On Wednesday 31 December 2008 18:23:54 ?iga Lenar?i? wrote:
> Maxima is actually lacking a proper fitting facilities (like many
> other things). lsquares is a poor excuse for a fitting package in
> many aspects:
> -slow
> -complicated to use
> -unintuitive
> -cryptic name (should be LeastSquares/least_squares, not
> lsquares_mse and what not)
> I think the latter three (if not all four) aspects of lsquares
> package applies well to a lot of maxima functions, if they even
> exist (still no ODE integrator or fast numerical linear algebra
> functions)... Especialy cryptic names are the plague of maxima.
> What can a user dechipher from 'create_list' and 'makelist'. It
> makes no sense..
> ...
Thank you for your comments and suggestions.
Some of my personal notes:
1. Predominantly, Maxima is a system for the manipulation of symbolic
expressions. Of course, any suggestions and contributions are welcome,
but you should keep in mind that there are some special software for
numerical computations, like GNU Octave.
2. "Unintuitive" or "cryptic" is a very personal opinion. For example,
I'd find "Fit" quite confusing and not informative (yes, much less
informative than "lsquares_estimates").
3. Maxima naming conventions (foo_bar_baz) are differ from the
Mathematica naming conventions (FooBarBaz). Maybe you don't like them
just because you're more experienced in Mathematica.
4. I don't think we need to re-implement the numerical linear algebra
stuff which is already available from LAPACK. It is always better to
work on improvements of Maxima-LAPACK interoperability.
5. Solving linear systems of equations numerically as A^^-1 . b
is a bad idea. Why not to use something like linsolve_by_lu on
floatfield?
--
Setting Orange, Aftermath 73 YOLD 3174
Alexey Beshenov http://beshenov.ru/