> Keep in mind that as long as the document format is "non-procedural,"
> that is, less expressive than a turing machine, you maintain the
> possiblity of translating the document to another representation
> reliably.
We're clearly dealing with different goals here. I want a document that
is able to ask you questions, and execute semi-arbitrary processes based
on your answers. If you state the goal that broadly, you really do need
this kind of expressiveness, and you definitely sacrifice
translatability, as you point out.
In fact, this gets to the nub of MY concern for a single scripting
language: truly extensible documents require compuationally powerful
languages that simply CAN NOT be translated because of the halting
problem. (It's a fact of theoretical computer science, you can look it
up.) This is precisely why we need to converge on a single safe network
extension language as soon as possible -- it is because translating
between multiple such languages is permanently infeasible. -- Nathaniel