Responses to RDF mapping LC issues

Dear all, 

this email summarizes my views about how we should respond to the LC
issues raised for the RDF mapping document and recorded in our issues
list at [1].

[1] http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/issues/wsd-issues-condensed.html


291: agree with Eric, will incorporate XML C14N into the next version of
the RDF mapping


292: agree with Eric, unrecognized extension attributes should be
canonicalized.
Currently, when turning an extension attribute foo:bar="baz" into RDF,
the result is (in N3):
        
        <parentComponentId> wsdl:extensionAttribute <genId>
        <genId> rdf:type wsdl:ExtensionAttribute
        <genId> wsdl:attributeName <qname>
        <qname> rdf:type wsdl:QName
        <qname> wsdl:localName "bar"
        <qname> wsdl:namespace "http://example.com/foo"
        <genId> rdf:value "baz"
        
This loses any XML context information that might be used in the value
of the attribute; for instance namespace bindings. Instead of that, we
should produce something like this:
        
        <parentComponentId> wsdl:extensionAttribute 
            "<quote:attrContainer 
                xml:base='...' xml:wsdl='...' 
                xmlns:foo='...' foo:bar='baz' />"^^rdf:XMLLiteral


293: agree with Karl that the text is unclear; clarify that the
normative part of the RDF mapping do not consider the case where
multiple WSDL documents are merged and that section 4.2 points out some
problems


294: agree with Karl, will add text that the RDF mapping maps from valid
WSDL files to RDF graphs; the mapping can be implemented by generators
of WSDL files which can at the same time generate the RDF form; or as
plugins in Semantic Web applications, enabling them to understand WSDL
data on the Web as RDF


295: the RDF mapping document uses no capitalized RFC 2119 keywords, it
only repeats (in lowercase) the restrictions from WSDL itself. However,
I will try to reformulate things like "must" and "should" to clarify
what is a new requirement from the RDF mapping spec and what is
guaranteed (required) by WSDL


296: will add some text (in conjunction with previous resolutions) that
will clarify what it means to implement the WSDL RDF mapping, but unsure
about calling it a "Conformance" section


297: I can take action (unsure if I can fulfill it before the call) to
contact Karl about what he means - it seems he's requesting that we tell
somebody how we handled generic XML extension elements in RDF, but I'm
unsure; this will clarify the issue and then we can move on to try and
resolve it


298: editorial comments will be taken under advisement for the next
version


289: we need to check this issue after all the others are resolved and a
new version is ready


The resolution emails (for the WG and the originator of the issue) for
all issues except 297 and 289 should be written when the issues are
believed tackled in the editor's draft, I'd do that as appropriate. 
297 needs clarification, and 289 will be re-checked.

Now for timing: I've talked about a new version, and I believe it could
be ready by the end of November (can it be tracked as an action?), which
could give us enough time to publish it as WD (or we could call it next
LC) before we publish the doc as WG Note together with WSDL 2.0 Recs.

It doesn't seem very likely that the SAWSDL WG will take this document
on to Rec - this is an option in their charter.

Hope this is useful for the telcon tomorrow,

Jacek

Received on Wednesday, 27 September 2006 15:05:18 UTC