Resolving Public Libraries at Amazon

Peter Brantley -- September 4th, 2013

Ohio Univ's Online CatalogThe end of summer and the start of school; a last round of easy dinner parties and sleepovers for the kids. Such times lead me into conversations about ebooks, ereaders, and why things are the way they are, instead of some other way. Explaining to friends that no, they can’t lend an ebook to their wife in most cases; no, there’s no easy way to set up a shared or group account; and yes, they might be able to get a book from their library, but the selection is incomplete and the process sometimes difficult – it all seems like an old refrain at this point.

The explanations to academic friends make the contrast with that environment jarring. The way it should work for consumer books and libraries is analogous to the way it works for academics. A reader should be able to identify themselves with a particular library, which has in turn licensed access to a variety of digital books and journals, and then regardless of where the reader searches, they should be directed to the copy that they have access to through their institution. At the root of this reference linking service is a link -naming syntax and protocol called OpenURL, invented by Herbert Van de Sompel and Patrick Hochstenbach when they were working at Ghent University. It became widely available through the SFX service marketed by Ex Libris, and it is now supported by every academic aggregator.

Essentially, OpenURL is a mechanism by which URLs for target resources, like articles or digital books, are presented as a sequence of known metadata, such as ISBNs, author names, titles, and publisher information. This data is then looked up in a central repository called a Knowledge Base (KB) held by the university or corporate library that keeps track of which publications it has licensed access to, and from where. This is helpful because it is not be unusual in a large library for a user to be able to obtain the same article via different aggregating services with overlapping licenses – e.g., it might be available in full-text in one location, but only as an abstract from another. The Knowledge Base works with the OpenURL resolver to show the user the best possible location.

Of course, this works only so far as the user taps into a resource discovery service that is OpenURL-aware. In the academic software market, OpenURL has become such a fundamental construct that its support is ubiquitous. However, should a user from the University of California, for example, search Google, then they would be directed to whichever copy of the article that Google wanted to promote to the user, rather than the copy that the University of California has already purchased on behalf of its faculty and students.

Except that Google, Microsoft, Yahoo and many others support OpenURL in their academic search services. Google Scholar partners with academic institutions to deliver the most appropriate copy to their users. The library grants permission for Google to index its KB, and Google permits its users in turn to identify themselves with one or more institutions. For example, I might have a joint appointment at Berkeley and Stanford, and I could select both those institutions in Google’s Scholar configuration. Then, when I search for an academic article, Google Scholar consults both the UC and Stanford KBs and shows me the best places to obtain it.

This “works” for libraries and publishers because I cannot illegitimately claim to be from Stanford and get access to the Stanford library. Google connects me to the best article source negotiated by the Stanford Library, but I still have to demonstrate to Elsevier, Springer, or ProQuest that I am actually a member of the Stanford community: either by being on Stanford’s internet, or by logging in through a proxy or authentication service that provides the equivalent assertion.

There’s no reason this couldn’t work for consumer books. I should be able to identify myself as a member of the Berkeley Public Library (BPL) community to Amazon, and when I search for a Kindle book, Amazon could tell me if I could check it out from the Berkeley Public Library. Just like Google Scholar, Amazon could permit its Kindle users to affiliate themselves with their community library, and if the public library agreed, Amazon could index a list of the digital books and resources that the public library has licensed. This might be in the same KB format that academic institutions use, or it could be a simple OPDS list of titles with links in the OpenURL syntax. If I searched for “Treacherous Paradise” by Henning Mankell at the Kindle store, Amazon would tell me that I could borrow the copy at the Berkeley library. I would then have to authenticate myself to the Berkeley library in order to borrow it. Berkeley could signal Amazon once this was accomplished, and Amazon would provide me the book under the licensed terms (such as a two week borrowing period).

This doesn’t break any economic models because enabling discovery of the library’s digital collection doesn’t suddenly enlarge it. If Berkeley Public has licensed only one copy of Treacherous Paradise, then I may have to be put on a wait list if someone else has already checked it out. If Amazon endorsed OpenURL KBs, it would merely raise the visibility of the public library’s collection and the breadth of its use, but it would not negatively impact the publisher. Indeed, the net effect could be positive, because Amazon and BPL would both know when a particular book was in high demand, and could obtain additional licenses for it.

Of course, this could also work for other ebook retailers, such as Barnes & Noble, Kobo, or an independent purveyor. It is most tractable for Amazon because of Amazon’s decisions in library provisioning. Generally, Overdrive, 3M, and other intermediaries provide the equivalent of a direct resolution service for digital books. Readers are forced to search at either their library’s aggregator site, such as Overdrive, or on the library’s web site to located digital books. However, Overdrive passes borrowers directly through to Amazon. Overdrive merely acts as a licensing agent for Amazon books and doesn’t provide access to them – a situation unique among ebook vendors. A library should therefore be able to provide a list of its licensed collection to Amazon, permitting access to users searching there.

Of course, ebook aggregators such as Overdrive might prohibit the release of a library’s digital catalog to third parties, but that would be both awkward and potentially unenforceable. It also does nothing to protect reader privacy, since Amazon will wind up with data on which books its users are borrowing – a concern poignantly raised by Gary Price in InfoDocket. However, using OpenURL is one example of how public libraries and commercial partners can better use existing Internet protocols to deliver better services without additional costs.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>