> you didn't highlight it but -1 is a double so math:pow($x,-1) is quite a tricky function, math:pow($x(xs:int(-1)) with an integer power is much simpler to define.

In case the IEEE 754-2008 Standard for Floating-Point Arithmetic implements pown() with division when the exponent is a negative integer, then they actually transform expression (2) to expression (1), therefore it seems meaningful just to use expression (1).

Dimitre