Identifiers
From The Big Picture Wiki
Identifiers
I’m consulting to a company that distributes e-audio books, and onsite we’ve been talking a lot about how useful ISBNs are. I’m also chairing the BISAC Identifiers Committee, where we engage in rambunctious conversations about what the best identifier is for this product and that – conversations that get more rambunctious with the introduction of digital products for sale. And over the last week or so, all of these discussions coalesced for me into some cogent thought (or, at least, I hope it’s cogent).
As more and more book content is available to consumers, the question of how to identify it becomes more important. And here’s why: so long as a company is offering digital products for sale on its own website, identifiers are only important insofar as that company can track sales. But when that company begins exchanging information with other companies – sending out data feeds for distribution on other websites – it’s crucial that the products get identified in ways everyone can understand.
That sounds kind of obvious, but in practice it appears to be anything but. For example, Amazon is offering short stories and other small digital works for sale. How are these identified? By Amazon’s proprietary ASIN – an alphanumeric identifier that identifies everything on its website from tee-shirts to shoes to electronics…to “The Hope of Elantris”, a short story by Brandon Sanderson.
But that ASIN is only good on Amazon’s website. What if Brandon Sanderson wants to sell his story on other websites? How will his publisher identify it? What if Lightning Source wants to provide “The Hope of Elantris” to Barnes & Noble? How will the publisher track those sales?
The same is true for coursepacks – compilations of digital material sold to students to accompany textbooks or online courses – or chapters of an audiobook.
As content distribution becomes more and more digital (and there’s no forcing that horse back into the barn), the question of identifying that content in the supply chain becomes ever more crucial – and we are in the early stages, so of course that conversation’s pretty fragmented. Some folks have brought up the DOI (Digital Object Identifier) as a possible identifier for these chunks that don’t comprise a whole book in themselves.
And the DOI is quite useful. It’s a great tool for identifying something at any level of granularity – a book, a chapter within a book, a picture within a chapter within a book. But the DOI does something that other identifiers do not do – it also resolves. It points to the location of that digital work, or summons metadata about that work – it doesn’t simply identify the work. It functions. It acts.
And as far as e-commerce is concerned, that’s a little confusing. You don’t want an identifier in your database doing anything but identifying. In the words (which I aim to make immortal) of George Wright, Sr. (of PIPS), “Let the identifier identify; let the database describe” – each point in the supply chain does different things with identifiers for given works; their databases describe relationships using those identifiers in different ways. If instead of an ISBN, which merely says, “This edition of this book is recognizable by this number”, you have a DOI which says, “This edition of this book has this metadata associated with it and can be found at this URL” – quite a number of e-commerce websites are going to start screaming: “Too much information! We just need to know the number that goes with this piece of content!”
(In my opinion, the very term DOI is a misnomer – it should be called a Digital Object Locator or Digital Object Resolver rather than a Digital Object Identifier.)
Meanwhile, it’s worth remembering that ISBNs should be assigned to all formats of a title – even the various digital ones. In other words, according to the ISBN standard, a PDF gets a different ISBN from an HTML version of the same document. Yes, that means publishers have to buy more ISBNs. No, there are no shortcuts. And yes, different compressions of downloadable audio get different ISBNs, just as the CD version and the cassette version do.
That’s about all we know regarding identifying digital book content at the moment. No obvious solutions leap out, and the Identifiers Committee can look forward to a lot of interesting discussion about it.
--Ljnd 14:17, 11 June 2007 (EDT)
