RE: DATATYPES: mental dump.

> >I.e., I've understood that my X proposal was quite
> >similar in principle to your P++ proposal in that it
> >defines urirefs and literals as just labels of nodes,
> >and nodes are unique in the graph,
> 
> I am puzzled, since this is just true of all RDF graphs: urirefs and 
> literals ARE labels of nodes, and nodes ARE unique. Right. So what 
> does it mean to say that the X proposal 'defines' this to be the 
> case? All of RDF assumes this to be the case, surely (??)

Not in terms of "tidying".

Of course, nodes themselves are unique. What I meant is that
while nodes with the same uriref labels might get merged,
nodes with the same literal labels (or no labels) would
not.

Thus, each literal is "hosted" by its own node and those
don't ever merge. This allows literals (or rather their nodes)
to serve as subjects. 

> >and one would expect
> >to merge any two nodes having the same uriref label,
> >but not merge two nodes having the same literal label
> >(which is why there are distinct UNodes and LNodes
> >in the X model)
> 
> I confess to not really understanding the X model. It seems to be a 
> proposal to describe RDF graphs in RDF, a kind of meta-RDF . 

That's exactly what it is. Sort of ;-)
 
> I don't 
> really see what the point of doing this is, or how you would get back 
> to the original assertions you started with.

But the assertions are what are being modeled. X model
constructs are themselves not RDF statements, thus we
are not making assertions about assertions in X graphs.
We are just modeling assertions.

> >Thus, in fact, P++ is expressable in RDF/XML but requires
> >a means to clearly denote the node identity and the
> >label.
> >
> >So while there may not be an "elegant" way to express
> >P++ in RDF/XML, I think it is at least expressable
> >according to a given idiom.
> 
> If you had said just 'expressible in XML' then yes; but RDF/XML 
> already is an idiom.

I thought an 'idiom' was a way to consistently express something
in RDF which is not apparent by the graph structure it self, such
as e.g. the DC approach where special understanding of certain
subgraph "templates" are needed to interpret the "idiom".

How were you using the term 'idiom'? ;-)

Patrick

Received on Thursday, 15 November 2001 16:46:19 UTC