Re: Big integer library

From: Richard Peters (r.a.peters_at_[hidden])
Date: 2004-04-29 10:05:05

"Thorsten Ottosen" <nesotto_at_[hidden]> wrote in message
news:c6quof$dtl$1_at_sea.gmane.org...
> I have some questions. Will it be possible to say
>
> 1. typedef boost::rational< big_integer > big_rational.

Yes. In fact, the big_integer_example contains an example where
boost::rational<big_integer> is used.

> 2. Why is faculty() a part of the interface. I have a generic version of
> faculty in some prob&stat code I'm doing that I hope would work
> with big_integer.

Your generic version will most probably work. There is one difference, my
version returns an expression instead of a value, which saves the creation
of one temporary value. In this case, the gained performance is probably
minimal. I'm happy to refactor this function out of the library, throw it
away, or come up with a hybrid solution where the expression-returning
version is included when both big_integer and your library is included.

> 3. Are you planning big_float? I would be happy with focus on interface
> specs and then better implementation later.

Unfortunately not. I can see that big_float would have its uses, but I
rarely even use float. My interests ly in algebraic computations such as
cryptology.

>     The C++ standard could really use a "big" library and it's always nice
> with a version in boost first.
>
> br
>
> Thorsten

Thanks for the response,

Richard Peters