JISC meeting on institutional repositories

JISC has issued a summary on the conference I have just attended, About Digital repositories: Dealing with the digital deluge (Manchester, June 5-6, 2007). The summary deals with the plenary sessions – I might have commented if this blog had been awake. So these are fuzzier thoughts from two days on.
There was a welcome emphasis on the thesis. This is the one area in which institutions have complete control over the scholarly process. (Not true if the thesis – as in some European countries – consists of bound published papers, so this is a Britocentric viewpoint). So – and we have a very small window of opportunity – the instutitions can really show how publishing – scientific communication – SHOULD be done. I’ve blogged about this before  but it means really cutting adrift from the printed page, the static moment of publication, even the role of the team rather than individual, the value of data … etc. Can I be optimistic? Please let’s try something soon.
There was a major shift from repositories of text (exemplified by ePrints/DSpace and other current technologies where only holy PDF matters) to DATA. Examples were Keith Jeffery’s presentation and a fine session from David Shotton of Oxford on image repositories. Not surprisingly chemistry got a prominent airing (Simon Coles and ourselves).
I am still very worried about our gift of intellectual property to publishers. In discussion about images some people assume that an image is a published paper is the copyright of the creator. Well we know that Wiley doesn’t take this view and I suspect few other non-OpenAccess publishers do. BUT there are technical solutions. If we use rich formats (XML, SVG, even GIF) we can embed rights information in the object. So here is a simple idea:

Adapt our image generation and processing software so that it always contains a license specifying the author and asserting Science Commons or Creative Commons licenses. Allow the author to change this away from default – I suspect few will do so. Then, when the images reach the publisher they can only reassign the image copyright by specifically asking the author. This will become rather unpopular…

And while in discursive mode: “why do we need institutional repositories?” I used to think it was obvious – now I think it’s rather confusing. And I tried to answer the question “why does a scientist need institutional repositories?” and haven’t come up with a very good answer. So here are some facets – see what you think:
Political:

• Universities wish to advertise their success to applicants, funders, reviewers (RAE, etc.). Fair enough – they pay our salaries, and I have an obligation within reason – but it doesn’t directly help me.
• Libraries need to reinvent themselves and repositories is an important area where they can do this. Generally not very compelling for the average scientist.

Financial:

• Repositories are new and so there is funding. That’s valuable for us (thank you JISC) in developing new informatics strategies and tools. But not very general.
• Repositories will generate a new funding opportunity for data curation/archival etc. Perhaps.

Scholarship and scientific practice

• Scientists will be able to archive their own data, funded by the institution. That’s compelling. Often the users of public databases are those who originally deposited the data, lose it, and then find it in repositories. But are universities the best place to put scientific data? Can they all support chemistry, archaeology, astronomy, etc.? These are hard. They are much more likely to happen in domain-specific repositories. No reason why these shouldn’t be supported by LIS staff, but you won’t find them in every institution.
• Liberation of data (a al SPECTRa). I argue that this is much more likely to happen at a departmental level than at institutional. There was a general agreement that mandates don’t work well and that incentives are critical. The normal unit of loyalty is the research group or department
• Sharing data. Yes, but this is primarily cultural and is much more likely to be encouraged at domain level.

And I had a general worry. We are seeing the enormous force of world-centric technology – Google, Flickr, and the new RDF/SPARQL enthusiasts. This is on a scale magnitudes above what institutions can do. Also the blogosphere can often move much more quickly than the average IT project. So we are always in danger of being overtaken and overwhelmed by the collaborative tools and practices “out there”. We need to do exactly those things – and no more – that we are specialists at – education, scientific research, domain-specific ontologies, etc.
So at the end of the meeting the audience was asked for questions and suggestions. It’s too late for the current JISC funding round (I think) but I suggested that we award some modest bursaries directly to graduate students and undergraduates to build their ideas of information management systems for their own requirements (e.g. thesis creation). After all, wasn’t there a fairly large Internet company founded by students? – give them their chance.