At eResearch Australasia 2009 in November in Sydney, I'm going to to be co-facilitating a Birds of Feather session with Anna Gerber from UQ, and Peter Murray-Rust and Jim Downing from Cambridge; Jim Downing will be involved but he won't be able to join us at Manly. The topic is Boundaryless eResearch: use the Web, use Linked Open Data. There's an extended abstract in PDF format which I'm hoping we can build into a paper. I'm supposed to be starting that work now, hence this blog post to try to get my thoughts in order.
The three things I most care about for this session are:
Getting tools into the hands of researchers so they can 'do' linked data. They're never going to read Tim Berners lee's Design Issues page or wade through a tutorial so we need to give them tools that help them do their work, back up their data, publish data and both informal and formal commentary. Right now Linked Data is Too Hard. I think we're going to need some online wizard services and other tools that let people do semantic web stuff without all the hair-tearing.
Anna's team, in Jane Hunter's group are working on this with their Firefox Add In for creating compound semantic-web objects, for literary scholars.
Getting the web into the scholarly communications process as a first-class citizen, as seen over and over on this blog, now branded as Scholarly HTML. A lot of the discussion of Linked data is about giving things names, and describing relationships in RDF, but how will real people talk about stuff? How will they email and blog and wave at each other, with those links embedded inscholarly discourse in a meaningful way?
Meaning matters. It's not enough to link to something; you have to be able to indicate why you're linking to it. Am I linking to someone to say their an author on this paper? To cite them? Or to indicate that I'm talking about them.
I think that working out how linked data fits with documents is very important and services like the Australian National Data Service (ANDS) will ignore this dimension at their peril. I mean, how are we going to bootstrap the data-citation economy if we don't think about documents? That's, like, where people cite things.
Bringing the web to the desktop. Linked Data is all about the web but desktop operating systems are still mainly web consumers and it's not easy to move desktop or lab generated stuff to the web; the work we're doing with The Fascinator and Jim's team is doing with Lensfield is trying to let people see all their stuff (including data and documents) as part of the web from the moment they create it, by providing a web-view of local files which is integrated into a system that can move stuff to repositories, and publish, and back things up, and serve as a platform for analytical tools.
© Peter Sefton 2009. Licensed under Creative Commons Attribution-Share Alike 2.5 Australia. <http://creativecommons.org/licenses/by-sa/2.5/au/>