W3C home > Mailing lists > Public > xml-dist-app@w3.org > September 2000

RE: Schematron schema for SOAP 1.1 Envelopes

From: Rick <ricko@gate.sinica.edu.tw>
Date: Fri, 22 Sep 2000 04:37:31 -0400 (EDT)
Message-ID: <39CB1A12.FE99360D@gate.sinica.edu.tw>
To: xml-dist-app@w3.org
> From: Andrew Layman <andrewl@microsoft.com>
> Two points deserve to be made:
> (1)
> Name calling and conspiracy theories are not appropriate in rational
> discussion and do not advance technical consideration of the issues.

Err, isn't calling my coments "conspiracy theories" a kind of name

> (2)
> What the specification says is "It is not a goal that it [a namespace name]
> be directly usable for retrieval of a schema (if any exists)."
> That is, the specification says that SOME namespace names are not usable
> directly for resource retrieval.  It does not say that ALL namespace names
> are not usable directly for resource retrieval.

You have made the transition from resource to schema: this is the same
transition that Henryk makes.  But it has not been established that
structural/datatyping schema languages are what should define a
namespace: semantic schemas fit the bill better.  And it has not been
established that we should make namespaces correspond to single
resources: to the contrary, because there is disagreement it suggests
that a namespace should correspond to a directory (which could include
an XML Schema) or be queryable.  

The position that a namespace is not a schema is not the same as that
the namespace URI reference cannot be used to retrieve resources which
could include XML Schemas.  What is missing are the conventions to allow
a plurality of uses for namespaces: namespace URI references as the base
for resource discovery. For example, there is no reason why elements in
a namespace should not have (default) style as part of their
meaning--HTML is a good example of this...XML Schemas' annotation
element (though good as far as it goes) is the wrong design to be the
practical base for this: from my side of the world, there is a strong
premium on being able to fetch resources directly rather than at the end
of a chain of references.  

Equating namespace with schema does not create a way to simply download
the some important information about a namespace, it blocks off
systematic, direct access to plural resources, each of which may be
important: an XML Schema, a CSS default stylesheet, an RDF schema, a
schematron schema for some kinds of business logic, a human readable web
page, an ActiveX applet provided to edit the thing, etc.

> This is useful, but interoperability is not helped by deconstructist readings 
> of the existing specification. 

If someone says that the ability to find XML Schemas at the other end is
"why a namespace identifier is a URI", how is it deconstructionist to
attempt to look in the Namespaces spec for the slimmest evidence of
this, given that there is an explicit section on "motivation"?  Surely
the purpose of the spec is to allow questions like this to be settled.

Rick Jelliffe

(Opinions not given as representative of employer or any working
Received on Friday, 22 September 2000 04:39:17 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 22:01:10 UTC