----- Sergio Rossi wrote: -----
>apparently this simple linear system cannot be
>solved with Maxima 5.9.3 under WinXP :
>
>(%i1) eq1: a0 + x1*a1 + y1*a2 +
>x1^2*a3+y1^2*a4+x1*y1*a5 = f1;
>
>(%i2) eq2: a0 + x2*a1 + y2*a2 +
>x2^2*a3+y2^2*a4+x2*y2*a5 = f2;
>
>(%i3) eq3: a0 + x3*a1 + y3*a2 +
>x3^2*a3+y3^2*a4+x3*y3*a5 = f3;
>
>(%i4) eq4: a0 + x4*a1 + y4*a2 +
>x4^2*a3+y4^2*a4+x4*y4*a5 = f4;
>
>(%i5) eq5: a0 + x5*a1 + y5*a2 +
>x5^2*a3+y5^2*a4+x5*y5*a5 = f5;
>
>(%i6) eq6: a0 + x6*a1 + y6*a2 +
>x6^2*a3+y6^2*a4+x6*y6*a5 = f6;
>
>(%i7)
>linsolve([e1,eq2,eq3,eq4,eq5,eq6],[a0,a1,a2,a3,a4,a5]);
I suspect that a solution would be such a lengthy
expression that it would be worthless. But I could be
wrong. Maybe some carbon-based computing could determine
a simple solution. Further, I suspect that the solution
would not be suitable for numerical evaluation
(subtractive cancellation.)
If ultimately you want the solution for numerical values
of x1 through y6, you'll be *much* better off re-solving the
linear system for each set of parameters. Maxima should
have no trouble solving these equations for numerical values
of x1 through y6.
Barton