ISBN or EAN-13 as e-book identifier?
By Roger Sperberg, New York Editor for TeleRead
In the United States, small publishers have a significant levy placed on them if they obtain an ISBN for every format in which they issue an e-book. A publisher like Random House might pay only five or ten cents for each ISBN it assigns. On the other end of the spectrum, a new e-book publisher must either pay $1120 upfront for one hundred ISBNs or $325 for ten at a time as it goes along. As we all know, there are vastly more than ten e-book formats so this is a sticky point.
Bookstores have long declined to sell print books without an ISBN, a reality of entering the book-distribution chain that new commercial ventures have simply had to accept. For physical objects — items that pass across a checkout counter’s barcode scanner — there’s no getting around it.
But inventory management is irrelevant to e-books, which aren’t barcode scanned, and so an ISBN isn’t required to provide a permanent and globally unique identifier. Many e-book publishers have found it simpler to piggyback on the internet’s domain-name system that guarantees unique web addresses and just utilize a URI as their GUID. After all, registering a domain ensures that no one else is using the prefix you begin with and URIs can be valid without referencing an actual web page.
Unfortunately, URIs are not inherently permanent and domain ownership lapses all too readily so this approach has problems. Perhaps if permanent URIs, such as those provided by PURL.org had been widely utilized, we would not today be hearing a clamor for applying different ISBNs to every format an e-book is issued in. But they weren’t, and the squawking of new and small publishers at the expense entailed more than matches that clamor.
For some years, Jon Noring has suggested establishing a registry for e-books. which would of necessity assign suitable GUIDs. Getting mandatory registration, though, is impossible and voluntary compliance from every publisher everywhere unlikely.
Monday, Teleread ran a piece by Elizabeth Burton decrying the unequal effect of requiring a small press in the U.S. to churn through these expensive ISBNs (whereas next door in Canada, publishers pay no charge at all for ISBNs). Discussion about ISBNs broke out anew in other places, including on the Reading 2.0 list.There Jon brought up a new twist to his suggestion that has a distinct stroke of brilliance.
The current ISBN-13 code is a conforming EAN-13 code, Jon pointed out. By utilizing EAN-13 similarly, e-book identifiers from a new registry would be compatible with ISBN-13.
And being compatible, an e-book-only identifier could be placed in the field intended for ISBNs in a sales or inventory program without causing anything to break in a bookseller’s database. (Even though such overloading isn’t necessarily advisable.)
Additionally being an EAN-13 number, this identifier provides the guarantee of uniqueness and vendor registration a retail business needs (the famous “if it doesn’t have an ISBN, they won’t be able to sell it in bookstores” dictum actually means EAN-13 today).
What’s more, a U.S.-ebook EAN-13 identifier would never duplicate an ISBN code. So publishers could stick with an all-ISBN system or mix the two as they find most convenient.
Let’s see: retailer friendly, affordable, complementary to and not overlapping the existing ISBN registry, globally unique and permanent, non-proprietary — this sounds like a good answer to the need for e-book identifiers to differentiate format within the larger “electronic” designation.
And with it, then booksellers, libraries, content repositories, and us content consumers could reasonably demand that all content producers and packagers provide the central piece of metadata that all of us want to be included in any book distributed electronically. Use the ISBN system or use this alternative, we can say, or don’t be downloading your untrackable files into my e-reader.
As per Jon, a completely new registry agency could be set up to distribute EAN-13-conformant e-book identifiers. However, as part of its intended role to further e-book standards and interoperability and keep our industry from factionalizing, I wonder why we do not expect IDPF to take this on.
Other than its post-dotCom-era lassitude, is there any reason we shouldn’t turn to IDPF? As a believer in metadata’s pre-eminent role in internet discovery, I’m convinced that permitting e-books to be issued without permanent and globally unique identifiers portends disaster.
IDPF should act promptly to forestall this. Who better to cope with this situation? It’s what a trade organization is for.
June 4th, 2009 at 2:52 pm
I think we need to look much more closely at this situation.
As has been pointed out, e-books are not physical matter, so most of the constraints and requirements of physical sales do not apply. Most of the rationale for ISBNs is directly related to physical sales… and although they could be applied to electronic sales, they are not necessarily the best tools for them.
Physical sales also tend to be more centralized and interconnected, again requiring tools like ISBNs to keep things straight. In today’s more distributed web-based sales world, this is not a constraint. For example: I sell my e-books from my own website; I could arrange for ads or referrers from other sites; and only I need to worry about tracking my sales. And with a fairly simple system in place, a Paypal storefront and a delivery script, I can do exactly that right now, without needing ISBNs, or worrying about whether anyone else can read my figures.
Common identifier usage, based on centralized sales of physical content models, would mostly benefit traditional sellers using traditional methods to sell non-traditional products. Fine for them… but they are not everybody, and their needs do not need to be imposed on those who are looking forward.
I am more in favor of developing new selling tools and methods that better fit the less centralized, less constrained Digital era of the future, not using jury-rigged tools and methods from the Physical era of the past.
June 4th, 2009 at 6:11 pm
I’m sure I dont’ grasp all of the ramifications of ISBN/EAN-13 usage. However, assuming that someone like the IDPF does setup an EAN-13 registry and that the cost for this service is either free or very minimal, then using EAN-13’s for ebooks sound like a good idea to me. If it did nothing else, it would gurantee the unique identifier that epub and even older OEBPS formats (mobi, lit) need.
This may not be something that Steve Jordan or others think they need, and that is their choice not to use it. Myself, I have used a web-based UUID generator on the few public domain books that I have converted to ebooks. But this method certainly would not suit the purposes of many others.
I haven’t looked into the intricacies of ISBN or EAN-13, but I do have a question. Is there a particular numeric sequence, or some other easy way to distinguish an ISBN from an EAN-13? If so, then it should be fairly easy to make a software change to allow the use of one data field for both, while still separating them out for other processing.
June 5th, 2009 at 5:15 am
ISBN-13s are EAN-13s. The EAN authority have allocated prefix 978 and 979 to ISBNs – any EAN-13 starting 978 or 979 is also an ISBN.
The old 10-digit ISBNs (9 significant digits and a check digit) all have an equivalent ISBN-13 that is the original 9 significant digits with 978 prefixed and a new check digit at the end.
So you might be able to persuade the EAN authority to allocate a new prefix for eBook EANs, but then you’d also need to set up a numbering authority for ebooks to allocated numbers to publishers, essentially duplicating the ISBN authorities, although perhaps using just one global authority rather than lots of country specific ones.
It might be possible. But the new ebook EAN-13s are unlikely work with existing systems expecting ISBN-13s. (I would expect many of the systems that accept ISBNs will reject any that don’t begin 978 or 979.)
I think that small (or large) publishers worried about the cost of ISBNs for multiple formats of ebooks would be better off abandoning the practice of selling each format separately, and instead selling customers a bundle of all available formats. Then only one ISBN is required per ebook, not per format.
And if DRM costs involved with each format make this uneconomic, perhaps they’d also be better off not wasting their money on DRM systems that only annoy their paying customers, not the people who download books for nothing.
June 5th, 2009 at 4:27 pm
Last time I looked, a single ISBN could be used as a base identifier, with the various formats separately identified by a local modifier that was controlled by the retailer.
Bowker has always claimed that each format needs a separate ISBN, but the actual marketplace has been going in the opposite direction. Amazon, for example, has now entirely stopped using them for Kindle books. This is not to say that Amazon-style proprietary solutions are the answer.
The price of ISBNs has been rising much faster than inflation for many years. I doubt the current pricing model would survive a letter from a class-action lawyer identifying monopoly-pricing issues.
June 7th, 2009 at 1:07 pm
As I understand the basis for the unique-identifier policy, it’s driven mostly by non-retail participants, especially libraries, who rely on ISBNs for cataloguing purposes. However, it seems as if the issue wasn’t really studied, but rather that the existing system was expanded to include it because that was the easiest route.
So, simply bundling all formats and selling in, say, a ZIP file, wouldn’t resolve the problem because it goes beyond just sales.