RE: Request for XSLT2.0 / XML Schema triples

> It seems to be a kind of type-inference engine. I wonder whether
> Michael Kay is not already doing something like this (of course I
> can't know as I've never tried his SA Saxon).

Yes, Saxon does this kind of thing. But in general it only makes inferences
that are guaranteed correct, which limits the scope considerably. A
lint-style checker that is prepared to generate some false alarms can
probably spot a lot more errors. 

Actually, Saxon does also generate warnings for some things which are almost
certainly mistakes, but not illegal according to the language spec, such as

declare function f ($a as xs:int*, $b as xs:string*) as empty-sequence() {
  $a eq $b
}

This is almost certainly an error because it can only deliver a non-error
result when both $a and $b are empty sequences. 

Michael Kay
http://www.saxonica.com/