Re: XML Schema WG comments on RDF documents

Thanks very much for these detailed, thought through and polite comments.

This is a simple acknowledgement of the comments about RDF Concepts and 
Abstract Syntax; with a small amount of reply on comments 2.2, 2.4.

RDF Core coordination: [[
Brian please assign four issues for the four points

           + 2.1. [5]Mapping from lexical forms to values
           + 2.2. [6]Values without lexical forms 
           + 2.3. [7]Lexical forms, strings, and character sequences
           + 2.4. [8]Strings for natural-language data

May I suggest keeping the XML Schema WG's numbers if possible.
e.g. xswg-2.1 thru xswg-2.4

The fifth is editorial and is accepted; Graham please ensure this is tracked 
on the editors' page.
           + 2.5. [9]Typos and minor editorial notes
]]

I find your statement of the issues clear and comprehensible and omit trying 
to rephrase the issues.

2.2 Values without lexical forms

I believe we consciously decided to allow values without lexical forms; IIRC, 
it was motivated by union datatypes in XML Schema. I will check with the WG 
for other input on this, but if my memory is right then we need to have cross 
WG dialogue on this one.

2.4 Strings for natural-language data

This is actually an editorial rather than a substantive issue.
We agree wholeheartedly with your comment that simple strings do not suffice 
for natural language texts; and ruby is one of the use cases the WG has 
considered for the rdf:XMLLiteral datatype. We were trying to capture that 
the language tag is intended for use with natural language text in the 
lexical form; I failed editorially. I will think again and get back to both 
you and the RDF Core WG with alternative text.

FYI There is an example of ruby embedded within RDF in the OWL Test Cases 
editors draft.
http://www.w3.org/2002/03owlt/editors-draft/snapshot#proposed-misc-200-xmlliteral

The triples format below the RDF/XML, while unreadable, does show that the 
language tag is combined with the XML structure to form the ruby text.

I am not aware of any ruby tests within the RDF Test suite; I suppose it would 
be straightforward to repackage the OWL Test to only exhibit RDF features.

Jeremy

Received on Tuesday, 11 March 2003 04:39:48 UTC