WHAT DO WE WANT? *Immediate decisive action*
WHEN DO WE WANT IT? *Sometime quite soon I should think…*
It’s always the things you think are going to be straightforward that end up being riddled with detail and complexity. Choosing a software platform for the repository was a case in point.
It’s kind of like re-inventing the wheel. You know most institutions with a repository have been through this scenario. Having a quick reference guide comparing the software available would be fantastic – not just technical specifications but real selling points of one system over the other. This came up at the RSP summer school as a lot of people were in the same boat.
Just for the record, we found these documents to be really useful:
At any rate, we short-listed EPrints and DSpace for comparison, as we wanted something ‘out of the box’, open source and with a good user community. Fedora would have offered a terrific solution, but unfortunately we ruled it out as we wanted something fast with minimal person-power to get it underway. I notice that some of the more mature repositories in Australia are moving to Fedora with various interface options, so something to keep an eye on.
At the end of the day one of the key criteria that swayed our decision was integration with existing IT systems and expertise. Apparently we’re java based with better Oracle infrastructure and a SOAP web interface would map well to other applications (I feel like I’ve just started speaking another language..).
The ideal solution would have been a platform with a robust user community (in the UK?), good integration with existing systems, and one to which the newly developed Scholarly Works Application Profile could be applied. I think we’ll need a FRBR based model as some point, particularly if we look at linking up datasets down the road.