[
Lists Home |
Date Index |
Thread Index
]
Your description of what you want out of the XQuery static type system again
differ from what I've heard other XML Query working group members state.
Naturally. If we all agreed on everything we wouldn't still be writing
drafts, would we?
But actually, all the arguments boil down to two desired benefits:
correctness of queries, and optimization of queries. I think most of us
agree (a) that those are desirable goals, and that (b) some degree of static
typing has the potential to deliver benefits in both areas. Any differences
are in the relative importance of those two objectives, in the detail of how
they are defined.
If you look inside the WG archives you'll find that when it comes to static
typing, I'm rather critical of our current offerings because (a) the rules
are too prescriptive to be useful for optimization (implementors can easily
do better than the current type inference rules) and (b) they don't actually
flag the most common user mistakes as errors (because in many contexts void
path expressions are quite legitimate). But that doesn't mean I don't agree
with the aims.
Michael Kay
Software AG
home: Michael.H.Kay@ntlworld.com
work: Michael.Kay@softwareag.com
<<attachment: winmail.dat>>
|
