Lists Home |
Date Index |
Uche Ogbuji wrote:
> Wow. I'm a fan of RDF, but I find this somewhat of a counter-intuitive idea.
> I've always advocated RDF for closed systems and wondered about
> its scalability in open systems.
I think that what Roger is really talking about is *interoperability*,
not really scalability (at least not in a performance sense).
> In the terms you put it, RDF might help the user agent understand
> the vocabulary more efficiently, as long as the agent has all
> the data it needs to fully process the RDF. but if the agent
> does not, and thus has to hop over to another Web site to
> fill in a missing piece (maybe a site where an RDF schema or
> DAML ontology is stored), then has to hop over to yet another
> site to clarify some missing pieces there, and so on, couldn't
> this cause all sorts of network congestion, and notable delays
> in the agent's processing?
If the alternative is having a programmer EXPLAIN the semantics to a
program, this delay will seem slight. ;)
And don't forget about client-side schema/ontology caching.