Real numbers accuracy.

["Maciej Radziejewski" <maciejr@amu.edu.pl>]

Hi,

I tried the following evaluations:

1+0.E0-1
\\ 0.E-9
1+0.E0-1+1-1
\\ 0.E-19

The claimed accuracy keeps getting higher through arithmetic operations.
The effect is the same under the experimental and stable Pari.
Is it meant to work like that, or is it a bug?

If it is supposed to be like that, is there a Pari feature, or extension,
where one could get a rigorous accuracy estimate? Such a feature would be
very useful for theorem proving using analytic methods.
If I wanted to define arithmetic operations on pairs (z, r), where z is a
complex made of reals and r is the upper estimated absolute value of an
error, does it require tampering with Pari's source code?

Thanks,

Maciej.

__________________________________
Do you Yahoo!?
Yahoo! SiteBuilder - Free, easy-to-use web site design software
http://sitebuilder.yahoo.com