Re: response to issue pfps-09

From: pat hayes <phayes@ai.uwf.edu>
Subject: response to issue pfps-09
Date: Tue, 4 Feb 2003 17:49:00 -0600

> I believe that there is a missing part of datatypes in the RDF model
> theory, and, moreover, that this missing part makes datatypes unusable
> in RDF.  The missing part is a mechanism for tieing a URI reference to
> a datatype.
> 
> ----
> 
> The MT assumes that datatypes are defined externally to RDF, and that 
> part of this defining process involves associating a uriref with each 
> datatype which can be used as its 'name', ie the semantic assumption 
> is that the denotation is defined externally. 

The MT provides mechanisms for communicating the lexical space, value
space, and L2V map of these externally defined datatypes to datatyped
interpretations.  However, it is lacking a mechanism for communicating the
RDF ``name'' of the datatype to datatyped interpretations.

> I feel that this is 
> entirely appropriate for a semantic specification, and that the 
> general issue of how meanings can be associated with urirefs is 
> beyond the scope of this WG.

I disagree.  Datatyped interpretations need to know which uriref denotes a
datatype.  

In the absence of this connection in a datatyped interpretation, the only
way to make datatyped interpretations useful in the model theory is to
create semantic extensions that provide for the connection.  Thus any
useful theory of datatypes will be a semantic extension of datatyped
interpretations, just as RDFS is a semantic extension of RDF (but on a
considerably smaller scale).

> A more practical answer to this comment is that the inference rules 
> for datatypes each specify a certain kind of information about the 
> datatype, and clearly state that the rule can be applied only when 
> that kind of information is somehow made available to an inference 
> engine. Since the use of urirefs as URLs which provide access to APIs 
> is well established on the Web, this seems to provide an clear guide 
> to implementers as to how to proceed.

Sure, if I want to create a datatyped version of RDF, I know how to
proceed.  I define my datatype, say octal integers <O,V,L2VO>, and say that my
datatyped version of RDF has this datatype.  I also need a semantic
constraint that says something like
	IS(my:octal) = <O,V,L2VO>
I thus have created not a form of D-interpretation, but a semantic
extension of a D-interpretation.  Without this semantic extension, my
datatype is useless, as there is no way of determining that 
	I("10"^^my:octal) = 8

However, if RDF datatypes included information about their ``name'', and
this was made part of D-interpretations, then I could just say that my
datatype was <my:octal,O,V,L2VO> and then D-interpretations that included
this datatype would have
	I("10"^^my:octal) = 8
without the need for any semantic extension.

> I therefore propose that the WG takes no action on the basis of this 
> comment , except possibly to refer the issue to some other authority 
> (eg the TAG group?)

I would view this as a failure of the WG to provide usable datatypes.

> Pat

Peter F. Patel-Schneider
Bell Labs Research
Lucent Technologies

Received on Tuesday, 4 February 2003 19:40:49 UTC