Lists Home |
Date Index |
Elliotte Rusty Harold wrote:
> At 11:17 AM +0100 6/10/04, Bill de hÓra wrote:
>> You really don't understand what this thread is about, do you?
> Actually, Bill I'm afraid it's you who is confused. Apologies if I
> confused you. I did wonder how in one sentence you'd say I was totally
> wrong and in the next sentence you'd say you agreed with me.
Probably because you're talking about many things in different
layers. I'm less confused than you think. But even though I find the
lack of clarity and critical thinking annoying, I should be less
testy in my posts (if we were talking about binary XML or Unicode
arcana in this way, would you put up with it?).
> The question at hand is when one is publishing data should it be
> published in XML+Namespaces or in RDF.
Ok. Obviously XML+Namespaces - you can't publish anything /in/ RDF -
it's entirely abstract and has no syntax.
However if we were to suppose RDF/XML, then RDF/XML is inscribed
using XML+Namespaces (encoding URIs in XML for RDF was a use case
for Namespaces as I recall).
So I suspect you're asking the wrong question. The question perhaps
is whether we should /model/ data using RDF and not with traditional
programming types or the less formal, ad-hoc or implicit notions we
use today and encode in XML. Personally I don't much care whether we
target RDF for modeling for much the same reasons you and Walt Perry
don't much care about draconian schemata and error handling - if RDF
triples are what is wanted out of an XML grammar a local mapping
will suffice (for example as I've done for Atom).
If XML provides syntactic interop then RDF provides relational
interop. I won't be surprised if RDF is eventually targeted as a
database interlingua rather than an ontology/gofai language. To me
the interesting thing is not how RDF relates to XML, but how it
relates to Object, UML and business middleware. The prime motivation
to model in RDF is surely to reduce the amount of code you need to
process data, not to outmode XML...
> Are there high quality RDF tools that
> enable one to do things with RDF data that can't be as easily done with
> the high quality XML tools operating on the XML data?
With the possible exception of Jena and Redland, I don't think so.
The most interesting technology imo for RDF is RDQL. But I'm not too
concerned about adoption rates right now. For all I know RDF is akin
to SGML and not XML in terms of "adoptionality" and the kind of
things I find it useful for don't require mass adoption :)