Re: The FragId issue

At 09:09 PM 2/21/03 +0100, Massimo Marchiori wrote:

>Another issue for consideration.
>
>Issue FragID's:
>Section 7 in the RDF Concepts doc would seem to imply that
>"any RDF URI reference consisting of an absolute URI and a fragment 
>identifier identifies the same thing as the fragment identifier
>does in an application/rdf+xml"

Yes... and I think that needs be asserted normatively somewhere ... I think 
the application/rdf+xml MIME type registration is that place.

>First point (normativeness): it's unclear in the whole section whether 
>there are actually normative statements (there are no "must",
>"should", "may" or so, but wordings like "consider", "assume" etc). This 
>makes hard to understand what real statements are mandated,
>and what is just normal discourse (and as such, with no normative power).

I think the essential normative information is provided by the formal 
semantics.  The purpose of this section was to show how the formal 
semantics is consistent with other uses of fragment identifiers.

I'm not sure whether this is or is not normative;  I don't know of another, 
contrary, view that is consistent with the formal semantics of RDF and 
which would fail to interoperate with the view expressed here.

But I think it's correct, so I'll respond as if it is normative...

>Second point (indication): in any case, assuming normativity, this would 
>seem to mandate that any URI present in RDF has to
>"indicates a Web resource with an RDF representation". I would be very 
>unhappy with this mandate, as URIs don't necessarily have to
>be dereferenceable (it's not all URLs....), and in any case, they 
>shouldn't be required to have an RDF representation (think of an
>image).
>And to further clarify, what does it formally mean for a URI to "indicate 
>a Web resource with an RDF representation"? Some
>clarification is needed...
>
>Note that obviously the second point is moot if the answer to the first 
>point is: no normative statements.

I don't have a formal meaning for "indicate a Web resource with an RDF 
representation", though I thought the intent was clear enough.  Maybe I 
should say "identify a Web resource with an RDF representation", or even 
just "Identify a web resource, which is presumed to have an RDF 
representation".  (I'll assert that *any* resource has an arbitrary number 
of RDF representations, so this doesn't create any new constraints.)

The point about URIs not necessarily being dereferencable as RDF is 
explicitly addressed:
[[
eg:someurl#frag means the thing that is indicated, according to the rules 
of the application/rdf+xml MIME content-type as a "fragment" or "view" of 
the RDF document at eg:someurl. If the document does not exist, or cannot 
be retrieved, then exactly what that view may be is somewhat undetermined, 
but that does not prevent use of RDF to say things about it.
]]

Can you explain if this does not address the concern you raise?

>ps On normativeness, the same issue would apply to Section 4 as well 
>(Meaning of RDF), but that would be on a different scale: [...]

This is being dealt with under the heading of social meaning, so I'll 
assume I can let this matter be handled there.

#g


-------------------
Graham Klyne
<GK@NineByNine.org>

Received on Tuesday, 25 February 2003 12:19:27 UTC