Re: properties as nodes etc.

From: "Danny Ayers" <danny666@virgilio.it>

> Presumably the edge (==arc) can have as an attribute the URI of a property
> (e.g. dc:creator), but that this wouldn't be unique. The edge itself would
> have an identity status akin to that of bNodes, only existing in relation
to
> the nodes to which it is connected.
> Is there any good reason within RDF MT why the edge shouldn't have it's
own
> (instance) URI?

... probably because that's the way they designed it.  The standard
structure,  defined by RDF which we now call a triple or edge, has 3 and
only 3, attributes: (<SubjectUri>, <PropertyUri>, <ObjectUri>).   It *is*
it's own instance; but it does *not* have a URI;  at least not one defined
by RDF.   Me thinks giving it a URI, in the sense that a URI is supposed to
identify the same unique thing to everybody in the world, would be folly.
These triple thingies only become unique in that sense when they are seen in
some context.   If your application or file defines that context, well then
give each edge a number ... that's what lots of API's actually do .... but
then don't call it a URI, because everybody will have their own different
number for it.

The way I would like to see RDF extended is to give the context of a triple
a URI.  That way the triple becomes a quad: (<ContextUri>, <SubjectUri>,
<PropertyUri>, <ObjectUri>).  Unique triples can now flow into a context
from whatever source;  not just one file.  They can flow into the same
context because now we have a name for it.  This <ContextUri> identifies the
same thing that is delineated in TimBl's N3 by { .... } enclosures.  But
TimBl doesn't call it a context (which is a dirty word in some circles), he
calls it formulae.  I think programmers and the man in the street will know
what we are talking about just fine if we just call it a context.

Seth Russell

Received on Wednesday, 19 June 2002 09:43:24 UTC