goofy usage of rdfs:domain in infoset.rdf?

[RDF IG folks: note that XML Infoset is in last call,
and it includes an RDF schema that, while normative,
could be awfully useful. Please review it. I found
a few bugs.]

My understanding* of rdfs:domain is:

If
	prop(subj, obj)
	and rdfs:domain(prop, X)
then
	rdf:type(subj, X).

By that logic, this is almost certainly not what is
meant:

/////////
<rdfs:Property ID='baseURI'>
  <rdfs:domain resource='#Document'/>
  <rdfs:domain resource='#Element'/>
  <rdfs:domain resource='#ProcessingInstruction'/>
  <rdfs:domain resource='#ExternalEntity'/>
  <rdfs:range resource='http://www.w3.org/2000/01/rdf-schema#Literal'/>
</rdfs:Property>
\\\\\\\\\\
 -- http://www.w3.org/2001/02/infoset#
	http://www.w3.org/TR/xml-infoset/#rdfschema
	http://www.w3.org/TR/2001/WD-xml-infoset-20010202/#rdfschema

that says that anything with a baseURI property
is a Document *and* an Element *and* ... .

I think what you mean is that anything with a baseURI
property is a Document *or* and Element *or* ...;
i.e. the domain is a superclass of all those
(and a subclass of InfosetItem). I plan to follow
up with specific replacement wording/code.



* there's some wording in the spec that suggests otherwise,
but I think it's bogus and lots of other folks do too;
cf
"Semantics of Domain and Range constraints
http://lists.w3.org/Archives/Public/www-rdf-interest/2000Jun/0018.html
http://lists.w3.org/Archives/Public/www-rdf-interest/2000Jul/0032.html
http://lists.w3.org/Archives/Public/www-rdf-interest/2000Sep/0132.html
"
from
http://lists.w3.org/Archives/Public/www-rdf-interest/2000Sep/0240.html
linked from
http://www.w3.org/2000/03/rdf-tracking/

btw... DanBri, I searched for 'domain' in
http://www.w3.org/2000/03/rdf-tracking/

and was shocked that it doesn't occur; please do
incorporate the "More material to be incorporated".

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Thursday, 22 February 2001 02:22:16 UTC