RE: A proposal for clarifying the definitions of datatype maps

Hi Boris!

Boris Motik wrote on Thursday, April 30, 2009:

>Hello,
>
>Michael pointed out in one of his review of conformance that the
>definitions of
>datatypes might benefit from further clarification. In fact, I have
>noticed that
>there is a slight inconsistency between the Syntax and the Conformance
>documents
>regarding datatypes in OWL 2 DL. Conformance explicitly says that an OWL
>2 DL
>ontology must not contain datatypes other than the datatypes mentioned
>in the
>OWL 2 datatype map; however, this is not reflected in the Syntax
>document, where
>Section 5.2 implies that an OWL 2 DL ontology could include datatypes
>that are
>not in the OWL 2 datatype map. As Conformance is more restrictive its
>conditions
>on datatypes are the ones that actually hold.

[...]

>In Direct Semantics:
>
>Clarify that the datatype map D (used throughout the document) must
>coincide
>with the OWL 2 datatype map on the definition of the semantics of the
>datatypes from Section 4.

I'm not sure what the word "coincide" refers to here. There are two aspects
I can see: 

Aspect (1): Do you meant that if, for example, the datatype name "xsd:float"
is used in a datatype map as defined in the Direct Semantics, then the
semantics of this datatype must *coincide* with the definition of xsd:float
given in the Structural Spec. This would indeed be an obvious requirement to
me, and I would agree that it should be said explicitly in the Direct
Semantics. So no issue with this on my side.

But I wonder whether you ask for more than just (1). I can see that the
definitions of datatype maps, vocabularies and interpretations in the Direct
Semantics are stated in a very general way, and do not make any reference to
the concrete set of datatypes listed in the Structural Spec. So do you want
to change this in the way that

Aspect (2): the set of datatype names in a datatype map must *coincide* with
the set of datatype names listed in the Structural Spec, i.e. a datatype map
used in the Direct Semantics must contain those and only those datatypes
that are given in the Structural Spec?

I would consider (2) a strong restriction compared to the current situation,
because it wouldn't allow implementers to provide new datatypes (theorem in
Section 3 would become redundant then, AFAICT).

Michael

--
Dipl.-Inform. Michael Schneider
Research Scientist, Dept. Information Process Engineering (IPE)
Tel  : +49-721-9654-726
Fax  : +49-721-9654-727
Email: michael.schneider@fzi.de
WWW  : http://www.fzi.de/michael.schneider
=======================================================================
FZI Forschungszentrum Informatik an der Universität Karlsruhe
Haid-und-Neu-Str. 10-14, D-76131 Karlsruhe
Tel.: +49-721-9654-0, Fax: +49-721-9654-959
Stiftung des bürgerlichen Rechts, Az 14-0563.1, RP Karlsruhe
Vorstand: Prof. Dr.-Ing. Rüdiger Dillmann, Dipl. Wi.-Ing. Michael Flor,
Prof. Dr. Dr. h.c. Wolffried Stucky, Prof. Dr. Rudi Studer
Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus
=======================================================================