W3C home > Mailing lists > Public > www-xml-linking-comments@w3.org > January to March 2002

Re: No requirement for XML Base support = no interoperability!

From: Paul Grosso <pgrosso@arbortext.com>
Date: Thu, 21 Mar 2002 17:24:47 -0600
Message-Id: <>
To: "Roger L. Costello" <costello@mitre.org>, www-xml-linking-comments@w3.org, costello@mitre.org
At 13:03 2002 03 14 -0500, Roger L. Costello wrote:
>As far as I can tell, XML Base is a technology that is "available, but
>there's no requirement for tools to support it".  
>For example, with XML Schemas you can use a relative URL in
>schemaLocation, with xml:base providing the base URL:
>  <BookCatalogue xml:base="http://www.example.org"
>                 schemaLocation="BookCatalogue.xsd"
>                 ...
>An XML Base-aware schema validator tool should create this as the URL
>for schemaLocation:
>   http://www.example.org/BookCatalogue.xsd
>The problem is that there is no "requirement" for schema validators to
>support XML Base (and thusly no schema validators support it).  There is
>no statement in the XML Schema spec saying "for a schema validator to be
>conformant it must support XML Base". 

That is an issue for XML Schema.

>XLink, XInclude seem to be in the same predicament as XML Schemas, i.e.,
>they state that XML Base can be used but there is no requirement.

Not true.  The conformance clauses make it clear.

For XLink, http://www.w3.org/TR/xlink/#N1022 says XLink processing
depends on XML Base (and, being in the conformance section, that
"depends" is meant to mean normatively).  To make it clearer, in
http://www.w3.org/TR/xlink/#link-locators it says "If the URI reference 
is relative, its absolute version must be computed by the method of 
[XML Base] before use" and *must* is italicized to show it is a
normative requirement.

For XInclude, http://www.w3.org/TR/xinclude/#application says
"An application conforms to XInclude if it supports XML 1.0, 
XML namespaces, the XML Information Set, and XML Base."

>This lack of clear direction will kill interoperability.
>The problem seems to stem from the fact that they (XML Schemas, XLink,
>XInclude) all "pass the buck" to XML Namespaces, which states:
>"Several information items have a [base URI] or [declaration base URI]
>property.  These items are computed according to [XML Base]"

I think both XLink and XInclude are clear.  It sounds like XML Schemas
is the one with a problem.

>Note that it does not say, "These items MUST BE computed according to
>[XML Base]".  As a result, there is no requirement to support XML Base,
>and everyone is left hanging.

You are correct that there is no requirement that all specs require
support for XML Base.  The problem is that XML Base came out after
many other specs, so it could not be required after the fact.  

>XML Base is a great idea, but it sure doesn't seem very useful unless
>there is a hard requirement for tools to support it.  Can someone please
>clarify this issue for me?  Am I missing something?  /Roger

I agree that there may be some interoperability problems, especially
once we get into pipelines of processes, some of which support XML Base
and others that don't.  

It sounds like XML Schema should make it clear that conformance to it
requires conformance to XML Base.  But that isn't something that
XML Base can say, that's something only XML Schema can say.

Received on Thursday, 21 March 2002 18:29:15 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:32:24 UTC