RE: Owl 1.1 and queries

> Dear All,
> 
> I remember that there was a paper from Manchester about query 
> answering in DLs by using a "rolling-up" technique.

Well, there are more then one, but the latest one using the "rolling-up
technique" is: 
Birte Glimm, Ian Horrocks, Carsten Lutz, and Ulrike Sattler. Conjunctive
Query Answering for the Description Logic SHIQ. Proceedings of the Twentieth
International Joint Conference on Artificial Intelligence (IJCAI 07).
In this joint work with Carsten Lutz from Dresden we show decidability of
and complexity results for conjunctive query answering in SHIQ (OWL Lite). 
It is not yet known, to the best of my knowledge, whether conjunctive
queries for OWL DL (SHOIN) are decidable. One could, of course study the
problem of conjunctive queries for OWL 1.1 without nominals and see what
additional effects the other additions of OWL 1.1 have. 

> I was just wondering if anyone knows whether this would work 
> for OWL 1.1? The reason I ask is that negated roles are now 
> allowed in the A-Box, but (if I remember correctly) the 
> procedure involved taking an A-Box query and transforming it 
> into a new class with existential quantification in the 
> T-Box, and then using existing techniques to infer class 
> membership, so wouldn't you end up trying to make class with 
> a negated role in the T-Box definition?

There is no standardised query language for OWL so far. If you consider
conjunctive queries (CQs), their basic form as known from databases is a
conjunction of atoms of the form A(x) or r(x, y), where A and r are
(positive) concept and role names respectively. Due to the power of DLs, one
can rewrite a CQ with conjuncts C(x), where C is a complex concept such as
(¬A_1 and A_2)(x) into a query that contains only atoms by replacing (¬A_1
and A_2)(x) with an atom A_q(x) for a fresh concept name A_q and extend the
TBox with A_q = (¬A_1 and A_2). For convenience one can, therefore, directly
allow complex concepts in the query. This is, however, not the case for
roles, neither in OWL Lite, DL, or 1.1 since also in OWL 1.1 negations of
roles are allowed only in the ABox. Your example is, therefore, not a
trivial extension of conjunctive queries in their standard form. 

I hope this is helpful for you. Best regards, 
Birte

> E.g:
> 
> People(x) & hasBrother(x,y) & OldMan(y) --> NewQueryClass 
> \equiv People \sqcap \exists hasBrother.OldMan
> 
> but then:
> 
> People(x) & hasBrother(x,y) & OldMan(y)  & ¬ hates (x,y) --> ???
> 
> Apologies if this is obvious, but I was just curious....
> 
> Thanks,
> 
> Matt
> 
> 
> 
>

Received on Thursday, 8 March 2007 09:18:41 UTC