Friday, January 21, 2011

eBook ISBN Uncertainties - or, Is the ISBN Outdated?

With the rapid transition from print to digital underway, a wide variety of issues have arisen - from multiple product formats to digital rights management and licensing - many of which have yet to be worked out in anything but rudimentary form. One element underlying all these issues that has bubbled to the surface lately is the waning relevance of the International Standard Book Number in the digital age. In an online scenario where no physical barcode is actually scanned and the definition of what even constitutes a "book" spans the gamut from simple text to high-tech multimedia presentations, not only is the usefulness of the ISBN as an identifier debatable, but its capacity to even fulfill that function is coming into question.

In 2007 the ISBN was updated from a 10 to a 13 digit code in order to accommodate a wider range of data. However, this may very well be a case of far too little far too late. Amazon, for example, had already by this time given up on the traditional identifier for its own uses in exchange for an internally generated ASIN (Amazon Standard Identification Number), seen not just on ebooks, but on every product in its store. And while most items on Amazon carry both an ISBN and an ASIN, no Kindle edition has, in fact, ever had an ISBN assigned to it.

In part, this is due to the inability of the ISBN to distinguish variant ebook formats of the same title from one another in a segment of its code without having to assign an entirely unrelated ISBN to it, as if it were a different title. This is, of course, how the hardback / paperback / audiobook division was handled, but with an almost unlimited number of variations of format and content possible with digital, this becomes burdensome at best. Not only are there a half dozen major proprietary and open-source software formats in use, but there can be additional interactive content in each one, such as embedded audio and video, hyperlinked and touch-interface capabilities, different rights management encryptions employed (or not), lending-enabled editions, scanned pages versus reflowable text formats, and a wide open horizon of endless future possibilities as new media are developed and incorporated. The question then becomes what exactly is an ebook, and at what point is one version significantly different from another so as to require a unique identifier. Without a clearly defined set of criteria, and the means to differentiate them in code, no set of numbers can mean anything useful to a broad overall group involved in its production and distribution.

An additional concern is the role - and even identity of - the publisher of an ebook edition. Digital titles published for the Nook format, for example, are given an ISBN which identifies Barnes & Noble as the publisher, regardless of who actually published the print edition (mine was given a new one even though I'd already assigned one to it). This technically means that B&N "owns" the ISBN for that edition, and not me as the author/publisher. This is important because what, in fact, separates a true self-published author from an author whose book is produced by any of the many ebook distributors who reformat digital editions is the actual ownership of the ISBN: whoever owns the ISBN is the "publisher of record" (even though the rights to the work itself may remain with the author or other rights holder). Consequently, a single title in ebook format can now have many separate publishers, some of which are little more than retailers while others are megalithic literary conglomerates, a sign of how the roles are blurring in the digital age. The distinctions are obscured even further by the fact that a buyer of an ebook does not technically own the book they buy, but only purchases a license to read it.

The usefulness of an internal system by an online retailer is understandable, but the assignment of random unrelated identifiers to what is essentially the same title brings up a range of problems for everyone else, from libraries who need to catalog a title to the author who wants to track their sales across multiple platforms. Metadata conflicts are a growing concern, as voiced in the recent findings by the Book Industry Study Group (BISG), who have undertaken a research project to identify and define ebook classifications with an eye to revamping the ISBN system. For more on the issues involved, you can download a pdf Summary of BISG Report Findings (1-13-2011), which outlines their initial results and recommendations. As an interim solution, in November of 2010 the International ISBN Agency released an updated set of Guidelines for the Assignment of ISBNs to eBooks (including "apps"), in the form of a 13-point F.A.Q. which is useful for any author or publisher to read. 

Ultimately, as with many things, it may simply come down to a question of cost, as ISBN blocks are not only relatively expensive to purchase outright from Bowker, the U.S. agency tasked with selling and recording ISBN data, but are now often given away for free by ebook retailer-publishers (such as Barnes & Noble) or for a minor fee (as the $19 fee charged by BookBaby), or they are simply eliminated altogether (e.g. Amazon, the major ebook publisher worldwide, with whom Bowker is understandably unhappy). But unless the price of an ISBN is all but eliminated (and soon), Bowker's days as caretakers of the univeral literary cataloging system are numbered (and in single digits, too).

Current costs (or "processing fees" as they're called) to purchase ISBNs outright from Bowker are $125 for a single number (which does not, by the way, include the actual barcode graphic - that costs $25 more), or $250 for a block of ten (which is what I bought), and $575 for 1000 (it goes up from there, but at that point you're not wasting your time reading this). Believe it or not, these are the recently reduced prices designed "to accommodate the digital identification needs of authors, publishers, libraries and the supply chain at large." A year ago a block of 1000 was $995. But even at a nearly 50% price reduction for that many, who's going to buy something that someone else is giving away for free? Especially when the highest prices are being charged for the smallest quantity. If anything, they should all cost the same regardless of how many you buy, since (like ebooks) there isn't anything physical actually being produced. But then, if you want to be the actual publisher of the book, you have to pay their fee.

In the end, the only identifier that will really work is one that incorporates all the new relevant data that surrounds a digital edition of a book, both external (production data) and internal (content data). Anything short of that is irrelevant.