Tuesday, March 5, 2013

Open Access: A World Without Database Vendors?

As a thought experiment, let's say we "win." Professional and academic associations go open access, as much of physics has. The Directory of Open Access Journals is able to capture the far majority of these newly free works, and in turn these are snapped up by library catalogs thanks to link resolvers and discovery services. The same happens with the Directory of Open Access Books with regards to chapters in edited volumes.

But there's a catch: DOAJ's search function is not, to put it politely, robust. And there's a larger problem behind search functionality thanks to incomplete metadata. Link resolvers and discovery services that pull from that search, culling that metadata, will lead to frustrated end users who cannot access and discover what they're looking for.

In addition, the DOAJ is overrun with new items to catalog in this scenario, creating a backlog of epic proportions.
Stack Of Books
This is not my desk, but it feels like it sometimes. 
There are roles for vendors in this universe: generating better metadata for these newly open access items; designing stronger, more relevant search functionalities; and creating attractive and user-friendly platforms; among others.

There's a less maximalist, more realistic, "winning" option here, too: more journals, and more publishers, allow for pre-prints to be housed in institutional repositories, which are cataloged by member institutions, and perhaps shared via consortia, and with a wider audience via interlibrary loan.

However, these repositories do not solve access problems, in fact, they exacerbate them by creating not only a patchwork network of databases, but also at least two discrete classes of items: pre-prints and final products, each with a different symbolic value attached to them despite containing the same information. Vendors can solve the first of these, a coordination game, by designing, creating, and implementing databases that allow access to the pre-prints among and between libraries, including negotiating licenses and usage rights with publishers. Only vendors have, as of now, been able to create inter-library databases with robust network effects and positive externalities. That is, the more libraries, the more repositories, that join an inter-library database containing items that might otherwise only be found in a repository, the stronger, the more useful said database is. Some of those vendors in our current information ecosystem were founded by libraries, such as OhioLink, JSTOR, and OCLC, as well as The Digital Public Library of America (perhaps), though they have now taken on lives of their own.

As for the latter, pre-prints versus final products, I have no solutions other than to hope that academics "get over it" by incentivizing open access. Giving preferences to OA publications as part of tenure and promotion, for example, would be an important and powerful signal.

Thus, academics hold their chains in their hands, but there are no worlds, no futures, without database vendors. I write this not only to reassure vendors, but also to argue that we librarians are inexorably tied to vendors, even knowing that vendors do not always behave as we wish them to.

1 comment:

  1. Interesting observation regarding "two discrete classes of items: pre-prints and final products," I'd never really thought of that as a potential problem down the road as more final products become OA. Reading the link now.

    I've been thinking lately that one could build a pretty sizable OA search engine out of all the disparate sources, but so many don't expose their data in an API (most egregiously DOAJ, unless I'm mistaken). Even if they did, normalizing metadata would be a valuable service. OAIster does a lot, but perhaps not everything, here.