> > xdt:anyAtomicType is in fact an abstract type, and this 
> cast should perhaps
> > fail.
> 
> If it is abstract, why should it /perhaps/ fail? Shouldn't it 
> then always fail?
> 
> Also, where is it normatively asserted that the type is 
> abstract? I can't find 
> it here, for example: 
> http://www.w3.org/TR/xpath-datamodel/#types-predefined
> 

I've raised these points on the public comments list. (The public comments
period has closed, but we're meeting later this month and there should be a
chance to tackle issues raised up to that date.)

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