Pete,
I changed the subject line since we've moved a bit off the original topic.
> I guess I'm suggesting this still raises a barrier for generic tools
> which rely on HTTP requests and on following
> rdfs:seeAlso/rdfs:isDefinedBy relations, and have some
> builtin knowledge
> of DC. To that tool, this non-en-US data is simply not available. It's
> not just that the tool can't take advantage of the registry to perform
> advanced queries on the data (I'd be quite happy with that!)
> - the data
> is not available at all. The non-en-US data is available
> _only_ to tools
> in which the developer embeds this additional code specific to ths
> application, even if that code is autogenerated.
I disagree. It's not more work - just different work. With the HTTP
approach you still have to implement code to process the HTTP request. If
you want to take advantage of rdfs:seeAlso/rdfs:isDefinedBy relations then
you need to parse the data sent back (after creating a model from the
serialized version). Your assertion that the data is not available to tools
is not accurate. The data is available. Tool developers simply need to
follow the protocol.
> Maybe that's OK or unavoidable, and I haven't thought this through
> properly, but it seems to me we're effectively saying that the barrier
> for access to non-en-US data is higher than that for en-US data.
Perhaps so, from a broader metadata perspective. Non-English versions of
the schemas are not available on the DCMI Web site (however, there are
numerous links to translations). This is a DCMI policy decision, not a
registry decision.
Regards... harry
> Pete
>
|