W3C home > Mailing lists > Public > www-ws-desc@w3.org > November 2002

Re: Using alternative schema languages

From: Amelia A. Lewis <alewis@tibco.com>
Date: Thu, 21 Nov 2002 12:15:31 -0500
To: Jacek Kopecky <jacek@systinet.com>
Cc: www-ws-desc@w3.org
Message-Id: <20021121121531.75278e03.alewis@tibco.com>


On 21 Nov 2002 18:01:14 +0100
Jacek Kopecky <jacek@systinet.com> wrote:
> quoted below is your issue. It seems to be concerned about ns1:type1
> being XML Schema definition of the same XML type as ns2:type2 in other
> type system. I don't think this is a problem.

No, actually, it covers both cases, potentially.

> What I'm concerned about is the case where ns1:type1 is an XML Schema
> definition of a type, ns1:type1 is also a RelaxNG definition of a
> (possibly different or same, doesn't matter) type. I think I heard you
> say on the call that you'd disallow this case.

This all depends upon whether multiple embedded or imported schemas are permitted to declare that they describe the same namespace.  Possibly we wish to allow several embedded or imported schemas to express that they do so.  If so, the problem arises.  If the schemas do not claim to represent the same namespace, then they cannot produce the same QName.

So perhaps my issue was phrased too infosetty.  But it does, I think, touch this question.  If two schemas can claim to define the same namespace, then there's a conflict, and we have to specify how to resolve the conflict.  If they are not allowed to define the same namespace, which I believe is the current situation (although I might be wrong), then the problem cannot arise, except as in the original example that you give in the first para, above.

> Referring to both types using the unqualified attribute 'type' would
> effectively disallow this case and I think this would be a serious issue
> with our support for different type systems. If we require that the
> attributes for referencing definitions in other type systems be
> namespace-qualified, this problem goes away, and IMHO much confusion is
> averted, too.

I don't believe that this is an ideal solution, in fact.  If we allow two schemas of different languages to declare that they describe the same namespace, then why should we forbid two W3C XML Schemas to declare the same namespace?  And if we permit the latter, then how do we differentiate between conflicting definitions?

Amelia A. Lewis
Architect, TIBCO/Extensibility, Inc.
Received on Thursday, 21 November 2002 12:15:39 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:06:26 UTC