> I would prefer the whole spec to be wrapped in "as if",
> rather than introducing it for specific points. The ISO C
> standard works that way.
Obviously a spec defines the outcome and not the implementation, so
anything that's specified in terms of a processing model has an implicit
"as if". The problem arises when "as if" is used in the absence of an
explicit processing model - it tends to introduce ambiguities because it
very often leaves it unclear what order of operations is being mandated,
for example, whether a particular rule referring to the treatment of
whitespace is to be applied before or after CDATA sections are expanded.
Michael Kay
