Karim BELABAS on Mon, 12 May 2003 15:08:51 +0200 (MEST)


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

Re: CVS: seriesprecision too large


On Sun, 11 May 2003, Ilya Zakharevich wrote:

> On Sun, May 11, 2003 at 07:24:42PM +0200, Bill Allombert wrote:
> > I think it is 2.2.5 F22:
> >    22- for transcendental function f and polynomial p, f(p) only gave
> >        seriesprecision significant terms when val(p) = 0
>
> It is not clear why good behaviour w.r.t. multiplication is prefered
> to good behaviour w.r.t. addition.  Of course, one needs to make a
> choice, but *why* this one?

I tried to make it consistent with the handling of t_REALs:
'[series | real]precision' is the number of significant [terms | digits] one
computes when converting an exact quantity [polynomial | rational number]
before an inexact computation.

> Did not you notice
>
> /     1 2   1  4    1  6     1   8      1    10     11  \
> | 1 - -y  + --y  - ---y  + -----y  - -------y  + O(y  ) | x +
> \     2     24     720     40320     3628800            /
>
> with seriesprecision==9?

After the above conversions, computations with inexact terms yield exactly as
many terms as can be obtained from these initial (converted) objects. We miss
something analogous to 'format' to truncate series to a given number of
significant terms on output. It's trivial to implement [ 'seriesformat' ?
setting it by default to 'seriesprecision' terms ? ]. I'm not sure how useful
it would be.

    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              http://www.math.u-psud.fr/~belabas/
F-91405 Orsay (France)            http://www.parigp-home.de/  [PARI/GP]