Re: RADL

Thanks for the response Marc, I'll be thinking on your points.

But I just realised there was something missing from my last post. I
can't actually think of an application where RDF would be more
appropriate that Web description (Site Summary NG?). But I believe
there are advantages in having a constrained XML syntax, rather than
allowing general RDF/XML.

DOAP [1] has shown that it's possible to get the benefits of the RDF
model, use RDF/XML syntax and still have something that doesn't look
unnecessarily complicated, and can be parsed by regular XML tools. The
fact that it makes it harder to serialise to from RDF-based systems is
a drawback, but at least in the near future I expect most systems
would be looking through plain-XML lenses anyway.

There is the alternative of using a plain-XML format and separately
having a mapping to the RDF model, but as the cost of integrating
compatibility into the format only demands minor syntax adjustments it
would seem perverse to do it any other way.

Aside from it simplifying parsing (and XML schema validation), I do
have an ulterior motive for suggesting a constrained subset of
RDF/XML. I've got some server-side Python I put together to make it
easier to build RESTful service. If the Web description language was
plain XML I could run it through XSLT to generate skeleton code for
services. (I also have an ulterior motive for RDF, because that'll be
my fastest route to service discovery & interop, but that's another
story).

Anyhow, time to mull over WDL...

Cheers,
Danny.

[1] http://usefulinc.com/doap

-- 

http://dannyayers.com

Received on Wednesday, 25 May 2005 19:11:50 UTC