On Mon, 2003-12-22 at 20:27, Tupshin Harper wrote:
Certainly they are susceptible to criticism, and not making a turing
complete validation mechanism is one of them. But at least recognize
that (for better and for worse) they tried to take a "middle ground"
approach. XSD got a huge amount of criticism for being too complex
(compared to DTDs), and adoption of it has been slow, largely for that
reason. They were justly leery of adding even more complexity.
This is drifting quite quickly off-topic, but I'm not sure that making a
data interchange format with a turing-complete validation language
necessarily involves so much complexity -- think LISP, which is
potentially quite handy for both halves of the problem.