Re: RDDL again

Dan Connolly wrote:

> Hmm... I still think the way purpose is handled isn't what
> you want/mean... e.g.
> 
> <> rddl:related <http://www.ietf.org/rfc/rfc2396.txt> ;
> <http://www.ietf.org/rfc/rfc2396.txt> rddl:nature <http://www.isi.edu/in-notes/iana/assignments/media-types/text/plain> ;
> <http://www.ietf.org/rfc/rfc2396.txt> rddl:purpose <http://www.rddl.org/purposes#normative-reference> ;
> 
> RFC2396 is a normative reference *for rddl*. I'd expect that to be
> written:
> 
> <> <http://www.rddl.org/purposes#normative-reference> <http://www.ietf.org/rfc/rfc2396.txt> .

This has come up a few times and it's not a slam-dunk either way.  These 
  kinds of discussions are really hard without a whiteboard to draw 
graphs on...  your assertion immediately above contains a little bit 
less information than my version: it doesn't tell you explicitly that 
the #normative-reference property is a rddl:purpose.  You could infer 
that from the fact that it's attached to rddl.org/purposes, but that 
doesn't work because anyone should be able to make up their own purpose 
and name it via their own URI.  For example, if Antarctica wanted to 
define a bunch of resources that, given a namespace URI, invoked code 
that produced on-screen maps of documents in that vocabulary, it's 
unlikely that that purpose are going to be rddl.org/purposes, they'd be 
over at antarctica.net/whatever.

Now if you *know* what purpose you're looking for, you can just look for 
a property of the form

<namespace-I-care-about> <purpose-I-seek> <related-resource>

and there's no further problem.  But this bothers me, because it weakens 
the central RDDL notion of lookup by nature & purpose.

The question is whether you think it's worthwhile to know that a 
property you're asserting about a related resource is a RDDL 
nature/purpose without having to do any inferring.  My formulation 
retains that information simply and directly.  If you decide you don't 
care to retain that, then your formulation is fine.
  -Tim

Received on Wednesday, 11 June 2003 05:07:16 UTC