Re: let's allow the same terms everywhere, esp in triple patterns

On Sat, 20 Nov 2004 13:01:59 -0600, Dan Connolly <> wrote:

> w.r.t.
> The set of triple patterns is
>     (RDF-U union RDF-B union V) x (RDF-U union V) x (RDF-T union V)
>   $Revision: 1.137 $
> Let's make that
>     (RDF-T union V) x (RDF-T union V) x (RDF-T union V)

AndyS commented later about as URI-based relationships as being
fundamental to interoperability of the semantic web.  Best not change
that lightly.

> True, literal subjects and bnode predicates won't match any graphs
> made from RDF/XML documents, but the RDF Core WG
>  "noted that it is aware of no reason why literals should not
>   be subjects and a future WG with a less restrictive charter may
>   extend the syntaxes to allow literals as the subjects of statements."
>    --

One problem I see we have with such a change is that we could build
results that cannot be serialized since we use graph pattern (set of
triple patterns) in CONSTRUCT.

This would be a legal query with your change
  CONSTRUCT  ( "foo" _:a "bar" )

which is not an RDF graph result and none of RDF/XML, N-Triples or
Turtle allow it to be serialised.  There are a variety of ways to fix
this, all not appealing - extend one or all of the syntaxes, create a
new RDF syntax, make one syntax required and the only way to
serialize such extended rdf graphs, drop the CONSTRUCT format.

[My software could handle RDF+ models, but I don't let it, since
it's an RDF system]


Received on Wednesday, 24 November 2004 12:31:34 UTC