Re: A slight issue with datatypes in OWL 2 RL

Boris,

before we do this... let me just raise this issue: just because we _can_
does not meant that we necessarily _should_.

In my view, one of the goals of RL is a possibility for an easy
implementation, too. With my limited implementation experience the
datatype handling of RL is by far the most complex part of an
implementation. Sure, if one goes for a very efficient implementation
then taking care of things like owl:sameAs becomes also more complex,
but that is not 100% necessary for a compliant thing. Datatype handling
is. (I actually did not even have the time to implement it, I just rely
on the underlying RDF/Python environment and do whatever it can do. I
can see many implementations doing just that.) Oracle has already
indicated that they are not really in favour of an owl:rational
inclusion in OWL RL, and I think their reaction reflects the same concerns.

Based on this I actually do _not_ believe that this is just an editorial
 comment but would definitely warrant a new LC round because it would
significantly add to the complexity of implementations. My personal
interpretation (maybe wrong!) of that comment in the document is that
some datatypes (like rational) may actually be dropped from the list and
not add all other datatypes blindly...

My 2 cents...:-)

Ivan

Boris Motik wrote:
> Hello,
> 
> Here is a Last Call comment about datatypes in OWL 2 RL. This issue was pointed out by Jos de Bruijn during the RIF integration
> meeting, and I remembered it today after a private discussion about datatypes with Zhe. Thanks to both of them!
> 
> Currently, OWL 2 RL disallows certain datatypes on the grounds that reasoning with them would not be polynomial. Now we could
> actually relax this restriction and allow all OWL 2 datatypes to occur in OWL 2 RL ontologies.
> 
> This is actually an oversight of mine, caused by the following technical issue. OWL 2 EL and OWL 2 QL have existential quantifiers;
> hence, you can state existence of concrete objects whose values is not known precisely. But then, if you allow combinations of
> datatypes such that the intersection of possibly negated datatypes is finite, you really do get into problems: your reasoning
> suddenly becomes NP-hard because you need to start guessing the appropriate value of existentially implied object. To prevent this
> from occurring, I selected the set of allowed datatypes in OWL 2 EL such that each intersection of possibly negated datatypes is
> either empty or infinite; then, I merely copied this set to all the profiles.
> 
> As Jos rightly pointed out at the RIF integration meeting, however, OWL 2 RL *does not* have existential quantifiers; consequently,
> the value of each concrete object is fully known. But then, there is no need to actually restrict the set of datatypes: to support a
> datatype, you just need a procedure that recognizes whether some literal is in the range of a particular datatype (which is easy to
> do for all of OWL 2 datatypes).
> 
> 
> The fix to this comment would be to revise the datatypes section for OWL 2 RL and allow all OWL 2 datatypes to occur in OWL 2 RL
> ontologies. Since we already have a note saying that the set of supported datatypes might change, I believe that this change would
> not warrant another Last Call round.
> 
> I'm really sorry about this oversight!
> 
> Regards,
> 
>  Boris
> 
> 

-- 

Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf