> i know its late in the day re XSLT / XPATH 2.0 drafts,
> 
> but doesnt this scenario sort of 'scream out' for a boolean function
> that tests for type, both basic (node, string, number) and schema ?
> 

I would "scream" for an

   xdt:numeric

type.

Having such type have an immediate positive effect -- it will
significantly reduce the size of the F & O spec.

I realize that there might be different interpretations -- what I
propose is not a common base type (that probably is not possible to
"insert" in the type hierarchy), but just a convenient "union" of all
numeric types.

I tried to look for problems that this would create -- couldn't find any.


Cheers,
Dimitre Novatchev.