Boost logo

Boost :

From: Fernando Cacciola (fcacciola_at_[hidden])
Date: 2001-06-21 17:06:14

----- Original Message -----
From: <helmut.zeisel_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Thursday, June 21, 2001 11:44 AM
Subject: [boost] Re: Unlimited integer

> --- In boost_at_y..., "Fernando Cacciola" <fcacciola_at_g...> wrote:
> >
> > ----- Original Message -----
> > From: <helmut.zeisel_at_a...>
> > To: <boost_at_y...>
> > Sent: Wednesday, June 20, 2001 2:28 PM
> > Subject: [boost] Unlimited integer
> >
> >
> > > Is there any interest in an unlimited integer class?
> > >
> > I might be relevant to look at the following widely used number
> packages:
> >
> >
> >
> >
> >
> >
> >
> Fine, since many people already invented wheels,
> this means that there must be some interest for wheels generally
> and for a standard wheel particulary :-)
> Which of these libraries could be
> a starting point for inclusion into a future C++ standard,
> by offering a complete C++ interface
> (compatible e.g. to boost::rational and boost::operators)
> and providing a boost compatible license?
There are at least two more C Libraries for numerical computations along the
lines of GMP:

The LiDIA library: (which I
forgot when I made the first list)
The PARI project:

The three of them, GMP, LiDIA and PARI, are {currently} Unix-only, so I
don't think they can be standardized (or boostified).
(they can be installed on a unix-on-windows host, such as mingw/cygwin; but
not on win32)

The CORE library ( is a C++ wrapper for
LiDIA, and is my personal choice (due to its interface design).

In my opinion we have two alternatives, either write something from scratch
which is standard (portable) and boostified, but following closely the above
libraries; or we find a way to get some people porting some of these
packages so they can be included in boost; but can't tell which is harder.

Fernando Cacciola
Sierra s.r.l.

Boost list run by bdawes at, gregod at, cpdaniel at, john at