Is there any significance to the patterns used by various publishers when they print the ISBN numbers on their books? They all (now) use a 13-digit number, but there seems to be no consistency across publishers as to how that number is broken up.
Some do it like this: ###-#-####-####-#
and some do this: ###-#-##-######-#
and others do this: ###-#-#######-#-#
and I’ve seen a few other variants.
So … why?
There’s is a significance to those patterns, but only by convention. Publishers get a designated sequence (a number of digits following the country/language code) and can use the entries within that sequences as they see fit, except for the last number, which is a checksum. The dashes are just there to make it easier to read the code by eye. The barcode accompanying the numeric code does not have any equivalent to the dash - the ISBN number is purely determined by the digit sequence.
See ISBN - Wikipedia
ETA: I forgot that in 13 digit EAN codes, there is an additional leading 3-digit code that specifies the industry. Those numbers are used for more than just books - they’re universal barcodes. See International Article Number - Wikipedia
The number is notionally composed of a publisher prefix and an item suffix (plus check digit) - there are long publisher prefixes, which have a small number of digits to denote the item - and these are issued to small publishers (because a smaller number of different items can be described in the possible permutations of the shorter item suffix)
and there are short publisher prefixes which have a larger item suffix to describe a larger range of different items.
The dashes supposedly separate the different portions by their meaning.
Except that I don’t think the above system was adhered to in the end, and partial number ranges just ended up being issued as required.
For the most part, we don’t even use the dashes anymore, but…
The original 10-digit ISBNs (International Standard Book Numbers) were assigned in blocks. They were broken into four parts. The first digit identified country of origin. The second block identified a particular publisher. The third identified the specific book. The fourth was a single check digit (only this final “digit” could be an X–all others were actual digits).
Big publishers got a very short ID, leaving them lots of space for books. If the publisher ID was 3 digits long, that allowed 5 digits, capable of identifying 100,000 different books. When I set up my first tiny little publishing company, they gave me a 7-digit publisher ID, and only one digit to identify my entire product line (2 books, at the time).
The industry ran into a problem because (a) it was a pain having barcodes for both ISBNs and EANs (the 13-digit codes) on books that were sold in grocery stores and gift shops and (b) we were running out of ISBNs. So three more digits were added, and the ISBN system was folded into the EAN system. You’ll notice, if you look at any book still containing both ISBN and EAN, that the EAN will start with 978, then have the first 9 digits of the ISBN, and then end with a check digit (typically different from the ISBN check digit).
If that doesn’t answer the question, Bowker (the company that assigns ISBNs) has a whole lot of information on their website, and I could probably dig you up an explanation of the formulas used for the check digits pretty quickly if you need it.
You also have to take into consideration the length of the country code. If it has two digits a publisher ID with three digits leaves four digits to the book ID and so forth.
Thanks for the responses. I’m confused, though, by the part about giving differently-sized publishers different numbers of digits for their publisher ID and book ID. That makes sense at first glance, but … if the dashes were really just there to visually break up a long number, and the actual 10-digit number was all that really mattered, couldn’t that lead to duplication of numbers?
Say we have a big publisher with an ID of 123, and a very small publisher with an ID of 12345. Without the dashes, what’s to distinguish between the big publisher’s book # 45678 and the small publisher’s book 678? With the dashes they might look like this:
Big pub: 1-123-45678-1
Small pub: 1-12345-678-1
… but as plain 10-digit numbers, they’re both 1123456781. It just seems like it would be too easy for something like that to happen.
There are no long publisher IDs assigned which are the beginnings of short publisher IDs.
For example, for 978-0 (part of the English Language range) it’s
978-0-00 to 978-0-19
978-0-200 to 978-0-699
978-0-7000 to 978-0-8499
978-0-85000 to 978-0-89999
978-0-900000 to 978-0-849999
978-0-8500000 to 978-0-9999999