Perhaps the bot operator could explain why a few citations had this postscript added to them. – Jonesey95 (talk) 23:31, 30 March 2014 (UTC)
{{Inconsistent citations}} marks a page as needing human verification after such a change. I don't know why it is being escaped to display the wikicode, though. —PC-XT+ 05:35, 31 March 2014 (UTC) This search seems to find the affected pages; at least, the ones I checked were affected. —PC-XT+ 08:15, 31 March 2014 (UTC) 08:39, 31 March 2014 (UTC)
It appears that the bot is correct in attempting the change to the predominant form (in this case citation to cite xxx). The use of |postscript= is intended to ensure that no visible change is made to the citation postscript ("." or " " as the case may be). This could have the unwanted effect of an edit with no visible effect. The hidden wikitext added to draw human editor attention to the postscript was recently changed to UTF-8 encoding, but I'm not sure why it was rendered visible. Clearly a bug, but the human editor invoking the bot should catch it easily if they're paying attention. LeadSongDogcome howl! 15:36, 31 March 2014 (UTC)
The message is evidently generated by line 2523 of objects.php in the current trunk. Not sure why {{inconsistent citations}} isn't inside the hidden text string postscripted. Somehow, the { and } are being escaped as #123 and #125, and this leads to them not serving to trigger transclusion of the template. LeadSongDogcome howl! 17:15, 31 March 2014 (UTC)
I think a better fix for this problem would be to simply not attempt to convert templates. To do so is a huge assumption, and a bot should not be trying to count the "dominant" form and then force everything to conformity. E.g., there is a place where I deliberately used a {cite} template where otherwise I was using predominately {citation} (at this point I don't recall why, but I felt it was justified). It is not for a stupid bot to do a mindless conversion based on mere arithmetic. If mixed citation/cite is really a problem then the most that the bot should do is collect a list for real editors to examine. ~ J. Johnson (JJ) (talk) 20:23, 5 April 2014 (UTC)
It looks like maybe the bot is failing to check CrossRef for citation parameters. Just a guess. This is a new problem with dev575. The previous released version was working, except for the multiple authors bug. – Jonesey95 (talk) 04:00, 30 July 2014 (UTC)
As I said in a section above I have seen many of those recently. Some examples:
Attention: This template ({{cite doi}}) is deprecated. To cite the publication identified by doi:10.1109.2F32.126773, please use {{cite journal}} (if it was published in a bona fide academic journal, otherwise {{cite report}} with |doi=10.1109.2F32.126773 instead.
Attention: This template ({{cite doi}}) is deprecated. To cite the publication identified by doi:10.1145.2F800152.804919, please use {{cite journal}} (if it was published in a bona fide academic journal, otherwise {{cite report}} with |doi=10.1145.2F800152.804919 instead.
Attention: This template ({{cite doi}}) is deprecated. To cite the publication identified by doi:10.1016.2Fj.scico.2009.10.007, please use {{cite journal}} (if it was published in a bona fide academic journal, otherwise {{cite report}} with |doi=10.1016.2Fj.scico.2009.10.007 instead.
You can find more like these on the pages that link to those templates. Hope this helps. JMP EAX (talk) 08:10, 1 August 2014 (UTC)
I don't want to nag, but a week later the problem is still here:
Attention: This template ({{cite doi}}) is deprecated. To cite the publication identified by doi:10.1145.2F2366145.2366191, please use {{cite journal}} (if it was published in a bona fide academic journal, otherwise {{cite report}} with |doi=10.1145.2F2366145.2366191 instead.
Just one more reason not to use cite doi, as if we needed another. LeadSongDogcome howl! 20:30, 13 August 2014 (UTC)
It worked fine for my purposes before. I find filling out citation templates immensely boring, so this was a good tool. FunkMonk (talk) 20:32, 13 August 2014 (UTC)
Consider the one-digit error impact. You cite the wrong paper, but don't find out about it until the bot gets around to filling in the details. If you forget to check back ... LeadSongDogcome howl! 20:37, 13 August 2014 (UTC)
Well, I always just push "jump queue", which instantly shows me if anything is wrong. I sometimes have to fix an error or two, but that sure beats filling out the entire thing from scratch. FunkMonk (talk) 20:39, 13 August 2014 (UTC)
This discussion is irrelevant to the bug. It applies to cite journal templates directly in articles too. There is nothing magic about cite doi. In fact, if the DOI matches a PMID, then the bot seems to work. AManWithNoPlan (talk) 23:27, 13 August 2014 (UTC)
I blocked this bot as it was creating empty pages like here and here. Ruslik_Zero 18:55, 15 August 2014 (UTC)
Ruslik0: The bug has been reported (see above). There is no reason to block this bot because of a single bug that does not do any harm other than failing to fill in citations. The editors that create the citations are responsible for ensuring that they are filled in. If the citations are not filled in, blocking the bot does not help.
The bot was filling in and updating other references just fine (with the exception of a few minor bugs listed above). Please remove your block. – Jonesey95 (talk) 19:51, 15 August 2014 (UTC)
Arrg. This is a very useful tool. I hope it can be returned ASAP. (I've used it here recently, and the page fails to update: [162]) Tal Galili (talk) 22:17, 16 August 2014 (UTC)
Can't this just be temporarily replaced by an earlier version that actually worked? This is getting ridiculous. FunkMonk (talk) 18:52, 17 August 2014 (UTC)
Other than for dois, this is the best version of the bot in quite some time, arguably since the advent of Lua template use broke an earlier version. It would be better to change the preloaded template to generate a more explicit diagnostic until it is filled in.LeadSongDogcome howl! 19:29, 17 August 2014 (UTC)
The new empty template aren't really any worse than the expand by hand or jump the queue stuff AManWithNoPlan (talk) 20:14, 17 August 2014 (UTC)
[ec] Is that to say that implementing Lua prevents reversion to an earlier working version? I have found this bot sufficiently annoying in some of its results that I would rather do without it. Is there any great and pressing need for which we can't wait till the bot is fixed? ~ J. Johnson (JJ) (talk) 20:23, 17 August 2014 (UTC)
What happened to the good one where you could just write "citation-dev" in the URL to make it work? FunkMonk (talk) 21:52, 17 August 2014 (UTC)
The shutting down of the tool server didn't help either. What I would really like it the citations button to work in the edit window. That way the changes are checked by a humor and it circumvents the block. AManWithNoPlan (talk) 23:02, 17 August 2014 (UTC)
I would definitely support the implementation of this option. Dispenser's tools do this. Normally, the bot taking responsibility for its edit is good, but I think this option should be available in case of block, perhaps with an editnotice telling the human to review the changes before saving, because it will be in their name (with a note that they used the bot automatically placed in the edit summary, as with other tools.) —PC-XT+ 03:18, 21 August 2014 (UTC)
This tool is very useful to cleanup references in articles (eg google books links). It has some bugs, especially when the DOI are ill-defined. Please consider removing your block. Alberto Fernández Fernández (talk) 10:14, 20 August 2014 (UTC)
If this continues, can't the functions be split off into different bots, so we don't lose everything wheneve one function stops working? FunkMonk (talk) 17:04, 20 August 2014 (UTC)
The downside would be that those of us that would rather not put up with this bot at all would have to add additional deny lines. Still, I think this suggestion is excellent. ~ J. Johnson (JJ) (talk) 20:14, 21 August 2014 (UTC)
One could certainly use the same block for all bots. Multiple bots though sounds like a lot more work to maintain. I think a better move would be a config file that turned off things. Such as turning off completing {{cite doi}} AManWithNoPlan (talk) 02:05, 22 August 2014 (UTC)
Completing cite doi is performed by Citation Bot 2. All bot accounts use the same code because a performance enhancement that affects one may affect all of them. But they perform separate functions. Martin(Smith609 – Talk) 09:50, 25 August 2014 (UTC)
So is it working now? FunkMonk (talk) 09:53, 25 August 2014 (UTC)
Yep—as far as I can see. Any problems, let me know—e-mail is best for major problems like this. Martin(Smith609 – Talk) 10:54, 25 August 2014 (UTC)
This user's unblock request has been reviewed by an administrator, who accepted the request.
The bug that caused a user to block the bot has been fixed (revision 577). Martin(Smith609 – Talk) 09:02, 25 August 2014 (UTC)
Accept reason:
I've unblocked the bot. Otherwise, for what's it worth, I consider blocking the bot to be a legitimate response to the malfunction. PhilKnight (talk) 10:56, 25 August 2014 (UTC)
Since the bot was moved from the Toolserver, I've managed to expand citations only a few times. In 99% of other cases I get a white screen (immediately), or some timeout screen (after some time). After the migration I've never managed to run the bot on a large article; the bot expanded a few cite doi templates for me, but even this function halts these days. Am I doing something wrong? I have the bot activated in gadgets, and in User:Materialscientist/monobook.js + User:Materialscientist/vector.js. Thanks. Materialscientist (talk) 06:04, 1 September 2014 (UTC)
It works for me when I replace citations with citations-dev in the url. So the answer is no and yes... FunkMonk (talk) 06:08, 1 September 2014 (UTC)
Great, thanks, expanding cite doi/cite pmid works with me with this substitution. The bot still hangs on articles, which are not so long by the past measures, like synaptic vesicle. Materialscientist (talk) 06:18, 1 September 2014 (UTC)
Fixed again (19:57, 1 September 2014 (UTC)). Hopefully the bot is converging on a stable state. Martin(Smith609 – Talk) 19:57, 1 September 2014 (UTC)
I don't know, maybe try it again on that page? Or, maybe I'm just confused?
We can't proceed until
Input from editors
Anyone know why this happened? Am I missing something? I reverted the edit, and it seemed to work fine in preview. I then also converted the template to {{cite news}} before saving, in case that matters. —PC-XT+ 04:56, 8 August 2014 (UTC)
The bot appears to have made an error. That dead link is fixable. 79.67.248.171 (talk) 05:46, 8 August 2014 (UTC)
Publisher and work were misattributed by humans. It should have been the WSJ via nethope.org, though it's much simpler to just cite the WSJ directly. LeadSongDogcome howl! 14:59, 8 August 2014 (UTC)
Ah, maybe the bot picked up on the publisher/work disagreement, then. Thanks a lot for fixing the link! For some reason(s), I was having trouble getting the archive to work, and the few searches I tried in WSJ archives didn't find results, either. When the bot flagged this error, I figured I'd take a break and see what you guys thought. I'm glad I did. Thanks again! —PC-XT+ 05:19, 9 August 2014 (UTC)
Enable the Citation expander gadget, edit a page, and click on the Citations button.
We can't proceed until
Agreement on the best solution
How frustrating. Until I get the chance to look into this, you can replace 'citations/' with 'citations-dev/' in the URL to access a working version of this tool. Martin(Smith609 – Talk) 07:31, 21 October 2014 (UTC)
can you explain me how to change the url that the button "citations" leads me to? - Ihaveacatonmydesk (talk) 11:17, 25 October 2014 (UTC)
The dev version is down too now, or at least non-responsive. AManWithNoPlan (talk) 23:40, 28 October 2014 (UTC)
This is a feature, not a bug. There were two |date= parameters in the same citation. Citation bot renamed one of the parameters, which puts the article in Category:Pages with citations using unsupported parameters for a human to fix. Citation bot should not simply delete one of the two parameters, because people sometimes put the access date in a duplicate date field. They sometimes do even stranger things. – Jonesey95 (talk) 22:12, 10 November 2014 (UTC)
You are correct. I thought it was confused by "archivedate", but I was :) I will look into the problem. Thanks. Secondarywaltz (talk) 22:18, 10 November 2014 (UTC)
This is a feature, not a bug. There were two |publisher= parameters in the same citation. Citation bot renamed one of the parameters, which puts the article in Category:Pages with citations using unsupported parameters for a human to fix. Citation bot should not simply delete one of the two parameters, because people sometimes put something useful in a duplicate parameter. – Jonesey95 (talk) 15:31, 11 November 2014 (UTC)
A specific edit to the bot's code is requested below.
Not a bug. That paper has two authors but whomever added the citation neglected to correctly identify them both. Citation bot has made that edit twice. Correctly. Perhaps you should look more carefully at what it is telling you.
Bob K: This is a feature, not a bug (see the two essentially identical reports immediately above this one). There were two |last= parameters in the same citation. Citation bot renamed one of the parameters, which puts the article in Category:Pages with citations using unsupported parameters for a human to fix. Citation bot should not simply delete one of the two parameters, because people sometimes put something useful in a duplicate parameter. In this case, I changed the second |last= to the proper |last2= to fix the underlying problem. – Jonesey95 (talk) 00:36, 12 November 2014 (UTC)
Enable the Citation expander gadget, edit a page, and click on the Citations button.
We can't proceed until
Agreement on the best solution
How frustrating. Until I get the chance to look into this, you can replace 'citations/' with 'citations-dev/' in the URL to access a working version of this tool. Martin(Smith609 – Talk) 07:31, 21 October 2014 (UTC)
can you explain me how to change the url that the button "citations" leads me to? - Ihaveacatonmydesk (talk) 11:17, 25 October 2014 (UTC)
The dev version is down too now, or at least non-responsive. AManWithNoPlan (talk) 23:40, 28 October 2014 (UTC)
This is a feature, not a bug. There were two |date= parameters in the same citation. Citation bot renamed one of the parameters, which puts the article in Category:Pages with citations using unsupported parameters for a human to fix. Citation bot should not simply delete one of the two parameters, because people sometimes put the access date in a duplicate date field. They sometimes do even stranger things. – Jonesey95 (talk) 22:12, 10 November 2014 (UTC)
You are correct. I thought it was confused by "archivedate", but I was :) I will look into the problem. Thanks. Secondarywaltz (talk) 22:18, 10 November 2014 (UTC)
This is a feature, not a bug. There were two |publisher= parameters in the same citation. Citation bot renamed one of the parameters, which puts the article in Category:Pages with citations using unsupported parameters for a human to fix. Citation bot should not simply delete one of the two parameters, because people sometimes put something useful in a duplicate parameter. – Jonesey95 (talk) 15:31, 11 November 2014 (UTC)
A specific edit to the bot's code is requested below.
Not a bug. That paper has two authors but whomever added the citation neglected to correctly identify them both. Citation bot has made that edit twice. Correctly. Perhaps you should look more carefully at what it is telling you.
Bob K: This is a feature, not a bug (see the two essentially identical reports immediately above this one). There were two |last= parameters in the same citation. Citation bot renamed one of the parameters, which puts the article in Category:Pages with citations using unsupported parameters for a human to fix. Citation bot should not simply delete one of the two parameters, because people sometimes put something useful in a duplicate parameter. In this case, I changed the second |last= to the proper |last2= to fix the underlying problem. – Jonesey95 (talk) 00:36, 12 November 2014 (UTC)
The bot replaces an instance of |isbn= with |DUPLICATE_isbn= in {{cite book}}, but |DUPLICATE_isbn= does not appear to be a valid parameter, and an "unknown parameter" warning is shown (in the ref list).
This is a feature, not a bug. There were two |isbn= parameters in the same citation. Citation bot renamed one of the parameters, which puts the article in Category:Pages with citations using unsupported parameters for a human to fix. Citation bot should not simply delete one of the two parameters, because people sometimes put something useful in a duplicate parameter. In this case, the proper fix is to remove the redundant 10-digit ISBN in favor of the 13-digit ISBN. – Jonesey95 (talk) 04:18, 16 November 2014 (UTC)
I seem to recall sources with only the 10-digit ISBN in the printed version where a 13-digit ISBN was added later in the on-line cataloging information. In such cases the 10-digit ISBN is not redundant, as that is what will identify the hard-copy version. Hard-cover and soft-cover editions with identical content also have different ISBNs. Therefore multiple (not "duplicate") ISBNs seem reasonable. I presume citation bot objects only because the template takes only one ISBN. I usually finesse the matter by trailing the second ISBN after the template. Perhaps citation bot could attempt something like that? Even better, perhaps the template should revised. ~ J. Johnson (JJ) (talk) 20:38, 16 November 2014 (UTC)
I understand the reasons for renaming the duplicate parameter to an invalid name, but it still seems to me that something that changes a cite so that it displays a reader-visible warning (where previously there was no such warning, and an ISBN was displayed) looks like a bug. I suggest that, if nothing else, the bot should insert a comment immediately after the renamed parameter, for the benefit of any human editor who sees the warning or the change in the article other than by seeing it listed in Category:Pages with citations using unsupported parameters. That comment should include a reference to the relevant section of User:Citation bot, which in turn should mention that particular "feature". Currently the documentation doesn't mention this feature. Mitch Ames (talk) 12:04, 17 November 2014 (UTC)
The bot replaces an instance of |isbn= with |DUPLICATE_isbn= in {{cite book}}, but |DUPLICATE_isbn= does not appear to be a valid parameter, and an "unknown parameter" warning is shown (in the ref list).
This is a feature, not a bug. There were two |isbn= parameters in the same citation. Citation bot renamed one of the parameters, which puts the article in Category:Pages with citations using unsupported parameters for a human to fix. Citation bot should not simply delete one of the two parameters, because people sometimes put something useful in a duplicate parameter. In this case, the proper fix is to remove the redundant 10-digit ISBN in favor of the 13-digit ISBN. – Jonesey95 (talk) 04:18, 16 November 2014 (UTC)
I seem to recall sources with only the 10-digit ISBN in the printed version where a 13-digit ISBN was added later in the on-line cataloging information. In such cases the 10-digit ISBN is not redundant, as that is what will identify the hard-copy version. Hard-cover and soft-cover editions with identical content also have different ISBNs. Therefore multiple (not "duplicate") ISBNs seem reasonable. I presume citation bot objects only because the template takes only one ISBN. I usually finesse the matter by trailing the second ISBN after the template. Perhaps citation bot could attempt something like that? Even better, perhaps the template should revised. ~ J. Johnson (JJ) (talk) 20:38, 16 November 2014 (UTC)
I understand the reasons for renaming the duplicate parameter to an invalid name, but it still seems to me that something that changes a cite so that it displays a reader-visible warning (where previously there was no such warning, and an ISBN was displayed) looks like a bug. I suggest that, if nothing else, the bot should insert a comment immediately after the renamed parameter, for the benefit of any human editor who sees the warning or the change in the article other than by seeing it listed in Category:Pages with citations using unsupported parameters. That comment should include a reference to the relevant section of User:Citation bot, which in turn should mention that particular "feature". Currently the documentation doesn't mention this feature. Mitch Ames (talk) 12:04, 17 November 2014 (UTC)
Surnames misinterpreted as given names and vice versa
Deleterious: Many bot edits require manual fixing.
What happens
The bot ends up adding last names into first name fields, requiring manual cleanup. For example, the linked edit led to this wikitext being included: |last12=Jäger |first12=Jäger
What should happen
The bot should not add incorrect parameters—no action is better than incorrect action.
These edits will be a pain to identify and fix. Note that the problem is intermittent: the edit on Hassium includes both adding parameters (apparently) properly and adding parameters badly (i.e. this bug). {{Nihiltres|talk|edits}} 20:25, 22 November 2014 (UTC)
This does not look like a bug to me. The bot correctly added last names to the citations, leaving first names alone. The incorrect first names were added in these edits by a human editor. Asking the bot to replace incorrect first names with corrected first names, most likely changing the style of some of the citations in the article (from initials to full first names, for example), is not reasonable. – Jonesey95 (talk) 21:31, 22 November 2014 (UTC)
You're right; I mixed up which parameters the bot added. Sorry about that… {{Nihiltres|talk|edits}} 23:01, 22 November 2014 (UTC)
Does the bot add DOI if JSTOR/OtherId and standard info already exist
I was wondering if this bot adds the DOI if another id (e.g. JSTOR) and the standard information already exist. For example, several of the cites at Bargaining problem have JSTOR (and other info) but not DOI. If not, I would encourage the addition of this feature as the DOI is probably a safer long-term ID than JSTOR. Often times JSTOR doesn't list the DOI through its API for older articles, but crossref/guestquery has it. Thanks for the great bot! --Bequw (talk) 22:09, 6 January 2015 (UTC)
There's a default JSTOR to DOI mapping, but it really should not be used, as there often is a real DOI for the original document and a different one for the JSTOR copy of that document. Populating the DOI parameter with the JSTOR equivalent DOI obscures this fact. LeadSongDogcome howl! 22:43, 6 January 2015 (UTC)
This happens when the citation template in question exists on a page. You can often look in the template's history or in What Links Here to find a link to the page in question. In this case, see User:FeatherPluma/tools#cite_ref-XxYYYY_2-0. – Jonesey95 (talk) 14:44, 21 January 2015 (UTC)
[ec] What is the bug here? That editors use "year=" rather "date="? That citation bot chokes on usage that we used to allow? I have never understood why "year=" was deprecated, in that there are dates that are validly only the year. As "date=" (presumably) can handle can such cases, why isn't "year=" just aliased to "date="? ~ J. Johnson (JJ) (talk) 21:28, 26 January 2015 (UTC)
Oppose – This is a non-starter. |year= has not been deprecated. Furthermore, the Template:Citation Style documentation#date clearly documents cases where |year= is required. Even in cases where it is not required, why fix something that isn't broke? Boghog (talk) 21:21, 26 January 2015 (UTC)
Does Citation bot's code produce any citations that explicitly require |year= per the documentation? If not, then I don't see a reason to keep using |year=. ~Tom.Reding (talk ⋅contribs ⋅dgaf) 21:52, 26 January 2015 (UTC)
There may be other citations in the article that require |year= in which case it would be easier for editors to integrate the citation bot edited citation. Where was the consensus to change the documentation? Furthermore, what harm is caused by using |year=? If the parameter value is a year, it is reasonable to use the more specific |year= parameter. The bug is in the documentation, not the bot. Boghog (talk) 22:55, 26 January 2015 (UTC)
|date= is preferred, but |year= is not deprecated and is in heavy use. Older implementation of the templates could not handle just a year in |date=, but that was resolved with the updates. --Gadget850talk 23:07, 26 January 2015 (UTC)
So it appears there is no reason at all to "fix" yyyy type entries in either date= or year=. Is that the bug being complained of here, that citation bot does such pointless fixes? ~ J. Johnson (JJ) (talk) 23:21, 26 January 2015 (UTC)
No, I think the complaint is that the bot doesn't make these pointless fixes. I think this should be marked "not a bug". —David Eppstein (talk) 23:54, 26 January 2015 (UTC)
Actually, the complaint (< cosmetic improvement) is that CB properly use each parameter per the documentation, though because this particular case doesn't produce an error, this might be asking too much (asking more than the typical bug/complaint). ~Tom.Reding (talk ⋅contribs ⋅dgaf) 19:05, 27 January 2015 (UTC)
I have boldly marked this as "not a bug" per the discussion above. If, at some point, |year= were to be deprecated and there were an active effort to convert |year= to |date= (via some other bot, presumably), it would be reasonable to ask this bot to stop adding |year= to articles. – Jonesey95 (talk) 01:09, 27 January 2015 (UTC)
As I noted |date= is preferred over |year=; |year= is really only needed to create unique dates in anchors. If the bot were to use |date= it would still work. --Gadget850talk 19:25, 27 January 2015 (UTC)
This bug was fixed long ago. The edit history shows that these poorly-encoded author names were added in 2010. I blanked the authors and ran the bot again, which filled in the author names. – Jonesey95 (talk) 16:42, 25 January 2015 (UTC)
Awesome, thank you! --MZMcBride (talk) 18:52, 27 January 2015 (UTC)
To resolve this error, provide a value for |url= or remove |access-date=. Editors should try to determine why the citation has |access-date= without |url=. For example, the citation may never have had a |url=, or |url= may have been removed because it links to a site that violates the creator's copyright (see WP:COPYLINK), or because |url= was deemed to be dead and (mistakenly) removed. If the citation never had |url= or it was removed for copyright violations, remove |access-date=. When a dead |url= has been removed, restore the |url= and if possible repair it (see WP:LINKROT).
Various discussions at (I think) Help talk:Citation Style 1 and probably other places led to that above-quoted text and the determination that this category is not one that easily lends itself to repair-by-bot.
probably best to only do that on pages that automatically after the page has been unedited for a month at least to give humans a chance to fix it. AManWithNoPlan (talk) 04:01, 17 February 2015 (UTC)
Humans will not know there is an error, because the error message is hidden by default. This category does not appear to be bot-fixable, at least on a grand scale. There may be edge cases that a bot could address. – Jonesey95 (talk) 05:09, 17 February 2015 (UTC)
True. The edge case that the bot could fix is if there is a pmid, doi, hdl, jstor, or pmc in the citation. Then the citation has a clear non-url link out. AManWithNoPlan (talk) 15:23, 18 February 2015 (UTC)
That sounds reasonable. This bot is not maintained regularly, but a bot or AWB user could probably do this work. I suggest that you start a discussion at Help talk:Citation Style 1, and if that results in a consensus, post a request at Wikipedia:Bot requests. – Jonesey95 (talk) 17:28, 18 February 2015 (UTC)
Discussion started. Closing this ticket. AManWithNoPlan (talk) 04:51, 20 February 2015 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
The bot screwed up a reference to OMIM, conflating it with the PubMed entry with the same number. Admittedly this was a while ago (2010-01-07) and may have been fixed already
Admittedly I don't know whether the error springs from the bot or from Rjwilmsi whose username appears in the edit summary. Feel free to nuke this bug report if I've logged it in error.
--Mathieu ottawa (talk) 04:45, 21 February 2015 (UTC)
It appears that this bug no longer exists. I copied that citation to my sandbox and ran Citation Bot on it. It changed the unsupported parameter |source= to |laysource= (which is another matter, and not a problem, since there are only ever one or two citations on en.WP with |source= erroneously entered by a human), but it did not replicate this old bug. – Jonesey95 (talk) 14:46, 21 February 2015 (UTC)
{{notabug}}. This is not a bug. There were two sets of identical parameters and values in that citation. Citation Bot removed the duplicates. – Jonesey95 (talk) 17:18, 27 February 2015 (UTC)
When [164] an edit is done, the diff software is itself not sure if the new sentence was appended or was inserted. (I thought this was a bug in the diff software itself, but actually it is a genuine ambiguity. See the discussion I had about it here: [165]). The bot, in its subsequent edit ([166]) instead of replacing the arxiv link, added the reference to the preceeding sentence (!) and botched the whole thing.
What should happen
Should replace arxiv citations properly in such special cases.
This is not a citation bot bug. You are seeing an inadequacy in how diff works. Citation bot correctly populated the Cite arxiv template, as you can see in the "link showing what happens" that you provided. The Cite arxiv template does appear to have a small bug in it, however, in that it showed the "jump the queue" prompt even after the citation had been populated. I worked around that problem in the article by deleting some of the empty parameters. – Jonesey95 (talk) 04:35, 21 March 2015 (UTC)
A specific edit to the bot's code is requested below.
This really isn't a bot problem that if you include an invisible character AFTER the number that is not handled by {{cite pmid}} or {{cite jstor}}. AManWithNoPlan (talk) 03:04, 16 July 2014 (UTC)
although it would be cool if the bot fixed these. AManWithNoPlan (talk) 03:07, 16 July 2014 (UTC)
Citation bot dev532 did not convert first names to initials in cite doi template
A specific edit to the bot's code is requested below.
Requested action from maintainer
Restore previous, documented behavior of the bot.
Complete disagreement. A sizable fraction of the Wiki community believes strongly that first names should not be abbreviated in the citation data, but rather in the rendering template if so desired (e.g., {{Cite web}}, {{Cite journal}}, etc.). Sometimes, a lot of detective work goes into finding out what an author's full first name is (nobody wants the sad story of Laurent Cassegrain to be repeated), and the bot should not increase entropy by abbreviating the citation data. Until this behaviour is removed (and note that it is not part of the bot's documented function), editors will be forced to add bot-denial instructions to the citations needing protection. Urhixidur (talk) 14:11, 8 September 2014 (UTC)
The formatting of author names in the {{cite doi}} template is documented at Template:Cite_doi#Formatting. The bot should fill in the {{cite doi}} template in accordance with the documentation. This bug report does not address the formatting of {{cite journal}} or other similar templates that the bot might work on within articles. – Jonesey95 (talk) 17:52, 8 September 2014 (UTC)
Just don't use cite doi. It's a **BAD** idea, highly vulnerable to typos, linkrot, and trivial vandalism. Make a full cite journal entry and be done with it. LeadSongDogcome howl! 18:58, 8 September 2014 (UTC)
In some fields you are only supposed to use initials so that you do not know if the author is male or female or black or white or jewish, etc. AManWithNoPlan (talk) 14:28, 18 September 2014 (UTC)
This looks like a problem in the source database that the bot draws the data from. Sorry that nothing can be done! Martin(Smith609 – Talk) 06:06, 19 September 2014 (UTC)
If the source database used was PubMed, then it appears that the problem is not with the source database (see PMID12712314). Boghog (talk) 06:41, 19 September 2014 (UTC)
Changed the "coauthors" parameter to "author2" in a {{cite book}} template and this caused an error with the harvard referencing (citation style used {{sfn}}) used within the article.
What should happen
Any changes shouldn't break the article's referencing system, regardless of whether the parameters are considered redundant.
No idea, but obviously any changes shouldn't be causing problems. Ref system was working before this error.
Just another note, while I appreciate the work done by the operator of this bot, maybe in the future if changes are made to an article on the main page (such as the WP:TFA in this case), it may be best to have any bot edits checked manually. Especially as any errors will be so visible. -- Shuddetalk 02:17, 10 October 2014 (UTC)
Harvard references are brittle, unless an explicit Harvard reference name is given in the {{cite journal}} template. You never know when another reference is going to be added that has the same author and year or when a typo in a name or date is going to be fixed. AManWithNoPlan (talk) 14:30, 10 October 2014 (UTC)
This was not caused by another book being added to the bibliography, and it wasn't caused by any of your examples of the "brittle[ness]" of the harv style. It was caused by the bot changing a parameter. Hence my bug report. Had the bot not made any changes, there would not have been a problem. -- Shuddetalk 02:36, 13 October 2014 (UTC)
The bot should probably change the ref=harv to be the full version. ref=harv is prone to breakage, since it can break when things are fixed. That would fix the problem long-term. AManWithNoPlan (talk) 23:02, 13 October 2014 (UTC)
That would be a solution. But as I don't maintain the bot, or code them, I can't say what the best solution is. The bot just came through and made the same buggy edits ([169] -- which I've reverted). Why is this bot running if it's not being maintained, or at least some eye being kept on this page? -- Shuddetalk 02:44, 16 October 2014 (UTC)
I just reverted some buggy edits. References that list the names of authors now include all the last names following the first entry of their names. I am One of Many (talk) 05:33, 16 October 2014 (UTC)
Inconsistency on adding second names
Diffs on my corrections on what the bot just did. When adding a second name, adding "Author2" instead of "last 2" and "first2" threw the Harv Ref into error mode. Also, on the Time Inc. article, there was no author, as writers don't always get a byline. The bot decided the publisher was also the author. — Maile (talk) 18:50, 23 November 2014 (UTC)
The Time citation edit was clearly wrong. Perhaps though, the better fix is |author=Staff writer instead of |author=Staff|first=writer so that the citation renders without an extraneous comma between Staff and writer. Alternately, it is common to do this: |author=<!--Staff writer(s); no by-line.-->
{{wontfix}}
Hi. I just filed a bug here and the pre-filled form is really rough. We should either dramatically improve the wiki bug reporting or we should switch to somewhere else (such as Phabricator). Pre-filling the editing text area with {{bot bug}} and then mixing obscure template markup with HTML comments in a single blob really isn't acceptable. --MZMcBride (talk) 04:52, 25 January 2015 (UTC)
This is the wiki standard way of doing things. Discuss the bot on the bot page. Also, many so-called bugs are not bugs. AManWithNoPlan (talk) 20:13, 11 May 2015 (UTC)
Missing page numbers?
I'm not sure if this is a bug or just me misunderstanding something, but pretty much every time I use the cite DOI template, the full page ranges are not given, only the first number, and I have to change it manually, as in here: [170]FunkMonk (talk) 13:49, 1 May 2015 (UTC)
do you see this if you do cite journal with doi= in it? Cite doi is going away by the way. AManWithNoPlan (talk) 17:36, 1 May 2015 (UTC)
I'll try next time. What do you mean by going away? Will it be replaced? It is an extremely helpful tool, one I use the most. FunkMonk (talk) 15:35, 1 May 2015 (UTC)
It is being eliminated. Transcluding citations directly into the article is what will be done. They want the bot here to start converting cite doi without a template into cite journal's. Soon a bot will transclude all existing cite doi/pmid/jstor/isbn/hdl. AManWithNoPlan (talk) 17:36, 1 May 2015 (UTC)
Ouch, what will happen if the cite doi is used after that, will it continue to be converted, or will it just not work? And where is this discussed? FunkMonk (talk) 17:53, 1 May 2015 (UTC)
at some point it will be gone or it will be bot expanded in place. Content in templates does violate Wikipedia policy, etc. cite ISBN was worse since ISBN are not very unique in practice. AManWithNoPlan (talk) 20:16, 1 May 2015 (UTC)
Didn't work when I tried with =[171]FunkMonk (talk) 19:12, 4 May 2015 (UTC)
Never mind, seems that DOI didn't work at all... FunkMonk (talk) 19:13, 4 May 2015 (UTC)
{{cite journal|doi=10.XXXXX}} AManWithNoPlan (talk) 00:00, 5 May 2015 (UTC)
@AManWithNoPlan: when will it start doing so? Can you point me to the relevant discussion, please? Thanks.--RoadTrain (talk) 10:37, 20 May 2015 (UTC)
no idea when the bot will start doing it, but you can manually run it to expand {{cite journal}} . In mean {{cite doi }} still seems to work AManWithNoPlan (talk) 00:22, 24 May 2015 (UTC)
Seems to depend upon the quality of the metadata. Many have full range. AManWithNoPlan (talk) 20:15, 11 May 2015 (UTC)
{{notabug}}
This is a case of garbage in, garbage out. These "authors" are listed at crossref.org, which is where citation bot is pulling its data from, I believe. This is a crossref.org bug, not a citation bot bug, unfortunately. I sent an e-mail to crossref on the off chance that they can fix it. – Jonesey95 (talk) 12:44, 14 May 2015 (UTC)
"Removed accessdate with no specified URL"
In this edit the bot has removed accessdate= from a number of citations with the edit note "[579]Removed accessdate with no specified URL". Whilst the citations altered do not have url= parameters, they do have external links formed using doi=. Should the bot, therefore, be removing access dates in these situations?--DavidCane (talk) 11:12, 15 May 2015 (UTC)
Yes. |access-date= applies to ephemeral urls and only when those urls are identified in |url= or |chapter-url= and its aliases. External links created by identifiers like |doi=, |pmc=, |pmid=, etc. are considered permanent so |access-date= is not needed. And, when there is more than one of these permanent identifiers in a cs1|2 template, as is often the case, to which identifier would |access-date= apply if it did apply to these identifier parameters?
I don't know if this has been fixed but citation bot used to add Template:dead link into the format parameter of Template:citation like here which would cause an error in the capitalization of the categories it put things in.
What should happen
The dead link reference should not be in the format category but after the citation and before the ref tag closes.
Editor Ricky81682 wrote: However, if the update to Module:Citation/CS1 has happened, there is no bug at all. This is false. {{dead link}} should not be placed in |format=. That is not a function of Module:Citation/CS1 but is or was a function of Citation bot. An update to the module will have no effect on the bot.
The diff linked above is from 2009. The bot does not do this anymore. I copied the citation in question to my sandbox, deliberately broke the URL, and ran the bot on the page. No "dead link" template was added. I am marking this one as "fixed". – Jonesey95 (talk) 22:16, 18 May 2015 (UTC)
The bot is removing access dates from sources that are wikilinked (obviously to Wikis other than Wikipedia) when it probably shouldn't be
We can't proceed until
Bot operator's feedback on what is feasible
Requested action from maintainer
Check for wikilink in the source before removing access date parameter
Can you provide an example of an edit? AManWithNoPlan (talk) 22:50, 9 May 2015 (UTC)
This is not a bug. Users should link to a specific version of the wiki, not the wikilink, if they want an accessdate. AManWithNoPlan (talk) 21:44, 12 June 2015 (UTC)
First stop these edits. Second if these edits are to be resumed make sure that the bot edit the parameters using information obtained from the indicia of the books in question and leave an HTML comment explaining why the edit was made.
99.9% of the time what the bot is doing is correct. People almost never are linking to amazon content, they are linking to the book. I suggest you use [http:|text] instead and include a comment that this is a link to Amazon and not the book. Or, just block the bot on those pages. AManWithNoPlan (talk) 20:34, 24 May 2015 (UTC)
One more thing. Human editors will often do what the bot is doing. AManWithNoPlan (talk) 20:41, 24 May 2015 (UTC)
The use of the cite web template with a URL creates a strong presumption that the reference is to a web page not a book. The fact that human editors do this too is not germane. There should at least some attempt to verify that it is in fact a book being cited. That Amazon.com (including variants) is being left as the publisher indicates to me that it is otherwise. – Allen4names (contributions) 20:58, 24 May 2015 (UTC)
True, but editors often (way too often) link to google books and then put google as the publisher, so there really is common for people to do it wrong. Also, I often find people linking to amazon when what they really mean to link to is the book. Sloppy, but common. AManWithNoPlan (talk) 21:04, 24 May 2015 (UTC)
Perhaps I should clarify. In the diff linked above the bot left "Amazon, Inc" as the publisher. As far as I know Amazon does not publish books so the resulting references are questionable at best and much more likely corrupt and if not corrected quite removable. – Allen4names (contributions) 21:18, 24 May 2015 (UTC)
The bot partially corrected these references. The references were showing the existence of a book, so {{cite book}} was used and the ASIN was standardized. As for |publisher=Amazon, it is sometimes legitimate, so the bot should not remove it wholesale. A human editor can do an search for insource:/\|\s*publisher\s*=\s*Amazon/ and fix the ones that are wrong; I get over 3,000 hits on that search string, so it could keep someone busy for quite a while. – Jonesey95 (talk) 14:28, 25 May 2015 (UTC)
I guess there is only one thing left for me to do and hope it finally gets through. It looks like a duck to me – Allen4names (contributions) 20:43, 25 May 2015 (UTC)
That ASIN is also an ISBN, as is the usual case for print books. Looking it up on WorldCat finds it at OCLC818787175 for the four volume set if that is preferred. Using a proprietary service is something we try to avoid when an open one is available. If however there is a sound reason to insist on keeping it, as agreed at the article, then simply use wp:bot exclusion to tell this bot not to mess with it. LeadSongDogcome howl! 13:56, 26 May 2015 (UTC)
Inconvenience: Humans must occasionally make immediate edits to clean up after the bot
What happens
mw:MediaWiki 1.25/wmf4 added detection for duplicate parameters; these pages are now placed in a tracking category. Thus, duplicate detection by the bot is now redundant.
For what it's worth, the bot locates and tags the duplicate citation parameters, causing the citations in question to emit red error messages. Until the duplicate template parameter errors emit their own error messages (which is in the works, apparently), the bot's tagging makes it a lot easier to locate these otherwise hidden errors. – Jonesey95 (talk) 01:47, 22 November 2014 (UTC)
I have an AWB script that is removing {{PDFlink}} from CS1 citations. Most times {{PDFlink}} is used as the value in |title= but occasionally, editors have used it in |url= so when the external link portion of {{PDFlink}} is disassembled into |url= and |title=, there are now two titles. Initially, I simply commented out the title portion of the {{PDFlink}} parameter but that draws no attention. So, on consideration I have adopted Citation bot's |DUPLICATE_title= so that the change can be noticed and fixed.
try processing any page (i was in the middle of editing Latex allergy and now it looks like i have to manually fill in all the citations instead of any bot completing them)
We can't proceed until
Bot operator's feedback on what is feasible
Even expanding PMID and DOI is dead. AManWithNoPlan (talk) 14:51, 12 June 2015 (UTC)
claims to not be able to find page. Expand doi claimes bot is logged out AManWithNoPlan (talk) 14:56, 12 June 2015 (UTC)
Bot is logged off? Cuzkatzimhut (talk) 19:09, 12 June 2015 (UTC)
OK, good – so it's not just me... --IJBall (contribs • talk) 00:33, 15 June 2015 (UTC)
Thanks for the report; I'll try to find time to investigate the cause. Martin(Smith609 – Talk) 06:58, 15 June 2015 (UTC)
The bot appears to be working now; I jumped the queue to test if a citation worked and it looks good. No errors or anything --Rfilmyer (talk) 16:39, 16 June 2015 (UTC)
The sidebar Expand citations link and the edit window Citations button are both still down. Edit window has been down for many months though as described above in a separate bub report. AManWithNoPlan (talk) 16:49, 16 June 2015 (UTC)
Bot blocked
{{resolved}}
I have just blocked the bot for misbehaving. Block can be lifted when cause is determined and bot is fixed. --NeilNtalk to me 03:49, 17 June 2015 (UTC)
We have fixed all of the articles that the bot blanked. Now we are waiting for Smith609 to fix the bot. – Jonesey95 (talk) 03:58, 17 June 2015 (UTC)
Please specify misbehavior; are there any page features that make the bot misbehave? I have used the Citation_bot for years without problems (except that citations often take a while to be inserted). It's the only bot that I use regularly and probably one of the most useful bots in Wikipedia. Peteruetz (talk) 20:04, 23 June 2015 (UTC)
Peteruetz, please look at the contribution history before the bot was blocked. [174] --NeilNtalk to me 21:25, 23 June 2015 (UTC)
See above. The bot was blanking pages. I have proposed a possible reason for the bug here. – Jonesey95 (talk) 00:41, 24 June 2015 (UTC)
It's difficult to see a pattern in the contribution history.[175] I use the bot exclusively for PubMed (PMID) references and never had a problem. Can the bot be unblocked for certain uses? Maybe there are certain types of citations which the bot doesn't like? Most of the pages listed in the contribution history seem to be unrelated to (bio-)medical issues, hence the problem does not appear to be caused by PMID citations. Peteruetz (talk) 11:46, 24 June 2015 (UTC)
the pattern is me running it on a whole bunch of pages-and others I assume . I usually go back and check changes, but since the bot claimed to not find the pages I just left Wikipedia and did other things. Little did I know I had just cluster bombed Wikipedia AManWithNoPlan (talk) 12:48, 24 June 2015 (UTC)
I activated the bot on a bunch of pages as well. The pattern is that the bot completely blanked all of the pages, starting shortly after WP was changed to https-only. It could be coincidence, but the bot's code was not changed, and the https-only change caused a few other problems. The bot appears to have hard-coded instances of "http://en.wikipedia.org..." in its code, so it seems like a reasonable guess that the two events are related. – Jonesey95 (talk) 14:06, 24 June 2015 (UTC)
There's something else going on too. Manual activation with "commit edits" disabled leads to no preview output. Does this work properly for someone else? LeadSongDogcome howl! 16:28, 25 June 2015 (UTC)
Well, the bot is blocked, as far as I know, so I would not expect any of its functions to work properly. – Jonesey95 (talk) 02:14, 26 June 2015 (UTC)
Isn't that the point of the "commit edits" switch? Turning it off should mean nothing is saved, so the block has no effect. How else could the operator test fixes? LeadSongDogcome howl! 03:18, 26 June 2015 (UTC)
Manual activation from the edit window always gave me a completely blank result, although now it just sends me to an error page. It always worked for me from the doibot page. There were some glitches such as being unable to detect citations past certain comments, but I never experienced page blanking. I work mostly with cite doi and cite journal(doi and bibcode) references. Lithopsian (talk) 15:43, 7 July 2015 (UTC)
{{resolved}}
This block has been lifted as of my time stamp. – Jonesey95 (talk) 00:23, 22 September 2015 (UTC)
Bot Cat579 added unsupported |eprint=, which was supposedly fixed in r572
It looks like the bug that was fixed in r572 was the removal of a valid URL for citations sourced in the arXiv database. The bot is still adding |eprint=, however, which I believe is valid in {{cite arxiv}} but not in {{cite journal}}. – Jonesey95 (talk) 06:08, 20 October 2014 (UTC)
It appears that the bot no longer does this. I copied the citation in question from the above article into my sandbox and ran the bot on it. – Jonesey95 (talk) 03:37, 22 September 2015 (UTC)
In this edit the bot deleted the entire article text. Zerotalk 03:44, 17 June 2015 (UTC)
I manually ran the bot on some of these articles and the bot basically said article not found and then did nothing. Obviously, it did something! AManWithNoPlan (talk) 14:19, 17 June 2015 (UTC)
Ah, yes, thank you. Those had {{cite doi}} or {{ref doi}}, again without the subpage having been expanded. It's past time that the preload page was changed to generate an error diagnostic message. LeadSongDogcome howl! 08:29, 24 June 2015 (UTC)
Removes accessdate for citations with chapterurl instead of url
Citation bot removes the accessdate for citations that don't have URLs (which makes sense). However some book citations have chapterurl instead of url, and in those cases, it should leave the assessdate.
A specific edit to the bot's code is requested below.
Requested action from maintainer
Only remove the accessdate if the citation is missing both the url and chapterurl parameters.
This seems like a feature request. I would not expect an accessdate unless url and only url were used, since a "chapterurl" is not necessary (even in an ebook) for identifying the work. Even for a "bookurl" I'd be awfully skeptical. This one requires further discussion regardless IMO. --Izno (talk) 00:52, 23 September 2015 (UTC)
Maybe you're right. Consider the bug report revoked in that case :) Kaldari (talk) 01:08, 23 September 2015 (UTC)
No. If the source is a chapter in some work, it is quite appropriate to have urls to both the chapter (if it exists separately) and work. Or even to only the chapter. Either way, 'chapterurl' clarifies what is being linked. And either way, an accessdate may be appropriate, and should not be bot-removed. ~ J. Johnson (JJ) (talk) 21:30, 23 September 2015 (UTC)
I can confirm, encountered this today.--RoadTrain (talk) 10:30, 20 May 2015 (UTC)
For now you can just do the edits, save the page, then run the bot from the sidebar, but that leaves the page saved in a bad state for a while. Probably best to use a sandbox. :-( AManWithNoPlan (talk) 13:38, 12 June 2015 (UTC)
Cite DNB to Citation error
Back in March 2014 this bot change {{cite DNB}} to {{Citation}} this was a mistake see diff], and if this Bot is to resume this bug needs fixing. -- PBS (talk) 18:50, 18 August 2015 (UTC)
I just ran my sidebar Citation Bot link (called "Expand citations") to try to expand {{Cite doi/10.1063.2F1.455515}}. It ran, but the Results page has lots of links to test.wikipedia.org instead of to en.wikipedia.org. This is the link in my sidebar when I am on that template's page. – Jonesey95 (talk) 00:26, 22 September 2015 (UTC)
{{fixed}} Fixed now, thanks for the report. --Fhocutt (WMF) (talk) 00:55, 22 September 2015 (UTC)
In the square brackets at the start of the bot's edit summary, it says "[Revision]". It should have a revision number here, I believe. – Jonesey95 (talk) 02:48, 22 September 2015 (UTC)
That's actually a holdover from the old version control system and refers to the version of the bot, not the page revision. Confusing! References to the old version control system should be removed from the code. --Fhocutt (WMF) (talk) 18:47, 22 September 2015 (UTC)
Sorry, I knew that and should have made it clear. The bot used to show the revision number of the bot's code. This helped in bug reporting; if someone found an old edit and reported it as a bug, the developer could determine if the bug had been fixed in a later revision of the bot's code. – Jonesey95 (talk) 23:50, 22 September 2015 (UTC)
Unfortunately, that system relied on SVN which is long gone. A new version system would be have to be created from scratch. Kaldari (talk) 01:10, 23 September 2015 (UTC)
See also the "Volke" citation here for another example. In this example, the bot capitalized the letter following the incorrect character. – Jonesey95 (talk) 03:30, 22 September 2015 (UTC)
Bummer. The bot could try to detect these characters and refuse to modify a citation with these characters in the metadata. I have submitted a bog report to Crossref. Thanks. – Jonesey95 (talk) 03:24, 27 September 2015 (UTC)
I disagree that the bot shouldn't use it. The citation with a � character is still much better than nothing. Unlike most other errors it is obvious to the reader that it needs fixed. AManWithNoPlan (talk) 15:24, 1 October 2015 (UTC)
The bot really needs to take what it can get. These sort of characters are very common in scientific journal cross-references and leaving out titles wholesale wouldn't be good. So long as the bot doesn't trample over the title after it has been manually fixed then put it in there, warts and all. There are CS-maint categories for tracking down these characters. Lithopsian (talk) 18:42, 1 October 2015 (UTC)
Fair enough. I am willing to withdraw this report, since it is not a Citation Bot bug. I have added a feature request at Help talk:Citation Style 1. – Jonesey95 (talk) 01:28, 2 October 2015 (UTC)
Deleterious: Human-input data is deleted or articles are otherwise significantly affected. Many bot edits require undoing.
What happens
At Stolen Childhood CitationBOT replaced <ref name=BIP-1995/> with <ref name=BIP–1995/> and so forth.
The bot doesn't change all the reference names, producing broken references. Since a hyphen is one of the allowed characters in a reference name but a dash requires quotes around the name, even changing all the hyphens can lead to problems. Best if the bot stays out of reference names altogether.
Also reported here: #Bot_unnecessarily_adding_last2.2C_last3.2C_last4.2C_..._parameters. Support for a new |vauthors= parameter will soon be added to {{cite journal}} (see discussion). If an author list contains only commas as punctuation, it is likely to be formatted in the Vancouver system. In these cases, it would be appropriate for the bot to rename |author= to |vauthors= instead of adding redundant author parameters. One of the arguments for not supporting multiple authors in a single author parameter in {{cite journal}} was that it was not documented. |vauthors= will soon be documented hence there are no more excuses for not fixing this bug. The problem of course is that this bot is no longer supported. Boghog (talk) 03:04, 12 May 2015 (UTC)
I added a caution alerting editors to the unfixed bugs in this edit. Boghog (talk) 03:17, 12 May 2015 (UTC)
flagging as {{Resolved}}, since this bug report is a duplicate and the bot is getting better. AManWithNoPlan (talk) 13:52, 3 October 2015 (UTC)
Remove display-authors=etal when inserting all the remaining authors
What the bot should probably do is to set display-authors to the number of authors which were in the article prior to "filling in the rest", in keeping with WP:CITEVAR. --Izno (talk) 23:11, 16 May 2015 (UTC)
Is there any other bot that can be used for inserting citations, even temporarily, as long as the Citation Bot is down? Especially PMIDs and DOIs? Thanks! Peteruetz (talk) 15:54, 4 July 2015 (UTC)
Citoid in the visual editor handles PMIDs and DOIs automagically, if you're willing to use VE for when you need to cite. --Izno (talk) 16:37, 7 July 2015 (UTC)
There's also the DOI Citation Generator. Neither this nor Citoid are as good as Citation Bot, but they might be some help in individual cases. Neither can handle Bibcodes which is almost essential for astronomical articles. I find Citoid fails to find valid DOIs on many occasions, or just returns a minimal entry with no details. It would be much more helpful to have Citation Bot available for manual use - it isn't like it is going to go out on its own and start blanking articles. Lithopsian (talk) 12:41, 8 July 2015 (UTC)
Yes, we need a new bot soon; I am not looking forward to having to write the code for this ref by hand. Unfortunately WMF is incompetent so their citation tool Citoid doesn't work for most references, and there's no other good replacements here that I've been able to come across. At least that DOI Citation Generator looks promising; I was not aware of it before! StringTheory11 (t • c) 02:26, 16 July 2015 (UTC)
Here you go:
Abramowski, A.; Aharonian, F.; Benkhali, F. Ait; Akhperjanian, A. G.; Angüner, E. O.; Backes, M.; Balenderan, S.; Balzer, A.; Barnacka, A.; Becherini, Y.; Tjus, J. Becker; Berge, D.; Bernhard, S.; Bernlöhr, K.; Birsin, E.; Biteau, J.; Böttcher, M.; Boisson, C.; Bolmont, J.; Bordas, P.; Bregeon, J.; Brun, F.; Brun, P.; Bryan, M.; Bulik, T.; Carrigan, S.; Casanova, S.; Chadwick, P. M.; Chakraborty, N.; Chalme-Calvet, R.; Chaves, R. C. G.; Chrétien, M.; Colafrancesco, S.; Cologna, G.; Conrad, J.; Couturier, C.; Cui, Y.; Dalton, M.; Davids, I. D.; Degrange, B.; Deil, C.; deWilt, P.; Djannati-Ataï, A.; Domainko, W.; Donath, A.; Drury, L. O'C.; Dubus, G.; Dutson, K.; Dyks, J.; Dyrda, M.; Edwards, T.; Egberts, K.; Eger, P.; Espigat, P.; Farnier, C.; Fegan, S.; Feinstein, F.; Fernandes, M. V.; Fernandez, D.; Fiaßon, A.; Fontaine, G.; Förster, A.; FüSSling, M.; Gabici, S.; Gajdus, M.; Gallant, Y. A.; Garrigoux, T.; Giavitto, G.; Giebels, B.; Glicenstein, J. F.; Gottschall, D.; Grondin, M.-H.; Grudzińska, M.; Hadsch, D.; Häffner, S.; Hahn, J.; Harris, J.; Heinzelmann, G.; Henri, G.; Hermann, G.; Hervet, O.; Hillert, A.; Hinton, J. A.; Hofmann, W.; Hofverberg, P.; Holler, M.; Horns, D.; Ivascenko, A.; Jacholkowska, A.; Jahn, C.; Jamrozy, M.; Janiak, M.; Jankowsky, F.; Jung, I.; Kastendieck, M. A.; Katarzyński, K.; Katz, U.; Kaufmann, S.; Khélifi, B.; Kieffer, M.; Klepser, S.; Klochkov, D.; Kluźniak, W.; Kolitzus, D.; Komin, Nu.; Kosack, K.; Krakau, S.; Krayzel, F.; Krüger, P. P.; Laffon, H.; Lamanna, G.; Lefaucheur, J.; Lefranc, V.; Lemiére, A.; Lemoine-Goumard, M.; Lenain, J.-P.; Lohse, T.; Lopatin, A.; Lu, C.-C.; Marandon, V.; Marcowith, A.; Marx, R.; Maurin, G.; Maxted, N.; Mayer, M.; McComb, T. J. L.; Méhault, J.; Meintjes, P. J.; Menzler, U.; Meyer, M.; Mitchell, A. M. W.; Moderski, R.; Mohamed, M.; Morå, K.; Moulin, E.; Murach, T.; de Naurois, M.; Niemiec, J.; Nolan, S. J.; Oakes, L.; Odaka, H.; Ohm, S.; Opitz, B.; Ostrowski, M.; Oya, I.; Panter, M.; Parsons, R. D.; Arribas, M. Paz; Pekeur, N. W.; Pelletier, G.; Perez, J.; Petrucci, P.-O.; Peyaud, B.; Pita, S.; Poon, H.; Pühlhofer, G.; Punch, M.; Quirrenbach, A.; Raab, S.; Reichardt, I.; Reimer, A.; Reimer, O.; Renaud, M.; de los Reyes, R.; Rieger, F.; Rob, L.; Romoli, C.; Rosier-Lees, S.; Rowell, G.; Rudak, B.; Rulten, C. B.; Sahakian, V.; Salek, D.; Sanchez, D. A.; Santangelo, A.; Schlickeiser, R.; Schüssler, F.; Schulz, A.; Schwanke, U.; Schwarzburg, S.; Schwemmer, S.; Sol, H.; Spanier, F.; Spengler, G.; Spies, F.; Stawarz, Ł.; Steenkamp, R.; Stegmann, C.; Stinzing, F.; Stycz, K.; Sushch, I.; Tavernet, J.-P.; Tavernier, T.; Taylor, A. M.; Terrier, R.; Tluczykont, M.; Trichard, C.; Valerius, K.; van Eldik, C.; van Soelen, B.; Vasileiadis, G.; Veh, J.; Venter, C.; Viana, A.; Vincent, P.; Vink, J.; Völk, H. J.; Volpe, F.; Vorster, M.; Vuillaume, T.; Wagner, P.; Wagner, R. M.; Ward, M.; Weidinger, M.; Weitzel, Q.; White, R.; Wierzcholska, A.; Willmann, P.; Wörnlein, A.; Wouters, D.; Yang, R.; Zabalza, V.; Zaborov, D.; Zacharias, M.; Zdziarski, A. A.; Zech, A.; Zechlin, H.-S. (2015), "The 2012 Flare of PG 1553+113 Seen with H.E.S.S. and Fermi-LAT", The Astrophysical Journal, 802 (1), Art. 65, Bibcode:2015ApJ...802...65A, doi:10.1088/0004-637x/802/1/65.
227. That's impressive. Might want a |display-authors= in that. --Izno (talk) 03:37, 16 July 2015 (UTC)
Citations with a large number of authors require a ridiculous number of |firstn=, |lastn= parameters. This can be written and displayed more compactly using the new |vauthors= parameter:
Abramowski A, Aharonian F, Benkhali FA, Akhperjanian AG, Angüner EO, Backes M, et al. (2015). "The 2012 Flare of PG 1553+113 Seen with H.E.S.S. and Fermi-LAT". The Astrophysical Journal. 802 (1). Art. 65. Bibcode:2015ApJ...802...65A. doi:10.1088/0004-637x/802/1/65.
A huge number of authors gets to be ridiculous regardless of how you package them. The problem with looking at |vauthors= as an answer for "HNoA" problems is you then run afoul of CITEVAR. While I favor giving coauthors some visibility, where there are more than about a dozen there is much to said for an "et al.". Or even "Abramowski and 226 others". ~ J. Johnson (JJ) (talk) 21:31, 18 September 2015 (UTC)
@Smith609: The WMF Community Tech team is interested in trying to help fix Citation bot (as requested at the All Our Ideas survey). Two questions: Would you actually like us to help fix it? Do you have any suggestions for where to start? Ryan Kaldari (WMF) (talk) 00:23, 8 August 2015 (UTC)
This is great news. There is a community of people, many of them on this page, who would be willing to help.
Editor Smith609 is mostly inactive on WP (try to e-mail him, which usually results in a response), but I believe that he has expressed a willingness to hand over development (I can't find a link to such a discussion at the moment). The source code is at https://code.google.com/p/citation-bot/ if you would like to take a look.
As for where to start, there are roughly 32 feature requests right here on this page. The place to start is the "Bot blocked" section immediately above and this discussion. If you can get the bot fixed so that it can be unblocked, that would be great. – Jonesey95 (talk) 01:17, 8 August 2015 (UTC)
Great news. The existing code has been exported from Google code to Github several times already, as shown. It might be best to work with one of those. LeadSongDogcome howl! 03:48, 8 August 2015 (UTC)
I'd be delighted if the bot could be back in working order. Unfortunately I'm no longer able to devote any significant time to it but am happy to help in small ways as I can. As I understand it the current problem arose because the API now requires https:// rather than http:// urls, and I may not have found and replaced every instance in the text. I got as far as exporting the current code to GitHub but had problems running git on the WMFlabs servers – I think that's what stalled my last attempt to get the bot going. It shouldn't take much to get that issue resolved, I wouldn't have thought. If you do need my input, the best thing is to send me an e-mail, as I don't often check WP. Martin(Smith609 – Talk) 07:09, 10 August 2015 (UTC)
I'd let it run on this wiki as soon as possible, provided it doesn't break articles/refs. The bot will likely need fine-tuning, and you'll get much better feedback when it runs on this wiki rather than testwiki. Materialscientist (talk) 02:08, 19 September 2015 (UTC)
Agreed; if it's only being run manually, then any outstanding errors ought to be spotted immediately by the bot's initiator. The bot can quickly and easily be blocked if there is any hint that its activity is causing damage. If the stated changes have been made, I suggest the immediate submission of an {{unblock}} request.
And can I register an enormous thanks to Fhocutt (WMF) for all her work addressing the outstanding issues and bringing the bot back to life! Martin(Smith609 – Talk) 06:58, 19 September 2015 (UTC)
I'm going to follow the programmer's taboo against deploying anything that might break things on a Friday, as I won't be around to fix anything over the weekend. It'll be set up for work on enwiki on Monday. --Fhocutt (WMF) (talk) 07:27, 19 September 2015 (UTC)
Maybe just me, but this was running against real articles when I tested it. Not that I'm complaining, it did a very nice job. 90.205.107.87 (talk) 15:05, 19 September 2015 (UTC)
Seems to be running against testwiki now, but here is a diff from yesterday:
My bad. Handling git branches on Tool Labs is always fun. Citations is now set to enwiki, and I've requested the associated account (Citation bot) be unblocked. Citations-dev is still set to testwiki, with the CitationBotDev account. All should be in order now. --Fhocutt (WMF) (talk) 21:14, 21 September 2015 (UTC)
{{cite journal}} is the wrong template for the citation in that diff (it is a paper in a conference proceedings). This began as a user error (someone added it as cite journal with a bibcode and no other detail) but I'd hope that the bot could be smart enough to look at the bibtex for the bibcode and notice it is @inproceedings not @article. —David Eppstein (talk) 20:46, 19 September 2015 (UTC)
I manually changed "cite book" to "cite journal" to get citation bot to fill it in, then forgot to change it back. "cite book", and I think "cite conference", are ignored by the bot. Maybe another area where enhancements could be made. There are other glitches with this sort of citation, for example the bot will add illegal chapter tags to "cite journal" templates that should really be "cite book". Lithopsian (talk) 22:01, 19 September 2015 (UTC)
From what I remember of my experience, trying to change the template used to the 'correct' template can really frustrate users who had a good reason to use a template that the bot doesn't recognize as correct in a particular instance; in most cases, to minimise the burden placed on the bot's maintainers, my sense is that it is usually best to leave components alone where a particular course of action may not be altogether straightforward. If this is a case where a simple rule can be applied in all cases to identify the (unambiguously) correct template, I would suggest that you request this as a new feature on the 'bugs' page, and perhaps (in view of possible controversy) file a request for bot approval for the task. Martin(Smith609 – Talk) 07:54, 20 September 2015 (UTC)
I'm not (strongly) suggesting that the bot should change the template type, although it would be on the xmas list somewhere. Just pointing out that it does create expansions that cause CS1 errors when a book is listed as a journal and so chapters are present which are not allowed for journals. It is always tempting, or habit-forming, for users (ie. me!) to use "cite journal" for a book or conference because the bot ignores "cite book" templates. Lithopsian (talk) 13:39, 20 September 2015 (UTC)
Good to see the bot back and alive and getting it up is now {{resolved}} cool! AManWithNoPlan (talk) 15:19, 2 October 2015 (UTC)
Cite PMID
Hello, when I try to use the cite PMID shortcut, I receive the following message (though the citation does not expand)
Thank you for your report. There appears to a be an open RfC about whether to deprecate Template:Cite pmid. As the functions were the same for cite doi and cite pmid and both appear very likely to end up being deprecated, it was easiest to remove that functionality for now and decide how to handle it in the code when the requirements are clear and somewhat stable. I'd suggest temporarily removing the "Jump the queue" option from Cite pmid until there is some consensus on how this template should interact with the bot. --Fhocutt (WMF) (talk) 20:31, 24 September 2015 (UTC)
Ok thank you for the update. I sincerely hope they are not deprecated, as I find the shorthand an extremely useful way to provide reliable peer reviewed references, particularly with larger scale revisions. Could you direct me to the RfC page? I do not see it at Wikipedia:Requests for comment/All. Plumpy Humperdinkle (talk) 08:44, 25 September 2015 (UTC)
Small update, it looks like the cite PMID bot is working perfectly. This is great news. Thanks for the input regardless. Plumpy Humperdinkle (talk) 08:56, 25 September 2015 (UTC)
See Template_talk:Cite_pmid and similar cite doi discussions. Sore point, much history. cite doi was originally deprecated, rather quietly, about a year ago and cite pmid got caught up in that deprecation. Now there is a fight about it. Effectively it has already happened, bots have been created and partially employed to convert the templates to cite journal form, which has brought the matter to wider attention. Lithopsian (talk) 10:16, 25 September 2015 (UTC)
It was hardly done "quietly". See this for instance. It was also discussed on the bot's talkpage. There's no reason for a fight, all that is going on is cleanup of a longstanding mistake. LeadSongDogcome howl! 19:27, 25 September 2015 (UTC)
Exactly, and who goes there or has it on a watchlist? First I heard, like most people, was when I was getting nagged about using a deprecated template. I'm not arguing, just explaining that the decision was taken over a year ago without 99% of the users of the cite doi and cite pmid templates knowing anything about it. And now that people know, there are arguments. Lithopsian (talk) 21:04, 25 September 2015 (UTC)
Will there be an alternate way of automatically filling citation templates? Which is mainly what I used it for. FunkMonk (talk) 13:22, 3 October 2015 (UTC)
Yes, exactly like that. And then when you or someone else run(s) the bot (by clicking Expand Citations in your Tools bar on the left side of the page, for example), it will do this. – Jonesey95 (talk) 00:23, 6 October 2015 (UTC)
Ah, thanks, didn't get that part! Any chance of this working for jstor again? FunkMonk (talk) 00:25, 6 October 2015 (UTC)
Thanks! On a related note, how does one determine whether a JSTOR article has a doi or not? It doesn't seem apparent on their own website. FunkMonk (talk) 01:02, 6 October 2015 (UTC)
There is no way to know until you try. Some work, some don't. Either the data is in crossref or not. Remember that {{cite jstor}} was just a wrapper for {{cite doi}} AManWithNoPlan (talk) 01:57, 6 October 2015 (UTC)
RFCs on citation templates
There are ongoing discussions (mostly parallel but since each one is argued as separate consensus, separate) regarding the use of (A) Template:Cite wdl (which creates subpages for a wrapper of cite web) here; (B) Template:Cite pmid (which is a wrapper for cite journal either in-article or via pages at Category:Cite pmid templates) here; and (C) another RFC at Template:Cite doi (a cite journal wrapper with almost 60k pages at Category:Cite doi templates) here. There are unique wrinkles to each one but basically all three discussions concern whether to deprecate these templates or not. Please comment there if everyone could. -- Ricky81682 (talk) 06:24, 8 September 2015 (UTC)
Clicking "jump the queue" in an article runs Citation Bot on the whole article instead of just one template
{{cite doi}} templates are deprecated (see discussions above), so the bot does not create them any more. I don't have permissions to do it myself, but I would suggest taking out the "jump the queue" link on Template:Cite doi so unexpected behavior does not occur. --Fhocutt (WMF) (talk) 00:57, 22 September 2015 (UTC)
I'll leave the template deciphering to an admin as well, after the second RfC is closed. In the meantime, I replaced the reference with {{cite journal|doi=10.1093/ref:odnb/36519}} and then ran Citation Bot against the article. It did this, which is not quite right. It should have put "The Oxford Dictionary of National Biography" into |work= and "Henry Thuillier" into |title=. I don't know if that is possible given the metadata that is available. – Jonesey95 (talk) 02:31, 22 September 2015 (UTC)
I suppose a valid use case of editing on a talk page might be for something of the form "Talk:Example/draft" where someone is working on a draft of the page "Example". Other than that though.... If someone goes and puts a bad citation or one to be expanded in the mainspace from work done in the talk page, the bot will take care of it then. All other cases will probably end up emitting an error and thus be taken care of in the (usually active) discussion in the context of the citation. @Jc3s5h:. --Izno (talk) 02:56, 10 October 2015 (UTC)
If someone needs to experiment with manual activation of the bot to see what it does with a bad citation, there are better places, such as the draft or user namespaces, rather than the talk namespaces. I see no reason for the bot to ever operate in the talk namespace. Jc3s5h (talk) 22:52, 10 October 2015 (UTC)
If someone wants to run the bot in talk space, it should be limited to the citations button in the edit window, that way they have to preview the changes. Even then it is questionable. Never just clicking expand citations, and certainly never automatic AManWithNoPlan (talk) 23:15, 10 October 2015 (UTC)
Always questionable. Let's completely disallow "correction" of Talk space. If someone wants to test something, use Draft space. ~ J. Johnson (JJ) (talk) 20:34, 11 October 2015 (UTC)
Create a blank citation with vertical parameter layout, then run the Citation bot on it
We can't proceed until
Bot operator's feedback on what is feasible
The new bot puts parameter values on the line following the parameter when the parameters are laid out vertically. It should put the values on the same line as the parameter, leaving the existing line breaks in place. – Jonesey95 (talk) 02:42, 22 September 2015 (UTC)
Even though it's cosmetic, this does seem like it would be important to fix (and probably not difficult). Kaldari (talk) 20:50, 22 September 2015 (UTC)
I was making an edit and had pressed the Citations button to see if it would help. It blanked the edit box as has been described previously, so I pressed back to return to the edit with my text. When I pressed save page at the end, Citation bot kicked in and sent me to the empty edit box again. The page was List_of_most_massive_stars_known.
What should happen
Citation Bot shouldn't get involved when trying to save a page.
We can't proceed until
Agreement on the best solution
If this is a bug then it is a bug in the citation JavaScript that puts the button there and not the bot. The back button in JavaScript and editing windows is pure evil and I have seen this with other tools. AManWithNoPlan (talk) 23:53, 16 October 2015 (UTC)
Have a doi that is not in CrossRef database, but still resolves to a url using dx.doi.org
We can't proceed until
A specific edit to the bot's code is requested below.
Requested action from maintainer
Add code to check if doi works using dx.doi.org before marking as bad, even if crossref does not have doi in it.
Each of those DOIs works fine. I don't know why the bot would have marked these DOIs as broken. Looks like a bug, maybe a timeout value that is too short? – Jonesey95 (talk) 23:12, 13 January 2015 (UTC)
Clearly a timeout problem or a throttle problem on the other end. Running the bot again made them go away on another page. AManWithNoPlan (talk) 15:12, 14 January 2015 (UTC)
That sounds re-assuring, thanks. DrKiernan (talk) 19:20, 14 January 2015 (UTC)
Re-running does not repair the broken_doi errors (with me). Blocked to avoid potential damage, hoping this will be fixed/resolved soon. Materialscientist (talk) 00:20, 15 January 2015 (UTC)
Unblocked - seems like the problem is gone. Materialscientist (talk) 04:37, 21 January 2015 (UTC)
That doi is not in the crossref database. AManWithNoPlan (talk) 02:57, 25 January 2015 (UTC)
Thanks. Suggestion: before marking a doi as inactive, the bot could verify it at dx.doi.org/[doi number] rather than the crossref database. Materialscientist (talk) 07:14, 25 January 2015 (UTC)
This is still happening. HereEaldgyth - Talk 12:08, 28 April 2015 (UTC)
I verified that crossref and doi.org were both up and then tried again. Still messed up. AManWithNoPlan (talk) 14:06, 28 April 2015 (UTC)
It's STILL doing it. diff. Can we fix this or stop doing a task or something... it's pretty annoying. Ealdgyth - Talk 15:51, 11 May 2015 (UTC)
A specific edit to the bot's code is requested below.
Not a bug. The citation previously had a vertical bar separating a blank doi parameter from another invalid parameter named by the paper's doi. The bot correctly ignored the invalid parameter and filled in the blank doi parameter. —David Eppstein (talk) 20:48, 22 November 2015 (UTC)
missed that. Amazing that the bot did anything AManWithNoPlan (talk) 21:17, 22 November 2015 (UTC)
Unresolved
Resolved
Moved from User talk:Brainy J § Please clarify
Done As above, URLs like "http://www.amazon.com/dp/0123456789" should be tidied up to 
Fixed in GitHub Pull 346
Unresolved
Fixed in GitHub Pull 334 
Resolved
Invalid
Martin (Smith609 – Talk) 19:49, 20 August 2012 (UTC)
benzband (talk) 21:04, 23 July 2012 (UTC)
Won't fix - can't replicate.
Hello, I'm BracketBot. I have automatically detected that your edit to Kakkonto may have broken the syntax by modifying 1 "[]"s. If you have, don't worry, just edit the page again to fix it. If I misunderstood what happened, or if you have any questions, you can leave a message on my operator's talk page.
Moved from User talk:Brainy J § Please clarify
Done. While you two bots were flirting, I fixed the bracket problem. Carry on. Don't let me interrupt. – Jonesey95 (talk) 20:04, 26 September 2013 (UTC)
Template:Cite doi/10.5665.2Fsleep.1378 has been nominated for deletion. You are invited to comment on the discussion at the template's entry on the Templates for discussion page. Ricky81682 (talk) 08:43, 6 June 2014 (UTC)
{{resolved}}
Not fixed
It looks like a duck to me – Allen4names (contributions) 20:43, 25 May 2015 (UTC)
