Re: Semantic content negotiation (was Re: expectations of vocabulary)

>> Rubbish. They may *optionally* provide these facilities, and
>> no extension is needed beyond those which HTTP allows.
> Please, let's not use "may" or "various" sort of phrasing. If an  
> "option" is
> rarely be honored, what is the point to discuss it.

It seems to work for Accept and Accept-Language, which are optional.

> Imagine yourself trying to describe something in RDF.  Are you  
> honestly going to start an RDF engine behind just for that?

If I'm going to be serving to resource-limited clients, then yes. I'd  
have to anyway for a SPARQL endpoint, mm?

>> I expect smart servers to be more feasible than smart clients.
>> Servers have the advantages of better hardware, caching, and
>> more available information.
> A server can be a sophisticated web service, can't it?  And it will  
> be case
> in SW.  Software agents trying to things for human.  We use these  
> agent to
> handle RDF, not ourself.

Good, I'm glad you agree.

Received on Sunday, 30 July 2006 18:49:15 UTC