Skip to Content

DuraSpace

This post was originally published at http://talatchaudhri.wordpress.com/2009/05/15/duraspace/

The recent announcement of the merger of DSpace and Fedora Commons as DuraSpace is potentially a very significant advance in the repositories sector. Although the two platforms will continue to exist as separate entities, they will no doubt collaborate to their mutual benefit in technical development. In addition, new software products such as DuraCloud are described.

I have personally found DSpace to be an effective and flexible platform, although until version 1.5 it was missing some fundamental functionality that meant it was overall an inferior product to EPrints. However, I have always said that if certain issues were sorted out, such as a more granular permissions system, versioning and so on, it is otherwise as good overall as EPrints. (I particularly appreciate how easy it is to see full metadata records in DSpace, from the point of view of research, though this is an entirely trivial technical point - it just happens to suit my work in my present and previous post!)

DSpace is clearly second to EPrints in terms of market penetration, but it is the only other major competitor to enjoy such a sizeable market share. Fedora is third not on its technical merits but largely because it is not a "packaged" product and requires much more customisation. It is evident that both platforms have much to gain from the collaboration. I would bet that the EPrints people may well have cause to worry about their future market dominance, given this development.

I'm particularly interested because Fedora makes much greater use of RDF, a technology that has its supporters and detractors, but has not been the basis for a wholesale change to the promised Semantic Web that might have been hoped. However, one can see the potential application within content management systems such as repositories. One stumbling block seems to be that triple stores are not particularly efficient databases and need significant optimisation efforts before they rival traditional relational databases, a subject on which I am not a great expert at present. (I thank a colleague at UKOLN for educating me on this.) It is particularly interesting, then, to note the reference to efforts to improve the triplestore-based storage layer Mulgara.

I'm awaiting further developments with considerable interest, noting the new version DSpace 1.5.2 and recent references to the planned versions 1.6 and 2.0. I wonder how much the repositories community will have changed in a year's time? Things seem to be moving fast right now.



Dr. Radut | blog_post_tf