RE: proposal to resolve issue-114 - Which combinations of punning should be allowed?

Hello,

I am not sure I understand this e-mail. Please note that, in RDF, all possible combinations of punning are allowed: you can use the
same URI for all different types of entities. For example, in RDF (and OWL Full) there is nothing that prevents you from using the
same URI as both a class and a datatype; furthermore, you can use the same URI as a class and a property.

In OWL 2 DL, we need to restrict that because of the two-sorted nature of our logic and the decision not to disambiguate class
constructors involving properties in the RDF syntax (see Issue-17). Therefore, in OWL 2 we cannot pun between classes and datatypes,
and object and data properties. Thus, restrictions on punning in OWL 2 are *stronger* than in RDF (where they are actually void).

Thus, it seems to me that your proposal for resolution actually means we should leave the spec as it is right now (which I fully
support).

Regards,

	Boris

> -----Original Message-----
> From: public-owl-wg-request@w3.org [mailto:public-owl-wg-request@w3.org] On Behalf Of Alan Ruttenberg
> Sent: 19 August 2008 07:47
> To: OWL 1.1
> Subject: proposal to resolve issue-114 - Which combinations of punning should be allowed?
> 
> 
> I propose that we resolve issue 114 by allowing punning that occurs
> in RDF, and no more. That is: Punning between individuals, classes,
> and properties, but not between types that are distinguished in OWL,
> but not in RDF.
> 
> This latter restriction matches much of the punning we currently
> disallow - that between the various kinds of properties, and between
> classes and datatypes. At the same time it does allows the requested
> punning between properties and classes.
> 
> I am unaware of user requests for any other type of punning.
> 
> -Alan
> 
>

Received on Tuesday, 19 August 2008 16:41:32 UTC