Re: is XSLT 2.0 implementable? (was: N : M transforma

  One suggestion for conformance testing (not as silly as it seems) is
  that you should construct the source data model using XSLT. That way,
  it's well defined what data model you get from a given input.

I can see that that is a reasonable suggestion for nailing things down
for conformance reasons, although the input-generating stylesheet will
have to be careful not to assume anything about _its_ input, even
just a single top level template match="/" can't be relied on as the
effective PSVI obtained from (any) input might be a (possibly empty,
possibly of length > 1) sequence of nodes, more than one of which
matches /) I think, I must check the latest definition of / and initial
input...


If such precautions were just to avoid theoretical edge cases to tie
down a test suite for predicatble results then that would be reasonable,
but my very real fear is that these problems are not going to be rare
edge cases, they are going to be every day incompatibilities.
It's not going to be possible to answer any question on xsl-list without
first ascertaing what XSLT system and (especially) what parser is being
used. This will be bad, it may not be as bad as I fear (but I fear it
will be).

David


________________________________________________________________________
This e-mail has been scanned for all viruses by Star Internet. The
service is powered by MessageLabs. For more information on a proactive
anti-virus service working around the clock, around the globe, visit:
http://www.star.net.uk
________________________________________________________________________

 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list