Lists Home |
Date Index |
> Getting XPath 1.0 expressions right is hard enough.. Personally, I
> regard XPath expressions as atomic: they get me something out of a
> document. There is no control flow in them, simple predicate tests
> aside. With XPath 2.0, control flow will be scattered between the
> languages built around XPath and the expressions themselves.
Yes, that's a good way of making the distinction. Of course it depends
what you mean by "atomic". In XPath 1.0, you might consider node sets
as atomic, so similarly you should regard sequences in XPath 2.0 as
atomic. And you could view a for expression as an expression that
returns a sequence.
The line is definitely blurred which is why, I think, different
people's intuitions about where it should be drawn are different, and
hard to articulate. Life would be simpler if we were in a position
where we didn't have to draw the line at all, and we were using XQuery
or Lisp. But we're not.