indicating meta and type?

Bill Kearney wkearney99 at h...
Sun Dec 8 11:38:37 UTC 2002

> From: "Phillip Pearson" <pp at m...>
> Would it make sense to just keep adding predicates? They wouldn't
> necessarily need to be in the FOAF namespace ...

The trouble, of course, is what other namespace to use in a manner that aids
foaf? Just inventing another one is usually something to avoid when possible.
I'd want the data to be of 'obvious' use in other places like PIMs.

> <rdf:Description rdf:ID="person">
> <im:mesengerAccount>
> <im:Account>
> <im:AccountType rdf:resource="(etc)/MsnAccount"/>
> <im:login>login name</im:login>
> </im:Account>
> </im:mesengerAccount>
> </rdf:Description>
> Starting to get a bit verbose, isn't it ... hmm.

Yes, and that's always something to avoid. RDF proponents understand the
semantic value to the verbosity. But to everyone else it looks like a ton of
baggage that doesn't "do anything" for them. So being simple is worth

The flipside, I suppose, would be to just use elements from other efforts.
vCard and SWAP come to mind:


Trouble is, neither have a way to indicate an instant messenger address. Nor do
they have a 'generic' metadata type. vCard's ability to indicate which is the
preferred value is handy. But it's TYPE fields don't appear extensible. So it
wouldn't really be correct to overload one of vCard's existing elements to
shoehorn in an IM address.

-Bill Kearney

Several other links to interesting works:

More information about the foaf-dev mailing list