Re: RDF API / URIs of reified statements

Jonas Liljegren wrote:

> Is it the same triple if two models is exactly the same? Is it the
> same tirple if both triples have an equal literal? If yo reifie a
> triple in one model; does that count as a reification of identical
> triples in other models.

I think that equality should be context free, i.e. triples with the same content
should be equal independant of where they occur. Reification should not
propagate, although I'm not sure what it should do as reification is a tricky
area.

> I think that it is the answer to this that should guide in the
> question of wether to choose triple URIs based on the model or not.
>
> I argued in another message [1] that you maby would like to let the
> triples have identical URI's if they is composed of only URIs, but
> diffrent URIs if they have a literal, because that literal could maby
> change. And it is not always the best thing to change the triple URI
> if the literal changes [2].

You bring up alot of good issues that I think should be addressed as part of
figuring out higher layers in the RDF stack. I don't know if we should
distinguish between Resource triples and Literal triples as far as URI
generation. In [2] you bring up issues as far as how to handle change
management. I'd like to see us define an unambiguous set of basic layers that
could then have  higher level policy layered on top of. These higher layers
would need to deal with all the issues of temporal and semantic aliasing and
equivalence.

Gabe

> [1]
> http://lists.w3.org/Archives/Public/www-rdf-interest/1999Nov/0115.html
> [2]
> http://lists.w3.org/Archives/Public/www-rdf-interest/1999Dec/0012.html
>
> ---
> / Jonas  -  http://paranormal.o.se/perl/proj/rdf/schema_editor/

Received on Wednesday, 8 December 1999 21:52:50 UTC