W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > July 2001

Re: A use case for anon nodes - action from telecon

From: Graham Klyne <Graham.Klyne@Baltimore.com>
Date: Thu, 19 Jul 2001 20:13:47 +0100
Message-Id: <5.1.0.14.2.20010719200132.03e47ec0@joy.songbird.com>
To: Jan Grant <cmjg@mercury.ilrt.bris.ac.uk>
Cc: Brian McBride <bwm@hplb.hpl.hp.com>, Aaron Swartz <me@aaronsw.com>, rdf core <w3c-rdfcore-wg@w3.org>
At 12:09 PM 7/19/01 +0100, Jan Grant wrote:
>Heh. I've used anonymous nodes for (at least) three different things:
>
>- placeholders for things that don't have a URI (eg, people)
>- objects that I don't particularly care to name (eg, inventing a
>superclass of two classes after the fact)
>Both of these uses are pretty "standard"; also:

I think they're variations on the same basic case.

>- acting as placeholders in a query template.

This is the one I'm trying to be clear about.

>And I see this last use as legitimate. My querying process is simply
>this: the server offers to locate bits of a graph that match a pattern;
>you describe that pattern to it using anonymous nodes as variables. Yes,
>it might be a semantic overloading, but when I looked* there wasn't any
>one true method for expressing queries in RDF, so I made one up.

That kind-of says it.  Out charter (for RDF 1.0) isn't to add to RDF those 
things that people made up along the way, but to clarify what RDF 1.0 
*does* define, or to add those things for which a compelling case has been 
made to add now rather than to wait for RDF 2.0.

Query variables are clearly a useful feature (I assume that's the sense in 
which you meant "legitimate").  But RDF 1.0 says nothing about query 
variables that I can see, and as yet nobody has come forward and made an 
argument for *adding* them to RDF 1.0

>There is clearly a distinction here - one use is an existence claim of
>somethign without a URI; the other is using an anon node to describe
>something that I don't know the URI of. Logic only knows if these two
>things are really the same.

I am currently convinced that they are logically different.  Pat has 
described the difference as that between existential and universal 
quantification (an argument I can only partially sustain).

#g
--

PS:  we've been discussing whether to deal with anonymous variables as 
variables or Skolem constants.  I've just remembered something about the 
process for converting arbitrary FOL expressions to Horne clause 
form:  transformations are applied so that all the universally quantified 
variables are moved to outermost scope, all the existentials replaced by 
Skolem constants, and the universal quantifiers deleted because all 
variables are implicitly universally quantified.

My point is that the two alternative mechanisms we have been thinking about 
are used here for very complementary purposes.



------------------------------------------------------------
Graham Klyne                    Baltimore Technologies
Strategic Research              Content Security Group
<Graham.Klyne@Baltimore.com>    <http://www.mimesweeper.com>
                                 <http://www.baltimore.com>
------------------------------------------------------------
Received on Thursday, 19 July 2001 15:52:40 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:38:13 EDT