Re: XML 1.1 Last Call Comments from the XSL WG

At 10:35 2002 07 17 -0400, Norman Walsh wrote:
>The XSL WG makes the following comments on XML 1.1.
>
>[XSL]We find the extension to XML name and whitespace characters
>uncontroversial.
>
>[XSL]However, we find that the requirement for full normalization is
>overly burdensome and introduces significant technical problems.
>
>  - XSLT constructs result documents and therefore could be required to
>    perform significant additional computation. For some normalization cases,
>    it isn't clear that the additional computation is even well defined.
>
>  - Even when new nodes are not constructed on the fly, they might be
>    copied from different documents, which could potentially be a
>    mixture of XML 1.0 and XML 1.1 documents with different
>    normalization properties.
>
>  - Intermediate result trees that might or might not be serialized
>    would also need to be considered with respect to normalization
>
>[XSL]In as much as the normalization requirement is transitively
>applied from the Character Model specification and given our
>objection[1] to normalization in that document, we cannot agree to the
>mandate for normalization in XML 1.1 until such time as our issues
>with the Character Model specification are resolved.
>
>[XSL]The XSL WG wonders if the XML Core WG has a position on the
>normalization of 1.0 documents to make them 1.1 compliant, and if so
>what role do they envision for XSLT in that process?
>
>                                        For the XSL WG,
>                                          norm


We have recorded this issue at [1]
wherein our resolution is listed as follows:

  Resolution: Noted. The Core WG doesn't see any reason why the XSL WG 
  would need to define XML 1.1 result trees, so the only issue is 
  serialization of 1.1, and algorithms for doing that in the presence
  of sensible data are well-defined. A note will be added to XML 1.1
  about how to serialize 1.1 documents (given normalization).


The above mentioned note in our latest draft at [2] is:

  XML applications that create XML 1.1 output from either XML 1.1 or
  XML 1.0 input should ensure that the output is fully normalized; it
  is not necessary for internal processing forms to be fully normalized. 

Please let us know if you are not satisfied with this resolution of
your comments.

paul
Paul Grosso, co-chair for the XML Core WG

[1] http://www.w3.org/XML/Group/2002/01/blueberry-comments#XSL1
[2] http://www.w3.org/XML/Group/2002/08/CR-xml11-20020820

Received on Wednesday, 21 August 2002 14:02:30 UTC