Greetings,
[This message is posted to the semantics, dm and dal lists. To avoid duplication, I suggest any followups go to a single list, and I suggest the semantics group (I've set the reply-to field accordingly, though this might not make it through the list manager).]
I've published a Note on how one might link UTypes and URIs, following up (finally!) on one of the talks I gave at the Moscow Interop. The note is at <http://www.ivoa.net/Documents/latest/utype- uri.html>, and I'd much appreciate comments.
The `elevator pitch' is this: The definition of UTypes in the VOTable document makes them look much like URIs. If you tighten up the VOTable definition and require that UTypes actually _are_ resolvable URIs, then you get (a) documentation for the UTypes, and (b) standards-based machine-readable declaration of the relationships between UTypes, via a mechanism I describe in the Note. That makes it feasible to have more (ie, more precise) UTypes in the wild, with a lightweight standardisation process, without sacrificing interoperability. (phew!)
The document describes (in section 4.2) a lightweight reasoning service, which would allow applications to key into this network of UTypes in an almost trivially easy way. If you want to play with that, there's a resolver currently (and temporarily) sitting at <http://thor.roe.ac.uk/utype-resolver/>.
The language of the Note is fairly close to a proposed standard, and I'd like to suggest that the Note move towards becoming a standard in fact, with some WG's blessing. Would the Semantics WG be most appropriate?
All the best,
Norman
-- ------------------------------------------------------------------------ ---- Norman Gray / http://nxg.me.uk eurovotech.org / University of Leicester, UKReceived on 2007-03-22Z16:56:25