S&AS review: general remarks

OWL - Semantics and Abstract Syntax
Version of 20 March 2003

I re-reviewed all the normative parts.
I mail these comments separately: one mail about
Sections 1 to 4 and one mail about Section 5.

I believe that all of the comments I give in these
mails can be processed quickly. 
Most corrections that I think are needed deal with 
Section 5, RDF-compatible model-theoretic semantics,
and arise from the need to make the document up to date
with the RDF Semantics.
Several comments are editorial, in connection with new
text.

I did not yet completely review the remaining, informative 
part (the appendix).  As was noted earlier,
the normative parts are the central parts in connection 
with the decision to go to last call.
However, I looked far enough in the appendix to keep it
consistent with the changes I describe for Section 5.

The S&AS document to be reviewed has no Section 4.3 on
RDF descriptions of OWL DL and OWL Lite.
I believe that if WebOnt decides to go to last call without this
section in the S&AS document, then the clear intention should be 
confirmed to add this section later to the document.
Strictly speaking, as I noted earlier, the S&AS document without
Section 4.3 is not consistent with the RDF Semantics spec,
which requires that:
>Specifications of such syntactically restricted semantic 
>extensions MUST include a specification of their syntactic 
>conditions which are sufficient to enable software to 
>distinguish unambiguously those RDF graphs to which the 
>extended semantic conditions apply. 


I did not review Section 4.3 as it appears in Jeremy's recent
mail to Webont [1].  Any complete proposal for 
Section 4.3 should be accompanied with a proof that the given
standalone RDF description of OWL DL and OWL Lite is correct
given the normative description of OWL DL and OWL Lite as
the outcome of the mappings from the abstract syntax.
[1] does not contain such a proof.
Is there an idea how to complete this version of Section 4.3 
with such a proof, given that the mapping rules and
abstract syntax in S&AS differ from those in Jeremy's earlier
version?

Herman ter Horst

[1] http://lists.w3.org/Archives/Public/www-webont-wg/2003Mar/0161.html

Received on Wednesday, 26 March 2003 09:09:24 UTC