W3C home > Mailing lists > Public > public-rif-wg@w3.org > May 2008

Re: importing RDF and OWL

From: Jos de Bruijn <debruijn@inf.unibz.it>
Date: Mon, 12 May 2008 18:34:58 +0200
Message-ID: <482871B2.7050502@inf.unibz.it>
To: Dave Reynolds <der@hplb.hpl.hp.com>
CC: Sandro Hawke <sandro@w3.org>, public-rif-wg@w3.org


>>> * I'm a little concerned about your notion of promotion in a
>>>   hierarchy of entailment regimes.  If we imagine a world with NAF, then
>>>   promotion might be incorrect.  Maybe add a flag, so users have to
>>>   explicitely "import with context = X" or "import with context >= X" ?
>>
>> When importing RDF/OWL there can be only one context for all imports. 
>> So, any ruleset with both "import with context= X" and "import with 
>> the context=Y" be rejected.
>>
>> Perhaps it would make sense to just specify one context for all imports.
> 
> Not sure I follow that but if it means what I think it means then I 
> agree :-)
> 
> My preference would be to only allow one RDF/OWL context for a given 
> rule set. If you import with two different contexts that's a error (or 
> an undefined situation) rather than defining a promotion mechanism.

Thinking a bit about it, it is actually not possible/feasible to specify 
just one context for all imports, because in the RIF rule set you can 
import another rule set, which may also import RDF graphs with a 
different profile.  It is possible to have this situation be an error, 
but I'm not sure this is desirable.

> If you want to mix multiple entailment regimes then use the future 
> modules mechanism.

Yes, with our current technology we cannot mix regimes.

> 
> Part of my motivation for this is that the context list should be openly 
> extensible. For example, I suspect the OWL-R fragment currently being 
> defined in the OWLWG may be very important for RIF. We might even want 
> to predefine an IRI to refer to it (so long as we can do so without 
> entangling the specs). That seems easier if there isn't a predefined 
> promotion hierarchy to fit into.

The list of profiles is meant to be extensible, and I will update the 
text to reflect that (probably just before the telephone conference 
tomorrow :-)).
But we still need to deal with the situation where different profiles 
are specified, if only in the above-mentioned scenario of rule set imports.

Best, Jos

> 
>>> * ... I think maybe "Context" can be thought of as "Language".  You're
>>>   loading some data/rules/knowledge, and naming the language it's
>>>   written in, in case it's not properly labeled (as RDF data is not).
>>
>> Yes, that is probably be a better name.
>> Other opinions?
> 
> I guess I'm OK with "Language" and prefer it over "Context" but it is 
> more than just the language that is being referred to. How about 
> "Entailment Regime"?
> 
> Dave

-- 
Jos de Bruijn            debruijn@inf.unibz.it
+390471016224         http://www.debruijn.net/
----------------------------------------------
An expert is a person who has made all the
mistakes that can be made in a very narrow
field.
   - Niels Bohr
Received on Monday, 12 May 2008 16:34:58 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:33:48 GMT