Content moved to wikipedia talk:ISBN

See http://www.isbn-international.org/converter/ranges.htm for details.

Broadly:

0 English speaking areas
1 English speaking areas
2 French speaking areas
3 German speaking areas
4 Japan
5 Former USSR
60x various
7 China
8x various: see the cite above for details
9x various: see the cite above for details
92 International organizations as publishers

I know that ISBNs are 10 digits: language digit, dash, publisher code, dash, book number, dash, check digit. Is a book uniquely identified if dashes aren't given? I.e. could there be different books with ISBNs:

• 0-123-45678-whatever
• 0-1234-5678-whatever
• 0-12345-678-whatever? Bubba73 (talk), 20:28, 16 January 2007 (UTC) reply

The short answer: yes, the book is uniquely identified without the hyphens, and no, your three different examples are not possible. There is only one standard way of hyphenating a given ISBN. The hyphens are (so far as I can tell) just a way of making the number more human readable, the way we partition ten-digit phone numbers into area codes and exchanges. Here, in your last example, 0 is a region code (denoting English speaking countries) and 12345 is a publisher ID. The 678 is used by the publisher to identify a specific book, and the 'whatever' is the check digit. If the ISBN authority issues -12345- to a specific publisher, then they will not issue any of its prefixes. E.g. 1, 12, 123, 1234 would be off limits as publisher codes. You seem to be a computer person, so you see the logic of this! Unfortunately this means that only an entity that knows all the publisher codes of the world (like isbn.org) can hyphenate an arbitrary ISBN properly. In practice one can use isbn.org/converterpub.asp to hyphenate an ISBN that does not have the hyphens already. EdJohnston 21:58, 16 January 2007 (UTC) reply

OK, thanks. Yes, I do see the logic, and it works the way I thought it should work. Then (to just grab a book at hand), with ISBN 0-385-51010-1, publisher 385 (Doubleday) book 51010, there would not be a publisher 3855 with a book 1010, because that would result in the same digits. Bubba73 (talk), 23:17, 16 January 2007 (UTC) reply

It's different to what I thought. I have definitely heard before that people "must be careful to ensure the groupings are retained" when writing ISBN's.

Actually, I was going to ask a similar question prompted by the article statement:

Barcode format compatibility is maintained, because (aside from the group breaks) the ISBN-13 barcode format is identical to the EAN barcode format of existing ISBN-10s.

i.e. could the EAN barcode ever misidentify a book (because of ambiguity)?

I see now in the ISBN Users' Manual (International Edition, Fifth Edition) that

The elements must each be separated clearly by hyphens or spaces when displayed in human readable form: [...]

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

(Italics are original.)

—DIV ( 128.250.247.158 ( talk) 04:14, 20 February 2009 (UTC)) reply

Here in Wikipedia I suggest that, when adding a book reference, you use either an unhyphenated ISBN or a hyphenated one that you're sure is correct. Certainly you can't make up your own hyphenation! If you copy the hyphenated number from the book itself or from a catalog it is probably correct. Ambiguous EANs aren't possible since one digit string corresponds to at most one book, regardless of hyphens. As to whether hyphens are compulsory, Amazon shows ISBNs on their site with either zero or one hyphen and we can't do anything about that! EdJohnston ( talk) 04:36, 20 February 2009 (UTC) reply

Thanks for the input, EdJohnston.

I had also read:

An ISBN should be printed on the reverse of the title page [...]. It must be printed exactly as given somewhere inside the book for it to be valid.

from https://www.enett2.com/ThorpeBowker/ISBN/ (ISBN Agency Australia).

—DIV ( 128.250.247.158 ( talk) 04:35, 23 February 2009 (UTC)) reply

The ISBN numbers are all numerically unique, and for most people's purposes, the hyphens are irrelevant. Any "lookup" software that can find individual books by their ISBN number won't need the hyphens, and should be able to pull up all the associated book info lodged on the central database(s) for that ISBN number.

However, the numbers are originally "sold on" to publishers in consecutive blocks of ten, or a hundred, or a thousand, etc. in escalating powers of ten. By only selling blocks in powers of ten, the issuing agencies ensure that all the members of a given "sold" block will have the same initial sequence of digits, and if that shared initial sequence is separated off from the remainder of the number with a hyphen, it can be used as the block's "ID number". That ID can then be tied to the publisher who happened to buy that run of numbers.

The advantage of the separating hyphens is that it lets a book industry company's computer system recognise common recurring block IDs and cache the info for which publisher those block IDs correspond to without having to do a complete database search on the full number each time. You can run through your sales list and find which publishers are generating how much income for you, by using the block IDs as publisher IDs (although a single publisher can have multiple IDs if they've bought multiple blocks). Publishers are encouraged to "buy big" rather than buying multiple smaller blocks, to keep the size of the lookup tables manageable and encourage the use of the blocks as publisher identifiers. So for instance, the ten-digit ISBN number for my 2005 copy of "Freakonomics" is 0-141-01901-8 ... the hyphenation tells me that Penguin Books seem to have bought the entire "141" block, which consists of a hundred thousand consecutive ISBN numbers. If I was a bookseller, this'd tell me that I could probably spot other Penguin titles on a printed spreadsheet by looking in the ISBN column for a "-141-", and because the company had bought so many numbers, it'd suggest that unless Penguin had bought additional separate blocks for special purposes, they'd be unlikely to have used up all 100,000 numbers, and that I could probably spot every other Penguin book on the spreadsheet by its -141- tag.

So the hyphens do (did?) have some purpose for some companies in the supply chain, for stock control. If you sell a lot of Penguin titles, the -141- is probably comparatively easy to recognise on a spreadsheet printout. But for the final customers and for academic and online citations, where it's assumed that the reader hasn't memorised a set of common publisher blocks IDs (and is more interested in the author and title info than the publisher), the ISBN number is used to uniquely identify a //book//, and the hyphens probably just serve to pad out a thirteen-digit EAN to an annoying seventeen characters. :(

Short answer - the hyphens might still be useful to some people in the book industry, but they're not really much use to the end-customer. They'll help a savvy bookshop staff member to find your book by searching by publisher even if you've gotten one of the last digits wrong, but that's probably about the only useful thing you'll see from it. ErkDemon ( talk) 01:51, 11 September 2010 (UTC) reply

Penguin didn't just buy the 141s - they bought the 140s too. They may have bought the 142s through 149s inclusive. -- Redrose64 ( talk) 18:24, 11 September 2010 (UTC) reply

The article uses the phrase 'Language code', but there is no recognition of that phrase at isbn.org. I propose replacing it with 'Group identifier', which is their official term. Some extra words would be added to explain how it's generally used. Let me know if anyone objects. In fact, it's not really a language code, it's a code for a group of countries that (in most cases) happen to share a language. Any book issued by a publisher in that set of countries gets the given group identifier, regardless of the language of the book. That's how Springer can issue English-language books in Germany with group identifier '3'. And how a publisher in Iran (with group identifier 964) can issue a book in the Azeri dialect of Turkish [1]. This clarification might benefit Wikipedia's ISBN-fixers when searching for books in languages other than English. EdJohnston 22:22, 17 January 2007 (UTC) reply

Why is this article opted out of bot attention entirely? -- nae' blis 01:02, 18 January 2007 (UTC) reply

I think it has to do with the possible transfer of ISBN-10 on wiki to ISBN-13. Maybe too because we need to preserve code that is seen by bots as bad coding that needs fixing? Circeus 16:05, 18 January 2007 (UTC) reply

• ISBN generator and converter on official site of International ISBN Agency

194.154.66.232 22:34, 26 January 2007 (UTC) reply

This tool appears to be intended for use by publishers. (Generates batches of ISBNs from a given publisher number). It's not clear how it would benefit the average Wikipedia reader. The tool isbn.org/converterpub.asp is already cited in the article for validity checking of individual ISBNs and for conversion of an ISBN-10 to ISBN-13. EdJohnston 02:16, 28 January 2007 (UTC) reply

How does ISBN-13 solve the "pending shortage in certain ISBN categories" if all it does is prefix an ISBN-10 with 978 or 979? Njál 21:52, 16 February 2007 (UTC) reply

The official documents don't shed much light on this. However, there is an advantage in the 979 category that they won't use the existing publisher codes, so there is some hope that the 979 range will hold a lot more numbers than the existing 978 range. If the present codes were used, adding 979 would provide a factor of two increase. EdJohnston 23:16, 16 February 2007 (UTC) reply

Yep, "979" serves as an overflow for "978".

It's not necessarily the case that all the 978 numbers have been //used//, but they've increasingly been bought and reserved by publishers, and if a new publisher appeared who was liable to want to buy a vast block of consecutive numbers, that possibility was getting increasingly difficult to deal with using the remaining 978 codes.

Suppose that Google Books eventually decides to start offering reprinted hardcopies of its massive scanned collection of out-of-print, non-copyright titles ... and wants to apply EAN barcodes to their reprints. Since EAN is a product code linked to the publisher, if a reprinted book from a new publisher is going to have an ISBN, it needs a new number. So Google might hypothetically decide that to future-proof their numbering system, they want a block of a million consecutive ISBNs. Heck, they might decide that they want more, in which case, they either buy consecutive blocks of a million, or stretch to a "ten million" block. ErkDemon ( talk) 02:18, 11 September 2010 (UTC) reply

At present this article has 20 external links. (Probably too many to have it be accepted as a featured article). Does anyone have time to carefully study the remaining links and recommend here, on the Talk page, which ones we can do without? As an example, I think we have too many different links to 'conversion aids'. If we are going to keep all those links we at least we should at least add some text to the article about conversion. EdJohnston 16:20, 19 March 2007 (UTC) reply

Is ISBN 13 less capable of detecting errors, or equally capable but less capable of detecting transposition errors which are perhaps one of the most common forms of errors (i.e. it's better at detecting other errors)? I presuming it's the later so I've clarified the article accordingly. Obviously if transposition errors are more common then other types of errors then by not being so good at detecting transposition errors ISBN 13 may be less likely to detect errors but it would still arguably be equally capable Nil Einne 15:24, 22 April 2007 (UTC) reply

The article said:

the 10-digit ISBN check digit,[8] which is the last digit of the 10 digit ISBN, is calculated on a modulus 11 with weights 10 to 2, using X instead of 10 where 10 would occur as a check digit (and 0 instead of 11)

0 instead of 11? No number mod 11 could possibly result in 11 in the first place. —The preceding unsigned comment was added by 203.217.22.128 ( talk) 05:05, 10 May 2007 (UTC). reply

I think it is consistent. The language may be a bit loose, but the point is that any number that is a multiple of 11 will give a check digit of 0. Can you think of a better way of saying it? EdJohnston 17:53, 12 May 2007 (UTC) reply

The check-digit substitution in the article is incorrect and should be fixed. The algorithm is to substitute a 0 for a 10, and substitute an "X" for 11.

http://www.bisg.org/isbn-13/conversions.html#Algorithm%20for%20checking%20the%2010-digit%20ISBN —Preceding unsigned comment added by Dzurn ( talk • contribs) 05:45, 23 November 2008 (UTC) reply

What's the source on the pronounciation for "ISBN"? I've never heard anyone pronounce it this way. It's not a word or an acronym, it's an initialism. So, it should be pronounced "I-S-B-N." —The preceding unsigned comment was added by 138.217.163.48 ( talk) 07:16, 16 May 2007 (UTC). reply

I was just going to add the same comment. In the UK, I have only ever encountered the pronunciation "I-S-B-N". Can any definitive sources be provided supporting one or the other position? Hassocks5489 11:41, 23 May 2007 (UTC) reply

Ditto - I've never heard this in the U.S. It seems Rossumcapek added this here. I'm adding a note in his talk page, and for the time being removing the pronunciation. - hc5duke 2007-06-18 08:59 (UTC)

When I worked in bookstores, some members of the staff would pronounce ISBN as "is-ben." I'm pretty sure this is a case of bookstore industry jargon/lingo. What is the best way to source this? (Thanks for the heads-up about this discussion, hc5duke.) — Rossumcapek 17:05, 18 June 2007 (UTC) reply

Not really sure how you would reference that - I think the way you phrased it was changed over time so that it looked like "is-ben" was the main pronunciation. I don't think anyone would have a problem with "sometimes pronounced as" or something.- hc5duke 2007-06-19 17:35 (UTC)

If some smart aleck pronounced U.S.A. as "us-sa", that would not be enough evidence to claim that it is sometimes pronounced that way. There is no vowel in the second syllable of ISBN and ISSN, so they were intended to be spelled out. GUllman 22:09, 28 June 2007 (UTC) reply

I see how edits over time changed to imply that the abbreviation was always pronounced so, but I certainly didn't intend that. Also, there is plenty of precedent for some abbreviations to be pronounced as though they had vowels: SCSI ("scuzzy"), SQL ("sequel"), Tcl ("tickle"), and so forth. It's certainly not universal, but not uncommon. What about using the phrase "sometimes pronounced (IPA symbols for is-ben) in the book industry" instead? That should be clear. — Rossumcapek 17:17, 1 July 2007 (UTC) reply

Is it correct that there is no way to see the correct hyphenation of a given ISBN without consulting an external source, like the book itself etc.? (I.e. given 012345689 may it theoretically be hyphened 0-12-34567-89, but also 0123-4567-89)? Thanks. Jakob.scholbach 03:15, 11 June 2007 (UTC) reply

This is discussed somewhat in previous entries on this Talk page, like #Question about ISBN. An oversimplified answer is that you need a list of all the publisher codes in the world to be able to hyphenate an arbitrary ISBN. In practice I think there are a few shortcuts, so a successful algorithm may not need the complete list. Open up the web form at http://www.isbn.org/converterpub.asp and enter the number, if you want to determine the right placement of the hyphens in a particular ISBN. EdJohnston 04:44, 11 June 2007 (UTC) reply

There is a list of ranges which tells you where the hyphens should be placed, but I have temporarily mislaid my link. In the meantime, there are several people who fix ISBN formats: maybe you could ask them? HTH HAND Phil | Talk 11:19, 18 June 2007 (UTC) reply

The best list I have found is at http://www.isbn-international.org/converter/ranges.js (pretty useful since it is continuously updated and machine-readable). Maybe the link should go into the article? -- CyHawk ( talk) 14:38, 5 January 2009 (UTC) reply

no, it's not correct. the hyphen placement can be determined from only the number. [2]  — Chris Capoccia ^T⁄ _C 10:36, 29 May 2009 (UTC) reply

Unfortunately, this article doesn't tell, which lead me to find the answer from

http://www.bisg.org/isbn-13/faq.html

http://www.collectionscanada.ca/iso/tc46sc9/isbn.htm

And

which prompts me to request the article writer to either append these two links or amend the content.

... or you could make the edits yourself. This is an open source encyclopaedia. You are the article writer. Have fun! ElectricRay 23:54, 27 June 2007 (UTC) reply

I have come across a book with only one page and is of ISBN number. I would like to check if it is a valid number assigned to the book. Please add the knowledge of that. Thanks. —Preceding unsigned comment added by 218.82.228.37 ( talk) 11 July, 2007

Please sign your comments using four tildes before pressing 'Save page', i.e. ~~~~, and ask the question more clearly. What book are we speaking of? EdJohnston 14:08, 11 July 2007 (UTC) reply

They don't have the book, only the ISBN. They should Google on "ISBN <number>". Unfortunately, the positions of the hyphens matter (omitting them worked best for me). HairyWombat 22:53, 6 August 2007 (UTC) reply

Book Industry Study Group suggests a certain 5-digit price add-on structure for the ISBN-13 with price add-on.

* First digit is "5" for $US
* First digit is "4" for $CAN 

Add-on seems to be mandatory for several US retailers.

Question: are there any other examples for other countries than US/CAN in this respect? Has anyone seen price add-ons say in Germany or France?

Gs1mo 13:49, 12 July 2007 (UTC) reply

I have been told (BSIG) that price add-on is specific to US market only. To be specific it's mandated only by Barnes & Noble

Gs1mo 14:52, 23 July 2007 (UTC) reply

There's not much need to have such an add-on for Germany and France (or for the rest of the Euro-zone), since those countries use the same currency (Euro) - as distinct from the US/CAN situation of using the same name ('Dollar') for different currencies (USD/CAD) 87.194.60.87 20:24, 12 August 2007 (UTC) reply

I just wanted to thank whoever wrote the section Check digit in ISBN-10. I needed to cite a publication with the old ISBN 2-88145-076. 9-digit ISBNs are not recognized by the Wiki ISBN magic but, using the article, I was able to manually calculate the check digit to give ISBN 2-88145-076-8. Again, thank you. HairyWombat 22:39, 6 August 2007 (UTC) reply

It is unlikely this book was ever issued by a commercial publisher, so they got away with having a defective ISBN (one digit too short). I looked up that ISBN, as you corrected it, in the Library of Congress, the British Library and Amazon but it's not for sale or catalogued anywhere. Good thing the book is available online! Your calculated check digit seems correct to me. When you encounter an actual 'old' 9-digit ISBNs you can usually convert it by putting a 0 in front. The one printed in this book is not old, but bad, in my opinion. EdJohnston 02:26, 7 August 2007 (UTC) reply

Yep, looks like the ISBN is defective. The publication was by the International Committee of the Red Cross in Geneva. Although the publication is in English, Geneva is French speaking which would explain the leading "2". I found a contact page for the ICRC's Library and Research Service, and have asked them for the correct ISBN. Until they get back to me I have listed the defective one. HairyWombat 23:05, 8 August 2007 (UTC) reply

I think that this article would be greatly improved of the explanation of the format of ISBNs and the method of their calculation/assignment were simplified. As it is, I found the mathematics (expecially the detailed discussion of modulo and modulus) to be lacking clarity and truly confusing. The tone, with the lack of deferential explanation to the non-mathematician, indicate an orientation other than towards readers, editors, or even librarians, which is to say ISBN users. When I read this article, I felt a bit like I had looked up telephone area codes and instead of seeing a list, I read a technical explanation of why the middle digit once could only be a '1' or '0' and how phone systems were badly fouled up for more than a year after other numerals were introduced. That wouldn't be helpful or appropriate for such an entry, and this one fails on the same account.

I don't propose that this information be removed, but rather that a more direct, practical, non-technical (in a mathematical sense) explication of ISBNs be added, and that perhaps the detailed math moved further down into the article. As it is, after the "Overview" section, the math associated with ISBNs is put forth as the central and most vital information on the subject. I don't think this is the case. For Wikipedia users (as for most encyclopedias, except specialized or technical ones) who will most often be readers, booksellers, or others using/buying/selling/caring for books, the most sought information is simply how to parse or interpret an ISBN, followed by why they are needed or required, their history, the assigning body or authority, and finally by the more special interest information of the formulas and math/math theory that is here.

I am not myself a good candidate to do this editing. That person would more ideally be a librarian or bookseller who better understands and is more experienced in the pragmatics of ISBNs (and they are what else other than practical?), and who perhaps also understands the math at least well enough to re-write the article to include the mathematical info in a more clear and appropriate (in tone, position, and proportion) manner. Googlyelmo 19:29, 11 August 2007 (UTC) reply

Thanks for your comment. One way to handle this would be to split off a separate article that did the math bit. It would assist us in making this reform if you would identify a block of material that you think could be handled elsewhere. We need your input on what you consider to be the 'pragmatics of ISBNs', since that's what you apparently feel is missing. Or is it present here, and just obscured by the math? EdJohnston 19:35, 11 August 2007 (UTC) reply

It is the second through fourth sections/paragraphs, following the initial section under the heading "Overview". These go into great detail about the formulas for arriving at the check digits of both the 10 and 13 digit ISBN. Yes, splitting this to another article, with appropriate pointer/link, might help. Most of the information necessary to interpret ISBNs seems already to be present, though it is scattered over several sections (not at all uncommon for an article with multiple, successive writers and editors, none of whom truly collaborate, in the contemporaneous sense). Some amplification and elaboration of how ISBNs are used, by whom, and for what purpose might improve the article. Just imagine why someone, a non-specialist, might visit Wikipedia to look up this topic. To the limited extent I have edited or added to articles, I have always had that in mind, especially where I do have some level of expert knowledge of a subject. So, someone comes to this article after seeing an ISBN in a book ad, on an Amazon page, or on a copyright page, and is curious or unclear what it signifies. Also, someone is given an ISBN or told to look a book up via an ISBN as an employer assigned task, when asking for a good book to read from a friend, or when a family member expresses an interest in a particular, exact edition of a book (as part of a gist wish list perhaps). So just the basics (the details and real arcana can go at the end), like: What is an ISBN? What do the parts mean? What is it used for? (this one varies according to next item) By who? Who chooses it? Is it mandatory? Are they unique? Are they the same all over the world? How can I get one for my self-published book? Does it cost money? When did they start issuing them? What about books published (and not altered/revised) before then? Why did publishers/booksellers decide they needed ISBNs?, and so on. Many of these are already answered, but even these are scattered about the article, where they might all be addressed in the first overview section, and then elaborated and amplified later. It's really more of an organizational issue, compounded with the distraction of the math discussion being featured early. I see this a lot with articles about prescription drugs, among other things, which often talk over the heads of even MDs (written by/for biochemists/pharmacologists, it would seem), to say nothing of laypeople. Googlyelmo 21:01, 11 August 2007 (UTC) reply

Another remark about the math and number theory: the predominant discussion on this talk page is about number theory and the math associated with the ISBN numbering system, and in combination with the emphasis (too much and too early in my opinion) on this in the article itself, it begs the question: Who in practice uses this to verify the validity of a given ISBN? What I mean is, why would one suspect a given ISBN of being invalid? What would prompt that suspicion? Second, once one does suspect an ISBN of being invalid, is there a straightforward method for checking the number? Maybe that was covered in the article, but I couldn't work it out amid all mathematical discussion. I'll add that to my suggestions: in addition to re-focusing on user-oriented infoout the ISBN and system as I propose above, the math and validity checking procedures might be rewritten to enable someone without more than good arithmetic and very basic algebra to actually and practically check an ISBN for validity. Googlyelmo 23:46, 11 August 2007 (UTC) reply

In practice you wouldn't do the math yourself, you would submit the questionable ISBN at http://www.isbn.org/converterpub.asp, or some other site that offers free checking. I had heard that our own Special:Booksources had a checking link you can click on but I looked just now and didn't find it. EdJohnston 21:36, 12 August 2007 (UTC) reply

Recent edits have replaced 'group identifier' with 'group identifier code number.' Looking at the web site http://www.isbn.org, I see only the simpler phrase 'group identifier'. I would like to remove 'code number' from the article unless there are objections. EdJohnston 02:58, 20 September 2007 (UTC) reply

This borders on Trivia (which is discouraged by Wikipedia) for someone interested in finding out about ISBN numbers. It smacks of anti-Chinese politics which I think has no place here Ray3055 ( talk) 11:17, 12 December 2007 (UTC) reply

If the section had a reference, I'd be OK with keeping it. Since there are none, if you want to remove it, go ahead. EdJohnston ( talk) 16:10, 12 December 2007 (UTC) reply

The suggestion about using very large weights with a very large modulus misses the point: ISBN only uses 10 symbols (along with the X in ISBN-10). Commerce is based on 0,1,2,3,4,5,6,7,8,9, so something with modulus 100000000000 is not sensible.

As for modulus 20, it is false that alternating weights 1 and 3 will detect all adjacent transposition errors. A system of weights will detect all adjacent transposition errors only when the difference of consecutive weights is relatively prime to the modulus in use. Since 3-1 = 2, there's a problem here with modulus 20, which is not relatively prime to 2.

The choice of weights has nothing at all to do with the length of the weights in terms of their digits (using 1, 10, 100, 1000, and so on, say) and everything to do with the divisibility relations among the weights and the modulus. At the very least, a sensible system of weights should be relatively prime to the modulus to get the most common error (single-digit errors) always detected. This follows from Bezout's identity in number theory.

The point of using a prime modulus is that all positive integers less than it are relatively prime to it, and that makes the detection of all adjacent transposition errors feasible without much work needed. This is impossible for the standard weight-based modular check digit methods if the modulus is even. Of course the prime 2 is useless. But the *mathematical* features of primes make a difference, so the choice of modulus 11 (not too far from 10, which has ad hoc historical relevance to humans) is sensible.

There are some articles by Joe Gallian about check digit methods which explain the math behind this.

Concerning the footnote about Springer codes in English and German having the same check digit, I have no idea how publisher codes are assigned but it had to have been planned by Springer in some way to get the publisher codes they did; maybe they made a request or something. Because if their publisher codes in English and German were just randomly selected for them then there is no way their check digits in English and German with the same product code would lead to the same check digit every time. I noticed this systematic feature when looking at ISBNs on books on a shelf at a university library. For Springer, their language+publisher code a-bcd in ISBN-10 always has the feature that 10a + 9b + 8c + 7d = 3 mod 11. A random choice of digits for a,b,c,d would not behave like this! —Preceding unsigned comment added by 137.99.17.121 ( talk) 5 April, 2007.

"In fact, no system of weights used to compute a check digit based on modulus 10 (or any even modulus) can detect all single digit errors and adjacent transposition errors."

Well how about weights 1000000000 100000000 10000000 1000000 100000 10000 1000 100 10 1 modulus 100000000000000? And we talk about primes, 2 is a prime, but not very useful. Do we mean "co-prime with and larger than 10"? Rich Farmbrough, 15:47 4 September 2006 (GMT).

And in fact, a little WP:OR shows that 3,1,3,1... with mod 20 will pick up all single digit errors and adjacent transposition errors. Rich Farmbrough, 16:11 4 September 2006 (GMT).

Wouldn't it be good to say something re. the origins of SBN codes. I had always assumed that ISBN codes came into operation in 1970, because it is on this date that you start finding the numbers appearing on the books themselves. As to the SBN codes, I notice, for instance, that an SBN code appears on HSMO publications as early as 1959. Cf: http://www.amazon.co.uk/exec/obidos/ASIN/0116700750/026-0325148-9310042.

When and by whom were SBNs conceived? What is the story behind prefix '1'?

The ISBN page is a magnet for spam links. It happens that people will add items to the External Links section with no history comment and with no discussion on the talk page. When this happens, those links may be removed by other editors in regular cleanups. Please propose your new link here and explain its value (with reference to what is already available) if you don't want it to be cleaned out. ( You can help!)   EdJohnston 21:38, 25 October 2006 (UTC) reply

ISBN10 to 13 batch conversion There is a good free ISBN10-ISBN13 Conversion tool online. Difference between this tool and others is that it can handle batch ISBN conversions, unlike the others, that simply do them one-by-one. I thought it might be a helpful link. Dariyam 21:01, 1 May 2007 (UTC) reply

Mass conversion of ISBNs is unlikely to be of much importance to ordinary Wikipedia readers. (Have you ever needed to do a mass conversion, yourself?) Since we are all vigilant about spam, I think this one should be excluded. EdJohnston 21:14, 1 May 2007 (UTC) reply

List of ISBN publisher codes Here is a list of ISBN publisher codes for the English language (from http://blog.openlibrary.org/2009/07/20/isbn-publisher-codes/ where there is also a link for the publishing houses starting with 2 and 3. Fabrivelas ( talk) 13:36, 29 March 2010 (UTC) reply

Blogs are generally frowned upon, see WP:ELNO item 11. -- Redrose64 ( talk) 15:03, 29 March 2010 (UTC) reply

Since the ISBN check digit is a means of finding and correcting errors, it is of interest to know if and when the actual checks get performed. Does anyone object if the following words are added to the article?

Some book-ordering systems, like R.R.Bowker's Books In Print, will not even attempt to search for a book if given an invalid ISBN. It is clearly to the publishers' advantage to issue their books with valid ISBNs, although not all bookselling web sites will do the check. For instance www.amazon.com, as of October 2006, listed some books for sale that were published with invalid ISBNs. Essentially they are passing along to the customer without further ado whatever ISBN was assigned by the publisher. EdJohnston 20:05, 10 November 2006 (UTC) reply

I find this section very POV, with plenty of "shoulds" and even a violation of WP:SELF. Is there a source for all of these "shoulds" or "should" this section be removed from the article? Cleduc 00:36, 18 December 2006 (UTC) reply

Let me get this straight. A book merchant should profess ignorance when a potential customer asks them for a well-known (e.g. published) but technically invalid ISBN number? Are they in the business of serving the customer, or in being ideologically pure? Is there an Merchant Code of Conduct for Handling ISBNs published somewhere that says merchants should thus abuse their potential customers? Forgive the rant, but this is ridiculous. Cleduc 00:47, 18 December 2006 (UTC) reply

The point you are making, I think, is, 'Bad data conceivably still has some value'. It's hard to disagree with that, but the management of the bad data can become very tedious. Certainly the U.S. Library of Congress has attempted to do that, off and on, but the net result is you can never trust an ISBN you see in a record at the LOC to be valid. A clear policy of expunging the bad data may be easier to live with long term. If you add a check digit to the ISBN, what use do you anticipate making of it? If you let invalids flow through the system and go everywhere, how do you benefit from having a check digit?.. The 'shoulds' that you refer to above were (for the most part) not included in the main space article. EdJohnston 00:55, 18 December 2006 (UTC) reply

Yes. Obviously it has some value, people use ISBNs to find actual books. The records exist to find the books, not vice versa. Expunging bad data is all well and good -- but is the "final solution" to burn (sorry, "expunge") the books that offend the checksum? My point is that this argument places the cart in front of the horse: information systems exist to serve the purposes of their users. In either case, this doesn't belong in this article because it is pure opinion. I could reply with my own opinions, but no original research is something we should stick to. I guess there's no source for this? No "code of conduct" as I mentioned above?

I did overstate the "shoulds" -- there is only one in this article, in the second sentence: The next place where errors can be caught is at the bookseller: they should keep bad ISBNs out of their catalog, and not accept orders that request bad ISBNs. The final section which preaches how encyclopedias "can" make the world a better place carries the same tone, with "can" holding the place of "should." Furthermore it violates WP:SELF.

So, the section is definitely POV in content and tone, and does not cite sources for any of these opinions. Finally, though the policy surrounding invalid ISBNs is doubtless a very important issue, the Wikipedia article is a place to report verifiable information, not a place to try to influence public policy at the LOC. Cleduc 01:32, 18 December 2006 (UTC) reply

See what you think of my attempt to remove the 'shoulds', which happened in the latest version. The ISBN article (before my addition of the 'Actual Use' section), seemed to have no concept at all of what you might use a check digit for, which I thought left it somewhat unbalanced. Whether the balance is now redressed is something that can still be discussed. In terms of sourcing, I'm not sure how much further it's reasonable to go. If we assume that our users can click on a link, they can verify most of the observations for themselves. I could be more specific if you think that's useful. EdJohnston 02:07, 18 December 2006 (UTC) reply

The recent changes are a good start. The article still needs sources for each statement, because it is all still original research. No other encyclopedias include ISBNs that I know of -- this is an oblique WP:SELF so I'm going to pull it. Cleduc 02:01, 18 December 2006 (UTC) reply

It's not really a SELF becasue it refers to WP in a way that an article in another encyclopedia could. Rich Farmbrough, 23:55 9 January 2007 (GMT).

I trimmed it to include only short, relevant facts and reordered them in terms of usage -- there are many more libraries that don't filter their ISBNs (mostly running Dynix, unless I'm terribly wrong) versus three online booksellers. I removed all conclusions about what "should" be done -- perhaps someone has a source in a library journal about how invalid ISBNs should be treated as a matter of information hygiene? Cleduc 02:22, 18 December 2006 (UTC) reply

I thought that the format was one digit for the language, four digits for the publisher, four digits for the item number, and one digit for the check.-- Luke Elms 14:59, 22 November 2006 (UTC) reply

No, it's more complicated. See [3] for the region codes, known as group identifiers. 0 and 1 cover the English-speaking countries. The less common language areas get longer and longer codes. For instance, Armenia gets its own group identifier, which is 99930. This leaves fewer digits to distinguish the publishers, but there are not as many of them for the smaller language groups. EdJohnston 04:34, 29 November 2006 (UTC) reply

When adding an example two months ago an editor used the following language:

Some books have more than one language area code, e.g. A.M.Yaglom, Correlation Theory... , published by Springer Verlag has ISBN 0-387-96331-6 and ISBN 3-540-96331-6. Note that in this case Springer has publisher codes for both language areas, and chooses to use the same item number. The check digit happens to coincide.

There is more discussion of this on User_talk:Pascal666. Unless anyone objects I am planning to remove the phrase 'The check digit happens to coincide'. I actually don't see the problem with it, but I also don't see why it would have been considered advantageous by Springer in the first place. A coinciding check digit might increase the chances of mixing up two distinct ISBNs. EdJohnston 19:19, 8 December 2006 (UTC) reply

Actually, one advantage is that the item number and check digit don't depend on which prefix they get paired with. It's then possible to compute the check digit in isolation regardless of what prefix it's with, and so the check digit doubles as a check digit for the item number regardless of the language area prefix. A given pair of 4-digit prefixes will have this property if (10*x₁ + 9*x₂ + 8*x₃ + 7*x₄) mod 11 computes to the same value for both prefixes. -- 68.88.69.227 23:42, 21 April 2007 (UTC) Edit: Looks like someone else pointed this out a little further up. (I'm the guy that wrote the 68.88.69.227 comment.) -- Mr z 00:06, 22 April 2007 (UTC) reply

I added a link to OttoBib.com—–Free tool to generate an alphabetized bibliography from a list of ISBN numbers in MLA, APA, or Chicago/ Turabian format (with a permalink). I did not post it here first because it is an extremely useful tool for citing sources. Dhaluza 15:43, 17 December 2006 (UTC) reply

Has anyone ever thought of starting an ISBN Prefix List page with coresponding publishers? For example:

0-7645  : Wiley Publishing, Inc.
1-56592  : O'Reilly Media, Inc.
2-07  : Les Éditions Gallimard

And so on...

Bpg1968 01:49, 22 December 2006 (UTC) reply

I can see the interest of doing this, but imagine it could be criticized because of WP:NOT (a directory). Conceivably each publisher's ISBN prefix could be added to the Infobox shown in their respective article. See for example the one in John Wiley & Sons. That way, small publishers that might not be notable won't have to be included. Since the user can reach Special:booksources just by clicking on an ISBN, they can find out the publisher for a given ISBN in one or two clicks anyway. EdJohnston 19:33, 22 December 2006 (UTC) reply

This could certainly be useful in WP space. Also I would like to see 0-00 -> 0-19 in the WP article. Rich Farmbrough, 23:59 9 January 2007 (GMT).