RE: IRIs as node labels (proposals and counterproposals)

On Fri, 22 Mar 2002, Jeremy Carroll wrote:

> It does seem reasonable to suggest that anyURI in XML Schema datatypes is
> the current "standard web identifier".

[...]

> Sorry Dan I am not going to reply to your counter-counter proposal - it
> seems surprisingly out of scope!

You just did!

RDF supports datatyping and description of resource characteristics using
RDF properties. People *are* deploying identifier properties already (for
example dc:identifier and its sub-properties). And they're eagerly
awaiting our work on datatyping to bring more precision to their
vocabularies.

We shouldn't be suprised if they use the XML schema datatype anyURI.

This suggests a possible outcome: talk to the XML Schema WG and make sure
IRIs are on their datatyping agenda.

Since my proposal was about what future WGs might do, and about creating
modest rather than heavy work for them, I'm not sure it can be out of
scope!

Either the next WG comes along and says 'oh dear, we've got to allow IRIs
to label nodes', or they say 'oh dear, we need to invent a property URI to
represent IRIs'. We don't go out of scope by suggesting to this
as-yet-unchartered WG that we believe the latter approach is consistent
with the RDF Core 1.0 design. If we tried to do that work here, we'd be
charter stretching, I agree. But sketching how future WGs might build on
our stuff is perfectly OK.

Dan


-- 
mailto:danbri@w3.org
http://www.w3.org/People/DanBri/

Received on Friday, 22 March 2002 06:59:51 UTC