W3C home > Mailing lists > Public > public-owl-wg@w3.org > May 2009

Re: draft responses for four JC LC comments

From: Bijan Parsia <bparsia@cs.man.ac.uk>
Date: Wed, 13 May 2009 13:09:48 +0100
Message-Id: <11CCDBB4-09ED-4AD3-995E-26CF54C274C2@cs.man.ac.uk>
Cc: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>, <public-owl-wg@w3.org>
To: "Michael Schneider" <schneid@fzi.de>
First, I want to situate things a bit:
	http://www.w3.org/2005/10/Process-20051014/tr.html#doc-reviews

"""Reviewers should not send substantive technical reviews late on  
the Recommendation track. Reviewers should not expect that a Working  
Group will readily make substantive changes to a mature document. The  
more evidence a Working Group can show of wide review, the less  
weight substantive comments will carry when provided late on the  
Recommendation Track. Worthy ideas should be recorded even when not  
incorporated into a mature document."""


On 13 May 2009, at 12:32, Michael Schneider wrote:
[snip]
> * Reification Comment: I'm happy with this draft. And I agree that  
> a lot more could be said. If this should become necessary, I will  
> be happy to volunteer to say more on this. The WG may take me as a  
> champion for *not* (re|ab)using RDF reification as OWL's annotation  
> vocabulary.

I don't care one way or the other. This is a bone I'd be prepared to  
throw them.

Michael, as the champion, would it bug you terribly to give in on  
this point?

> * "Web-SROIQ" Comment: Maybe we should articulate that essentially  
> all the "problematic" stuff was already contained in the original  
> OWL member submission, where, to my knowledge, there was no  
> distinction between the "OWL" part and the "Web-SROIQ" stuff -- and  
> one of the submitting members was Top Quadrant!

I wouldn't try that bit of argument. After all, TQ can change its  
mind. A lot can change in a few years. Let's not give *more* weight  
to the fact they've changed their mind.

I think the current reply is fine. I hope that the director will take  
the time to write JJC and TQ to say that such frivolous comments are  
really not helpful and not valued.

> * NegProp Comment: The proposed change would have a large effect on  
> the RDF-Based Semantics document. Many parts would need to be  
> changed, and it would take me quite some thinking on how to treat  
> this encoding. And it would be a pain to express in OWL Full what  
> will then so easy in OWL DL: Having a negative property assertion  
> on the right hand side of an entailment query, which IMO will often  
> be a reasonable thing to ask for. And then there is also the round- 
> tripping issue, which would probably have a lot of ramifications.  
> So I deny this proposal.

Also, the *only* argument is the reducing addition vocabulary. It's  
quite clear that the current proposal is very very likely to be  
interoperably implemented as it's very easy to implement (e.g., see  
Ivan's implementation). It's also intention revealing, introduces  
symmetry into the language, etc. etc.

> * "owl:real" Comment: As a preliminary note (because it's hard to  
> tell without deeper analysis, and I don't have the time for this  
> now), I don't currently see a technical problem with the  
> cardinality of the universe of any OWL 2 RDF-Based interpretation  
> being uncountable. In particular, I don't understand why Jeremy is  
> concerned about a "dependency on the continuum hypothesis" (What  
> dependency?).


  The continum bit is completely off the mark. OWL is a fragment of  
first order logic, thus the (downward) Löwenheim–Skolem theorem  
applies, thus for any uncountable model, there is a countable model.  
Thus, who cares ;)

> I also don't buy the argument that owl:real "breaks RDF  
> conformance" and "is not a datatype". owl:real formally /has/ a  
> lexical space and a lexical-to-value mapping, though the lexical  
> space, and hence the argument space of the mapping, is empty. And  
> Section 5 of RDF Concepts states:
>
>  - Each member of the value space may be paired with any number  
> (including zero)
>    of members of the lexical space (lexical representations for  
> that value).

So it clearly is a datatype.

Actually, is its lexical space empty? Every owl:decimal is an owl  
real, thus quite a few elements of the value space are associated  
with a lexical space element.

> So this seems fine to me. But maybe we should contact one of the  
> editor's of the RDF Concepts spec to clarify how this is exactly  
> meant. :-)


Hard to see any other possible reading!

Cheers,
Bijan.
Received on Wednesday, 13 May 2009 12:05:57 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:42:12 UTC