From joeb@matematik.su.se Sun Jun 16 20:49:17 1996
Received: from pavidus.matematik.su.se (pavidus.matematik.su.se [130.237.198.6]) by insanus.matematik.su.se (8.7.5/8.6.9) with ESMTP id UAA00170; Sun, 16 Jun 1996 20:49:16 +0200 (MET DST)
From: J|rgen Backelin <joeb@matematik.su.se>
Received: (joeb@localhost) by pavidus.matematik.su.se (8.7.5/8.6.9) id UAA14833; Sun, 16 Jun 1996 20:49:14 +0200 (MET DST)
Date: Sun, 16 Jun 1996 20:49:14 +0200 (MET DST)
Message-Id: <199606161849.UAA14833@pavidus.matematik.su.se>
To: jans@matematik.su.se
Subject: bergmankoerning
Cc: jeroos@matematik.su.se, joeb@matematik.su.se
Status: RO

 Jag k|r nu }ter /usr/local/matte/bergman/Newbergman/tests/jansprep .
Jag har hittat & }tg{rdat EN lus i Hilbertseriehanteringen. Du f}r
g{rna kolla resultatens rimlighet f|r att se om du kan hitta n}gon mer.
(Utdata: ..../tests/jansout; logfil: ..../tests/janslog.)

 Processen k|r som 14801, omf|rtrevligad till 18, s} EN stor trevlig
bergman till kan Roos utan vidare k|ra. (Jag k|r ej mer {n en i taget,
och omf|rtrevligar f|rst efter en stund (om k|rningen kan antas ta
l}ng tid), s} detta st|r f|ga.)

	Mvh J|B

From jans@matematik.su.se Sun Jun 16 21:43:05 1996
Received: from localhost (jocosus.matematik.su.se [130.237.198.7]) by insanus.matematik.su.se (8.7.5/8.6.9) with ESMTP id VAA00271; Sun, 16 Jun 1996 21:43:03 +0200 (MET DST)
Message-Id: <199606161943.VAA00271@insanus.matematik.su.se>
To: joeb@matematik.su.se
Cc: jeroos@matematik.su.se, jans@matematik.su.se
Subject: Re: bergmankoerning
Reply-To: jans@matematik.su.se
In-Reply-To: Your message of "Sun, 16 Jun 1996 20:49:14 +0200 (MET DST)"
References: <199606161849.UAA14833@pavidus.matematik.su.se>
X-Mailer: Mew version 1.05 on Emacs 19.13 XEmacs Lucid
Mime-Version: 1.0
Content-Type: Text/Plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Date: Sun, 16 Jun 1996 21:43:03 +0200
From: Jan Snellman <jans@matematik.su.se>
Status: RO

Grad 10 st=E4mmer i alla fall, jfr med
~jans/problem/initial_ideals/automatic/3-3/bg5.out=



From jeroos@matematik.su.se Sun Jun 16 22:18:42 1996
Received: from pavidus.matematik.su.se (pavidus.matematik.su.se [130.237.198.6]) by insanus.matematik.su.se (8.7.5/8.6.9) with ESMTP id WAA01725 for <joeb@mailhost.matematik.su.se>; Sun, 16 Jun 1996 22:18:41 +0200 (MET DST)
From: Jan-Erik Roos <jeroos@matematik.su.se>
Received: (jeroos@localhost) by pavidus.matematik.su.se (8.7.5/8.6.9) id WAA15004 for joeb; Sun, 16 Jun 1996 22:18:40 +0200 (MET DST)
Date: Sun, 16 Jun 1996 22:18:40 +0200 (MET DST)
Message-Id: <199606162018.WAA15004@pavidus.matematik.su.se>
To: joeb@matematik.su.se
Subject: pavidus
Status: RO

Hej, Jag har tagit kaal paa min process paa pavidus eftersom
det gaar saa laangsamt nu. Om du ser till att candida
fungerar igen (bergman) saa skulle det underlaetta vaeldigt mycket
foer mig.
Haelsningar
Jan-Erik

From joeb@matematik.su.se Fri Jun 28 18:05:32 1996
Received: from homsan.matematik.su.se (homsan.matematik.su.se [130.237.198.14]) by insanus.matematik.su.se (8.7.5/8.6.9) with ESMTP id SAA14703; Fri, 28 Jun 1996 18:05:31 +0200 (MET DST)
From: J|rgen Backelin <joeb@matematik.su.se>
Received: (joeb@localhost) by homsan.matematik.su.se (8.7.5/8.6.9) id SAA06213; Fri, 28 Jun 1996 18:05:30 +0200 (MET DST)
Date: Fri, 28 Jun 1996 18:05:30 +0200 (MET DST)
Message-Id: <199606281605.SAA06213@homsan.matematik.su.se>
To: sanya@matematik.su.se, ufn@maths.lth.se
Subject: Serious bug
Status: RO

 I found out the root of the ReducePol trouble. It is due to the
integer calculation I do. In fact, (ReducePol augredand) doesn't
return the normal form of augredand but a constant times this
normal form; and since reduction of integer coefficient polynomials may
yield non-integer coefficient polynomial results, this seems to be a bit
hard to avoid. This is fairly serious.
 There are two principal ways I see to overcome this problem:

1) Introduce a new coefficient domain mode and file for calculating
  (internally, everywhere) with rational coefficients.
2) Keep track of denominators of the whole polynomials separately, in
  the anick tensor polynomial calculations.

 In both cases, there need to be new bergman routines written, and some
reprogramming of tenspol.sl to be done. Personally, I would believe 2) to
be the better method, but it would demand more revising of the file
tenspol.sl.

Irritating!
	Joergen

