probable bug in solve



On Sun, 2006-04-02 at 15:31 +0200, laurent couraud wrote:

> I try to sole this equation but Maxima function "solve" return only
> complex roots. 
>  
> the session is :
>  
> (%i1) x^3-8*x^2-197*x-502=0;
>                        3         2
> (%o1)          x  - 8 x  - 197 x - 502 = 0
>  
> (%i2) solve(%,x);
>  
> (%o2)  /* Maxima give me 3 roots. each one contaning the imaginary
> number %i */
You must simplify the result of solve, before you draw any conclusions:

(%i1) float(rectform(solve(x^3-8*x^2-197*x-502=0)));
(%o1) [x = - 3.082930217227974, x = - 8.370397997665673,
       x = 2.220446049250313E-16 %i + 19.45332821489365]

> but if I do :
>  
> (%i3) allroots(%o1);
>  
> (%o3)       [x = - 3.0829302172279736, x = - 8.3703979976656715, x =
> 19.453328214893645]

The two results agree, within the numerical accuracy.

Jaime