[scilab-Users] Problem in datafit

[Samuel Gougeon]

----- Message d'origine -----
De : paul.carrico at free.fr
Date : 18/03/2010 14:09:
> All,
>
> Because of number format (7 number max after the coma), the fitting procedure does not work.
>   
The number of digits displayed for numbers, and the floating point or 
scientific format
-- that can be tuned thanks to the format() function -- has no effect on 
calculation accuracy.
Calculations are always done at full accuracy. The relative accuracy for 
scilab calculations
is %eps.
For engineering format : format("e"). See help format

> a) How can I change this (in scientific format) so that the accuraty will be much better ? ... 
>
> .... this is as well a more general question since i noticed some differences between scilab & execl for example.
>
> b) Another quick question : how can I empty the memory when I load the program into scilab ?
>   
I don't clearly understand what you mean. Variables can be erased
(and then memory is disallocated) with "clear". See help clear
For graphical objects, delete() must be used.

SG