Some ideas for our friends in the UK for the JISC Rapid Innovation call.

JISC invites institutions to submit funding proposals for grants to fund technical rapid innovation projects addressing priority areas.

Proposals are sought under the following priority areas:

• Mashups of open data

• Aggregating tags and feeds

• Semantic web/ linked data

• Data search

• Visualisation

• Personalisation

• Mobile Technologies

• Lightweight Shared Infrastructure Services

• User Interface Design

Bids for projects dealing with other areas that are relevant to the Information Environment are also welcomed.

Funding of between £15,000 - £40,000 for 6 month projects for up to 30 projects is available

12 noon on Wednesday 22 April 2009The deadline for receipt of proposals in response to this call is .

We'd be happy to help out with projects like the ones I am rattling off here.

1 Mashups of open data

Why not give Tony Hirst £40,000 to go on holiday somewhere without a net connection, giving someone else a chance to (a) read all his blog posts about mashups, which would consume about 5 months and (b) spend the last month doing the one or two things Tony has not already mashed up.

2 Aggregating tags and feeds

Well, we could invest six months in the Lazy Lecturer thing (which was shortlisted for the dev8D prize). It's all about tagging, then aggregating and munging resources that were tagged, then re-publishing them back to the web.

3 Semantic web/ linked data

3.1 Authoring tools

How about a really simple approach to getting linked data going for documents. In six months you could easily do one or more of the following three things:

1. Add user pages to a repository and/or ePortfolio system and/or staff phone book that can be used to assert authorship. So, in addition to a page at ePrints like this one that shows publications1 there would be one which formally, explicitly states that linking to it asserts that I am an author. Then all I need to do in a paper is link whatever variant of my name I am using to that page.

2. Get as many onotologies online as possible and set up similar linkable pages for them. See the discussion here, sparked by Glyn Moody's concerns about the Microsoft Ontology add-in for Word. I said:

   Why couldn’t the ontology add-in (which is a good idea, don’t get me wrong) store data using the simplest most robust method possible; using a link?

   … Geonames.org provides a linked-data endpoint for Toowoomba. See what I did there? I linked to it. And look how you can get RDF. That will survive round tripping between Word and OpenOffice.org, saving as HTML, saving as .doc, sending to a publisher, and it’s perfectly usable without an add-in. Could someone explain to me why you would need to use custom XML here given the simplicity, robustness and interoperability of the alternative?

3. Make a new Zotero plugin for one or more of Word / Writer / GoogleDocs / WordPress which can turn plain old links into citations. Why do we have to go through all this finding a page, putting it in Zotero then using the insert citation button, why not teach people to find good endpoints and link to them then let the machine sort out all the rest of it, harvesting metadata if necessary and formatting citations and bibliographies. A modern journal could let the reader choose what bibliography style they want, and let them import all the references from a paper into their own reference manager.

3.2 Publishing tools

Or, for the publishers rather than the authors, why not put in place a demonstration issue of a Scholarly HTML journal, with linkable paragraphs and mineable data courtesy of embedded microformats. Maybe Ariadne at UKOLN? This would look a bit like the winner of the Elsevier Article 2.0 competition.

Or for a bit more of a challenge someone could have a go at integrating ICE with OJS as per this suggestion of mine.

3.3 Collaborative tools

See this proposal of mine for a general purpose annotation system. Could be very useful to JISC from what I hear about the marking process. Just saying. There are some nice interface ideas at the Article 2.0 winner, which does not appear to be available under an open source license.

4 Other

It turns out that the Moodle VLE/LMS we use at USQ doesn't have a proper full-text search system. One very useful project would be to hook up Moodle via some kind of feed system to a faceted search service which could index all the content, discussions and course metadata. The Fascinator would be a good base for this as it could cover security if you want to restrict people to just the courses they are enrolled in, but also potentially allow people to discover where there is content that they are not enrolled in. Even better, you could throw repository content in the index as well and associate it with course codes.