Time for a Digital Divide: DL top-down, OA bottom-up

Richard Poynder, the astute, eloquent chronicler of scholarly communication in the online era has done it again, with a shrewd, original and insightful review of the short history of the institutional repository movement.

His conclusions are surprising, but (I think) very apt. His analysis, among other things, goes some way toward explaining why on earth a "Repository Comparison" such as the one Rachel Heery cites below, would have left the first and most widely used Institutional Repository (IR) software (GNU Eprints) out of its comparison.

The answer is simple: Eprints is and always was very determinedly focused on the specific goal of 100% Open Access (OA), as soon as possible; it can of course also do everything that the other IR softwares can do (and vice versa!), but Eprints is focused on a very particular and urgent agenda: generating 100% OA to each institution's own research article output. Those who prefer leisurely fussing with the curation/preservation of arbitrary digital contents of any and every description will of course have plenty to keep them busy for decades to come. Eprints, in contrast, has an immediate, already-overdue mission to fulfil, and it is becoming clearer and clearer that -- with some prominent and invaluable exceptions -- the library community has found other rows to hoe.

Richard has accordingly proposed that it might be time for a parting of paths between the Generic Digital Curation/Preservation IR movement and the OA IR movement, and he might be right. One has a diffuse, divergent goal, the other a focused, convergent -- and urgent and immediately reachable -- goal, a goal that might now be hamstrung if it is subordinated to or subsumed under the diffuse, divergent goal of the other.

On Thu, 2 Mar 2006, Rachel Heery wrote in JISC-REPOSITORIES:

Thom Hickey's blog post on comparison of repository software features may be of interest, for the features considered as well as the 'scores'... He does this comparison on a feature by feature basis in categories: Data Support, User Support, and Miscellaneous Infrastructure. Thom invites feedback, some people in the repository programme may be able to provide that for Fedora and DSpace

I think the optimal strategy is latent in Richard Poynder's very timely and perspicacious article. We should especially recommend using Eprints modularly, at the departmental level, via computing-services and/or library support, along the lines CalTech are doing it with CODA:

Instead of building one monster-archive, Dspace/Fedora style, and then partitioning it top-down into "communities", CalTech have made natural and effective use of the OAI interoperability to create lots of Eprints modules, all harvested and integrated bottom-up into CODA. The rationale to be stressed in this is that this easy, light modularity can be used to get OA-specific archives going even in institutions that are slogging away at their own monster-archive, in parallel:

Let the OA-specific movement and focus proceed full speed in its very specific target direction (100% OA, ASAP), supported by institutional self-archiving policies/mandates; and plan on integrating the (one or many) OA archives with the monster if/when it becomes desirable to do so. But let the two proceed and grow at their own pace for now, and in their own direction, rather than letting the monster slow down, hold back or divert the OA modules and specific, targeted OA growth.

We certainly should not alienate the library community from this; there should simply be a division of labour: Let those in the library who are generic digital curation/preservation-minded devote their time and energy to the monster, and let those who are OA research-self-archiving-minded devote their time and energy to the OA modules.

(Of course what CalTech still lacks is an institutional self-archiving mandate! I will soon make the Southampton policy recommendations generic (removing the partisan puffery!) so other universities can use them in their own efforts to implement a mandate. With Arthur Sale's permission, I hope we can couple this with the masterly risk-assessment document he is currently drafting!)

Stevan Harnad