rational numbers

[=?iso-8859-1?Q?Jos=E9_Romildo_Malaquias?=]

On Mon, Dec 06, 1999 at 12:28:26PM +0100, Martin Wierich wrote:
> Hi Clifford,
> 
> We have a prototype of a module for exact rationals lying around here in
> Nijmegen. Such a module should be
> based on arbitrary length integers ("BigInts") which I am currently
> implementing. When this will be
> finished we will release these modules. For those who want to know: the
> BigInt implementation will be
> based on the GNU GMP library. To speed up calculation with small numbers
> we add new instrucions to the ABC
> machine, e.g. an instruction that returns both the result of adding two
> 32 bit integers and a Boolean that
> says whether an overflow occured.

Hello Martin,

I am working on project for doing Computer Algebra in Clean. The library
I am developing is to manipulate symbolic expressions literaly and
exactly. It does not use any floating aproximations for real numbers.
Its numeric capabilities depends on a good library for arbitrary precision
integers. Unfortunatly the current Clean implementation lacks such
a library. So we are going to develop our own integer library. It should
be efficient and well integrated into the language.

As you are working on an implementation for big integers, maybe my system
would rely on them. Do you have any idea when I could get into your
library? Which Clean version does it support? I hope it is/will be usable with
the current Clean implementation. Will it make into the language or
will be released as a separate library? I think that the Clean language 
deserves a standard type for arbitrary precision integers.

Romildo
-- 
Prof. José Romildo Malaquias <romildo@iceb.ufop.br>
Departamento de Computação
Universidade Federal de Ouro Preto
Brasil