W3C home > Mailing lists > Public > public-rdf-wg@w3.org > February 2012

Re: Another try.

From: Pat Hayes <phayes@ihmc.us>
Date: Tue, 21 Feb 2012 15:55:22 -0600
Cc: andy.seaborne@epimorphics.com, public-rdf-wg@w3.org
Message-Id: <7A43AC75-82EC-4A79-A577-AF46213F756B@ihmc.us>
To: William Waites <wwaites@tardis.ed.ac.uk>

On Feb 21, 2012, at 6:03 AM, William Waites wrote:

> On Tue, 21 Feb 2012 11:44:56 +0000, Andy Seaborne <andy.seaborne@epimorphics.com> said:
>> If I understand the quad proposal, then all existing
>> vocabularies are technically undefined because they never define
>> P(S,O,G), only P(S,O).
> Back in December, I experimented with a tuple-store that understood a
> variant of n-triples (mostly as an exercise in learning Haskell
> properly). It had datatypes of URI, Bnode, Literal and Nil and
> supported predicates of any arity.
> Is it useful define such a nil value and say:
>  For all (P, S, O), P(S,O) -> P(S,O,nil)

I thought about this. (Like a null value in an RDB?) We could try to do this, but it gets complicated and leads to all kinds of decisions that have to be taken and however we take them, someone won't like it. And I don't think there really is a demand for it, to be honest. So I would rather not try to make this more complicated, but just put the quads/context machinery out there and let people use it any way they like. The bare-bones proposal makes no necessary connection at all between the meanings of P(S, O) and P(S, O, G): they might be related in some way, but they also might not. It depends on how people want to use the particular vocabulary.

If we do legislate on this, the best rule would be, I suggest, to treat a missing context as meaning an existential quantification:

P(S, O) <=> (exists x) P(S, O, x)

which is like calling it a blank node. This corresponds to the natural English reading. Joe kissed Mary on Tuesday. Bill kissed Mary. When did Bill kiss Mary? Don't know, but it was at *some* time. Again: the TImes reported that Kissinger kissed Nixon. Which edition of the TImes reported this? Don't know, but it was surely one of them. 

> Does this connect to SPARQL's idea of a default graph in a useful way?

I dont think it does, actually. 


PS. Just noticed your email moniker. Didnt know you knew the Doctor that well :-)

> -w

IHMC                                     (850)434 8903 or (650)494 3973   
40 South Alcaniz St.           (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes
Received on Tuesday, 21 February 2012 21:56:00 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 22:02:03 UTC