>Well... the standard does in fact state the following signatures for pow (26.5.6):
Right Mario, that was my point to Citizen. But I still don't think he's figured it out yet.

I do agree that if the exponent is declared as a double, then use that signature.