John Cremona on Sat, 02 Feb 2008 10:49:37 +0100


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

Re: idealappr()


Dear Karim,

Thanks for the quick reply.  [3 days for me to remember that Dan
Bernstein's qsecretary response gets put into spam by gmail an find it
and reply to it;  minutes for your response!]

As you know, the number field functionality in Sage is based on the
pari library.  In Sage, the function uniformizer() uses isappr but was
not documented, so I was adding documentation and this led me to the
question.  I'm not the one who designed the number fields in Sage, so
I am not sure when it uses nfinit() and when bnfinit().  I was only in
that code because I found a different bug and had to fix it.  I think
it only uses bnfinit() when a function is called which needs it;  but
it would be reasonable for uniformizer() to be such a function.

Thanks again,

John

On 01/02/2008, Karim Belabas <Karim.Belabas@math.u-bordeaux1.fr> wrote:
> * John Cremona [2008-02-01 22:15]:
> > The documentation for idealappr(K,P) implies that in the case that P a
> > prime ideal in the number  field K, the element returned will be a
> > uniformizer for P and integral.
>
> True.
>
> [ P.gen[2] is another "simpler" way to get a uniformizer ]
>
> > Will it always be a generator for P when P is principal (as it might
> > then be)
>
> No.
>
> > If the latter, would it be relatively expensive to call
> > bnfisprincipal() to ensure that we have a generator when it exists?
>
> It's actually impossible: the argument to idealappr is an nf and
> not a bnf, which would be required for bnfisprincipal (idealappr is a
> rather trivial function, especially if the input is a prime ideal; it is
> *much* simpler than bnfinit + bnfisprincipal !).
>
> One could check whether the input is in fact a bnf, then test whether P
> is principal, and if so compute a generator; but it would complicate a
> simple function for a rather limited use.
>
> What's wrong with calling directly bnfisprincipal ? [ it will actually
> produce a "uniformizer" for any ideal, not necessarily a prime; and
> bypass a useless idealappr call ]
>
> What is your specific application ?
>
> Cheers,
>
>     K.B.
> --
> Karim Belabas                  Tel: (+33) (0)5 40 00 26 17
> Universite Bordeaux 1          Fax: (+33) (0)5 40 00 69 50
> 351, cours de la Liberation    http://www.math.u-bordeaux.fr/~belabas/
> F-33405 Talence (France)       http://pari.math.u-bordeaux.fr/  [PARI/GP]
> `
>


-- 
John Cremona