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

Re: Objective 4.6 -- additional semantic information

From: Eric Prud'hommeaux <eric@w3.org>
Date: Fri, 11 Jun 2004 22:10:07 +0900
To: Jim Hendler <hendler@cs.umd.edu>
Cc: Kendall Clark <kendall@monkeyfist.com>, Rob Shearer <Rob.Shearer@networkinference.com>, RDF Data Access Working Group <public-rdf-dawg@w3.org>
Message-ID: <20040611131007.GD15900@w3.org>

On Fri, Jun 11, 2004 at 08:31:27AM -0400, Jim Hendler wrote:
> 
> Eric-
>  I'm not against disjunction, but let me point out that the Algae 
> case points out exactly the problem with disjunction -- if I have 
> multiple disjunctive clauses then the cross product of all of them 
> must be looked at somewhere -- in your case, you only had to do one 
> query, but over on the server side your query engine had to run all 
> four

I'm not sure I follow. The Algae server had to recurse down the left
side of the disjunction, and, if it found nothing, down the right side
(it's treated as a shortcut OR, not a union). I think this had roughly
the overhead of introducing half of a new conjunction constraint.

When translated to SQL, I think the server has a similar burden.

I'm not dead-set for disjunction, but want to make sure we make the
decisions for the right reasons. I think that Alberto said that it
would be inefficient to impelement disjuntion in his (very slick)
database. But I'm not sure we want to keep users from haing ineffic-
ient tools at their disposal.

>      -- generally the problem w/disjunction is one of the many in 
> query language design where you can either have easy query processing 
> but the querier may have to do extra work or a more expressive query 
> language that allows more powerful queries that are harder to process 
> -- one way or the other the piper has to be paid...
>  -JH
> p.s. FWIW, if I had to commit this minute, I'd propose we use a 
> prolog like notion of allowing disjunction in the "where" clause -- 
> it's a compromise between the two above, so that objective 4.6 seems 
> to make good sense to me

Could you explain this a bit, either privately or on list? I don't
know anything about this. (If it's a pain, feel free to tell me to
read a book.)

> At 16:53 +0900 6/11/04, Eric Prud'hommeaux wrote:
> >On Wed, Jun 09, 2004 at 09:08:35PM -0400, Kendall Clark wrote:
> >>
> >> On Wed, Jun 09, 2004 at 04:59:18PM -0700, Rob Shearer wrote:
> >>
> >> > I must admit that upon re-reading the UC&R doc I am a bit surprised 
> >> that
> >> > disjunction has fallen off the radar. I certainly think users like 
> >> being
> >> > able to form arbitrary boolean constructions.
> >>
> >> FWIW, I was just thinking this today! Along the lines, "don't we need
> >> disjunction as an explicit requirement?" I don't know how or why
> >> disjunction use cases fell out of the doc; maybe this is one of those
> >> many mistakes I've made, but if so, it was totally unwitting.
> >>
> >> In other words, I think that disjunction should be an explicit
> >> requirement about disjunction, and I would be happy to help someone
> >> craft a use case that motivates it.
> >>
> >> > actually getting the benefit of the new system. If features like
> >> > disjunction are so rarely useful, then I fear a lot of us have wasted a
> >> > lot of time defining whole new languages like OWL and SWRL for
> >> > expressing things that are even more esoteric!
> >>
> >> I don't think that's the case at all -- rather, I think that
> >> disjunction just fell off the map. Let's get it back on.
> >
> >[1] finds an Annotea use case motivated by eliminating 3 extra queries
> >(in fact, why I put disjunction into Algae). I provided Algae syntax
> >and Alberto provided RDQL syntax.
> >
> >[1] 
> >http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/thread.html#247
> >--
> >-eric
> >
> >office: +1.617.258.5741 NE43-344, MIT, Cambridge, MA 02144 USA
> >cell:   +1.857.222.5741
> >
> >(eric@w3.org)
> >Feel free to forward this message to any list for any purpose other than
> >email address distribution.
> 
> -- 
> Professor James Hendler			  
> http://www.cs.umd.edu/users/hendler
> Director, Semantic Web and Agent Technologies	  301-405-2696
> Maryland Information and Network Dynamics Lab.	  301-405-6707 (Fax)
> Univ of Maryland, College Park, MD 20742	  240-277-3388 (Cell)

-- 
-eric

office: +1.617.258.5741 NE43-344, MIT, Cambridge, MA 02144 USA
cell:   +1.857.222.5741

(eric@w3.org)
Feel free to forward this message to any list for any purpose other than
email address distribution.
Received on Friday, 11 June 2004 09:09:58 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:19 GMT