RE: Banning relative - No real damage?

Please note that the relative schema references (XDR and not XSD) are
expressed using an absolute URI that conveys relative resolution semantics
(using the x-schema "protocol").

The only point where we use relative namespaces are for XSLT function
extensions. However, there the namespace is just used for identity and could
be anything. Also, I do not think that they are generated from our tools. 

Somebody showed an example of the extension mechanism earlier [1]:

<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
                xmlns:msxsl="urn:schemas-microsoft-com:xslt"
                xmlns:user="http://mycompany.com/mynamespace"
                version="1.0">
  <msxsl:script language="JScript" implements-prefix="user">
    function xml(nodelist) {
      return nodelist.nextNode().xml;
    }
  </msxsl:script>

  <xsl:template match="/">
    <xsl:value-of select="user:xml(.)"/>
  </xsl:template>
</xsl:stylesheet>

Here the user prefix is bound to an absolute namespace, but it could be
bound to a relative one. It is up to the author of the stylesheet to decide.

So the issue (as I repeat here for the 5th time at least) is not our tools
but more that people that use our and other people's tools may have used
relative namespaces in documents based on the literal interpretation and
built corresponding tools since all tools out there provide for literal
interpretation (more so than absolutization). As our mail today indicate, we
however assume that this currently constitutes a very small amount of cases.

As to your question below: We are still listening (however Jonathan is on
well-deserved vacation and I am catching up, having been at a W3C meeting).
We certainly still care and were internally looking for solutions to this
problem. A proposed solution has been forwarded by David Turner earlier
today. 

My apologies if we look mostly silent, but we just do not have the bandwidth
to indulge in 100+ mails/day discussions. We will however post factual
information and corrections such I tried to do in this mail and make
constructive proposals that we can live with.

Best regards
Michael 

> -----Original Message-----
> From: keshlam@us.ibm.com [mailto:keshlam@us.ibm.com]
> Sent: Wednesday, June 07, 2000 7:53 AM
> To: xml-uri@w3.org
> Subject: Re: Banning relative - No real damage?
> 
> 
> 
> >What Microsoft claims, and has claimed, is that their *tools*
> >generate such documents.  The actual documents are in the 
> hands of their
> >customers, who are probably not listening.
> 
> Yep.
> 
> The question then becomes: How many of these tool-generated 
> documents are
> _transient_? If it's the majority, one can argue that 
> migrating the tool to
> generate absolute namespace names and use something like 
> xmlns-binding: for
> the relative reference to the XSD Schema (which seemed to be their
> concern), might be possible without much adverse effect on 
> the users. Add
> support for the new indirection, start generating new 
> documents, then some
> time later phase out support for the old documents.
> 
> If the tool-generated documents are persistant, one can still 
> easily write
> a tool to convert them to the new convention... but the 
> problem is finding
> all those documents. If I were attempting to do this, I'd 
> suggest writing
> the conversion into the tools themselves... "A relative 
> namespace name has
> been encountered. This usage was permitted by the original 
> Namespace spec,
> but has been deprecated by the W3C. Do you want to reorganize 
> the document
> into the new recommended form or abort processing?"
> 
> 
> Are there still any Microsoft folks involved in the 
> discussion? If so, what
> say you -- is this a plausable approach to serving your customers?
>

Received on Thursday, 8 June 2000 20:38:11 UTC