Wikipedia talk:WikiProject Check Wikipedia/ISBN errors

2014-06-06 Current List inaccurate
✅

The update that just occurred contains items which do not and have not existed in pages for some time. Example: I changed these nearly a week ago. I, of course, just went back and verified that my edits, at least for the first article, are still in the current page. &mdash; Makyen (talk) 11:20, 6 June 2014 (UTC)
 * 12 x ISBN 992738758 : Ballaban's campaign of 1465, Battle of Mokra (1445), Battle of Mokra (July 1462), Battle of Oranik (1448), Battle of Otonetë, Battle of Pollog, Battle of Torvioll, Siege of Krujë (1450), Siege of Svetigrad (1448), Skanderbeg's Italian expedition, Skanderbeg's Macedonian campaign, Vrana Konti
 * Hi Makyen, the update I did 2 hours ago was only for adding interwikis, not updating the list itself. My bot should be running the update right now (it finished on frwiki 20mn ago), so the list should be updated here in no more than a few hours (I finally scheduled my bot to update the list every friday). I'll check to see if the update is correct. --NicoV (Talk on frwiki) 11:33, 6 June 2014 (UTC)
 * Hi Makyen, the update by the bot has just been done, your edits have been taken into account. --NicoV (Talk on frwiki) 14:27, 6 June 2014 (UTC)

Simple cases which would be correct with the "978" prefix added
I wrote a little bit of JavaScript to highlight the simple cases where the ISBN would presumably be valid if a "978" prefix were added: I've just been executing this in my browser's console (hence the terminating ".length" so my console doesn't get spammed with the list of changed items that would otherwise be returned), but I can write a user-script version if people are interested. Keep in mind that not every case is simple: be sure to double-check against WorldCat or a similar database to make sure there aren't other errors! {&#123; Nihiltres &#124;talk&#124;edits}&#125; 17:04, 27 June 2014 (UTC)
 * There are quite a few that are relatively simple cases of the numbers being mixed up in predicable ways, but from which the correct ISBN can be derived. One class of those is where the ISBN has been converted from a 10-digit ISBN to a 13-digit ISBN, or the reverse, without changing the check digit. Many of these can be identified by calculating/verifying the check-digit against the alternate.
 * However, I viewed it as insufficient to just perform the check digit verification and make a change based only on that. With only a single check-digit there is a significant chance that it will verify while still being the incorrect ISBN. Thus, I view it as a requirement to actually verify that the correct book is returned when using the ISBN to which you are changing to look-up the book in at least one, if not more, online database. Without verifying that the proposed ISBN actually returns the correct book you have at least a 1/10 or 1/11 change of being wrong. On the other hand, it is not unreasonable to use that check as a beginning to identify some which might be easy to verify.
 * BTW: Your calculation is incorrect. You are mixing the 13-digit and 10-digit algorithms. The presence of an "X" guarantees that it is a 10 digit ISBN; 13-digit ISBNs do not permit "X". The 3^(index%2) calculation is for 13-digit check-digits, yet you are including "X" as a possible value.
 * If you want it, I have JS code for ISBN hyphenation based on the April 2014 ISBN Range Message from the International ISBN Agency. Given the relatively large data structure, it is a bit long. &mdash; Makyen (talk) 18:28, 27 June 2014 (UTC)
 * Absolutely, I'm not advocating making any edits based on the number alone. I've been verifying all my changes with WorldCat entries, and I mentioned that in my blurb above.
 * I included the X value mainly as error prevention, since the code that follows it assumes that each item in the array is a number. For that matter, entries flagged that include an X may have been converted incorrectly somehow from a 13-digit ISBN (and there's presumably some problem with them if they're listed). I never intended this to be foolproof: it's a quick and simple way of finding candidates to be looked over manually, so a few edge cases aren't worth worrying about.
 * I'd love to take a look at the hyphenation code. I've been using the LOC converter a bunch, and it might be nice to expand on that tool with better hyphenation support, better error handling (its -based errors are annoying), and maybe some functionality that added a few guesses for fixing invalid ISBNs based on simple cases we've seen. {&#123; Nihiltres &#124;talk&#124;edits}&#125; 19:41, 27 June 2014 (UTC)
 * Hi, if you're interested, WPCleaner already handles some basic errors in manual mode by suggesting to search similar ISBN in online databases: ISBN-10 with the check digit from the ISBN-13 (or the opposite), ISBN-13 with an error in the first 3 digits (987, ...), ... It also gives indications about hyphenation. NicoV (Talk on frwiki) 20:41, 30 June 2014 (UTC)

November 13 update removed invalid ISBNs that are still present in articles
✅

I'm not sure what happened, but the latest update to this page removed many more ISBNs than usual. I was suspicious, so I looked up this entry:


 * 2 x ISBN 0-684-19176-9 : 2 x Chambersburg Raid

The ISBN is still in the article, and it is still invalid.

Should the November 13 update be undone, or redone? I will undo it based on the above information unless there is some change to the purpose of this page of which I am unaware. Anyone? – Jonesey95 (talk) 14:22, 13 November 2014 (UTC)
 * Hi Jonesey95, I'm also surprised by today's update done by Sfan00 IMG. WikiCleanerBot is scheduled to update the page every friday, so a new update should be done in less that 24 hours. I suggest either to wait until then without doing anything, or to undo the last update and wait for tomorrow's update. --NicoV (Talk on frwiki) 14:54, 13 November 2014 (UTC)
 * I'm wondering if it's because there's only a few hundreds pages listed for ISBN errors at CheckWiki, I thought there was a lot more before. Maybe Bgwhite or Magioladitis have an idea... --NicoV (Talk on frwiki) 15:07, 13 November 2014 (UTC)
 * For the moment, my guess is that somehow the ISBN errors lists at CheckWiki have been drastically reduced since last friday. As WPCleaner only uses Category:Pages with ISBN errors and the CheckWiki lists to build this page, it would be consistent with the current result. I will try to modify WPCleaner before next bot update to take into account the existing links on the page, but I'm not sure I will have the time. --NicoV (Talk on frwiki) 17:21, 13 November 2014 (UTC)
 * Still not working correctly... Some work to do to understand what's going on. --NicoV (Talk on frwiki) 12:31, 14 November 2014 (UTC)
 * Fixed ! And it helped me find a bug due to the new way of dealing with "continue" in API requests. --NicoV (Talk on frwiki) 22:32, 14 November 2014 (UTC)

,, Labs database was having a little oops while the English Checkwiki monthly dump was processing. As a result, all errors resulted in 0. I also do the processing at home and have the results for all errors, but they are in text format. If you want them, I can give them to you. FYI... I run it at home because my 5-year old laptop is 3x faster than running on labs and the results are updated to User:Magioladitis/AWB and CHECKWIKI. ISBN errors aren't listed at that page, because Magioladitis and I don't fix them (too many). Bgwhite (talk) 21:22, 13 November 2014 (UTC)
 * , thanks for the explanation and the offered help. I managed to modify WPCleaner so that it won't be disturbed by emptied lists in CheckWiki. --NicoV (Talk on frwiki) 22:32, 14 November 2014 (UTC)

Updates in the next months
✅

Hi, I think I messed up the configuration of my NAS just before going away travelling, so WikiCleanerBot is currently not updating the list and won't be updating it in the next 2 months at least. If someone wants to update it in the mean time, you can do it using WPCleaner. --NicoV (Talk on frwiki) 05:47, 17 December 2014 (UTC)
 * Automatic updates are working again, WikiCleanerBot should update the list every friday. --NicoV (Talk on frwiki) 15:55, 26 February 2015 (UTC)

Stop sabotaging the principal memes
I am engaged in editing a far more serious change than some quibbles on ISBNs, but this blasted bot decided to replace the link to a meme with one to here. Consequently some meme will end up with an orphaned link as you have no index on this page to the meme's you've dumped on. Well done! If you have concerns, add theme to the meme or Talk page, don't blank the enitre damned thing. This is very poor programming, 2/10. — Preceding unsigned comment added by 176.25.142.238 (talk) 17:07, 20 April 2015 (UTC)
 * , please post a link to the article, or the specific edit, with which you have a concern. – Jonesey95 (talk) 17:10, 20 April 2015 (UTC)
 * Doesn't seem to be related at all to this page, and WikiCleanerBot only updates this page (which is designed to be updated by bot, not by manual edits), not others, when checking for ISBN with errors on enwiki. --NicoV (Talk on frwiki) 08:19, 21 July 2015 (UTC)

Should this list ignore true and other non-errors?
✅

See L and M-class destroyer. true is used to indicate an invalid ISBN that appears in a work and is used by databases (e.g. Worldcat) to track the book. Should this list ignore invalid ISBNs marked with this parameter?

Also, the list shows non-errors like "ISBN 0906440351 96p." from James A. Mackay. The ISBN magic word rendering works fine, the ISBN is valid, and the link works, so this list should probably not report this formatting as an error.

Also see Matari and Papermaking for formats that render and link just fine but are reported as errors.

And Carry On Breathing, which has some invalid ISBN formatting wrapped in notatypo. – Jonesey95 (talk) 14:20, 24 May 2015 (UTC)

Jonesey95 we can easily fix cases like "ISBN 0906440351 96p.". We have some template solutions for the rest too. -- Magioladitis (talk) 14:47, 24 May 2015 (UTC)
 * For instance, the comma you used is a good solution. These books needs rework anyway. There is an inconsistency in the way they are presented. Thanks for starting working on them. -- Magioladitis (talk) 14:49, 24 May 2015 (UTC)

The list should definitely not ignore those cases. I've easily corrected a number of them. That said, it might be worthwhile if the status of "marked as not an error" were to show up in the list, because although they're worth checking, some are probably truly unfixable, and others difficult —it'd be worthwhile to deprioritize them. {&#123; Nihiltres &#124;talk&#124;edits}&#125; 18:01, 24 May 2015 (UTC)

Just an information: if you manually remove the pages with true from the list, WPCleaner won't add them back unless there's another ISBN problem in the page. That's because WPCleaner only scans pages which are either already listed here, detected by CW in errors #70 - #73, or categorized as having an invalid ISBN. --NicoV (Talk on frwiki) 09:11, 15 July 2015 (UTC)
 * Just to mark it as resolved. WPCleaner takes into account true if configured in  in User:NicoV/WikiCleanerConfiguration. --NicoV (Talk on frwiki) 08:22, 15 June 2020 (UTC)

Ignore ISBNs in pre tags and inside square brackets?
Damm algorithm is appearing in this list, even though the errors listed appear in the article inside square brackets. Should we ignore these valid wikilinks? I think we should.

Also, Extensible Resource Identifier is appearing in this list even though the ISBNs in question are inside  tags. Should we ignore this text that is not intended to be wikified? I think we should.

It looks like the check correctly ignores text inside  tags, so I have used those in a few cases to eliminate some false positives. – Jonesey95 (talk) 17:52, 8 June 2016 (UTC)
 * Jonesey95 I have added the  tags to ignore ISBN. I will see when I find time for Damn algorithm. There's always the possibility to add an article to the whitelists. --NicoV (Talk on frwiki) 20:26, 8 June 2016 (UTC)

A doubt whether it's incorrect
✅

I would like to know whether the ISBN 1-904994-10-5 is invalid or valid according to wikipedia rules. I have seen it under this page but while searching it in isbnsearch.org, it showed the name of book.Adithyak1997 (talk) 14:05, 25 January 2019 (UTC)
 * It is a valid ISBN. Sometimes valid ISBNs show up on this page because the template that is enclosing them is broken somehow. In this specific case, someone inserted text in the middle of a citation template, breaking the citation and putting isbn = 1-904994-10-5 in plain text in the article. I fixed the article. – Jonesey95 (talk) 14:22, 25 January 2019 (UTC)
 * Thanks for replying.Adithyak1997 (talk) 14:34, 25 January 2019 (UTC)

Why updating this page was returning a blank page
✅

In the past week, we had a problem where updating the page using WPCleaner was resulting in a blank page, and an Information screen in WPCleaner showing all of the output that should have gone on the page. Attempting to save the output to the page manually resulted in a URL blacklist error. The resolution was to fix the article in question so that the blacklisted URL was removed, or at least so that the article did not contain an ISBN error that would get it listed on this page. – Jonesey95 (talk) 19:38, 1 May 2019 (UTC)
 * I don't think it was returning a blank page, it was just not updating the page anymore. The blanking of the page was due to vandalism on WPCleaner's configuration, the problem with the page not updating appeared just after. --NicoV (Talk on frwiki) 20:13, 1 May 2019 (UTC)

New nowiki self-closed tag "error" detected
✅

I have been away for a few months, so please excuse this delayed report. Something appears to have changed in the detection code between 11 June 2019 and 24 June 2019, such that the construction "ISBN &lt;nowiki /&gt; 0-313-33340-8" is detected as an error. I have been inserting these nowiki tags into citation templates for many years in order to resolve false positives in this report.

A typical citation of a book review might look like this:

{cite journal|last=Dale|first=Pete|title=Punk Record Labels and the Struggle for Autonomy: The Emergence of DIY. By Alan O'Connor. Plymouth: Lexington, 2008. 145 pp. ISBN 0-7391-2660-1 Encyclopedia of Punk Music and Culture. By Brian Cogan. Westport, CT and London: Greenwood Press, 2006. 282 pp. ISBN 0-313-33340-8|journal=Popular Music|year=2010|volume=29|pages=176–178|doi=10.1017/S0261143009990638}

Note that there are two ISBNs in the title parameter. They should be processed as plain text, so inserting a self-closing nowiki tag between "ISBN" and the number used to work fine to remove the page from this report. Something appears to have changed, resulting in all of these articles appearing in the report.

Is there a better way to allow these valid, unlinked plain-text ISBNs to exist in the title parameter without causing any reports to flag these pages?

Before you ask, wrapping the full ISBN label and number in nowiki tags is not advisable, since there is at least one Visual Editor bug that results in ISBNs being wrapped in nowiki tags, and we need to be able to fix those. – Jonesey95 (talk) 21:47, 4 September 2019 (UTC)
 * Hi Jonesey95. Why do they need to be in plain text? Is it not possible to use ISBN inside the title parameter? --NicoV (Talk on frwiki) 08:10, 5 September 2019 (UTC)
 * That is correct. It causes a "stripmarker" error, as shown below:
 * In any event, we need to have a way to provide ISBNs without links, and now that wrapping them in  and inserting a self-closed are both flagged as errors by this page, I am asking for a better way to do that. – Jonesey95 (talk) 14:52, 5 September 2019 (UTC)
 * Hi Jonesey95. I think the question could also be asked to the designers of the Cite journal and other CS1 templates, to see if there are ways to find a solution with how the title parameter is parsed.
 * Otherwise a dedicated template can be used:  . The benefit is that this template can also be included in the template that are verified by WPC to check for incorrect ISBN.
 * --NicoV (Talk on frwiki) 15:58, 5 September 2019 (UTC)
 * I wondered if using Text works without negative side effects:
 * The citation metadata looks unharmed:
 * It renders a magic link, but bots won't try to fix it. I'll change a few articles to use it and see if the report still flags them. – Jonesey95 (talk) 23:03, 29 November 2019 (UTC)
 * Follow-up: It seems to work with no side effects. I replaced  with  . There is still a magic link in the title parameter, and the page still shows up in the magic link category, but magic links are supposed to go away at some point. Eventually. – Jonesey95 (talk) 16:43, 30 November 2019 (UTC)
 * --NicoV (Talk on frwiki) 15:58, 5 September 2019 (UTC)
 * I wondered if using Text works without negative side effects:
 * The citation metadata looks unharmed:
 * It renders a magic link, but bots won't try to fix it. I'll change a few articles to use it and see if the report still flags them. – Jonesey95 (talk) 23:03, 29 November 2019 (UTC)
 * Follow-up: It seems to work with no side effects. I replaced  with  . There is still a magic link in the title parameter, and the page still shows up in the magic link category, but magic links are supposed to go away at some point. Eventually. – Jonesey95 (talk) 16:43, 30 November 2019 (UTC)
 * The citation metadata looks unharmed:
 * It renders a magic link, but bots won't try to fix it. I'll change a few articles to use it and see if the report still flags them. – Jonesey95 (talk) 23:03, 29 November 2019 (UTC)
 * Follow-up: It seems to work with no side effects. I replaced  with  . There is still a magic link in the title parameter, and the page still shows up in the magic link category, but magic links are supposed to go away at some point. Eventually. – Jonesey95 (talk) 16:43, 30 November 2019 (UTC)
 * Follow-up: It seems to work with no side effects. I replaced  with  . There is still a magic link in the title parameter, and the page still shows up in the magic link category, but magic links are supposed to go away at some point. Eventually. – Jonesey95 (talk) 16:43, 30 November 2019 (UTC)

Possible false positive for ISBNs beginning with 9789999, 978-9999, 9796500
There are a set of ISBNs appearing on this report that begin with 9789999, 978-9999, or 9796500. They appear to be valid, and the articles in which they appear do not seem to have any syntax errors that would cause the articles to appear in this report. When you click through the ISBN from the article to Special:BookSources, the page does not show an error message, and the cite templates do not flag them as erroneous. This leads me to hypothesize that there is something wrong with WCW's validation code that incorrectly marks these ISBNs as errors. – Jonesey95 (talk) 17:55, 28 January 2020 (UTC)
 * The same thing is happening with 9798184241890 in Champin. – Jonesey95 (talk) 18:02, 28 January 2020 (UTC)
 * Hi Jonesey95. This is probably due to ISBN valid ranges : only a limited set of prefixes are valid for ISBN. You can get the current list at the International ISBN Agency. I've just updated WPCleaner with the latest list, does it change anything? But currently, nothing has been assigned for 978-9999 (the last with 978 is 978-9983 is for Peru, others above are not yet assigned), or for 979-6500 (970-12 is for Italy, 979-8 if for the US, but nothing between has been assigned), so it seems normal to me that they are being reported as errors. --NicoV (Talk on frwiki) 06:59, 29 January 2020 (UTC)
 * Hi Jonesey95. One possibility would be to put this kind of problems in another error number, if some wikis are not interested in fixing them. --NicoV (Talk on frwiki) 12:45, 31 January 2020 (UTC)

Should "Spezial:ISBN Suche" be detected?
✅

At this writing, there are 26 articles containing links that start with. They should probably be detected by this database analysis. – Jonesey95 (talk) 18:26, 13 June 2020 (UTC)
 * Thanks Jonesey95. I was already detecting  (with the dash), I'm adding   to the list, but I don't where it's coming from (it's not an alias for BookSources). --NicoV (Talk on frwiki) 19:24, 13 June 2020 (UTC)
 * Thanks. Some of them may be very old. It's possible that there was once a redirect. I fixed all of the existing ones that I could find; if new ones pop up, we can track them. – Jonesey95 (talk) 20:06, 13 June 2020 (UTC)

Invalid ISBNs which are listed in the book itself?
Does this effort account for - and, for that matter, would it even be worth accounting for - cases where an invalid ISBN is recorded in the book itself, and so people looking for the book may be searching for that invalid ISBN?

I have in front of me a copy of the 1990 version of Bill Bryson's "The Mother Tongue: English & How It Got That Way" (https://catalog.loc.gov/vwebv/holdingsInfo?bibId=16146235). Per the Library of Congress entry, the ISBN for this book is 0-688-07895-8, but the book's own information page reports the invalid ISBN 0-888-07895-8. (I'm guessing this is a simple misprint.)

It took me a fair bit of digging to track down the exact volume and find the correct ISBN, and being able to find the correct volume (and even identify the correct ISBN!) by searching for the incorrect ISBN could have been useful. While I have found no indication that this particular book is covered by Wikipedia in a way that would record the ISBN, the same principle could easily apply for other titles.

With such searches in mind, I would consider it unfortunate if an incorrect ISBN of this nature had once been recorded on Wikipedia, but had been removed out of an effort to clean up invalid ISBNs. Inverseparadox (talk) 15:19, 11 February 2022 (UTC)
 * See Listed Invalid ISBN. – Jonesey95 (talk) 17:25, 11 February 2022 (UTC)

ISBN flagged as an error when used correctly inside #invoke:cite book
✅

I don't know if this is a problem with CheckWiki or with Module:Cite book, but valid ISBNs are being flagged as errors when they are used in invocations of that module. See Bibliography of World War II for many examples. – Jonesey95 (talk) 03:51, 14 December 2022 (UTC)
 * Hi Jonesey95. It's simply that WPCleaner doesn't handle ISBN in functions : their use is usually discouraged on encyclopedic namespace, templates are usually favored. I will try a simple fix of ignoring detection in functions, but it may mask actual errors. --NicoV (Talk on frwiki) 19:15, 14 December 2022 (UTC)
 * Hi Jonesey95. I've ignored ISBN in #invoke functions, and did the same for ISSN. New release of WPCleaner will be available soon. --NicoV (Talk on frwiki) 21:06, 14 December 2022 (UTC)
 * Thanks. – Jonesey95 (talk) 21:11, 14 December 2022 (UTC)

ISBN as part of DOI
Is it possible to suppress flagging ISBNs that are embedded as part of a DOI? For example: Sabine Chaouche. Thank you. 76.14.122.5 (talk) 20:28, 1 December 2023 (UTC)