Re: bNodes as local constants

Ulrike Sattler wrote:
> 
> On 30 Apr 2007, at 15:40, Dave Reynolds wrote:
> 
>> Whilst the DL+log approach may be appropriate for OWL/DL integration I 
>> fail to see what benefit it brings to RDF integration. The DL+log 
>> weak-safeness condition seems to rule out most of the RDF processing 
>> use cases.
>>
> 
> Dave, can you explain why this is so? If I have some RDF data with some 
> DL+log rules and some OWL ontology, then the weak safeness roughly means 
> that the rules are only implied to the individual nodes in your RDF data 
> (and not to those whose existence is implied by the OWL ontology 
> only)...how can this rule out most of the RDF processing? Cheers, Uli    

Perhaps I'm misunderstanding but [1] says:

"""P is a set of Datalog rules, where each rule R has the form:

p1(X1) V ... V pn(Xn) <-
           r1(Y1), ... rm(Ym),
           s1(Z1), ... sk(Zk),
           not u1(W1), ... not uh(Wh)

such that n >=0, m >=0, k >=0, h >=0, each pi(Xi), ri(Yi), si(Zi), 
ui(Wi) is an atom and:
– each pi is either a DL-predicate or a Datalog predicate;
– each ri, ui is a Datalog predicate;
– each si is a DL-predicate;
– (Datalog safeness) every variable occurring in
R must appear in at least one of the atoms
r1(Y1)... rm(Ym), s1(Z1)... sk(Zk),
– (weak safeness) every head variable of R must appear
in at least one of the atoms r1(Y1) ... rm(Ym)."""

I read the last clause as saying "every head variable must appear in at 
least one positive Datalog atom".

That seems rather different from your paraphrasing above so no doubt I'm 
misunderstanding something.

I took Jos as suggesting that RDF properties would be treated as role 
names and thus be "DL-predicates" in Rosati's terminology and not 
"Datalog predicates" and so a rule which derives an RDF property (or 
class membership) atom and only involved RDF properties (or class 
membership) atoms would fail the weak-safeness clause.

Dave

[1] Riccardo Rosati, "DL+log: Tight Integration of Description Logics 
and Disjunctive Datalog"
http://www.tonesproject.org/index.php?option=com_docman&task=docclick&Itemid=45&bid=93

-- 
Hewlett-Packard Limited
Registered Office: Cain Road, Bracknell, Berks RG12 1HN
Registered No: 690597 England

Received on Monday, 30 April 2007 15:40:44 UTC