> A pragmatic answer "why": it uses the data model implied by
> SAX, which treats characters "quoted" by "<![CDATA[" ... "]]>"
> like any other characters (but without using '&' and '<' as
> markup delimiters).
Aha! I think this is the big difference in approach. The David's are saying
that CDATAsects are tags which switch in and out an effect, while I am
saying that the CDATAsect is markup which delimits a range and labels it.
Personally, I hope CDATAsects are removed from the mooted XML profile (does
it have a code name? EZX?), I have never thought they were a particularly
good idea. But I guess they wont be, in that it would make it possible to
generate EZX documents which were not WF XML documents. [I'd guess EZX would
remove DTDs (no entities!), make UTF-8 the only charset, allow but deprecate
PIs (in particular before and after fragments or root-elements), allow but
deprecate CDATAsects, build-in the ISO public entity sets with HTMLsymbols,
and build-in namespaces. I suppose this would then become the syntax used
for HTML 5, or whatever HTML+XML is called. That would be a nice little
language.]
Rick Jelliffe