Logged In: YES 
user_id=6380

Hm. There's something to say for making 3-arg pow() only
work for ints (and longs), and then the egcd would make
sense. But that would mean removing the 3-arg pow() for
floats. Why would anyone use 3-arg pow() with floats? I
don't know, but that doesn't mean it doesn't happen. *If*
there are no community objections against making 3-arg pow()
raise a TypeError on float or complex arguments, I'm OK with
the egcd interpretation. But this is PEP material -- that's
the only way to find out. phr, would you mind writing a PEP?
It can be short and sweet.