RE: XML Schema WG Comments on XInclude

The Schema WG discussed your responses to our comments on the XML Inclusions
Last Call Working Draft.  Unfortunately, we were only able to cover points (1)
through (3), but in the interests of a timely answer, we send you this message
on those three points now. We hope to consider the remaining points soon.

We are pleased with the thoughtful response from the XML Core WG on our issues.
We are satisfied with your answers for points (1) and (2).  

With respect to point (3), we continue to have concerns. We believe that there
are important use cases that require there to be some marker that XInclude
processing has occurred, and in the interests of interoperability we believe
that the XInclude specification is where the infoset property marking that fact
should be defined. We see an interaction between handling of extended infoset
properties, such as the PSVI, and the presence of a well-known marker. That is,
if the user option for preserving PSVI (or other extension) properties has been
selected in XInclude processing, it would be unfortunate for a processor
that understands how to "fix-up" the properties to have no way of knowing
where the fix was required. That is, such a property allows subsequent
processing that cares about infoset integrity to know where to work.

We therefore ask that the Core WG reconsider this point, and define an infoset 
property to act as a marker, with the suggestion that perhaps the provision of
such information might be under user or processor control.

Received on Monday, 14 January 2002 12:52:32 UTC