RE: RDFS implicitly included?

: Aaron Swartz:
: The RDFS spec states that:
:     The schema description language is[...]implicitly part of every RDF
:     model using the RDF schema machinery.

[Implementors hat on, not schema designers]

Can be interpreted as saying when at least one RDFS resource appears in your
model, then automatically all the RDFS resources (or complete graph) are
implicitly present your model. This addition is implicit (you don't have to
actually have them). So you get the self describing bits for free.

Is there a precise definition of "schema description language" anywhere, on the
offchance that it means something other than the substantitive matter (the RDFS
schema itself) of the RDFS?

: and
:     Every RDF model that draws upon the RDF Schema namespace (implicitly)
:     includes these.

Means if the RDFS namespace appears in your model, you've gone and included
every resource in RDFS.

: I do not understand what this means, nor how one should implement it.

It's all or nothing: when you send me a chunk of rdf that using some bits of
RDFS I can assume that all of RDFS is present with your chunk and leverage that
on my end. For a more operational form, one could invert the staments to say
that RDF processors/apps which see RDFS resources or the namespace in an RDF
model MUST behave as though all the RDFS resources were present in that model.
You can practically extend this to the M&S as well (not a lot of point being
fussy here). I don't recall if the RDFS says this intension extends to any RDF
schema: if not it should say so one way or another.

Note that Jena and Stanford create static java objects of the RDFS and M&S
resources for their APIs, so once you're running them you get the RDFS and M&S
resources as java objects for free (even if your model is empty :).

Bill de hOra

Received on Wednesday, 21 February 2001 21:37:08 UTC