RE: [serialization] XHTML output method

Max,

It seems to me to be a questionable approach to tie serialization of XSLT /
XQuery / whatever so tightly to a transitional technology - XHTML.

Even if one considers that XHTML has some significant value then one must
recall that the XHTML 1.0 Rec only defines XHTML documents which consist
solely of XHTML. The Rec recognises that mixed-namespace documents which
include elements from the XHTML namespace are also possible but fails to
define how they are to be structured or handled. 

It seems to me to be wholly inappropriate to tie XSLT 2.0 serialization
tightly to a narrow, backward-looking definition of XHTML 1.0.



You are making comments about XHTML here, not about XSLT.  
 
It would be wholly inappropriate for W3C to decide at the top level that
XHTML is needed, and for the XSL WG to ignore this decision on the grounds
that we don't personally like it.
 
Talking of XSLT being "tied tightly" to XHTML seems rather over the top;
this is about as loose a coupling between two specifications as you can get,
though as I mentioned in my other message we still need to do some work on
the conformance rules.
 
Michael Kay