Re: responses?

>According to my (possibly incorrect) records, I show the following 
>issues awaiting formal responses.
>
>   http://www.w3.org/2001/sw/RDFCore/20030123-issues/#qu-01
>   http://www.w3.org/2001/sw/RDFCore/20030123-issues/#qu-02
>   http://www.w3.org/2001/sw/RDFCore/20030123-issues/#williams-01
>   http://www.w3.org/2001/sw/RDFCore/20030123-issues/#timbl-02
>   http://www.w3.org/2001/sw/RDFCore/20030123-issues/#horrocks-01
>   http://www.w3.org/2001/sw/RDFCore/20030123-issues/#efth-01
>   http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-22
>   http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-23
>   http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-16
>
>I realise that some of these are awaiting updated text to refer to, 
>but it would be good to get the others closed off.

OK, some of these are mine.

Qu-01 is whether the domain of rdfs:member and its subproperties 
should be rdfs:Container rather than rdfs:Resource. I honestly don't 
know, and have no opinion on the matter, so am seeking guidance, and 
will editorialize the decision once we have it (or someone can 
document where it already happened.) Has the WG actually decided this 
question? Peter apparently thinks that the domain should be 
rdfs:Resource, based on email discussions about rdf:li.

A related question is whether rdf:first should be a subproperty of 
rdfs:member , so that member covers collections as well as 
containers. Dan C. requested this, but this wasn't my understanding 
and isn't in the semantics right now. It does make sense, however.

I think we need to discuss this and come to an actual decision.

----
qu-02 has emerged from a series of emails between Qu and me on the 
topic of whether the MT should specify as a semantic condition that 
membership properties be functional. I do not want to do this on the 
grounds that this would complicate RDF by introducing a hidden 
identity assertion, since it would require the following entailment:

_:xxx rdf:type rdf:Seq.
_:xxx rdf:_2 <ex:a> .
_:xxx rdf:_2 <ex:c> .
<ex:a> ppp zzz .
|-
<ex:b> ppp zzz .

and all other similar substitutions, including of property names:
...
aaa <ex:a> bbb .
|-
aaa <ex:b> bbb .

and for uses of these URIs inside typed literals, etc.

I think this would greatly complicate life for RDF reasoners and 
likely produce confusion among users, without adding effective useful 
functionality, and in any case it can be expressed directly in OWL.

Qu however did not accept my explanation, and requested that it be 
made non-editorial.

Over to you, guys.

------
timbl-02 gave rise to quite a lot of email.

The upshot is that an explanatory note has been added to the text of 
the reification section (3.2.1)
http://www.w3.org/2001/sw/RDFCore/TR/WD-rdf-mt-20030117/#Reif
to clarify that reification does not mean quotation.

Im not sure whether this is quite what Tim had in mind, however; and 
one could put a lot more explanation in there 
(http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Mar/0070.html), 
but it seems to me that this section is already too long, if anything.

Or, we COULD change what this section says altogether, but that would 
be a major change to the account of reification if not to anything 
normative. I have no technical opinions about reification, myself, 
and my nontechnical opinions are best kept off a public email. The 
message cited above basically lays out a 2x2 matrix, and I can write 
a semantics for any of the four options. The one we have right now is 
probably the most complicated.  Y'all choose.

-----
Horrocks-01 is decided, right?   I will draft a response.

Pat

-- 
---------------------------------------------------------------------
IHMC					(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola              			(850)202 4440   fax
FL 32501           				(850)291 0667    cell
phayes@ai.uwf.edu	          http://www.coginst.uwf.edu/~phayes
s.pam@ai.uwf.edu   for spam