So far as I know, solve does not check the roots it produces, but you
could easily write a program, say marino_solve,
that calls solve and then checks the roots.
One reason for solve not checking the roots is that sometimes Maxima is
not able to simplify the result of the
substitution into the original equation to zero. So the solution may be
correct but proving it
may be too hard. Roots of cubics and quartics can be very hard, for example.
RJF
Luigi Marino wrote:
> Andrew wrote:
>
> I noticed that solve does not check roots for tolerance range.
> A simple example of this is solve([1/x + 1/y =0 , x*y =0],[x,y]);
> Maxima gives the solution x=0,
> y=0. How to make Maxima check the roots? Maybe some global flags?
>
> Hi .
> A possible solution:
> declare x and y not equal zero
> and Maxima finds indeterminate system:
> x=%*r and y=-%*r
> infact (x+y)/(xy)=x+y (eliminate denominator).
> 2. Try k=x^2 and solve two equations.
>
> Luigi Marino.
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima
>