W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > April to June 2004

Re: Objective 4.6: additional semantic knowledge

From: Pat Hayes <phayes@ihmc.us>
Date: Mon, 17 May 2004 12:49:58 -0500
Message-Id: <p06001f0cbccea9395b5b@[]>
To: Eric Prud'hommeaux <eric@w3.org>
Cc: RDF Data Access Working Group <public-rdf-dawg@w3.org>

>On Tue, May 11, 2004 at 04:52:52PM -0700, Rob Shearer wrote:
>>  I appreciate the effort to include OWL in the objectives section, but I
>>  feel the current approach to these other semantic layers is a bit
>>  short-sighted. Some people think RDFS is neat, I think description
>>  logics (and OWL-DL) are pretty spiffy, and others like rules languages
>>  like SWRL. In time people may well come up with other ways of encoding
>>  knowledge. Importantly, only a very few of these languages/technologies
>>  have "structure" that can be sensibly and canonically realized in RDF.

Wait, I think I disagree with that last claim. RDFS is entirely 
encoded in RDF very naturally, OWL is encoded into RDF, albeit 
admittedly rather more awkwardly in places, but not impossibly so. 
SWRL admittedly doesnt have an RDF syntax, but SWRL rules apply to 
OWL syntax which has OWL/RDF as its official interchange standard, 
and usually its the results of the rules, rather than the rules 
themselves, which are relevant to query/answer exchange.

>I think this also underlies your objection to 3.5: subgraph results.
>What KR can you do in other languages that you can't do in RDF,
>however awkwardly?

Well, strictly speaking there's no way to naturally encode a 
universal quantifier in RDF without ignoring or breaking the RDF 
model theory (as N3 does at that point). But, hey, if we allow 
universal quantifiers in query results then we are doing 
arbitrarily-high-complexity FO inference, which is probably a bad 
thing to tackle for a simple query language.

>The presumption when writing the charter was that a large and
>practical subset of KR could be reallized as an RDF graph and
>querying that graph was the problem we were trying to solve.
>I attempted in [1] to pick an example disjunctive query over
>an asserted disjunction and show that since it had a canonical
>representation in OWL, it was easy to report it as a graph
>result. When won't that process work?

Yes, and I think we should take this presumption seriously. The world 
doesn't need yet another query language unless that QL is strongly 
related to RDF.


>  > I would simply recommend that we really address the "RDF as data model
>>  for the semantic web" notion on which all these other technologies are
>>  predicated. Some suggested text:
>>  4.6 Additional semantic knowledge
>>  It should be possible for knowledge encoded in other semantic languages,
>>  such as RDFS, OWL, and SWRL to affect the results of queries about RDF
>>  graphs.
>[1] http://www.w3.org/mid/20040505005012.GG9495@w3.org
>office: +1.617.258.5741 NE43-344, MIT, Cambridge, MA 02144 USA
>cell:   +1.857.222.5741
>Feel free to forward this message to any list for any purpose other than
>email address distribution.

IHMC	(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32501			(850)291 0667    cell
phayes@ihmc.us       http://www.ihmc.us/users/phayes
Received on Monday, 17 May 2004 13:50:00 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:00:26 UTC