[rdfweb-dev] yahooChatID, icqChatID and [Bug 14] foaf:chat sub-properties could be better described

[Graham Klyne]

At 07:06 09/07/03 -0400, Dan Brickley wrote:
>There is a tightrope to walk between an overly academic
>pursuit of cleanly modelled perfection (where we wouldn't allow things like
>aimChatID because it couples FOAF vocab to some particular service), versus
>pragmatism gone crazy where we keep on adding such ad-hoc properties because
>they are a cheap way of bootstrapping a large dataset. The former leads to
>non-deployment since this stuff, done "right" is super hard to agree; the
>latter risks giving the impression that this is a centralising technology,
>when in fact the opposite is true. There is no _need_ for yahooChatID to
>be in the FOAF vocab; Yahoo could have defined it themselves. But someone
>has to go first to show the potential, so I'm happy with this approeach
>even if it isn't the most elegant way to model things.

Having recently been looking at ABC [1], I do wonder of there's scope to 
invoke its "event model" here?

It may not be quite what you had in mind there, but the idea of (say) 
foaf:chatId being a generic chat id, and a foaf:ChatId event class being 
used to say more fine-grained things about the kind of chat system 
identifier.  Or maybe this is a too-academic approach in this context ?-)

#g
--

[1] http://www.ilrt.bristol.ac.uk/discovery/harmony/docs/abc/abc_draft.html
specifically:
http://www.ilrt.bristol.ac.uk/discovery/harmony/docs/abc/abc_draft.html#Multiple



-------------------
Graham Klyne
<GK at NineByNine.org>
PGP: 0FAA 69FF C083 000B A2E9  A131 01B9 1C7A DBCA CB5E