At a session today about the new version of the EPrints software – due for launch at the end of January 2007 (in San Antonio). We’ve been using EPrints at RHUL for 2 years as part of the SHERPA-LEAP project. We currently use EPrints v2, which is hosted for us by UCL.

We are currently looking for a piece of repository software which can be used for institutional research output (which is how we currently use EPrints v2), but also for materials related to teaching and learning – so we are interested in a product that can cope with the following:

Research papers
Theses
Digitised readings (course based reading specifically)
Digital Images
Digital Video
Digital Audio

The software needs to support workflow, and copyright/IPR issues. It needs to act as a digital library, an open access research repository, and a respository for learning material (underlying our Moodle VLE).

Since getting to the meeting, I’ve already been challenged as to why we want to implement a single repository in this way – why not plugin to web based services such as YouTube or Flickr etc.

This challenge reminds me a bit of a posting on eFoundations recently that considered Flickr and pondered what a Web 2.0 repository would look like – and surely YouTube and Flickr are examples of Web 2.0 respoitories?

Les Carr is emphasising that EPrints in not just an upgrade, but a complete redevelopment of the systm – with more functionaliyt, more flexibility, more interoperatility, and a move to put current ‘admin’ functions into the hands of ‘users’.

Some immediately noticeable facilities with v3 (we are now into the demo), is the links to Atom and RSS (1.0 and 2.0) feeds. The front-end looks pretty much like v2 – which is a bit of a shame, as the basic interface has always felt a bit clunky to me (apologies for this, but looks like it was designed by programmers, not designers). I guess this is something we can configure – and probably would need to put some effort into making it look nicer). Interestingly, appartently they did get a UI specialist to look at the interface (although not clear if this was just for the deposit interface)

The currently supported ‘browse by subject’ and ‘browse by year’ are still there – and probably we could do with investing in adding some extra views (e.g. by course)

Anyway, a few features:

• Atom and RSS feeds for whole respository, and for every search run
• A nifty preview for images and pdfs when rolling over thumbnails
• A ‘request a copy’ function – which allows a searcher to request a copy of an item which isn’t available in full-text. This triggers an email to the owner of the item, and they can approve, or deny, the request. This could be used when you can’t make an item openly available, but can supply copies on request from individuals.
• Re-ordering search results
• Export results in various formats – (ascii, BibTeX, EndNote, etc. etc.) (results in these output formats are URL addressable – so could build an interface on this by the sound of it). This is probably actually a bit more interesting than it sounds – it looks like this function gives the ability to view the results in various different interfaces – so, this is possibly (or actually?) the place where you have an API into the search interface – because you can address them in URLs which incorporate the search, and you can view the output in xml or other formats (e.g. Google Maps.

The Registered Users interface has changed quite a bit – with a ‘Manage Deposits’ function, to allow users to manage all their deposits, see which ones are under review, which ones are live, etc. The list of items shown can be filtered and configured by the user. Also new is a History of changes made to an item – which again can be filtered to changes made by a particular user etc.

There is a much wider range of default item types now supported (partly to demonstrate that EPrints is about more than textual content).

The deposit function seems much leaner than in v2. A clear 5 step workflow. Some nice things – like applying a license to the file from a drop down list. Also, you can adjust or add new workflows – this needs a bit more exploration to see how flexible it is.

A nice ‘auto-complete’ feature when filling in author names – taking information from the current authors entered in the repository. Really nice feature when filling in Journal or Publication title – uses Romeo as the authority source, so you can see immediately what status the journal is (Green, Gold, Grey etc.). The authority can be locally held in a file, so can use local sources. The authority files is very nice, but needs a bit more thought – Libraries have been handling this type of question for ages, and I’d like to compare how EPrints authorities works in comparison to some decent library implementations)

Some nice import functions – if you import a CrossRef DOI then metadata completed automatically.

Things that I didn’t see, that I would have liked to see (not to say they aren’t in there, but needs investigation):
Versioning
Technical metadata (and automatic extraction of this on file upload)
Ability to push items to workflow on specific triggers
More flexible and definable workflows – currently ‘workflow’ in eprints essentially defines how screens are presented in the ‘deposit of an eprint’ workflow
Notifications based on specific triggers
Ability to limit access to objects based on attributes related to a user (must be authenticated, must be a member of x institution, must be enrolled on x course)

Overall, there are some really nice features in EPrints v3, and much to be impressed by. Unfortunately, I think it is still very much aimed at ‘research output’, and I’m looking for something that is more engaged with ‘institutional respository’ in a broader sense. However, the development team are definitely interested in talking more about this, and I will do my best to involve them as we work towards selecting a product at RHUL.