ECS developers win $5000 repository challenge
Developers from ECS, Southampton, and Oxford University won a $5000 challenge competition which took place at the OR08 Open Repositories international conference.
Dave Tarrant, Tim Brody (Southampton) and Ben O'Steen (Oxford), beat a large field of contenders, including finalists from the USA and Australia, by demonstrating that digital data can be moved easily between storage sites running different software while remaining accessible to users (watch video). This approach has important implications for data management and preservation on the Web.
Repository sites have become a global phenomenon in higher education and research as a growing number of institutions collect digital information and make it accessible on the Web. There are now over 1000 repositories worldwide.
However, with the growth of institutional repositories alongside subject-based repositories, and in cases where multiple-authors of a paper belong to different institutions, it is important to be able to share and copy content between repositories.
Meanwhile the repository space has become characterised by many types of repository software - DSpace, EPrints and Fedora are the most widely used open source repository software - containing many different types of content, including texts, multimedia and interactive teaching materials. So although sharing content and making it widely available (interoperability) has always been a driver for repository development, actually moving content on a large scale between repositories and providing access from all sources is not easy.
The OR08 challenge, set by the Common Repository Interfaces Group (CRIG), had just one rule for the competition: the prototype created had to utilise two different 'repository' platforms.
The winning demonstrator showed data being copied simply from an EPrints repository to a Fedora repository, and then moved back in the other direction. What was striking is that among repository softwares, EPrints and Fedora are seen as being quite different in the way they handle data, so the approach used is likely to be just as useful with other repository softwarel.
This data transfer was achieved using an emerging framework known as Object Reuse and Exchange (ORE), a topic that attracted one of the highest attendances at OR08. ORE is yet to appear in beta form, but specifications are being developed that allow distributed repositories to exchange information about their digital contents.
According to Dave Tarrant, Ã¢â¬ËInteroperability is the innovation. We think it is a bad idea to reinvent the wheel so with the availability and support for ORE growing, this provides a very suitable technology to provide interoperability between repositories.Ã¢â¬â¢
The winning team are past and present members of the JISC Preserv 2 project that is investigating the provision of preservation services for institutional repositories, and will take this work forward in the project.