W3C home > Mailing lists > Public > www-rdf-interest@w3.org > November 2000

Re: A triple is not unique.

From: Seth Russell <seth@robustai.net>
Date: Sun, 19 Nov 2000 10:32:49 -0800
Message-ID: <3A181CD1.6FB96130@robustai.net>
To: "McBride, Brian" <bwm@hplb.hpl.hp.com>
CC: RDF-IG <www-rdf-interest@w3.org>
"McBride, Brian" wrote:

> The number 1 is not unique.  If you have a 1 in your computer
> and Dan has a 1 in his computer, and I refer to the number 1,
> then which 1 am I referring to?
> The point here is that the triple is abstract.  What you have
> in your computer is a representation of a triple, not the triple
> itself.

Well I would agree with you if you say that a triple will attempt to
represent an abstract thing, and that when the same triple is
asserted in different contexts, they might all refer to the same
abstract thing.  Of course we're happy when that happens because we
have communicated.  But when I talk *about* any triple, I am not
talking about the abstract thing - rather I am talking about its
representation in some context.  The point here is that there is no
triple that can be the actual abstract thing.  Sorry, ya really can't
do it !  So any triple is not abstract, rather it is a tangible part
of a model.

> Giving a URI to a triple will not help.  You'd have to decide if
> you the URI named the triple - i.e. the abstract thing - in which
> case you have changed nothing, or a particular representation of
> a triple, in which case you don'thave a means to refer to the
> triple.

There are two separate issues here:
1) Deciding if two representations refer to the same abstract thing ,
2) Deciding to which representation we refer.

I agree that giving a URI to a triple will not help with the former
at all.  But giving a URI to a particular statement will help with
the latter.  If I may add informal URIs to Jonas's example:

Model 1:

 S1: god notplayswith dice
 S2: S1 statedBy einstein
 S3: S1 statedAt 1950-04-10

Model 2:

 S4: god notplayswith dice
 S5: S4 statedBy jonas
 S6: S4 statedAt 2000-11-17

Then I can say something like:

S7:  S4 is plagiary

Because that statement as described by S5 and S6 is plagiary.

We could do the same thing without URI's if we make the assumption
that triples are unique within a particular collection of statements
(translation within a model); so by identifying the collection and
the triple we can refer to a particular statement.  But I find it
unclear after reading M&S how we can even communicate the fact that
a particular statement belongs in a particular collection.

...so I am really confused!
Seth Russell
Received on Sunday, 19 November 2000 13:31:34 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:46 GMT