W3C home > Mailing lists > Public > public-owl-wg@w3.org > August 2008

Re: Some editorial comments (Re: ISSUE-131: Preview of "unification")

From: Ian Horrocks <ian.horrocks@comlab.ox.ac.uk>
Date: Mon, 25 Aug 2008 15:34:23 +0100
Message-Id: <BB7FAA3C-E01A-4E48-B0DA-5ADB78DD9EF9@comlab.ox.ac.uk>
Cc: public-owl-wg Group WG <public-owl-wg@w3.org>
To: Ivan Herman <ivan@w3.org>

On 25 Aug 2008, at 14:12, Ivan Herman wrote:

> Hi Ian,
> first of all: thank you!
> I have some editorial issues/questions/comments on the profile  
> document.
> I may also have some more content questions at some point, but I  
> thought
> it would be better to get the editorial comments out of the way... So
> here they are.
> - In the definition of OWL 2 EL, you refer to "satisfiability,
> subsumption, classification, and instance checking" as being  
> polynomial.
> On the other hand, the introduction refers to "consistency,
> classification, instance checking, and conjunctive query" as the
> reasoning tasks used to describe the profiles. I think that the
> introduction should describe all of those. Having said that: aren't
> 'satisfiability' and 'consistency' the same, or am I missing the  
> point?

This clearly needs tidying up a bit and making more consistent across  
the profiles. Consistency and satisfiability are often used  
interchangably. I might be able to persuade myself that the  
unqualified use of consistency usually refers to ontology consistency  
(is there any model of the ontology as a whole) whereas the  
unqualified use of satisfiability usually refers to classes (is there  
a model of the ontology in which the interpretation of a given class  
is non-empty). However, I think that it is better to explicitly  
distinguish when we are talking about the ontology and when we are  
talking about classes (e.g., ontology/class consistency).

> - In EL, bulleted item under 'The following features...', 2nd item,
> there seem to be an extra ')' in the middle and an extra one at the  
> end, too


> - Section 3, introduction on QL, second paragraph refers to DL-Lite  
> and
> not QL, which is a bit out of the blue here (I suspect this is a
> leftover). Actually, I wonder whether that paragraph should stay at  
> all.

I changed it to say that QL is based on a variant of DL-Lite. I think  
this is useful as it allows us to explain the basis for QL, to point  
to the literature and explain why this particular variant was chosen.

> - Whereas the start up section of EL referred to the various reasoning
> problems for which EL is well suited for, this is not so explicitly
> stated for QL. I think it would be good to have a similar
> characterization of QL, too; if _all_ reasoning problems are in
> LOGSPACE, then it should be said explicitly.

Good point. Some more work is needed on the introductory and  
explanatory text throughout the document. It has lagged a bit behind  
the others due to the "planning blight" caused by the possible  
unification. Once this is resolved one way or the other we can get on  
with this work.

> - Section 4, introducing OWL RL, I actually have the same comment as
> before: it may be useful to be able to characterize (at least  
> partially)
> the profile with the reasoning problems. For example, isn't it correct
> that the OWL 2 RL/RDF rules are not complete vs. classification but  
> they
> are vs. conjuctive query and instance checking, for example?

I agree that we should be clearer about all this. Completeness  
depends mainly on the structure of the ontology. If the ontology  
satisfies the relevant syntactic conditions, then the rules are  
complete for ground entailment. This means that they can also be used  
for complete consistency checking (both class and ontology),  
subsumption and classification. E.g., C is a subClassOf D just in  
case asserting T(a rdf:type C) entails T(a rdf:type D) for some "a"  
that is not mentioned elsewhere in the ontology. This is a standard  
technique in rule based reasoning.

> (I think the reference to these reasoning tasks is a very good way of
> comparing the various profiles, hence these remarks...)
> - I do not want to reopen the profile naming discussion too much,  
> but:-(
> Is it necessary to add the '2' to all profiles? I would have thought
> that 'OWL QL', 'OWL RL', etc, would be o.k. It would also make the  
> rule
> representation a bit shorter: 'OWL RL/RDF' instead of 'OWL 2 RL/RDF'

This would be fine for me -- there is no confusion because OWL 1  
didn't have these profiles. If others agree, then changing the  
document is a trivial task.


> I guess that is for the moment after a first read through the text...
> Thanks
> Ivan
> Ian Horrocks wrote:
>> We (Alan and I) agreed that it would help to clarify this issue  
>> and to
>> inform our discussion on Wednesday if the Profiles document [1] were
>> updated to reflect the proposed "unification". This has now been  
>> done.
>> It should be read in conjunction with the (draft) conformance
>> definitions [2].
>> Regards,
>> Ian
>> [1] http://www.w3.org/2007/OWL/wiki/Profiles
>> [2] http://www.w3.org/2007/OWL/wiki/Conformance
> -- 
> Ivan Herman, W3C Semantic Web Activity Lead
> Home: http://www.w3.org/People/Ivan/
> PGP Key: http://www.ivan-herman.net/pgpkey.html
> FOAF: http://www.ivan-herman.net/foaf.rdf
Received on Monday, 25 August 2008 14:35:00 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:41:51 UTC