Karim Belabas on Fri, 02 Oct 2020 17:14:44 +0200


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: quadunit again


* Michael Hortmann [2020-10-02 16:43]:
> I should have looked at Henri Cohen's book on Computational Algebraic
> Number Theory first:
> The algorithm in "Chapter 5 Algorithms for Quadratic Fields, §5.7
> Computation of the Fundamental Unit" probably answers my question.
> 
> Still it might be interesting to know a priori which algorithms are used in
> the Pari functions. But of course that could vary from version to version
> ...
> So how would I have to proceed to gain that knowledge?

Writing to this list a good first step. :-) 

If a suitable published reference is available, then I agree it should be
listed in the documentation for the function; or at least a general
indication of the method when it's sufficiently well known (continued
fraction expansion for instance). It would be nice to add a complexity
statement whenever one is readily available as well. For quadunit, it should
go into src/functions/number_theoretical/quadunit, and I just did it starting
in 'master' from your and Bill's digging out a suitable reference :-).

I also suggested the (asymptotically far superior) alternative bnfinit +
bnfunits.

Given the current state of affairs (maybe 2% of all GP functions refer
to a publication and at most 10% have a complexity estimate or an indication
of the method), it is rather painful to do this systematically, though; I'm
open to all documentation patches that go part of the way !

Cheers,

    K.B.
--
Karim Belabas, IMB (UMR 5251)  Tel: (+33) (0)5 40 00 26 17
Universite de Bordeaux         Fax: (+33) (0)5 40 00 21 23
351, cours de la Liberation    http://www.math.u-bordeaux.fr/~kbelabas/
F-33405 Talence (France)       http://pari.math.u-bordeaux.fr/  [PARI/GP]
`