W3C home > Mailing lists > Public > public-rif-wg@w3.org > March 2006

Re: [UCR] RIF needs different reasoning methods

From: Ian Horrocks <horrocks@cs.man.ac.uk>
Date: Tue, 7 Mar 2006 15:18:30 +0000
Message-Id: <ceb865cf423a084550762aa09c9f9739@cs.man.ac.uk>
Cc: Francois Bry <bry@ifi.lmu.de>, public-rif-wg@w3.org
To: Bijan Parsia <bparsia@isr.umd.edu>

On 7 Mar 2006, at 14:28, Bijan Parsia wrote:

>
> On Mar 7, 2006, at 2:41 AM, Francois Bry wrote:
> [snip]
>> I gave concrete examples that need reasoning techniques as 
>> develioppped
>> and used in databases.
>
> But these correspond to well specified semantic/expressive subsets. 
> Why the *further* need to distinguish the reasoning technique used.
>
>>  General purpose reasoners arein pracxtice not
>> applicable to these cases.
>
> What if they are? Why aren't the efficiently parameters enough to 
> specify?
>
>>  If RIF does not ways to consider such
>> practical issues, then it won't be  successful in practice.
>
> I'm just trying to discern what the right ways are. As far as I can 
> tell, my slightly more abtract approach is pragmatically equivalent to 
> what I can figure out of what you what to do. But you don't seem to 
> agree, so I'm confused :)
>
> What I don't see is the reason for specifying particular proof 
> procedures *instead of* expressive subsets. Responsiveness 
> requirements are application dependent, not document dependent (as far 
> as I can tell). That is, you want to *in the context of a particular 
> application* specify which reasoner, given certain parameters, to use. 
> If two reasoners perform acceptibly and give the same answers...what 
> *more* do you need? But you *seem* to be saying that more is needed to 
> be practical (something above specifying the particular reasoner, but 
> below specifying the expressive subset + performance parameters), and 
> that this is something that goes with the document as part of the 
> specification of the ruleset.
>
> This is the part I don't understand. I read Ed's email, but I didn't 
> really get enlightenment, except maybe that you want to be able to 
> express reactive rules (as opposed to rules that may be processed 
> reactively). But, I'm sorry, I don't see why these are necessary to 
> meet your goals. I don't doubt the goals, I just don't see the 
> necessary connection with your means.

I am also baffled by the need for or utility of proof procedure 
specification.

Given that a particular prover has an adequate response time and 
provides correct answers (as defined by the semantics), then why would 
I care what procedure it uses in order to do its work? Moreover, how 
would I even be able to distinguish what different provers are doing if 
they all give the same answers? They could claim to be using any old 
technique, and I would have no way to tell if it was true or not.

Bottom line - what is the point of specifying something that has no 
*discernable* effect?

Ian



>
> Cheers,
> Bijan.
>
>
Received on Tuesday, 7 March 2006 15:18:37 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:33:27 GMT