On the other hand, the recent announcement at Open Repositories 2010 that the DSpace and Fedora software platforms are to be merged by 2011 seems to leave a fair amount of doubt about the future of DSpace as an independent platform, as Fedora is intended to be the underlying storage layer. Since the DSpace web interface (either JSP UI or XML UI) is ageing, this leaves relatively little in the longer term except perhaps the DSpace workflow, so that the new DuraSpace platform would have the packaged product and commercial presence that DSpace currently enjoys but the underlying flexibility and functionality of Fedora.

It remains to be seen how this will turn out in practice, but it seems like a reasonable bet that development resources will not be focussed on support for complex relationships and entity-relationship models unless it is shown that there is substantial demand. However, if all goes to plan by 2011, the new DuraSpace will have this functionality because it will be inherited from Fedora, which is a longstanding feature.

It will be interesting to see how this development will increase the competitive presence of DSpace > DuraSpace installations compared to EPrints installations in the HE sector, and the implications of radically increasing the accessibility of Fedora to a wider market. At present, it seems as though this can only be a good thing.