Re: Agenda and Logistics for Dec 18 telecon / Errata: XMLLiteral in OWL

>6.0 Errata
>Discussion of if/how we can handle errata at this time
>
>Discussion of XMLLiteral in OWL
>see thread starting with
> http://lists.w3.org/Archives/Public/www-webont-wg/2003Nov/0132.html

The first of the options that I described in [1] has become impossible,
now that RDF is a PR.  This leaves only option 4 from that list
as a possible option for making the OWL documents consistent.

Since the required correction is nontrivial, and involves many documents,
it seems appropriate to handle this correction in line with the 
dependencies between the documents:

1) S&AS is the source document (the W3C site http://www.w3.org/
has today on its frontpage the following description:
Semantics and Abstract Syntax: Normative definition of the OWL language)
2) Test is secondary to S&AS
3) Reference, etc. provide non-normative views on part of the 
material in S&AS

The quality and technical soundness of the OWL spec depends on the
care which has been exercised to keep these dependencies clear.
It seems desirable to consider the corrections required for XMLLiteral
in OWL in the order indicated by these dependencies:

1)  Correct misleading statements in S&AS
Jeremy introduced the word misleading in this context [2],
a very appropriate way to describe the situation.
- The main misleading statement is the statement in Section 3
that datatype maps map "xsd:string and xsd:integer to the
appropriate XML Schema datatypes", without mentioning
XMLLiteral. 
This sentence should add XMLLiteral, since each OWL entailment 
(Full or DL) works as if this was added here. 
This addition would not change one single OWL Full or OWL DL
entailment, and can prevent much confusion in the future.

2) Bring Test in line with S&AS
- correct misleading statements in Test, such as the one
cited by Jeremy in [2]:
>The datatype theory of an OWL consistency checker MUST minimally support 
at 
>least xsd:integer, xsd:string from [XML Schema Datatypes]. 
- examine and correct tests 201 through 205 to become consistent
with S&AS
- correct the word "datatype theory" to become "datatype map",
throughout the document (I mentioned this already in [3])

3) Correct the other OWL documents
- In particular, Reference would need at least one correction.


Herman


[1] http://lists.w3.org/Archives/Public/www-webont-wg/2003Dec/0058.html
[2] http://lists.w3.org/Archives/Public/www-webont-wg/2003Dec/0060.html
[3] http://lists.w3.org/Archives/Public/www-webont-wg/2003Dec/0051.html

Received on Wednesday, 17 December 2003 04:01:21 UTC