for example:

def calc(params):
    """ i am factoring numbers. """
    # an expensive CPU cost function but 
    # passin params and return result are both lightweight

cachedcalc = collections.defaultdict(calc)
result = cachedcalc[0xFFFFFFFFFFFFFFFFFFF0AC0FFF1]

-1. Besides compatibility issues, defaultdict is a dict: it contains data, and is not meant to consume CPU when accessing items. Its "default" function should return initial values, like 0 or an empty list.

I think what you want is "memoization"; a memoized function still looks like a function! there are many implementations for python, one of these is here: http://wiki.python.org/moin/PythonDecoratorLibrary#Memoize