Re: rdfms-resource-semantics

Hi Lee,

I'm sorry for the tardy response.  Email overload again.  Here are some
preliminary thoughts.

Lee Jonas wrote:
> 
> Are there no thoughts at all on this [1] from the RDF Core WG?

> 
> [1]
> http://lists.w3.org/Archives/Public/www-rdf-comments/2001AprJun/0105.htm
> l
> 
> regards
> Lee

[...]

>Why bother defining a special RDF notion of 'resource'?  After all, it
>conflicts with the definition in RFC2396 causing much confusion and a
>special RDF definition of 'resource' is probably unnecessary anyway.

Quite so.  I don't want to have a special RDF notion of resource either.
But I would like there to be a well defined and well understood one
for the entire web.  And I think there is some confusion about what
exactly a resource is which goes well beyond RDF.

If we are to be clear about RDF means, and RDF is about describing
resources, then its important to be clear about what a resource is.

Rather than shoehorn the notion of 'resource' into what the RDF M&S spec
requires as a primary building block, you could turn the problem around and
define RDF M&S in the existing terms of RFC2396.  Within RDF M&S:

>1) Describe RFC2396 as 'normative'.
>2) substitute all occurrences of the word 'resource' for the word
?'reference' (or something similiar).
>3) define 'references' as the union of URI references and anonymous
>constructs.
>4) refer readers to RFC2396 for the URI reference definition.

That may be a useful approach.  Does RFC 2396 tell me whether the
two different URI's can name the same resource?  Does it define a 
notion of equivalence?

That way, RDF's 4 sets become 'Statements', 'Literals', 'References' and
'Properties'.

>This has the following plus points:
>1) it avoids the name collision with RFC2396 resources altogether (rather
>than having to explain a special interpretation of 'resource').

I don't think we are looking to have a special interpretation of resource.
We just want to be clear what it means.

>2) it integrates RDF better with existing web technologies, adopting
>standard semantics for URIs and network entities.

Absolutely.

>3) it shifts the emphasis of RDF to the description of resource views /
>parts of resources (i.e. fragments).  (But, who cares whether metadata
>assertions are about RDF resources or RFC2396 resource views technically -
>in practise they still have the same effect).
>4) RDF can still be considered 'Resource Description' - only now it is
>description of resources and their views/parts in the RFC2396 sense.
>5) It won't break existing applications - the RDF syntax won't change as
>'Resource' is not part of the language.  It could even be considered just an
>editorial fix to the spec.

What you suggest is a change to the RDF formal model.  I don't think it
can be considered just an editorial fix.

[...]

>Just a thought!

I'm sorry this is a bit rushed. 

I'll circulate a pointer to your message on the WG to make sure everyone
has seen it.  Probably the best place for followup discussions is
rdf-interest.

Brian

Received on Tuesday, 5 June 2001 13:46:09 UTC