Wikipedia talk:ISBN

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Broken bookseller links at Special:BookSources[edit]

If you click on Special:BookSources/9781847242532#Booksellers (ISBN 9781847242532 is The Girl with the Dragon Tattoo), the generated links for at least the following two are broken: (talk) 08:54, 26 January 2012 (UTC)Reply[reply]

Borders Books has been taken over by Barnes & Noble, so I deleted this option. Strand Book Store should now be OK. --Robert.Allen (talk) 19:28, 14 February 2012 (UTC)Reply[reply]

WikiProject Unique Identifiers[edit]

I have created WikiProject Unique Identifiers for discussion and coordination of all UID related matters. Please join! Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:05, 14 February 2012 (UTC)Reply[reply]

Hyphenation: is it really meaningful?[edit]

Hi, I wonder whether it is really the best option to include the hyphenation. There are several reasons why to avoid it:

  • Wikipedia talk:ISBN/Archive 1#Current guidelines for ISBNs badly hyphenated at source? - it may be badly formed in the source itself and therefore wrongly written in the article.
  • Regarding the libraries, the Special:BookSources page parses the number correctly in any case.
  • Search engines, like Google, don't find books if the number searched for is copied from an article hyphenated but written unhyphenated in the indexed pages.
  • Google Books, which are increasingly used as a source of information and references, cite the numbers unhyphenated. It is much easier to copy the number without having to know exactly how ISBN have to be hyphenated, and it is much more important to include the number in any case, than to include it in a specific format. Editors will probably be more willing to add it if the format is not important, and in any case, the error of wrong hyphenation would be avoided.

The best option is to include the number unformatted to the article. The hyphenation, if necessary, should be part of the user preferences, and the software should parse the number, to avoid errors.

--Eleassar my talk 14:00, 13 March 2012 (UTC)Reply[reply]

I favor the unhyphenated forms for many of the same reasons. However, it would also be possible to modify the Wikisoftware to incorporate the unhyphenated ISBN-13s and the corresponding unhyphenated ISBN-10s as hidden text that would be found by search engines like Google. This might be a workaround which would allow hyphenation in the visible numbers, which could be either ISBN-10 or ISBN-13 format, and still provide both as targets for web search engines. (This would be automatic and would not appear in the edit window or be visible on the page. And while we're at it, we should also make the hyphenated forms nowrap.) --Robert.Allen (talk) 08:47, 14 March 2012 (UTC)Reply[reply]
This proposal does not solve many of the problems listed above. In addition, the hyphenated numbers are also copied from Wikipedia and inserted into Google search. --Eleassar my talk 18:10, 14 March 2012 (UTC)Reply[reply]
It's pretty easy to remove the hyphens for searching, just by clicking the link, and bots can add the hyphens, which is probably preferable to having editors do it, since the bot is more accurate. --Robert.Allen (talk) 18:51, 14 March 2012 (UTC)Reply[reply]
This (clicking the link) is not self-evident. I agree about bots inserting hyphens if necessary. To be exact, there already are bots doing it (e.g. User:Helpful Pixie Bot). --Eleassar my talk 18:53, 14 March 2012 (UTC)Reply[reply]
I don't believe we have ever had a consensus that the hyphens should be added. I think these bots should hold off until we have solved some of these problems and reached some consensus on it. I especially don't think they should be added until we solve the wrapping problem. --Robert.Allen (talk) 08:50, 15 March 2012 (UTC)Reply[reply]
Hyphenation is important for four reasons:
  1. It increases readability
  2. it decreases transcription errors from Wikipedia.
  3. It is the approved method of display, according to ISO, The International ISBN Agency and Wikipedia
  4. It conveys information about the book in human accessible form
Wikipedia is written primarily form humans, who jot stuff down on paper, are good with short runs of digits, but not with longer runs. ISBNs were designed to be human friendly, error detecting (though the 13 digit extension weakened that considerably), and conservative of the namespace (much like CIDR). We should stick to the ISBN standard. Rich Farmbrough, 11:54, 15 March 2012 (UTC).Reply[reply]
Please explain what information the hyphens convey. Who has shown that they increase readability? (After all, I just love to read ISBNs.) What do you mean by transcription errors, when its' actually easier to copy and past the unhyphenated form. I have never seen a consensus on the Wikipedia that this is the best format. The ISO "standard" was created long before the web was. Most of the web has chosen to ignore it. So how can the hyphens be considered to be "standard", when most websites ignore them completely? It should be described a nonstandard standard. --Robert.Allen (talk) 09:30, 17 March 2012 (UTC)Reply[reply]
The first segment (excluding any 978) is the language area, the second the publisher. Moreover you can get some idea of the relative size and age of the publisher. Remembering runs of digits is a standard psychological test on short term memory, dyslexics can generally remember shorter runs than mathematicians, but everyone is better at short runs than long runs. Cut and paste is easy in either form, some people do still use pens and pencils, or several devices, where transcription is needed. Most websites abuse apostrophes, spelling, grammar, HTML, and common sense. We try not to do that either. Nonetheless - Google hits for "1565920422" 2,700, Google hits for "1-56592-042-2" 9,000.
Moreover "1-56592-042-2 wikipedia" hits Wikipedia successfully as does "0596001320 wikipedia". Rich Farmbrough, 01:52, 20 March 2012 (UTC).Reply[reply]
It may depend on who is picking the ISBN. I ran a similar search with a different ISBN: 9780307269997 About 75,300 results [1] and "978-0-307-26999-7" About 2,990 results [2]. But it seems likely that in each of these searches, only a subset of pages with the relevant ISBN are getting picked up, because of all the variations in formatting. Providing four targets for each ISBN (ISBN-10, ISBN-13, with hyphens and without) would certainly improve the odds of finding all pages that cite a particular ISBN. --Robert.Allen (talk) 02:00, 21 March 2012 (UTC)Reply[reply]
ISBN-13s with dashes should be preferred. Doing it by bot is goodness; MediaWiki coercing this during a page save would be even better. Alarbus (talk) 10:02, 20 March 2012 (UTC)Reply[reply]
Books originally published with an ISBN-10 should be referenced by their ISBN-10. ISBN-13 only adds the 978 prefix, but changes the check digit, making searching for the book more difficult. Systems handling ISBN number can often convert between the two formats transparently, but when doing a plain text search (e.g. in digital or OCR'ed books) you have to search by the (often, but unfortunately not always, correctly hyphenated) ISBN-10. —Ruud 00:08, 21 March 2012 (UTC)Reply[reply]
Old, old comment, but ISBNs do not use dashes. (They will produce an error message.) Instead hyphens are acceptable. – S. Rich (talk) 02:24, 18 April 2023 (UTC)Reply[reply]
As someone who ran the bookstore of a university department for a year, I can attest that hyphenation of ISBNs is useful. I could often instantly recognize the publisher and sometimes title of a book by its (hyphenated!) ISBN. As noted above the hyphenation also adds some additional redundancy to the ISBN (an information-theoretic proof is left as an exercise to the reader.) There seems to be a common misconception that the hyphenation of ISBNs is done arbitrarily, but this is not the case. —Ruud 00:07, 21 March 2012 (UTC)Reply[reply]

Firefox Addon[edit]

Hi there,

I don't mean to self-promote here, but I have completed work on an add-on of some relevance to this page to allow, via right-click, the user to extract an ISBN out of text or links (as one may find often in references on Wikipedia) and go directly to (instead of first needing to go to Wikipedia's own page). Firefox is not permitting my add-on full release status yet because they say:

"It is unclear to us at this time if your add-on will be useful for a general audience so that it warrants public listing. We're granting you preliminary review only, but encourage you to promote your listing so that it garners more active users and user reviews. You may re-nominate your add-on once there is more proof of its usefulness."

If one is doing research on Wikipedia and they want to be able to quickly obtain books (yes, the addon currently only allows Amazon requiring a purchase), they have to go through an intermediate page each time, while my addon lets them skip that step.

So, would anyone kindly give me a quick review?

After (restartless) installation, you can just go to any page (whether on Wikipedia or not) with ISBN references and right-click the link (or highlighted text).

If you are interested in seeing the source code (which is quite simple), it is at .

(I had been hoping for a more browser-independent way that ISBN's could be flagged in markup to the browser--see , but that bug was rejected due to a lack of browser support--and still waiting on Firefox: .)

Thanks very much!

--Brettz9 (talk) 17:18, 31 December 2012 (UTC)Reply[reply]

ISBN Searching[edit]

I probably am not understanding something but it seems I can't quite fathom a few use cases based on the way ISBNs are current held. If I know the non-hyphenated version of an ISBN for the book I'm interested that may have been cited on wikipedia somewhere. How would I go about finding any references given that the ISBNs are recorded here only in hyphenated form?

1. via wikipedia search?

2. via google?

If I'm correct in saying this isn't covered, is there a case for either smartening up the search function a little, or actually changing the way the isbn is stored to include both hyphenated and unhyphenated versions in the markup? I'd be interested in knowing some thoughts, I'm probably missing something really obvious and simple so sorry in advance. ---- nonsense ferret 00:53, 1 February 2013 (UTC)Reply[reply]

In Wikipedia, ISBNs may be entered either hyphenated, spaced, or unpunctuated. Thus, ISBN 0-12-345678-9 ISBN 0 12 345678 9 and ISBN 0123456789 are exactly the same. If you click on any of these, you'll be taken to the same search page. --Redrose64 (talk) 01:12, 1 February 2013 (UTC)Reply[reply]
Sorry for asking a stupid question from stupidland, but is the point of that search page to search things outside of WP or search for all instances of that number within WP? ---- nonsense ferret 01:18, 1 February 2013 (UTC)Reply[reply]
so yeah, I'm probably doing something wrong - I click a number like ISBN 0-553-26315-3 and get to the search page, and then i click on section 11 find on wikipedia, and I'd sort of expect that to find the reference that I'd clicked on but it doesn't seem to when I tried. I guess unless I'm doing something stupid, the link is stored on the page hyphenated then it is converted to unhyphenated on the search page, thus depriving me of the ability to search for it in wikipedia ---- nonsense ferret 01:25, 1 February 2013 (UTC)Reply[reply]
The database used for that isn't updated in real time. If you added the ISBN to a page recently (say, within the last 24 hours), the search indexing spiders may not have updated the database, so it might not show. --Redrose64 (talk) 01:29, 1 February 2013 (UTC)Reply[reply]
Well I used an ISBN that was already there on the page for ages - it seems like you can't type in a non-hyphenated isbn and find it on wikipedia now that they all have been changed to the hyphenated versions. ---- nonsense ferret 01:31, 1 February 2013 (UTC)Reply[reply]

is there a problem?[edit]

maybe not much of one, but at least it seems like there is something a bit inconsistent about the current behaviour:

All ISBNs on pages have been converted to hyp version. I click on hyp version on an article I get sent to search page prepopulated with unhyphenated version. I click link to search wikipedia for all instances of the ISBN (using the prepopulated unhyp version) Expected: find the link on the page I originally clicked Actual: find no results - because there is a mismatch between hyp and unhyp.

Similarly, find a book on amazon which lists unhyp ISBNs, go to WP and search - Expected: find all references to that book where people have bothered to type in the ISBN Actual: find no references, ever, since they have all been converted by a bot.


  • I managed to use the search function correctly (BIG assumption)
  • A bot has been set to convert all ISBNs to hyp version (untested impression)
  • External sources tend to use unhyp versions of ISBN
  • It is difficult for a use to manually hyphenate an ISBN correctly and consistently for the purposes of searching
Not all articles have hyphenated ISBNs - see Rock 'n' Roll (John Lennon album) for one example of unhyphenated ISBNs. GoingBatty (talk) 19:20, 3 February 2013 (UTC)Reply[reply]

what would a solution look like?[edit]

Two obvious approaches:

  • smarten up the search function to hyphenate ISBNs for the user
  • enhance the way ISBNs are stored with some markup which will allow them to be searched in hyphenated or unhyphenated forms and apply consistency across the database.

Authors and titles?[edit]

So, when you type in an ISBN on Wikipedia's lookup system, there are two huge problems with it:

  • No link to the article if there is one
  • Author and title not listed

I get the first one - I can imagine that might be a technically challenging thing to implement, or lots and lots of labor. But the second? Really? When I type in a really long number, my first question is "did I type the right number". You're not answering this basic question here. Ego White Tray (talk) 13:00, 7 June 2013 (UTC)Reply[reply]

Are you referring to Special:BookSources? --  Gadget850 talk 13:28, 7 June 2013 (UTC)Reply[reply]
For which, see Wikipedia talk:Book sources for the discussion location. –Quiddity (talk) 18:42, 7 June 2013 (UTC)Reply[reply]

Check digits differ?[edit]

 Fixed We imply on the main page, and in some remarks above, that the check digit for ISBN-10 and ISBN-13 differs. Is that always true of valid numbers, so that coincidence of check digits represents an error by publisher or printer?

According to the official ISBN converter, this pair differ only by the prefix '978-':

  • 978-0-7134-8799-2 ; 0-7134-8799-2

I have supposed that the converter calculates check digits but wonder whether it refers to a database that incorporates publisher/printer errors that have been accepted.

--P64 (talk) 18:27, 25 May 2014 (UTC)Reply[reply]

They don't necessarily differ: where they are similar, it's a coincidence, and the chances of that happening aren't particularly small. The ISBN-10 system has 11 possible check digits (0-9 and X), so there is a 1 in 11 chance that a given check digit (for example, the same check digit that ISBN-13 yields) will be yielded under ISBN-10. Therefore I would expect that the chance that both ISBN-10 and ISBN-13 yield similar check digits would be 1/11 or 0.0909.
Example: ISBN 0-7110-0639-3 / ISBN 978-0-7110-0639-3 but not ISBN 0-7110-0638-5 / ISBN 978-0-7110-0638-6 or ISBN 0-7110-0640-7 / ISBN 978-0-7110-0640-9 --Redrose64 (talk) 10:21, 26 May 2014 (UTC)Reply[reply]
I reworded the "Types" section to hopefully deal with that ambiguity.
P64, none of converters that I'm aware of include a database of wrong ISBNs. The converters simply strip off the existing check digit, adds or removes the "978-" prefix, and recomputes the check digit using the ISBN-10 or ISBN-13 algorithm. Most often when a publisher makes a mistake it uses an ISBN that is a duplicate of one used on an earlier book. It's rare though has happened that a publisher has incorrectly computed an ISBN's check digit. Prior to the use of ISBNs paperback books, and a few hardcover books, often included a catalog number such as "F3838". Many book databases have a field for the catalog number which can be also used to hold a publisher's wrong or incorrectly computed ISBN. Thus, if a book record has a value in the ISBN field it most likely is a valid ISBN for that book. If the record has a number that looks like an ISBN in the catalog number field than most likely it's because that ISBN is either not the right one or has the wrong check-digit. --Marc Kupper|talk 18:41, 16 September 2015 (UTC)Reply[reply]

update available templates[edit]

I noticed someone had used {{Rowledge-Engines of the LMS}} which renders as:

Rowledge, J.W.P. (1975). Engines of the LMS built 1923–51. Oxford: Oxford Publishing Company. ISBN 0-902888-59-5.

Not sure if this is a standard method.

I added to Wikipedia:ISBN#Linking linking common templates - please review etc.Prof.Haddock (talk) 13:22, 5 June 2014 (UTC)Reply[reply]

That template is in Category:Rail transport book citation templates; see the text at the top of that category. --Redrose64 (talk) 14:57, 5 June 2014 (UTC)Reply[reply]

Misprinted ISBN[edit]

SusunW came about a rather bizarre case: According to Google Books this book misprints its own ISBN (correct is ISBN 978-0-8061-1758-4, as confirmed by WorldCat, not ISBN 978-0-8061-1458-4). Any chances of implementing a "redirect" so that entering the ISBN as printed in the book (assuming this isn't a digitalization error) will point to the correct ISBN? Huon (talk) 13:45, 31 January 2015 (UTC)Reply[reply]

No, because the links to apparent pages like Special:BookSources/9780806114584, like many other "pages" in Special: namespace, don't really exist; so, even if we could create a redirect (which we can't), it would be ignored. There is a note somewhere that ISBNs are occasionally misprinted, and that the true ISBN should be used and not the one that the book shows. --Redrose64 (talk) 15:09, 31 January 2015 (UTC)Reply[reply]
Thanks to you both. Guess it is a good thing that Wiki notifies one that there is an error. As clearly one would assume that in most instances the printed source correctly identified its own ISBN. SusunW (talk) 15:35, 31 January 2015 (UTC)Reply[reply]

ISBN font size[edit]

I noticed the ISBN is knocked down a size and put in parentheses on the French Wikipedia. I think it looks way better. Look at the Bibliography here: fr:Rammstein#Bibliographie vs Rammstein#Further_reading. Could we change it here? Thoughts appreciated. МандичкаYO 😜 03:44, 30 June 2015 (UTC)Reply[reply]

@Wikimandia: On French Wikipedia, that's only the case when the ISBN is wrapped in the appropriate template (i.e. {{ISBN|0-7119-7220-6}}); an ISBN produced using the magic link (i.e. ISBN 0-7119-7220-6) is in normal font size. See for example fr:Sandha Bhasa, ref. 5. The template concerned is fr:Modèle:ISBN which invokes fr:Module:Biblio; English Wikipedia does not have such a template. --Redrose64 (talk) 08:02, 30 June 2015 (UTC)Reply[reply]
Still, we must have some ISBN template because it automatically hyperlinks ISBNs (ISBN 0-7119-7220-6). Couldn't we make it happen? I just think it looks so much better. Thanks. МандичкаYO 😜 08:08, 30 June 2015 (UTC)Reply[reply]
That's not a template (you can tell because you didn't need to use double braces), it's part of the MediaWiki software. See Help:Magic links --Redrose64 (talk) 08:20, 30 June 2015 (UTC)Reply[reply]

OK to add my own tool?[edit]

Does anyone object to me adding the ISBN tool I wrote to this page? It converts ISBNs, hyphenates them if valid, and offers descriptive error messages and suggestions (with links to check against WorldCat) if invalid. IMHO it is superior to both the LOC and tools. {{Nihiltres |talk |edits}} 18:55, 17 February 2016 (UTC)Reply[reply]

No objections after a week; I'll go ahead and add it. {{Nihiltres |talk |edits}} 19:59, 24 February 2016 (UTC)Reply[reply]

Research study on the prevalence and validity of ISBNs in book citations[edit]

There is a recent academic paper that may be of interest here - it examined a sample of book citations to check how many had ISBNs, and how many of those were valid ISBNs. I have written a review for the Wikimedia Research Newsletter aka the Signpost's "recent research" series: "Test of 300k citations: how verifiable is 'verifiable' in practice?" (see also the talk page discussion there).

Regards, Tbayer (WMF) (talk) 21:15, 27 March 2016 (UTC)Reply[reply]

Future of magic links[edit]

Are people here aware that a decision has apparently been made to remove all magic links, which includes ISBN and PMID?

I can't find any discussions where consensus was established. There is Requests for comment/Future of magic links on Mediawiki. Five people have taken part, and it isn't what we call an RfC. There was a mailing list post by Legoktm in October; a thread on phabricator, T148274; and there is a request to approve a bot to remove them. Spinningspark left a note about the bot request on WP:VPP.

Leaving this here in case anyone is interested, and pinging Doc James because it affects PMID. SarahSV (talk) 16:01, 22 December 2016 (UTC)Reply[reply]

Was discussed here at WTMED[3] a little while ago. Have not used these links myself. Others however do. Not sure what the solution is. Doc James (talk · contribs · email) 17:04, 22 December 2016 (UTC)Reply[reply]
Editors who don't use citation templates use the ISBN and PMID magic links. If people object, the best thing is to leave a comment at WP:VPP#Removal_of_ISBN_magic_links and/or at the request for bot approval to remove them, Wikipedia:Bots/Requests for approval/Yobot 27. SarahSV (talk) 17:24, 22 December 2016 (UTC)Reply[reply]

ISBNs in mw:Citoid[edit]

Hey All; at the WMF we have been working with OCLC to make ISBN generated citations available, through using their ISBN database. We have deployed the feature on all language Wikipedias: you can learn more about it on the Wikimedia blog. Astinson (WMF) (talk) 19:29, 11 May 2017 (UTC)Reply[reply]


Hello All, has any thought been given to the new trend of texts as ebooks online with eIBSN references, please? William Harris • (talk) • 01:31, 19 January 2018 (UTC)Reply[reply]

@William Harris: Yes, many times at Template talk:Cite book and elsewhere. Most of the discussions will now be in the archives. In brief: use |isbn=. --Redrose64 🌹 (talk) 17:03, 19 January 2018 (UTC)Reply[reply]
Many thanks; this area is all very new to me. Regards, William Harris • (talk) • 19:51, 19 January 2018 (UTC)Reply[reply]

Magic links[edit]

With magic links deprecated, and with no plans to bring them back that I am aware of, would anyone object to the removal of information on magic links on the page? It takes up a considerable amount of text for something that is not used, and makes it hard to find the information you are looking for. Kees08 (Talk) 02:56, 1 February 2019 (UTC)Reply[reply]

I think that we should leave the magic link information here until they are actually disabled, if they ever are. It's been over two years since magic links were deprecated, and en.WP has been ready to disable them for a long time, but we haven't seen any movement. I would rather wait until it actually happens, if it ever does.
The "Linking" section should have subsections on the {{ISBN}} and {{ISBNT}} templates, however. – Jonesey95 (talk) 05:43, 1 February 2019 (UTC)Reply[reply]

Invalid ISBN because x is not X[edit]

Came across a reference using the cite book template which threw up an ISBN error because the terminal X was given as lower-case x. The book source had its ISBN written with lower case X and as it was in a printed book was obviously valid to the untrained human.

Is this a case where the error message could be more explicit when saying that the ISBN contains an invalid character as to what the invalid character is? Or even treat lower-case x as upper-case X? Or the various documentation could indicate that ISBNs have to be given with capital X to avoid the problem. GraemeLeggett (talk) 21:52, 29 January 2023 (UTC)Reply[reply]

When you clicked the "help" link, did you get to this section, where it says (amidst a lot of text) if the check digit is a lowercase 'x', change it to an uppercase 'X'? You are welcome to improve the help text to make it easier to parse. – Jonesey95 (talk) 21:57, 29 January 2023 (UTC)Reply[reply]
If we have a bot that can insert hyphens (see next topic), can't it fix trivial errors like that too? --𝕁𝕄𝔽 (talk) 10:07, 16 April 2023 (UTC)Reply[reply]
AManWithNoPlan advises that CitationBot repairs this type of error if a CS1/2 template is used. --𝕁𝕄𝔽 (talk) 19:32, 19 April 2023 (UTC)Reply[reply]

Hyphens in ISBNs[edit]

Our how-to page here has the sentence "Use hyphens if they are included, as they divide the number into meaningful parts; the placement of hyphens varies between books." Is this really the case? Let's try a book actually titled ISBN 9781909388062. If you go to Special:BookSources/9781909388062 and look you will get "results", but the various resources/external links listed -- WorldCat,, Google Books, HathiTrust, etc. -- do not list the book. But try OCLC 1065580670 -- yes, it is a real book, produced in a limited edition of 16. Is there a problem? Perhaps our ISBN 9781909388062 did not have the right hyphenation! Actually, NO. Why? Because BookSources says "Spaces and hyphens in the ISBN do not matter." What's up? When ISBNs were being developed 50 years ago the "meaningful" parts really were meaningful. This is no longer the case. When we click on How to Read a Book by Mortimer Adler, we can find publication data. The ISBN numbers listed in the article will give us links to WP:Book Sources, and those various external links all give ISBN-13 numbers. AND they are ISBNs that do NOT use hyphens! SO, I recommend that we remove the "Use hyphens if ...." sentence because it is out-dated and incorrect. Instead we say "hyphens may be included, but ISBN citations should seek a consistent style of hyphenation within each WP article." – S. Rich (talk) 02:17, 16 April 2023 (UTC)Reply[reply]

Change the sentence to leave out the "meaningful parts" bit, but not the recommendation. Don't require consistency. Printed books still provide hyphens for their ISBNs. It is convenient for databases to leave out the hyphens, but we have bots that fill in the hyphens if editors don't provide them. The tools we use for generating references provide hyphens. And they are easier to read if hyphenated when trying to see if a copy of a source is right version when checking references. I am not in favor of creating yet another "rule" to be followed or another consistency. (There is still the issue that ISBNs may not be unique, especially for pre-internet publications, since publishers have to pay money for them and small publishers reused them.) StarryGrandma (talk) 03:01, 16 April 2023 (UTC)Reply[reply]
I think this is one of these things that is non-cosmetic and useful enough to allow the bots to do and boring enough that we should not bother telling humans to do it. —David Eppstein (talk) 07:45, 16 April 2023 (UTC)Reply[reply]
The parts do have meaning, even if not to everybody. For example, the fully hyphenated version of the above ISBN, 978-1-909388-06-2, tells me that the book was published in an English-language country (the 1 as the 2nd element) and probably by a small publisher (the small number of digits in the 4th element). A second element of 90 says a book was published in the Netherlands, an ISBN starting with 978-0-521- is from Cambridge University Press, and so on. Kanguole 11:46, 16 April 2023 (UTC)Reply[reply]
We see "the placement of hyphens varies between books." Does this say a particular ISBN listing, with the same digits, will vary the results because of different hyphenation? E.g., {{ISBN|978-123456789-0}} and {{ISBN|978-1-234567890}} are different books? My point? – stressing the option of including hyphens leads us (especially me) into a world of confusion as to whether we want ISBN-hyphens in our citations. We shouldn't ban ISBN hyphens – we should stress citation consistency. – S. Rich (talk) 03:27, 17 April 2023 (UTC)Reply[reply]
If WorldCat, Amazon etc. don't list the book, all it means is that WorldCat, Amazon etc. haven't added the book to their database. It doesn't mean that the ISBN is in any way invalid. --Redrose64 🌹 (talk) 10:15, 16 April 2023 (UTC)Reply[reply]
There used to be a template, {{Format ISBN}} that used Module:Format ISBN and Module:Format ISBN/data. That template would correctly hyphenate an ISBN. {{Format ISBN}} was nominated for deletion by Editor Srich32977 at Wikipedia:Templates for discussion/Log/2022 November 18 § Template:Format ISBN partially on the grounds that [this] template is not used in any articles. The source for the template had this:
{{safesubst<noinclude />:#invoke:Format ISBN|main}}
which indicates that the template could be, probably always should be, subst'd which might account for its absence from articles. I'm inclined to resurrect the template and modules and enforce subst'ing by bot so that it is available to those editors who want properly hyphenated ISBNs in articles they edit.
In the TfD, Editor Srich32977 wrote: See my user-talkpage for some recent discussion on the topic. So I went looking for that discussion. I found several:
It appears that Editor Srich32977 is on a crusade to ban hyphenated ISBNs and continues to pursue that crusade despite repeated requests to stop removing hyphens from properly formatted ISBNs.
Trappist the monk (talk) 15:31, 16 April 2023 (UTC)Reply[reply]
It does, but at least Srich has sought wider input on this occasion.
I agree that {{Format ISBN}} would be useful, not least for helping with the aim of consistency. Kanguole 22:43, 16 April 2023 (UTC)Reply[reply]
Thank you so much, Trappist, for comparing me to General Eisenhower. With him in mind I've written a book titled Crusade in Wikipedia. And I have an ISBN for it: ISBN 9780801856686! (Please buy it via Amazon ASIN B00BABTBQ2.) Now only if I could figure out how to properly hyphenate my ISBN. (I'm going to ask for help from WP:GOCE. After all, they stress the "five Cs" – Clear, Correct, Concise, Comprehensible, and Confused.) – S. Rich (talk) 02:38, 17 April 2023 (UTC)Reply[reply]
There's 'forthright' and then there's 'tendentious'... XAM2175 (T) 17:15, 22 April 2023 (UTC)Reply[reply]
This argument appears to be based entirely on mechanical use, and to misunderstand what "meaningful parts" refers to.
Although Wikipedia interfaces to various websites and programs, it is ultimately read and written by people. Thus the requirement in the ISBN Users' Manual, section 5, remains relevant:

The elements should each be separated clearly by hyphens or spaces when displayed in human readable form

As the Manual notes,

The use of hyphens or spaces has no lexical significance and is purely to enhance readability.

It is basic usability that chunking of long strings of digits makes them easier for humans to handle. For example, here it helps when checking an ISBN against a book, and reduces errors when copying ISBNs. Moreover the chunks used contain information about the book (area of publication, publisher) that chunking makes readily accessible to (at least some) humans.
You may think these are small benefits, but they certainly outweigh any utility of uniformly formatted numbers. However, these two things are compatible. And the deleted template Trappist mentions above would be an efficient way to achieve consistency. Kanguole 22:43, 16 April 2023 (UTC)Reply[reply]
Most people find 13 digit numbers intimidating. They are also very hard to type in and/or verify because the eye keeps loosing track of exactly which position it is looking at. This is one reason why ISBNs, bank account numbers and phone numbers are almost universally printed as small groups with spaces or hyphens between them.
In the same that citation dates are reformatted, is it possible for the template to take whatever is in the isbn field, strip out everything except digits and then display the remaining string with hyphens according the the appropriate 13 or 10 scheme. This would give a consistent format in all cases, regardless of whether the editor put in hyphens, spaces or not. Also avoids the need for bots to modify the article - although this can still be done for neatness if desired.
Also, the template could reformat the isbn field into whatever format is required when passing to the various search engines (by stripping all non digits?). This avoids the entire argument of S.Rich about not finding books due to hyphens present or not present.  Stepho  talk  01:23, 18 April 2023 (UTC)Reply[reply]
True, but the placement of the hyphens has meaning. The third group in a 13-digit ISBN is the code for the publisher. Large publishers might have 3 or 4 digits like Cambridge University Press, while much smaller ones get many more digits. Personally, I would welcome reinstatement of the Format ISBN template as some ebooks and the like don't use the hyphens.--Sturmvogel 66 (talk) 16:29, 19 April 2023 (UTC)Reply[reply]
I would welcome reinstatement of the Format ISBN template Working on that...
Trappist the monk (talk) 16:31, 19 April 2023 (UTC)Reply[reply]
I have restored {{Format ISBN}} and restored modified versions of Module:Format ISBN and Module:Format ISBN/data updated to the ISBN RangeMessage as of 2023-04-19. The new version is different from the old so read the template's documentation before using. I generally suck a documentation so if there are improvements that can be made, please do, or let me know and I'll attempt to do better. The template is auto-subst'd by AnomieBOT.
It might be good to note the existence of this template on the project page and perhaps other places. I'll leave that to others...
Trappist the monk (talk) 22:37, 22 April 2023 (UTC)Reply[reply]
That's great stuff. I'm not sure it should convert ISBN-10s to ISBN-13s, though. This page says to use the ISBN-10 if that's all there is in the book, and not to convert it to ISBN-13. Kanguole 22:54, 22 April 2023 (UTC)Reply[reply]
That convert-10-to-13 was in the original module and I left it as is. There are bots running around that do that conversion (this edit for example) so I haven't done anything to preserve 10-digit ISBNs – why bother if Citation bot is going to convert it anyway? I can spend some time making the module format and preserve 10-digit ISBNs but with bots converting them to 13-digit in citation templates, that seems rather a wasted effort. I don't know if there are bots that convert-10-to-13 in {{ISBN}} or when the ISBN is stand-alone so, perhaps for those use-cases, preserving 10-digit ISBNs might be worth while. I suppose a specific template {{Format ISBN10}} would be the best way for editors to get this functionality.
Trappist the monk (talk) 23:26, 22 April 2023 (UTC)Reply[reply]
Forgive my ignorance, but is there any reason to preserve 10-digit uses? XAM2175 (T) 23:41, 22 April 2023 (UTC)Reply[reply]
I can imagine a reader looking for a book in the local library and finding an author/title/publisher/date match with a 10-digit ISBN but the citation at was a 13-digit ISBN converted from 10-digits. I can imagine the reader wondering if the book in hand is the same book or a different edition. Because we should not expect that our readers are familiar with the intricacies of 10- and 13-digit ISBNs, I think that we should cite the book as we see it: if the book's ISBN is 10 digits, we should cite using the 10-digit ISBN; if both 10- and 13-digit ISBNs or only a 13-digit ISBN, use the 13-digit form.
Having written this, I'm inclined to wonder if {{Format ISBN}} should be rewritten to format and return the ISBN in the form that it is given rather than to return only 13-digit ISBNs. If I do that, we can discard {{Format ISBN10}}. Opinions?
Trappist the monk (talk) 16:19, 23 April 2023 (UTC)Reply[reply]
I think that would be a good idea, per the example you gave. If an editor has a post-2007 book, they will have the ISBN-13 to supply to the template or to compare with the wiki entry. Kanguole 17:10, 23 April 2023 (UTC)Reply[reply]
Yeah, that's a fair point. Out of curiosity, is there a serious possibility of the problem presenting in reverse, though? (that is, a reader sees a 10-digit ISBN here but is confused by seeing online catalogues list the 13-digit version)
Either way, thanks very much for your efforts. If nobody has any objects, I intend at some point in the near future to modify the documentation for {{cite book}} to include guidance about using this template in the ISBN parameter. XAM2175 (T) 18:51, 24 April 2023 (UTC)Reply[reply]
I have created {{Format ISBN10}}. It works much the same way as {{Format ISBN}} except that 13-digit ISBNs as input are returned without modification compared to {{Format ISBN}} which converts ISBN10 input to ISBN13.
Trappist the monk (talk) 16:19, 23 April 2023 (UTC)Reply[reply]
I support modifying {{Format ISBN}} to leave ISBN 10s unchanged, and don't think two templates are required. It'd be consistent with the other ISBN templates which accept either form. I'd also support it handling SBNs (add a 0 to hyphenate as ISBN10, then remove the 0), but other templates do treat them differently, which makes sense for clarity. There could be a parameter that confirms the "as printed" digits, 9, 10 or 13? Since the ISBN should be entered as printed, the only value would be in confirming a 9 digit SBN is not a typo. Automatic conversion between forms seems unnecessary. I've been making a lot of ISBN hyphenation changes using the ISBN ranges as they are updated from . I'm interested in knowing how {{Format ISBN}} will be kept current, and could help with that as I know how the range files work (but unfortunately not much about Wikimedia template code). Salpynx (talk) 21:48, 23 April 2023 (UTC)Reply[reply]
Yeah, I'm getting closer to thinking that we don't need {{Format ISBN10}} and that n-digit input gets n-digit output. Hadn't given much thought to SBN formatting; adding a leading zero, formatting and then dropping the leading zero (and the hyphen) is simple enough. Not sure what you mean by a parameter that confirms the "as printed" digits.
I wrote a Lua module, Module:ISBN RangeMessage xlate, that reads a local copy of the RangeMessage.xml file and from that assembles the Lua table used by Module:Format ISBN. The local copy of RangeMessage.xml is hidden in html comment markup (<!--...-->) in Module:ISBN RangeMessage xlate/doc. There are instructions there that (I think) explain how to update Module:Format ISBN/data when International ISBN Agency updates RangeMessage.xml (with more than just changes to <MessageSerialNumber>...</MessageSerialNumber> and <MessageDate>...</MessageDate>). Does that answer your questions?
Trappist the monk (talk) 22:59, 23 April 2023 (UTC)Reply[reply]
Here is {{Format ISBN/sandbox}} which will format SNB, ISBN10, ISBN13. The template returns the formatted input:
  • {{Format ISBN/sandbox|035789360}} → 03-578936-0
  • {{Format ISBN/sandbox|1605206245}} → 1-60520-624-5
  • {{Format ISBN/sandbox|9781571816559}} → 978-1-57181-655-9
If we adopt this then {{Format ISBN10}} goes away.
Trappist the monk (talk) 17:53, 24 April 2023 (UTC)Reply[reply]
This discussion seems to have ground to a halt. Given that, I'm inclined to replace the current {{Format ISBN}} with {{Format ISBN/sandbox}} so that the template, by default, returns a formatted ISBN that has the same length as its input. I have made one adjustment: |out= will accept values of 10 and 13 so that editors may specify the format output. This does not work for SBNs or for ISBN13 with the GS1 prefix 979:
  • {{Format ISBN/sandbox|9781571816559|out=10}} → 1-57181-655-0
  • {{Format ISBN/sandbox|1-57181-655-0|out=13}} → 978-1-57181-655-9
  • {{Format_ISBN/sandbox|9791220008525|out=10}} → 979-12-200-0852-4
  • {{Format_ISBN/sandbox|035789360|out=13}} → 03-578936-0
Without objection, I shall replace the current {{Format ISBN}} and delete {{Format ISBN10}}.
Trappist the monk (talk) 15:19, 28 April 2023 (UTC)Reply[reply]
No complaints from me! XAM2175 (T) 15:23, 28 April 2023 (UTC)Reply[reply]
Module:Format ISBN has been updated from its sandbox and {{Format ISBN10}} is no more. I have tweaked Template:Format ISBN/doc to reflect the changes in Module:Format ISBN. The ~/doc could probably benefit from a third eye giving it a look-see.
Trappist the monk (talk) 18:15, 29 April 2023 (UTC)Reply[reply]
Ah, an argument over hyphens in ISBNs. This is what I visit Wikipedia talk pages for! :-)
Anyway, count me as another fan of hyphenation and its chunking benefits. I don't see the point of the example at the top of the thread, which is a limited-edition book whose title is ISBN 9781909388062. The WorldCat entry, at OCLC 1065580670, simply doesn't include an ISBN; of course the book won't come up if you search for the title in the ISBN field. WorldCat handles hyphens in search queries just fine. XOR'easter (talk) 13:19, 22 April 2023 (UTC)Reply[reply]
I just had to type a bunch of ISBNs today. Hyphenation absolutely makes sense and helps ensure accuracy snd legibility. Whenever one looks for such an identifier on a scholarly database or WorldCat hyphens are invariably used. Why on earth some seem to think removing hyphens from ISBNs is anything but pointless at best and a harmful distraction at worst is beyond my comprehension. MaxnaCarta (talk) 13:33, 22 April 2023 (UTC)Reply[reply]
Absolutely, I also support hyphenation for said benefits. In fact, I'd further support making it compulsory to include them. XAM2175 (T) 17:11, 22 April 2023 (UTC)Reply[reply]
I would support compulsory ISBN grouping for the same benefits, though with preference for spaces instead of hyphens. (Naturally, once you get three people in a room and ask them to create a standard, you end up with four standards; a fourth person then creates a new standard to replace all existing use cases and we now have five standards.) Ifly6 (talk) 05:03, 23 April 2023 (UTC)Reply[reply]
Whether you use hyphens or spaces, or even omit them entirely, makes absolutely no difference to the meaning of an ISBN, only the actual digits (or terminal X) matter. I think that I've pointed this out before.
As for the position of the hyphens, a thirteen-digit ISBN has four: one of these follows the first three digits (which are always 978 or 979), one precedes the last digit, and the other two have variable positions. A ten-digit ISBN has three hyphens in the same positions as the thirteen-digit form, if the 978- or 979- were to be omitted. Whilst two have variable positions, they serve only to split up the digits into convenient groups for country, publisher and individual book. The position of the hyphens is not critical: ISBN 978-0-12-345678-6 may be written ISBN 978-0-1234-5678-6 or ISBN 978-0-123456-78-6, or any other variation you may think of, they are all equivalent. --Redrose64 🌹 (talk) 17:38, 22 April 2023 (UTC)Reply[reply]
You could but you oughtn't. In your example ISBNs, 978 is the GS1 prefix and 0 is the registration group element. The 12 (or 1234 or 123456) is the registrant element. For the 0 registration group (English language), ISBN defines registrant elements in these ranges:
  • 00-19
  • 200-227
  • 2280-2289
  • 229-368
  • 3690-3699
  • 370-638
  • 6390-6397
  • 6398000-6399999
  • 640-644
  • 6450000-6459999
  • 646-647
  • 6480000-6489999
  • 649-654
  • 6550-6559
  • 656-699
  • 7000-8499
  • 85000-89999
  • 900000-949999
  • 9500000-9999999
Of your three examples, only the first, 978-0-12-345678-6 is correct because the registrant element (12) is the only one to fit into any of the defined ranges; 1234 and 123456 do not fit. You can check this for yourself by downloading the RangeMessage by group prefix document from (a direct download). ISBN has information and two other related documents available at ISBN Ranges (also direct downloads).
Trappist the monk (talk) 18:27, 22 April 2023 (UTC)Reply[reply]
My point is that since the hyphens are immaterial to identifying the published work, all of my examples can be reduced to ISBN 9780123456786 with absolutely no change in meaning. But it's not nearly so easy to read, so throw in some hyphens, ideally in the recommended positions, but if you misplace them it won't change the ISBN to a completely different book. I have seen Penguin books with a fifth hyphen (like 978-0-14-999-999-9), simply to split up the run of six into two threes, and not only is it more readable, it's also still valid. --Redrose64 🌹 (talk) 19:20, 22 April 2023 (UTC)Reply[reply]
Sure, maybe more readable to you but to someone else, additional or repositioned hyphens make it more difficult to decode the whole ISBN.
Trappist the monk (talk) 22:37, 22 April 2023 (UTC)Reply[reply]
Randomly placing the middle two hyphens is not a great idea. It won't refer to a different book, but it will lose two of the advantages of hyphenation for human readers: that the chunking matches the book and that the chunks are meaningful. Kanguole 08:36, 23 April 2023 (UTC)Reply[reply]
Are WP readers actually "reading" ISBNs? I venture they are not. They put the mouse pointer over the link and click. That gets them to BookSources, where they can then choose to get more info from WorldCat, Amazon, Google, etc. For WP editors the watchword is #5 in the Big Cs -- Consistency. When we add ISBNs we want a page where the reference and citation listings are in a nice, neat pattern. – S. Rich (talk) 16:30, 28 April 2023 (UTC)Reply[reply]
Putting hyphens in all ISBNs means that both of these objectives are satisfied, yes? Please don't start stirring the pot again. XAM2175 (T) 16:38, 28 April 2023 (UTC)Reply[reply]
I suggest that some readers do indeed "read" isbns and will appreciate the difference between one which indicates a large publisher and one which shows the opposite. And if a reader wants to copy the isbn manually, whether to write it down or type it into another resource, the hyphens make it easier to read and memorise. Usefulness is more important than consistency. PamD 06:53, 29 April 2023 (UTC)Reply[reply]