User talk:Citation bot/Archive 41

This is an archive of past discussions with User:Citation bot. 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.

Archive 35

←

Archive 39

Archive 40

Archive 41

Adds publisher/imprint as journal to book citation

Status: Fixed - the bot was not supposed to talk to Zotero when CrossRef works. I never actually verified that when being mostly in charge. That now actually is the case.
Reported by: —David Eppstein (talk) 01:37, 15 January 2025 (UTC)

What happens: Special:Diff/1262795625
What should happen: Not that. PS User:Dominic3203: this was an edit invoked by you and therefore you should have checked and caught this. See WP:ANI#User:Citation bot won't stop adding incorrect dates to articles re editors taking responsibility for the mistakes of citation bot.
We can't proceed until: Feedback from maintainers

Swedish Film Database / Svensk Filmdatabas (svenskfilmdatabas.se)

Status: Fixed - block dates for that webstie
Reported by: Nardog (talk) 14:05, 11 January 2025 (UTC)

What happens: The date of birth in a biography is misinterpreted as the date of publication.
Relevant diffs/links: https://en.wikipedia.org/w/index.php?title=Ninja_Thyberg&diff=1268497750&oldid=1262826678
We can't proceed until: Feedback from maintainers

Gardens of Castle Howard

In the Gardens of Castle Howard article, the bot has twice added implausible dates to website citations. I've reverted both occasions but should there not be a check that webpage dates are not earlier than the 1990s? Warofdreams talk 23:27, 18 January 2025 (UTC)

Fixed Added parksandgardens.org to list of websites that provide bogus dates. Such early dates are valid for many websites the re-produce works. AManWithNoPlan (talk) 16:53, 19 January 2025 (UTC)

Incorrect / overly specific dates using current day

Status: Fixed - anything older than 1990 will just get year.
Reported by: Salpynx (talk) 08:42, 17 January 2025 (UTC)

What happens: This edit shows the bot adding presumably incorrect dates and months, 02 December, to a number of different 17th century works from an edit made on 02 December 2024.

https://en.wikipedia.org/w/index.php?title=Aphra_Behn&diff=1260721229&oldid=1259140365

What should happen: The years are probably correct, but it's unlikely that all those works were published on the same day of the year as the edit. I don't know if this is user error or a problem with the bot.
Relevant diffs/links: https://en.wikipedia.org/w/index.php?title=Aphra_Behn&diff=1260721229&oldid=1259140365
We can't proceed until: Feedback from maintainers

That's crazy. I will add code to not add specific dates for things before 1900. Considering that a belief in precise dating is relatively new, and mulitple calendar systems, and the general published vs written vs date on the document issues. That is probably for the best. Even July 4 should be July 2, but no one cares. AManWithNoPlan (talk) 16:56, 19 January 2025 (UTC)

ANI

There is currently a discussion at Wikipedia:Administrators' noticeboard/Incidents regarding an issue with which you may have been involved. Thank you.Legend of 14 (talk) 14:40, 14 January 2025 (UTC)

Caps: TheoretiCS

Status: {{fixed}} with code that will stop this, even though I have no idea why this happened
Reported by: Headbomb {t · c · p · b} 06:29, 10 January 2025 (UTC)

What should happen: [1]
We can't proceed until: Feedback from maintainers

One other thing should also have happened here: When Citation bot converted this from a conference citation into a journal citation (a bad idea to do automatically in general but usually ok as a manual edit and ok this time; see #Converts conference citation to journal citation and changes title case) last November (Special:Diff/1256921655) it should have removed the url pointing to the conference version, to avoid having an internally inconsistent and garbled citation. —David Eppstein (talk) 07:13, 10 January 2025 (UTC)

I wonder if this is now fixed. I cannot reproduce this, and it is VERY odd. Sorry for being slow, but I am now a grandparent in the real world. AManWithNoPlan (talk) 17:08, 22 January 2025 (UTC)

Converting Mixed Case to CAPS

This edit replaced "publisher=The Last Magazine" with "work=THE LAST MAGAZINE". Maybe it scraped the website name from the web page, but I’m surprised it left it in CAPS. --Northernhenge (talk) 20:38, 12 January 2025 (UTC)

{{fixed}} with code that will stop this, even though I have no idea why this happened. I cannot reproduce. AManWithNoPlan (talk) 17:16, 22 January 2025 (UTC)

Leave page(s)=n.p. alone

Status: Fixed, and it will add the period back when missing
Reported by: Headbomb {t · c · p · b} 11:55, 24 January 2025 (UTC)

What happens: [2]
We can't proceed until: Feedback from maintainers

An exception to the dot removal rule. Headbomb {t · c · p · b} 11:55, 24 January 2025 (UTC)

Journal name initial letter is /lower case, bot mistakenly converts to upper case

Status: {{fixed}}
Reported by: Larry Koenigsberg (talk) 23:27, 24 January 2025 (UTC)

What happens: Citation bot incorrectly capitalized the initial letter of a journal name.
What should happen: It should leave it alone, as the initial letter is lower case.
Relevant diffs/links: See Edward A. Kramer#Selectedbibliography, reference for Fulton, Jim; Kramer, Ed (1 August 1997). "Can you ever be too thin?". The publication is netWorker <sic>. Citation bot incorrectly capitalized the initial letter. I reverted the change; a user had recently comitted the same error, after which I restored it and notified them.
We can't proceed until: Feedback from maintainers

Diff? Headbomb {t · c · p · b} 23:48, 24 January 2025 (UTC)

Special code add when journal title is "NetWorker". AManWithNoPlan (talk) 01:12, 25 January 2025 (UTC)

Page number replaced by page range of journal cite

Status: Not a bug, but thank you for mentioning. Better safe than sorry.
Reported by: Pol098 (talk) 15:12, 27 January 2025 (UTC)

We can't proceed until: Feedback from maintainers

Citation bot has edited a reference to a journal paper changing a specific page number to the range of the whole paper (the page number happened to be the first page of the paper). This is wrong. I don't know if the bot just did this, or if there was intervention by an editor. This is the diff. "page=213" was changed to "pages=213–244". While the page number given could have been wrong, it's not up to a bot to decide that (possibly a human editor was involved in this decision, I can't know that). I comment that in this particular case I checked the reference, and the two points sourced by it in the WP article text were in the first page (213) of the paper.

<ref name= tay>{{cite journal| author= Peter J. Taylor| doi= 10.1007/BF00146987 | title= Technocratic Optimism, H.T. Odum, and the Partial Transformation of Ecological Metaphor after World War II| work= Journal of the History of Biology| volume= 21| number= 2| date= June 1988| page= 213}}</ref>
was changed to
<ref name= tay>{{cite journal| author= Peter J. Taylor| doi= 10.1007/BF00146987 | title= Technocratic Optimism, H.T. Odum, and the Partial Transformation of Ecological Metaphor after World War II| journal= Journal of the History of Biology| volume= 21| number= 2| date= June 1988| pages= 213–244 | pmid= 11621655 }}</ref>

with summary

Add: pmid, pages, journal. Removed parameters. Some additions/deletions were parameter name changes. | Use this bot. Report bugs. | Suggested by Spinixster | Category:Schools of economic thought | #UCB_Category 69/95

It's standard for journal citations to give the full page range, not just the starting page. You can always specify the exact page with |pages=213–244 [213], or |page=213, if it's particularly important to specify the exact page. Headbomb {t · c · p · b} 15:40, 27 January 2025 (UTC)

"Standard" (enforced) or just "customary"? If a specific statement in the WP article is supported on a particular page of a paper, following the reference (either to check it or to find out more) in a long article leads to waste of time. Does Citation Bot only replace a specific page by a range if it happens to be the first page, or always?

It would seem to me that if CB changes the page, it should change "page=213" to "pages=213-244 [213]". That is what the editor specified; it's not up to CB to second-guess a human editor. Pol098 (talk) 15:59, 27 January 2025 (UTC)

I agree with Pol. The bot should not change the sense of a citation that a human has added. If one page has been cited, the bot should not change that. I am increasingly disturbed by the bot's behaviour, it is in danger of becoming a net negative. DuncanHill (talk) 16:03, 27 January 2025 (UTC)

The bot only replaces the first page with a page range because it's an ambiguous situation where what's intended is unclear (and several tools just use the first page instead of full page range, with the material supporting the claim not on the first page). In the case of a page in the middle of the range, it's clear what the intent is, so it doesn't touch it. The bot's always done this since as far back as I can remember, this is not new behaviour. Headbomb {t · c · p · b} 16:07, 27 January 2025 (UTC)

OK, not a bug, thanks, sorry to waste your time. Best wishes, Pol098 (talk) 16:38, 27 January 2025 (UTC)

Incorrect date added to dead/non-dead reference

Status: Not a bug - this is not a bot edit. This is a human edit and a bot edit. Thank you for reporting though.
Reported by: Orxenhorf (talk) 11:47, 28 January 2025 (UTC)

What happens: The bot added a date to a reference link that predated when the link was even added to the article, and the link is not actually dead. A reference was added to PASOK at 20:04 on 30 October 2024 that included "url-status=dead" in its parameters. At 20:21 on 5 November 2024 the bot changed that "{{dead link|date=May 2024}}". I don't know where it got the "May 2024" date from.
Relevant diffs/links: Added - https://en.wikipedia.org/w/index.php?title=PASOK&diff=prev&oldid=1254394036

Changed - https://en.wikipedia.org/w/index.php?title=PASOK&diff=prev&oldid=1255608757

We can't proceed until: Feedback from maintainers

Caps: AAPG

Status: Fixed
Reported by: Headbomb {t · c · p · b} 21:14, 29 January 2025 (UTC)

What should happen: [3]

Apostrophes are changed from source material to bot edits

Status: new bug
Reported by: Inonit (talk) 03:38, 7 February 2025 (UTC)

What happens: Not sure whether this is a bug or intentional, but the bot seems to replace a right single quote (U+2019) with a neutral apostrophe (U+0027). This ends up overriding the source material in things like headlines, causing source headlines to be changed from the headlines actually used by the source to a modified version produced by the bot. But maybe this is intentional in an attempt to simplify the characters used in citations? =
Relevant diffs/links: https://en.wikipedia.org/w/index.php?title=Birthright_citizenship_in_the_United_States&diff=1274312628&oldid=1274158512
We can't proceed until: Feedback from maintainers

@Inonit: per MOS:CONFORM, we're allowed to change punctuation to our own house style (in fact, it should be adapted... without comment). Module:CS1 displays curly quotes as straight quotes anyway, so I agree there's very little value to this kind of edit, but it does comply with MOS guidance and is {{not a bug}}. Folly Mox (talk) 12:02, 7 February 2025 (UTC)

Unable to log in to Citation Bot

Status: Fixed - this was something wrong with wikipedia
Reported by: Jay8g [V•T•E] 03:02, 31 January 2025 (UTC)

What happens: After allowing the OAuth connection, the bot gives the message Incoming authorization tokens did not work - try again please
We can't proceed until: Feedback from maintainers

I get the same, btw. Headbomb {t · c · p · b} 15:46, 31 January 2025 (UTC)

Something is seriously messed up with the toolserver infrastructure. Lots of networks stuff is not working.. I can run it just fine from my home machine. AManWithNoPlan (talk) 17:06, 1 February 2025 (UTC)

I will add that I rebooted the bot, and the logs are spammed with all sorts of errors still. AManWithNoPlan (talk) 17:15, 1 February 2025 (UTC)

Even the gadget API is giving "screw you sucka" errors. AManWithNoPlan (talk) 17:16, 1 February 2025 (UTC)

Seems like the bot is back up and running as of today. Jay8g [V•T•E] 21:14, 3 February 2025 (UTC)

Can you fix this

I added source information in this page, https://en.wikipedia.org/wiki/Pashtun_colonization_of_northern_Afghanistan#:~:text=Early%20colonization,-British%20Army%20Colonel&text=Before%20the%201880s%2C%20they%20numbered,called%20Pashtunization%20in%20northern%20Afghanistan.

But URL is stating a problem. 2402:E280:3D48:133:1CB2:C834:72C:618F (talk) 15:53, 2 February 2025 (UTC)

Bot messes up non-English orthography

Status: Fixed
Reported by: Phazd (talk

What happens: Bot applies English capitalisation to language (Croatian) with different capitalisation rules (correct: <Narodne novine>, converted to incorrect: <Narodne Novine>).
Relevant diffs/links: https://en.wikipedia.org/w/index.php?title=Croatian_language&diff=prev&oldid=1273517745

Not expanding bare URLs?

Status: Fixed - this was something wrong with wikipedia
Reported by: Jay8g [V•T•E] 23:43, 3 February 2025 (UTC)

Replication instructions: Run the bot on various pages with bare URLs, such as the ones here. Normally the bot is able to expand many of those, but right now it doesn't seem to be touching them at all.
We can't proceed until: Feedback from maintainers

I'm having the same issue - the bot is not expanding citations that previously it would have done. Ccferrie (talk) 09:38, 5 February 2025 (UTC)

Adds journal to cite books for SISSA citations

Status: {{fixed}} to deal with bad bibcode data
Reported by: Headbomb {t · c · p · b} 10:15, 12 February 2025 (UTC)

What happens: [4]
What should happen: [5]
We can't proceed until: Feedback from maintainers

Concerns doi prefix 10.22323, mostly. Headbomb {t · c · p · b} 10:15, 12 February 2025 (UTC)

Series: Current Topics in Behavioral Neuroscience / Curr. Top. Behav. Neurosci. / Curr Top Behav Neurosci

Status
Reported by: Headbomb {t · c · p · b} 22:45, 12 February 2025 (UTC)

What should happen: [6]
We can't proceed until: Feedback from maintainers

Not quite fixed, still need to do follow up cleanup, e.g. [7]. Headbomb {t · c · p · b} 16:28, 18 February 2025 (UTC)

User complaint about Croatian capitals

A user complained on my talk page about improper capitalization of a Croatian journal. This is the diff. They said that "standard Croatian orthographic rules are available at pravopis.hr" which may be helpful. Abductive (reasoning) 15:08, 19 February 2025 (UTC)

Putting |language=hr didn't fix this either. It should've. Headbomb {t · c · p · b} 15:29, 19 February 2025 (UTC)

{{fixed}} by adding title words to list of non-english words. AManWithNoPlan (talk) 01:59, 28 February 2025 (UTC)

Can you fix this added source

Added source to this page but link didnt come in the right way. Can you fix this.

https://en.wikipedia.org/wiki/Theories_of_Pashtun_origin#Saka_theory 2402:E280:3D48:133:85CB:4234:9E7:E962 (talk) 07:58, 22 February 2025 (UTC)

{{wontfix}} no idea what your want, nor do I have time. AManWithNoPlan (talk) 02:00, 28 February 2025 (UTC)

10.1002/leap (Learned Publishing) and 10.1002/aelm (Advanced Electronic Materials) are free access

Status: {{fixed}}
Reported by: Headbomb {t · c · p · b} 12:40, 23 February 2025 (UTC)

What should happen: [8]
We can't proceed until: Feedback from maintainers

Same for 10.22459/... Headbomb {t · c · p · b} 15:10, 25 February 2025 (UTC)

Caps bug

Status: {{fixed}}
Reported by: Bernhard Zelazny (talk) 15:33, 26 February 2025 (UTC)

We can't proceed until: Feedback from maintainers

On the page

https://en.wikipedia.org/wiki/Oryctes_elegans

the bot changed a citation of an article in the German journal

"Stuttgarter Beiträge zur Naturkunde aus dem Staatlichen Museum für Naturkunde in Stuttgart - Serie A Biologie"

to

"Stuttgarter Beiträge zur Naturkunde aus dem Staatlichen Museum für Naturkunde in Stuttgart - Serie a Biologie"

In "Serie A Biologies" the "A" refers to the first letter of the alphabet and should be written as a capital letter.

Regards Bernhard Zelazny (talk) 15:33, 26 February 2025 (UTC)

10.1017/nlp (Natural Language Processing) and 10.1016/j.nlp (Natural Language Processing Journal) are free DOIs

Headbomb {t · c · p · b} 16:34, 3 March 2025 (UTC)

{{fixed}} AManWithNoPlan (talk) 16:57, 8 March 2025 (UTC)

Bad title: "Client Challenge"

Status: {{fixed}} and removed from all existing pages
Reported by: Jay8g [V•T•E] 22:44, 6 March 2025 (UTC)

What happens: [9]
We can't proceed until: Feedback from maintainers

CAPS: MELUS

Status: {{fixed}}
Reported by: Headbomb {t · c · p · b} 02:17, 9 March 2025 (UTC)

What should happen: [10]
We can't proceed until: Feedback from maintainers

Removing DOIs

In this edit, the bot removes the |doi= parameter for an entry that also includes |jstor=. I'd think that DOI would be more authoritative than a JSTOR link, since DOI is specifically a permanent identifier whereas JSTOR is just a specific academic database (that we don't want to necessarily give priority to over others). In this case, the DOI went to an (admittedly broken) page for UC Press. Could someone explain why the DOI was removed here, and if needed could we consider the decisionmaking behind it? Sdkb ^talk 20:34, 12 March 2025 (UTC)

The issue here is that JSTOR has a bunch of DOI they never properly issued. In the cases where the DOI is broken and would point to JSTOR (DOIs starting with 10.2307), the DOI gets removed since it would (normally) point to the JSTOR page anyway.

Here though, the issue is that the DOI should have been updated to this page https://online.ucpress.edu/ch/article-abstract/45/3/279/28393/Carl-Irving-Wheat-1892-1966?redirectedFrom=fulltext, but wasn't. You can report it to UC Press, or to doi.org (at doi-help@doi.org) Headbomb {t · c · p · b} 22:00, 12 March 2025 (UTC)

{{wontfix}} - the DOI is not real. AManWithNoPlan (talk) 14:31, 26 March 2025 (UTC)

Bad title: "One moment, please"

Status: {{fixed}}
Reported by: Jay8g [V•T•E] 22:25, 12 March 2025 (UTC)

What happens: [11]
What should happen: no edit
We can't proceed until: Feedback from maintainers

Telegraph Obituaries is not an author

Status: {{fixed}}
Reported by: CorrectionsJackal (correct me) 03:32, 17 March 2025 (UTC)

What happens: Special:Diff/1280907329
What should happen: Don't add Telegraph Obituaries as an author
We can't proceed until: Feedback from maintainers

Web Desk is not an author

Status: {{fixed}}
Reported by: CorrectionsJackal (correct me) 06:30, 17 March 2025 (UTC)

What happens: Special:Diff/1280923128
What should happen: Don't add Web Desk as an author
We can't proceed until: Feedback from maintainers

Filipino News is not an author

Status: {{fixed}}
Reported by: CorrectionsJackal (correct me) 07:40, 17 March 2025 (UTC)

What happens: Special:Diff/1280928577
What should happen: Don't add Filipino News as author. If it's possible, if the bot detects "news" as part of the author's name, the "author" should not be included.
We can't proceed until: Feedback from maintainers

Bad title: "Please verify you are human"

Status: {{fixed}}
Reported by: Jay8g [V•T•E] 02:24, 21 March 2025 (UTC)

What happens: [12]
We can't proceed until: Feedback from maintainers

Am730

Status: {{fixed}}
Reported by: —👑PRINCE of EREBOR📜 11:08, 22 March 2025 (UTC)

What happens: "|author1=Am730" is added to am730 sources.
What should happen: am730 is not the author, it is a newspaper. Adding it to the citation template would trigger the green maintenance message. It should not be added to the
Relevant diffs/links: Special:Diff/1281715504, Special:Diff/1280774192, Special:Diff/1275721859
We can't proceed until: Feedback from maintainers

Why does CB standardize punctuation in citation titles?

As in this diff. It seems destructive to the point of a citation, which is to represent the cited material. ꧁Zanahary꧂ 06:53, 5 March 2025 (UTC)

See WP:MOSQUOTE. Headbomb {t · c · p · b} 07:39, 5 March 2025 (UTC)

That is for content, not citations. The MOS also (I am gonna guess) says not to misspell words, but policy rightly prohibits fixing spelling in citations, because it makes no improvement to the encyclopedia and defeats the purpose of a citation. ꧁Zanahary꧂ 03:58, 6 March 2025 (UTC)

Double slash in doi

Status: {{not a bug}}
Reported by: Isaidnoway (talk) 07:36, 5 April 2025 (UTC)

What happens: Double slash // in doi parameter creates CS1 error:external link. Bot added the doi parameter to citation with doi, despite the {{doi}} template already being used in the citation.
What should happen: Don't add the doi parameter to citation when doi template is already present in citation.
Relevant diffs/links: doi param added
We can't proceed until: Feedback from maintainers

Not a bug, this is the fix. See also [13] and Help_talk:Citation_Style_1#module_suite_update_12–13_April_2025. Headbomb {t · c · p · b} 00:23, 6 April 2025 (UTC)

BibCode search generates bad data

Status: Fixed with a more restrictive search
Reported by: Guliolopez (talk) 11:40, 3 April 2025 (UTC)

What happens: Bot added DOI and other details of a 1960s article in Nature magazine to a 2008 report by the Irish government. Seemingly keying (only?) on the name of the report?
Relevant diffs/links: https://en.wikipedia.org/w/index.php?title=Dublin_Institute_for_Advanced_Studies&diff=next&oldid=1278787155
We can't proceed until: Feedback from maintainers

Caps: aBIOTECH

Status: Fixed
Reported by: Headbomb {t · c · p · b} 05:05, 6 April 2025 (UTC)

What should happen: [14]

Not expanding bare URLs again

Status: Fixed for now. No idea.
Reported by: Jay8g [V•T•E] 18:31, 28 March 2025 (UTC)

What happens: Bot does not seem to be expanding bare URLs (specifically normal URLs that it relies on Citoid/Zotero for, not ones that it has special handling for). It seems that this is a bot problem as other tools that rely on Citoid/Zotero (such as ReFill) are working.
Replication instructions: Run on pages from this search; the bot should expand bare URLs on at least some of them but currently isn't working on any of them
We can't proceed until: Feedback from maintainers

@AManWithNoPlan: This was working earlier today, but it's now broken again. Jay8g [V•T•E] 22:15, 7 April 2025 (UTC)

I have no idea. It just suddenly fails on everything. And sometimes an hour or a week later it just starts working again. AManWithNoPlan (talk) 22:17, 7 April 2025 (UTC)

And when it stops working, then Zotero is down. AManWithNoPlan (talk) 22:18, 7 April 2025 (UTC)

"OUP Oxford" should be "Oxford University Press"

Status: Fixed
Reported by: —David Eppstein (talk) 23:47, 15 March 2025 (UTC)

What happens: Special:Diff/1280696238
What should happen: If we're going to cite the book, we should use the correct publisher name. But really this was never intended as a reference to a published book itself. It is a reference to the Google Books page about the book, which has a brief biography of its author. It is possible that the book itself also has the same brief biography but I don't know that because there is no preview available. So the book metadata cruft added by Citation bot over multiple passes including the earlier Special:Diff/924957588 should be removed. I have done so, and converted it into a manually formatted citation, because it appears to be impossible to prevent bots from misunderstanding the citation and breaking it. (In the distant future, all citations will be manual or broken by bots, but that's irrelevant to the wrong publisher title bug reported here.)

it appears to be impossible to prevent bots from misunderstanding the citation. See User:Citation_bot/use#..._the_bot_made_a_mistake? Headbomb {t · c · p · b} 00:24, 16 March 2025 (UTC)

It's a very general issue, that can be handled unsatisfactorily with very specific case-by-case workarounds. The general issue is that the bot is not actually intelligent, does not understand the purpose of citations, and cannot intelligently distinguish between references for closely related entities (a web page or database entry about a publication from the publication itself), as a human editor who is not deliberately trying to be as bot-like as possible could reasonably be expected to do. It should not be necessary for human editors, going about the day-to-day task of adding references to articles, to imagine all possible ways that a bot could misinterpret a citation and head them off one by one. —David Eppstein (talk) 00:41, 16 March 2025 (UTC)

URLS with "#about_author_anchor" will no longer be cleaned or expanded. Also, the publisher meta-data will get cleaned before being added. AManWithNoPlan (talk) 15:10, 26 March 2025 (UTC)

Adds generic place-holder names

Status: Fixed
Reported by: Worldbruce (talk) 18:26, 4 March 2025 (UTC)

What happens: Bot adds: |last1=Editor |first1=The
What should happen: Bot does nothing. (Or, if, in this case, "Editor, The" is valid and appropriate in terms of Help:CS1 errors#Cite uses generic name, some guidance would be appreciated regarding how users should determine that validity and appropriateness.)
Relevant diffs/links: https://en.wikipedia.org/w/index.php?title=Sir_Salimullah_Medical_College&diff=prev&oldid=1278062686

Oddly, in this case it is valid. We do not accept such authors from Zotero anymore, but it will now be formatted as "The Editor", instead of "Editor, The". When correct, the name should be ((The Editor)) to suppress the error message. AManWithNoPlan (talk) 15:17, 26 March 2025 (UTC)

Useless bibcodes redux

There has been another spate of Citation Bot edits adding completely content-free bibcodes and making no other change, "Suggested by Dominic3203". I have asked Dominic3203 before and they refuse to do any human checking of these edits / pass the blame onto Citation Bot. These edits are obnoxious and counter-productive, and tempt me to just block Citation Bot from the pages where they occur.

Some example diffs: 1277541897, 1277469361, 1277536067.

When a bibcode contains no information beyond the author, title, journal name, and maybe a DOI, it provides no benefit to readers, since we already relay the same information in our citation. The bibcode is pure line noise. Such bibcodes should not be indiscriminately added to articles by bots.

(Previously: Archive 40 § Can we please not add bibcode when it contains no useful information?)

–jacobolus (t) 11:42, 25 February 2025 (UTC)

@Jacobolus I would say just temporarily disable the bibcode adding function as a whole. There is no point accusing anyone about this. Dominic3203 (talk) 14:08, 25 February 2025 (UTC)

Bibcodes are highly valuable, not useless in the least. If you don't like them, or don't personaly have a use for them, just ignore them. They're plenty useful to me and to several other people. Headbomb {t · c · p · b} 14:36, 25 February 2025 (UTC)

Can you explain what is "highly valuable" about these examples? They contain no abstract, no copy of the paper, and the scarce scraps of metadata are redundant with information already in the citation. For me personally something like this Google Scholar link would be significantly more useful than Bibcode:2004MatGe..36..917K – should I be adding Google Scholar links to every citation on Wikipedia? Or how about this Internet Archive Scholar link? That one also looks significantly than the Bibcode in this particular case. Why don't we just add like 10 links to various citation indices for every citation? That must surely be better right? After all, if you don't have a personal use for a particular one just ignore it.... –jacobolus (t) 17:16, 25 February 2025 (UTC)

Using Bibcode:2004MatGe..36..917K for example, this has a link to the offical paper, which is extremely useful when DOI.org is down. For the abstract, it's just a matter of submitting a correction, wait a few days, it'll be there. You can also export the citation in various format, Bibtex, RIS, Endnote...

Or, using Bibcode:2012npro.book.....L as an example, you get a link to the paper. You yourself said you got the DOI from it, proving it's information is not worthless. Headbomb {t · c · p · b} 18:58, 25 February 2025 (UTC)

It doesn't have a link to the paper. All it has is a DOI (which links to the publisher's site, where the paper is paywalled). We can just as easily include the DOI in our citation. If we want a backup for the case where doi.org is down we should just put the link to the publisher's site directly in our citation and skip the pointless intermediary. All of the other information is also completely redundant; it doesn't even have an abstract. It would be significantly more useful to use a link to Google Scholar, which itself links out to multiple citation indices instead of this bibcode. Or if we're going ham on the citation indices to waste as much visual space as possible, why not also throw in

The latter five links even include PDFs. For that matter, we could also include links to search engine results pages at every major search engine with the author and title pasted in, all of which also immediately link to the publisher's website and also have links to PDFs. The bibcode is pure line noise, wasting space at readers' expense. Calling this "extremely useful" can't be a serious, earnest description; you are grasping at straws here. As for an "export citation" feature: most of the other citation indices include this feature, as does the publisher's site, but if you really care this would be relatively trivial to build a Wikipedia-integrated tool for, and then you wouldn't even need to go off-site, and you could actually rely on it being there instead of only finding it the same way in the small fraction of Wikipedia citations where a particular citation index happens to be listed. –jacobolus (t) 05:22, 1 March 2025 (UTC)

Some bibcodes are useful because they provide free views of content that is otherwise paywalled. This one isn't, though, I agree. If the bot cannot distinguish the useful from the useless ones then maybe it should not be doing this addition. —David Eppstein (talk) 06:54, 1 March 2025 (UTC)

@David Eppstein Agree. Dominic3203 (talk) 06:57, 1 March 2025 (UTC)

"All it has is a DOI" i.e. a link to the paper. And FWIW, I'd be in favour of adding most of those too. Headbomb {t · c · p · b} 14:40, 1 March 2025 (UTC)

"be in favour of adding most of those too" – This conversation is encouraging me to block citation bot from pages where it starts automatically adding this chaff. –jacobolus (t) 16:15, 1 March 2025 (UTC)

Bibcoded contain links to the CDS and SIMBAD database, which is pretty useful in astronomy articles for 1) Use SIMBAD objects to find other astronomical objects also mentioned in the paper and 2) Find data for astronomical objects in an external table. I use bibcodes a lot, specially when i need to know what astronomical objects are mentioned in a study, or when the relevant data isn't is in the paper itself, but in external tables that are linked in the bibcode. Plus, since it collects data for multiple journals, i can search for papers of which an specific author name is mentioned, which is helpful to find papers and cite them in Wikipedia. 21 Andromedae (talk) 19:27, 3 March 2025 (UTC)

I'm not an astronomer / don't read astronomy articles, so I've never come across this. I don't have a problem with bibcodes in a context where they contain non-redundant information not included already in the Wikipedia article citation. I interact with bibcodes in 2 different contexts: (1) I am looking for some old paper about the history of mathematics and the only online copy happens to be at ADS – here the bibcode is great! or (2) Citation Bot just added a bibcode about some random math paper to a math article but I try clicking it and then discover that it does not contain the paper or as far as I can tell anything else that wasn't already included in the Wikipedia article's citation or the publisher's site. –jacobolus (t) 19:34, 3 March 2025 (UTC)

I have never found a bibcode to be useful. Johnjbarton (talk) 17:59, 1 March 2025 (UTC)

Some bibcode pages have a scan of the full content available, in some cases the only full-text copy available online and in other cases the only full-text copy not behind a paywall. Such bibcodes have great value. For example, Bibcode:2006JHA....37..233K links to the PDF scan https://articles.adsabs.harvard.edu/pdf/2006JHA....37..233K of this review, which from a quick web search is only otherwise available at the paywalled publisher's site doi:10.1177/002182860603700209. –jacobolus (t) 19:03, 1 March 2025 (UTC)

I opened the DOI link then clicked on Google Scholar button, it has a link to the PDF.

I'm not saying there is no source uniquely available at harvard ADS, but bibcodes aren't needed to find them. Johnjbarton (talk) 19:19, 1 March 2025 (UTC)

Really? What did you click? I opened the DOI link and clicked the Google Scholar button and the result was "Sorry, we didn't find any articles that cite https://journals.sagepub.com/doi/abs/10.1086/isis.97.4125060." When I type the author/title directly into Google scholar, I get to 1320999009635041564 which has ADS as the only available full text link, https://adsabs.harvard.edu/full/2006JHA....37..233K (and it picks the web page of images rather than the PDF). In such cases we should link to Bibcode:2006JHA....37..233K, which is the landing page for this scan.

(This case is pretty weird though. ADS has 3 different records Bibcode:2006JHA....37..233K, Bibcode:2006JHA....37R.233S, Bibcode:2006JHA....37Q.233S. The first one has the book author and review author switched in the ADS metadata, and all three mix up the review title and the book title. The one with the incorrect author in ADS leads to a Google scholar cluster with the correct author, 1320999009635041564 whereas the one with the correct author in ADS leads to a Google scholar "cluster" with the incorrect author, 14551181599797032627. Lesson: ADS doesn't have particularly careful metadata QC, and something is going wrong when Google scholar ingests data from ADS.) –jacobolus (t) 20:33, 1 March 2025 (UTC)

Hmm. When I click the DOI link I end up on https://journals.sagepub.com/doi/10.1177/002182860603700209. My little blue Scholar button in my Chrome toolbar pops up a single entry kinda like this page but formatted in a small square. The PDF link is in the box.

I will say that Scholar gets these PDF / Harvard things wrong quite a bit. The link it offers often ends up being the HTML from the Harvard site again. At times the "All 3 versions" will be lucky. Johnjbarton (talk) 20:52, 1 March 2025 (UTC)

I have seen both useless and useful bibcodes. Is there a way a bot could tell their usefulness? Jo-Jo Eumerus (talk) 09:33, 2 March 2025 (UTC)

Bibcodes are extremely useful - I click them far more often than any other link in a reference (even more than the DOI). They link to a standard high-quality database (ADS) which contains useful metadata and links to the material - the official publication, archived open-access versions, and scanned copied of older material that often isn't available anywhere else. The bot should continue adding them. The very rare cases where there's some problem with the ADS entry can be dealt with on a case-by-case basis - in the vast majority of cases the bibcode & ADS link are useful. Modest Genius ^talk 12:38, 3 March 2025 (UTC)

A bibcode with a scanned copy of the paper is useful. A bibcode that merely links to the doi is a pointless extra step to reach a reference that already links to the same doi. A bibcode that doesn't even have a pointer to a reference elsewhere (like say Bibcode:1614mlcd.book.....N from Special:Diff/1278520244) is worse than useless: the only thing readers can do with it is waste their time. The bot should learn the difference and only add the useful ones. —David Eppstein (talk) 19:09, 3 March 2025 (UTC)

A bibcode like Bibcode:1614mlcd.book.....N is far from useless, it verifies that the book indeed exists, and that the biblographic information provided is correct. And if the book is found to be digitized, a url to the digitized version can be provided (I've added https://archive.org/details/mirificilogarit00napi/mode/2up to it, which should appear in a few days [and is now up Headbomb {t · c · p · b} 14:03, 4 March 2025 (UTC)].) Headbomb {t · c · p · b} 19:33, 3 March 2025 (UTC)

You know an even better link to verify that the book exists and has multiple available scans? Our article Mirifici Logarithmorum Canonis Descriptio here at Wikipedia. (Though it would be nice if scans and translations were linked in a separate "Editions and translations" type section near the bottom, instead of only in the footnotes.) The particular footnote that Citation Bot keeps trying to add a bibcode to in Logarithm already had (1) a link to our article, (2) metadata about the book and its sequel, (3) a link to a scan, and (4) a link to an excellent freely available English translation of both books which I added in special:diff/1153321451. In that case, the Bibcode is not adding any meaningful additional thing – it instead just distracts readers by the wikilinked word "bibcode" followed by a bunch of arbitrary symbols making an opaque identifier, which when clicked yields no new information. –jacobolus (t) 19:35, 3 March 2025 (UTC)

For every 'useless' bibcode like that, there are hundreds that are useful. Just because ADS isn't a perfect database doesn't mean it's useless. Don't let the perfect be the enemy of the very good. In the vast majority of cases, including the bibcode in a reference entry is useful. In the few cases where the bibcode isn't useful, it can be removed manually - that is a lot less work than manually adding all the ones that are useful. Of course the bot cannot make subjective judgements. Modest Genius ^talk 13:17, 4 March 2025 (UTC)

PS. I am assuming that there's a way for a bibcode to be manually removed without the bot adding it back e.g. by putting a hidden comment in the bibcode= field. If that doesn't work, then I do think the bot should provide that functionality. Modest Genius ^talk 13:19, 4 March 2025 (UTC)

Any identifier can always be supressed with |identifier=. Headbomb {t · c · p · b} 14:01, 4 March 2025 (UTC)

Not a bug AManWithNoPlan (talk) 14:23, 8 April 2025 (UTC)

Don't insert bogus ISBNs for books from the 1930s

Status: Won't fix, since citation was alread a mess
Reported by: Gnuish (talk) 23:04, 24 February 2025 (UTC)

What happens: CitationBot added an ISBN field to a book published in 1935, which is long before ISBN's were invented. The ISBN that it added was for a reprint of the book that occurred dozens of decades later, by a different publisher.
What should happen: Don't change citations for works from before the ISBN was invented, certainly not for works from before 1970. Also, compare publisher names, and don't add an ISBN if the citation in an article is to a publisher other than the one mentioned in the database that includes the ISBN.
Relevant diffs/links: https://en.wikipedia.org/w/index.php?title=Upton_Sinclair&diff=next&oldid=1276363842
We can't proceed until: Feedback from maintainers

The citation was a mess already, it purports to be to the 1935 edition published by Upton Sinclair, yet included a google books url for a 2023 POD issue of a 1994 University of California Press edition. This is not an uncommon situation alas. DuncanHill (talk) 23:17, 24 February 2025 (UTC)

It's not uncommon, but citations that mix up metadata from multiple publications of the same material are a very frequent trigger for Citation bot to start mangling the citation even more. Checking the isbn vs publication year seems like a basic and helpful sanity check to hold off some of this mangling. —David Eppstein (talk) 23:26, 24 February 2025 (UTC)

I agree with what I think @David Eppstein: is getting at. If Citation Bot detects a contradiction and can't resolve it with a high degree of certainty, it should do nothing. Jc3s5h (talk) 03:43, 25 February 2025 (UTC)

Some history of the example {{cite book}} template.

original: added 15 December 2016 by Editor Billmckern at this edit:

{{cite book |last=Sinclair |first=Upton |date=1994 |title=I, Candidate for Governor: And How I Got Licked |url=https://books.google.com/books?id=OqqpXJy-fRwC&pg=PA109 |location=Berkeley, CA |publisher=University of California Press |page=109 |isbn=978-0-520-08197-0 |ref={{sfnRef|''I, Candidate for Governor''}}}}

changed: 13 February 2025 by Editor Gnuish at this edit:

{{cite book |last=Sinclair |first=Upton |date=1935 |title=I, Candidate for Governor: And How I Got Licked |url=https://books.google.com/books?id=OqqpXJy-fRwC |ol=OL114964W |location=Pasadena |publisher=Upton Sinclair}}

changed: 18 February 2025 by Citation bot and Editor CorrectionsJackal at this edit:

{{cite book |last=Sinclair |first=Upton |date=1935 |title=I, Candidate for Governor: And How I Got Licked |url=https://books.google.com/books?id=OqqpXJy-fRwC |ol=OL114964W |location=Pasadena |publisher=Upton Sinclair|isbn=978-0-520-91352-3 }}

In the original template, ISBN 978-0-520-08197-0 is the thirteen-digit form of one of the two ten-digit ISBNs listed on the reverse of the title page (ISBN 0-520-08197-8). The ISBN added by Citation bot (ISBN 978-0-520-91352-3) is not a thirteen-digit form of one of those two. The OL 114964W identifier added by Editor Gnuish links to a copy at Archive.org. It also suggests that there is a 1935 edition somewhere and links to an OCLC search that shows a 1935 edition but with a different publisher. What a mess.

Seems to me that the original template is the best and should not have been fiddled with; either manually or by bot. Citation bot would not have touched the template had it no been futzed with by a human (the bot edited Upton Sinclair several times during 2023).

—Trappist the monk (talk) 01:25, 26 February 2025 (UTC)

If you go to the title page of the book I cited, you'll find a 10 digit ISBN. If you run that ISBN through a converter, the 13 digit result is the one I included in the citation. If the 13 digit ISBN is wrong, that's news to me. I guess I'll never again assume that a converted ISBN is correct. Thanks for letting me know. Billmckern (talk) 01:34, 26 February 2025 (UTC)

@Billmckern: I did not say anywhere that the 13 digit ISBN is wrong. I merely noted that ISBN that you provided is the 13-digit form of a 10-digit ISBN listed on the reverse of the cited book's title page. The general advice, as I understand it, is to use 13-digit ISBNs when both are provided. See Wikipedia:ISBN § Types.

—Trappist the monk (talk) 01:47, 26 February 2025 (UTC)

Where did 978-0-520-91352-3 (aka 0-520-91352-3) came from? -- GreenC 02:42, 26 February 2025 (UTC)

Did you not understand what I wrote? I'm agreeing with you. I assumed that when I converted the 10 digit ISBN from the book's title page, the 13 digit result was correct. I checked tonight, after your first post, and the 13 digit result is in fact not correct. Now I know better than to assume that converting an ISBN from 10 to 13 digits will yield a correct result. I'll double check in the future before I add one to a citation. Billmckern (talk) 02:56, 26 February 2025 (UTC)

The problem in the original citation is that it claimed the book was written in 1994, when actually it was written in 1935. (I happened to know that because I've read the book in the original edition.) Once I fixed that, I had to fix the publisher, etc, to match. The Open Library helped me fix that metadata.

Is CitationBot looking back in the page history to find earlier revisions of a citation? I was assuming that it would only look at the current version. If it's looking backwards at something that a human manually changed, how does it decide whether it was changed because the old version was actually WRONG? Gnuish (talk) 06:46, 26 February 2025 (UTC)

I don't think it looks at the article history. But it does look at the urls and links in the metadata. When you changed the year to the original publication year 1935, you left in place both a Google Books url (pointing to a 2023 edition) and an Open Library ol= id (pointing to a 1994 edition). The bot might have restored the ISBN from either of those. —David Eppstein (talk) 07:15, 26 February 2025 (UTC)

The fault is not the URL. The bot should not be inserting ISBNs.. at all. Generating reliable ISBNs is way more complex. It requires fuzzy matching the Author and Publisher and Title and Date in a database of books, making a reliable match that way, then it will give you the correct ISBN, for that edition. It's non-trivial work. The problem is once the bot makes an error, it becomes difficult/impossible to figure it out due to ambiguities introduced. -- GreenC 07:49, 26 February 2025 (UTC)

The problem in the original citation is that it claimed the book was written in 1994 Not true. The original citation claimed (correctly) that the cited edition was published in 1994. There was no reason for you to change that; especially since |url= linked to a google books scan of the 1994 edition.

—Trappist the monk (talk) 14:52, 26 February 2025 (UTC)

I was just about to reply with the same comment, changing it from 1994 to 1935 was in error. Unfortunately this happens a lot with citations were someone changes the edition from the one that was used for verification to the first edition, usually causing issues with verification as they do. If the original publication dates must be included then that's why |orig-date= exists. -- LCU ActivelyDisinterested «@» °∆t° 18:30, 26 February 2025 (UTC)

Unless the associated page numbers are the same across both the 1935 and 1994 edition, which seems unlikely, changing the edition is a bad idea. -- LCU ActivelyDisinterested «@» °∆t° 18:35, 26 February 2025 (UTC)

Page 109 for edition published in 1935, and page 109 for edition published in 1994, open to different page content, making the citation effectively unverifiable. By changing/adding ISBNs without verifying the book edition (by comparing Author + Title + Publisher + Date from the citation vs. the ISBN metadata) it creates apparently unverifiable content. Better off skipping. -- GreenC 02:42, 26 February 2025 (UTC)

More free DOI prefixes

10.14231/AG - Algebraic Geometry
10.4171/DM - Documenta Mathematica
10.4171/MAG - EMS Magazine/EMS Newsletter

Headbomb {t · c · p · b} 21:47, 27 March 2025 (UTC)

{{fixed}} AManWithNoPlan (talk) 14:28, 8 April 2025 (UTC)