W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > October to December 2010

Re: Proposed change to the OWL-2 Direct Semantics entailment regime

From: Bijan Parsia <bparsia@cs.man.ac.uk>
Date: Sun, 5 Dec 2010 18:21:48 +0000
Message-Id: <7082C679-4B0D-492E-91D7-BA5DE31D3634@cs.man.ac.uk>
To: SPARQL Working Group <public-rdf-dawg@w3.org>
On 5 Dec 2010, at 18:06, Enrico Franconi wrote:

> Your argument is all about the systems Manchester, Oxford, and your company are implementing and their users.

Racer systems as well. And actaully, it's not all about them, though they play a critical role.

> You are completely forgetting the support of the OWL profiles with lower expressivity (such as OWL-QL),

No, I'm not.

> in which distinguished variables are crucial in applications (DB queries, information integration)

Please point me to how this is so, exactly? I'm really not being silly about this: I don't see the *essential* connection.

> and the current spec would leave them out and the cost to adapt would be big. 
> 
> As I said, I also don't see how OWL-DL systems with UNA can easily adapt.

There are no such systems (OWL-DL that is)? I mean, Pellet has a mode switch. But UNA isn't part of OWL so I'm confused why that's relevant.

> I also got this comment from the RACER implementors: "The standard might explicitly allow implementation to impose restrictions on the use of non-distinguished vars, and then I think including them will not be harmful."

This point was made by both Birte and myself. If you restrict the patterns of bnodes for OWL DL then you can have implementable systems (see Pellet; I pointed out example queries). The argument against this from a design perspective is that it introduces more irregularity in what queries can be handled by what systems and that the same functionality can be gotten and (IMHO) more clearly expressed with class expressions.

That is, if you do something harmful (to users) it won't be (implemetnationally) harmful. We knew this!

> I will collect the opinions of all the system implementors of OWL-QL I know of (Rome, Bolzano, IBM, CNR).

Can I ask exactly what you are collecting? Preferences wrt regime? Implementation considerations? Are you going to present the full set of options in a neutral way?

> Be sure that they will agree with me.

Ah. I see.

> Should we really play this game, rather than trust each other?

Enrico, you are the only one bringing up trust (ok, I pushed back a bit above :)). And you are the only one expressing contempt toward your interlocutors both overtly (with explicit insults) and covertly (by ignoring key arguments and implying nefarious motivations or lack of expertise). I don't think that such dialogue is actually productive.

I think it would be helpful if you just took stock for a minute and consider my last email. You may think that the argument is, on balance, not as strong as yours, but it is, I think, reasonable and coherent and establishes a somewhat high cost for the regime you favor. To respond to it, you need to show that the cost estimate is wrong or that the benefits are reasonably high. 

Cheers,
Bijan.
Received on Sunday, 5 December 2010 18:22:19 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:44 GMT