Thanks Davide,
But I wrote the f_poly(v) few years ago and there were no troubles till 6965b.
It's strange that even minor update requires to review very old (considered reliable) worksheets.
f_poly() is pure numerical function and contains tons of evals .
How is it possible to corrupt posterior symbolic calculations?

Edited by user 06 February 2019 19:15:58(UTC)  | Reason: Not specified

Hi. Just a very small issue with subscripts:



Best regards.
Alvaro.

Hello Uni,
Recently I found another problem in symbolic calculations that probably related to maple plugin.
Let see the example in attachment.
This is simple example of constructing of system of equations for circuit of 3-rd order.
I need 2*3+2 equations some of them define frequency response, some define of the position of minimums/maximums or the transfer function.
The problem happens with taking derivatives. Smath can't calculate them without involving maple plugin. Try disabled equation for H1(n,k,b,w).
But using maple(simplify()) it does it right. After this step H1(n,k,b,w) became useless. Any symbolic expression that uses it returns "empty".
But if I use copy-pasted output of the H1(n,k,b,w), everything works fine. It looks like I need to force the symbolic result to be evaluated at some point,
similar to eval() for numeric calculations, but I don't understand how to do that.

equ3.sm (34kb) downloaded 37 time(s).

Alvaro,
Your version works. Do you have an idea how to generate sub-indexes from Range ? I need to generalize it for different n.
Best regards,
Igor

Edited by user 10 April 2019 18:18:05(UTC)  | Reason: Not specified

Hi Igor. Use this



equ3.sm (33kb) downloaded 36 time(s).

Best regards.
Alvaro

Hi. This is a workaround: use maple's 'value' function, which convert inert versions of some procedures to the actual procedures for evaluation. Usually, inert version have the same name, but with the first letter in uppercase.

Notice that for Diff you must to disable Maxima plugin.



Best regards.
Alvaro.

I'd like to clarify the problem. There were two issues in my worksheet.

1. Symbolic differentiation works strange: if you evaluate dA/dx symbolically and A is not a function of x it returns 0. That is correct.
But if A is an indexed variable it returns dA[k]/dX = dA[k]/dx, that's wrong, A[k] is not a function of x. That was the reason to change indexes to subscripts.

2. I found workaround using maple(simplify(dA[k]/dx))=0. I can get right result in this case, but I can't use it in futher calculations. This is the main issue.
If I'd be able to use the symbolic result as it was generated by maple plugin everything would be fine. Instead, Smath postpones symbolic evaluation and eventually
reports "empty" or something like that when the expression gets too complicated.

Edited by user 11 April 2019 16:33:01(UTC)  | Reason: Not specified

Jean,
Regarding issue #1 I attached sym_diff_issue.sm (4kb) downloaded 22 time(s)..
Where do you see "undefined upper limit" ?
For issue #2 I attached equ3+remarks.sm (34kb) downloaded 24 time(s). with remarks.
Regards,
Igor

For the other issues equ3_remarks, your programs are incorrect.
Set n=3 to see something

equ3_remarks.sm (67kb) downloaded 19 time(s).

Thanks Jean,
Don't waste your time, I already have solved the problem.
The n supposed to be odd number in the range 3,5,7,9,11. The practical case was 9.
The goal was to derive system of 22 equations. They looked too complicated for doing it manually.
I posted simplest case n=3 just to illustrate the issues I've encountered.
Sorry, I can't go into details, that is a kinda commercial project .

Edited by user 12 April 2019 15:23:03(UTC)  | Reason: Not specified