Subject: Re: replacement for bc(1), dc(1), diff(1), and diff(3)
To: Jason Thorpe <thorpej@shagadelic.org>
From: Phil Nelson <phil@cs.wwu.edu>
List: tech-userlevel
Date: 03/25/2005 13:37:26
=2D----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Friday 25 March 2005 12:05, Jason Thorpe wrote:
> There has already been some private discussion about replacing bc/
> dc. =A0I do not know the heritage of the OpenBSD versions, but Anders =A0
> Magnussen was looking at the now-freed-up old AT&T versions. =A0I'll =A0
> let him elaborate further.

The OpenBSD versions started with the old AT&T versions.  The math library =
is=20
a slightly modified version of the AT&T version.    The AT&T version that w=
as=20
recently made free still contains a bug that was documented in the paper:
Ingo Dittmer, ACM Signum, April 1993, page 8-11.   =20

The AT&T dc was modified by the OpenBSD people to now use the math routines=
=20
found in the openssl library (crypto/bn).   This fixed the bug listed above=
,=20
but the performance of OpenBSD bc/dc is still not as good as GNU bc for mos=
t=20
cases as shown in a different post on this thread.  In fact, there still ar=
e=20
places where GNU bc can do computations that OpenBSD dc can't.

I still believe that NetBSD shouldn't replace a GNU tool with an inferior t=
ool=20
just to get rid of the GPL. =20

=2D --Phil

=2D --=20
Phil Nelson                       NetBSD: http://www.netbsd.org
e-mail: phil@cs.wwu.edu           Coda: http://www.coda.cs.cmu.edu
http://www.cs.wwu.edu/nelson=20
=2D----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (NetBSD)

iD8DBQFCRISazbodwsP3RI0RAhuAAKCcNCXFJN1OZW202+UzBQEzjP6GrwCgx/52
heaiDeE6/MUurvLt+YiomyA=3D
=3Dojtw
=2D----END PGP SIGNATURE-----