Hello list members,
I'm Bill Whyte, project manager of ADVISER II, a Telematics Applications
Programme project concerned with providing better access to European RTD
results. (We have a pilot service at http://ADVISEReu.vws.co.uk.) It is our
aim to provide access to grey information - Deliverables, news etc - that
you don't get on CORDIS or PROSOMA. We do this via a variety of automated
and manual techniques, but I've also been ploughing a rather lone furrow in
trying to find an easy way for projects to mark up their Web sites, so that
machines can access them.
I've looked at many hundreds of Web sites in ACTS, ESPRIT, TAP and believe
that the information they display can be categorised, at top level at least,
into about 15 different types. (15 is coincidence, nothing to do with DC!).
My suggestion is to label the site's pages in some way so that, given a
start address (e.g. project home page), a machine could find these entities.
I am not yet proposing a mechanism for this, but note that it corresponds to
one tag per entity - not a heavy load to put on a project (and we have
developped prototype tools to make it very easy).
I also note that my proposal is contextual, rather than 'context-free' as
claimed by Dublin Core.
I have posted this proposal, in detail, at
http://vwsnt1.vws.co.uk/adviser/adviserhomepage/eurmi.html,
where I also tried to fire off some enthusiam for a 'European Research
Meta-data' activity. So far, response has been underwhelming, which is why I
am firing this off to you, to ask for any opinions.
Am I being sensible or off my tree?
Is it relevant?
How would this fit in with DC? (e.g through the TYPE element?)
Any other views, +ve or -ve, very welcome.
I have a deliverable coming up soon on this. If you reply, let me know if
you are involved in an EU project and I'll cite your cooperation.
Replies to me and I'll consolidate them for the list.
Thanks.
Bill Whyte
University of Leeds.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|