On 08/02/2011 08:45 AM, Alain Ketterlin wrote:
produce integers. And it will fail with overflow for big values.
If it would make you feel better I can use decimal.
Also, perhaps I can name my function billy_fibo(n), which is defined as
billy_fibo(n) +error(n) = fibo(n), where error(n) can be made
arbitrarily small. This runs in constant time rather than linear
(memoized) or exponential (fully recursive) at the cost of a minutia of
accuracy. I find this tradeoff acceptable.
--
Bill
--
http://mail.python.org/mailman/listinfo/python-list
