> BTW I think the next goal should be to reduce the cost of constructing
> a Fraction out of to plain ints by at least an order of magnitude. I
> believe this is possible.

I certainly hope so!  Here's a very simple (and simplistic) benchmark:

# start benchmark

from fractions import Fraction
from cProfile import run

def test1():
    return sum(Fraction(1, n*n-1) for n in xrange(2, 100000))

run("test1()")

#end benchmark

On my MacBook this reports a total time of 38.072 seconds, with 22.731 of 
those (i.e. around 60%) being spent in abc.__instancecheck__ and its 
callees.