Re: RDF changes

"R.V.Guha" wrote:
> In the spirit of considering changes to what exists, I'd like to propose
> that we  at least consider the following:
> RDF M&S puts too many things in one "layer". Maybe we should
> have a core which does nothing but define the triple model (and
> maybe a syntax). Then comes schema concepts like Class, domain
> etc. Then come layers for container, reification, etc.

That sounds like a good approach.  There has also been past discussion
on rdf-interest about separating model and XML syntax.  Would it make
sense to define a core in terms of graphs, with a simple triple syntax
suitable for formalization.  Then the RDF XML syntax can be defined in
terms of a translation to the simple triple syntax, and so can any
other syntax.

> This will help us put RDF on a firmer logical basis, help explain it,
> help implementations, ...


> Maybe the new thing is not called RDF, but xRDF or something like
> that.

I'm hopeful that we can restructure the specifications and the
model of RDF along the lines you suggest, without breaking so far from
the current specifications that we would want to rename it, or even
changing the namespace for that matter.


Received on Monday, 23 April 2001 03:03:31 UTC