W3C home > Mailing lists > Public > www-rdf-rules@w3.org > September 2001

Re: What is an RDF Query?

From: pat hayes <phayes@ai.uwf.edu>
Date: Tue, 11 Sep 2001 19:10:23 -0500
Message-Id: <v04210101b7c451d147b5@[205.160.76.173]>
To: Sandro Hawke <sandro@w3.org>
Cc: www-rdf-rules@w3.org
> > But there's no way to say 'tell me....' in RDF.
>
>There's no way to say anything useful in RDF without extending it.
>It's fairly straightforward (a reasonable exercise in engineering) to
>extend it to enable you to say "tell me".

Really? You must show me how the semantics goes. I'd like to see a 
model theory for speech acts.

> > graph really would mean something different when used as a query from
> > when used as an assertion. The English sentences "Joe is fat" and "Is
> > Joe fat?" are closely related in meaning, but they aren't the same
> > sentence. Writing a piece of rdf with <rqf>... </rqf> around it would
> > be like adding a question mark to an English assertional sentence to
> > make it into a query.
>
>Give the English sentence, in quotes, '"Joe is fat."' I can make
>assertions or questions.   I can say 'It is true that "Joe is fat."'
>and I can say 'Is it true that "Joe is fat."?'

All you have done is moved the question mark into the metalanguage. 
There is still the assertion/question distinction. You don't get rid 
of that by moving to the metalanguage; it has to do with 
propositional attitudes, not the content of the language.

>I imagine RDF playing the "Joe is fat." bit in both assertions and
>queries.

Sure, although why you want to use such a limited language as RDF 
escapes me (particularly as you seem to agree with this point, cf. 
your first line of this message.)

> I think it can play the other bit too, but I don't care
>enough to go on about that now.   Certainly not on day when you're
>playing the other side.....

This really isn't a matter for debate. Its like having an argument 
about whether or not mass and energy are really the same; not an 
issue any more, it was settled years ago.

>
> > > > Well, RDF has them, so relax :-)
> > >
> > >Thank you.   <big><sigh of relief /> :-) </big>
> >
> > I said in the *assertion* language, mind. <;-/>
>
>Does it have them enough to do the right thing (and turn into
>universal variables) if I put some RDF inside an "is it true that...."
>situation?

There are two issues here. First, using the truth predicate, which is 
just damn silly as it doesnt get you anything new and creates all 
kinds of problems; and putting something in a 'question' context, ie 
querying it instead of asserting it. If you query an existential then 
the variables 'act' like universal variables in an assertion. They 
don't actually turn into universal variables (unless you think of the 
query as a negated assertion, as is done in many inference systems 
but wouldn't work in a web environment), but they are treated the 
same way in processing, basically because the forward inference from
(forall (?x) (....?x...)) to (.....Name...)
looks *just* like the backward inference to
(exists (?x)(....?x....)) from (.....Name....)
In fact all the inference patterns have 'duals' like this (swap 
and/or, exists/forall, and treat negations inversely) which look the 
same backwards as the originals do forwards. But all this neat 
duality depends on keeping the directions straight: if you get them 
mixed up, confusion reigns.

Pat

PS.  BTW, that 'it is true that' is the notorious truth-predicate, 
the thing that produces paradoxes all over the place. As Peter has 
said, this is like using a nuclear reactor to heat your living room. 
Why do this insane thing just to keep RDF as the preferred language, 
when we have understood for years how to say what you want to say 
directly? There is a fully worked out theory - almost one of the most 
thoroughly worked-out pieces of mathematics in the last century, 
thanks to the fuss Goedel created - and an established, mature 
technology, about 15 years' experience with deployed systems, many of 
them in industrial use. The tradeoffs are well understood. We know 
how to do this stuff. Its *obvious* that RDF isn't good enough to do 
it. Its fairly obvious that something like Horn logic is good enough. 
Case closed. If the W3C is completely unable to understand this 
point, then someone else will make a useable standard which will 
become the lingua franca of the semantic web. It won't be RDF, for 
sure.

---------------------------------------------------------------------
(650)859 6569 w
(650)494 3973 h (until September)
phayes@ai.uwf.edu 
http://www.coginst.uwf.edu/~phayes
Received on Tuesday, 11 September 2001 20:10:28 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:53:09 GMT