Re: XML Schema WG comments on RDF documents

>At 2003-03-14 13:00, pat hayes wrote:
>
>>Thanks for your comments on the RDF semantics document.
>>
>>I will adopt your editorial suggestions, correct typos, and will 
>>try to remove uses of terminology with any potential philosophical 
>>penumbra from the technical prose in the document.
>>
>>I have one clarification question concerning your comment on datatypes:
>>
>>3.2. Types as lexical mappings (schema-related)
>>
>>       A datatype is an entity characterized by a set of character strings
>>       called lexical forms and a mapping from that set to a set of
>>       values.
>>
>>     We have a couple of reservations concerning this characterization.
>>...
>>      * The statement describes (with the exception of the problem just
>>         noted) simple datatypes, but not the class of complex datatypes
>>         which can be defined by XML Schema, nor all the types (or
>>         type-like constructs) definable in various other schema languages
>>         for XML.
>>
>>I have followed other authors in this characterization, and am 
>>somewhat surprised to discover that it is incomplete or inadequate. 
>>Can you suggest, or point me to a source for, a definition which 
>>does cover the more general case, or which clarifies what is wrong 
>>with this characterization? We are supposed to have a datatyping 
>>framework which can handle all the built-in XML Schema datatypes.
>>
>>Thanks for any help you can give.
>
>Thank you for your note. You are quite right: the term 'datatype' is used
>in our spec, as in yours, solely in a sense extensionally equivalent
>to 'simple type'.  The WG member who drafted our comments has what turned
>out to be an isolated minority view of the desirable usage, and the comment
>slipped by when the WG reviewed the draft comments without anyone noticing
>the problem.
>
>The WG has discussed the issue and confirmed that we do not wish to treat
>'datatype' as a synonym for 'type' or as covering both simple types and
>complex types.  So we withdraw the comment, with thanks to you for drawing
>the problem to our attention.

Thanks for your response, and to the WG for the effort which produced it.

BTW, the text of the relevant section has been rewritten in response 
to other comments. (See 
http://www.w3.org/2001/sw/RDFCore/TR/WD-rdf-mt-20030117/#dtype_interp 
)
In particular, the 'parameter' D of D-interpretations is now 
considered to be a functional mapping from URIrefs to datatypes, 
rather than a set of datatypes. This is in order to have a clear 
statement of the way that a D-interpretation treats datatypes as 
having 'names' without committing to the 'name' being in any sense a 
component or part of the datatype itself, an issue which arose rather 
acutely in OWL.

I have used the term 'datatype map' for this mapping, and although 
this is not intended to be normative, and is not currently used 
outside the RDF semantics document, I would welcome your opinion of 
whether this term is likely to produce any confusion with any 
terminology in your forthcoming drafts. The shorter word 'map' was 
chosen deliberately in order to be more distinctive, but it would be 
easy to change this terminology at this stage. I would prefer to 
sacrifice elegance to lack of ambiguity rather than the other way 
around, however.

Pat Hayes

PS. We are still discussing your suggestion about 'lexical mapping'. 
The text currently uses the longer term 'lexical-to-value mapping' in 
the interests of being thoroughly unambiguous, but this may be 
changed.


-- 
---------------------------------------------------------------------
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

Received on Saturday, 26 April 2003 20:15:32 UTC