1. Maxima 'solve' OK but very poor accuracy
2. Maxima 'Solve' not implemented Smath 5346

Though I can't compare with Mathcad Prime essentially Maxima,
I doubt Mathcad would not spit better accuracy than Smath/Maxima.
As it looks, 'solve' is implemented with a too low TOL [Tolerance].

Jean

Just a side note (unfortunately, unhelpful).
You keep using word "Maxima" in many places where it apparently doesn't apply (another example is here: http://en.smath.info/forum/yaf_postst7424_Maxima-X-Y-plot.aspx).
It looks like you mean that standard "solve" (all small letters) is a Maxima function. No, it is not, it is SMath own in-built solve().
In another post, you write:

Wrote

What I understand from Smath Studio: it is a stand alone big tool, a math tool.
2MB user interface working from associated plugins [total size 5346 = 235 MB].

Looks like you suppose that SMath itself is just a "shell" (interface) to its plugins. Well, that's not strictly true, these 2MB include its own load of math

Edit: maybe you'll find nonlinear solvers plugin useful?
It gives correct answers, though by carefully selecting ranges; you may directly control output precision.

Thanks Mike for Brent, I know about that stuff [Ref: Mathcad 11]
A much better candidate is the universal "Polyroots".
My remark cautionned Smath users about inaccuracy of "solve".


Brent("1:function", "2:condition", "3:condition", "4:condition")
- Brent's root-finding method of function "1:function", giving a
couple of delimiters "2:condition" and "3:condition";
calculation have at least "4:condition" function precision.

Jean




Integrate Quadratures Legendre Polyroots.sm (23 KiB) downloaded 48 time(s).