[
Lists Home |
Date Index |
Thread Index
]
From: <MDaconta@aol.com>
>
> While there are possible workarounds, I don't believe the current web-service
> tools or specs support using Relax NG. And when explaining this to policy makers,
> a custom workaround without significant benefits won't fly.
I think there are three possible good objections:
* WSDL does not allow other schema languages (wrong)
* WSDL implementations require WSDL tools (wrong)
* WSDL requires a type system and RELAX NG is not useable a type system (wrong).
On the first point: WSDL part 1 allows use of other schema languages.
" WSDL recognizes the need for rich type systems for describing message formats,
and supports the XML Schemas specification (XSD) as its canonical type system.
However, since it is unreasonable to expect a single type system grammar to be used
to describe all message formats present and future, WSDL allows using other type
definition languages via extensibility."
http://www.w3.org/TR/2001/NOTE-wsdl-20010315
On the second point, WSDL can be used as a specification for programmers.
I suspect that typically advertisement and discovery of a web service are
independent of its adoption, at least in the most simple case (Does anyone
use trusted repositories?) So the absolute requirement that you can go
from discovery to generated interfaces without human intervention looks
like a furphy. For example, that is our expectation with the Topologi
Validating Proxy: a programmer can use the WSDL as a spec, extract
the WXS or Schematron schema and configure the proxy with that;
we are happy to provide RELAX NG if anyone wants that, b.t.w.
On the third point, it is certainly true that Schematron is not a "type system"
and RELAX NG was not developed as such. However, unambiguous RELAX
NG schemas can be used for unambiguous typing. But when I read the
WSDL spec, it seems that readers can go too far with the terminology "type systems":
if the aim of the web service is merely to provide a computer readable
XML specification of the interface, then any expressive-enough validation
language will do.
Cheers
Rick Jelliffe
Topologi, Pty. Ltd.
|
