- From: Sandro Hawke <sandro@w3.org>
- Date: Wed, 10 Apr 2002 11:07:29 -0400
- To: msabin@interx.com
- cc: www-rdf-interest@w3.org
Miles Sabin wrote:
> Manos Batsis wrote,
> > Very simply, almost both sides are right. An HTTP URI represents a
> > document or fragment (after all, that's what it points to), which in
> > turn represents whatever. This should satisfy both sides. I don't
> > see the point of this argument.
>
> I'm happy to agree with this in some cases, but not in all.
>
> In some situations it makes sense to think of representation as
> transitive (ie. if the URI represents a document which in turn
> represents todays news, then the URI represents todays news), in which
> case the URI is still ambiguous without additional context.
Shouldn't we try to make that use of two steps explicit?
Sometimes you want to know about the two steps, sometimes you don't.
I think that means we should model both steps, and people/software can
later map to an easier-for-people model with implicit features. I
think this strategy also works for REST: let people define whatever
operations they want, THEN map those onto GET/POST/PUT as desired.
-- sandro
Received on Wednesday, 10 April 2002 11:09:53 UTC