Re: qflll() bug

[Karim BELABAS <Karim.Belabas@math.u-psud.fr>]

On Tue, 21 Jan 2003, Igor Schein wrote:
> ? m=[x^2-x+2,x^2-2*x-1,x^2+x+1;x^2-2*x+2,x^2+2,x^2-x-1;x^2-2*x+2,x^2+x+2,x^2-2*x-1];
> ? qflllgram(m,5)
> [[;], [3, -1, 23; 0, 1, 13; 1, 0, -1]]

This result is meaningless, m is not even symmetric.

> ? qflll(m,5)
>   ***   incorrect type in euclidean division (poldivres).

Funny, this already didn't work in release 1.39 ! Actually, I don't see how
it could possibly work over characteristic 0 domains, except in very
degenerate circonstances [ in each reduction, we replace an entry by its
primitive part, but don't update properly all data ].

Fixed by not taking the primitive part. If inefficiency of rational
polynomial arithmetic in this LLL variant becomes a concern in a genuine
application, I shall try to find a better way [ the only application I see
for this routine is over finite fields, to compute in the divisor class
group ].

    Karim.
-- 
Karim Belabas                    Tel: (+33) (0)1 69 15 57 48
Dép. de Mathématiques, Bât. 425  Fax: (+33) (0)1 69 15 60 19
Université Paris-Sud             Email: Karim.Belabas@math.u-psud.fr
F-91405 Orsay (France)           http://www.math.u-psud.fr/~belabas/
--
PARI/GP Home Page: http://www.parigp-home.de/