Eve Maler has published the first comments from a XSL and XML Linking WGs task force proposing fixes for issues raised at the borders between XLink/XPointer and XSL.
Posted on www-xml-linking-comments@w3.org, the message from Eve Maler is summarizing and motivating a note that should be made public soon.
The task force is asking to the XML Linking group to consider making a couple of minor changes
to the current CR and to consider three more important changes for a future release.
The minor changes are about the
"Behavior of Multiple onLoad/replace Links in a Document" currently specified
to follow the document order which is not realistic when the document is processed through XSLT and
"Non-Well-Balanced and Discontiguous Ending Resources" which representation
could be given more leeway than the only "unified concatenation" method currently specified.
The changes left for a future release cover enhanced control over the presentation of the
"ending resources", allowing to force the choice of a stylesheet for these
resources and differentiating between "Processing Context and Presentation Context"
to present them in a "larger context".
The issue of the support of the "ranges" XPath extension by XSLT and CSS
"whose data model granularity is coarser than the Infoset" is also falling into
this category and could rely on a subset of XPointer
"whose data model has a better match with these technologies".
Last week, James Clark had answered other concerns about the support of XPointer by
XSLT processors reminding that XPointer is the method chosen by the IETF specification to
define fragments for text/xml documents and that any text/xml compliant XSLT processor
needs to support XPointer through the document().