Mathcad Toolbox - Contains analogs of functions from Mathcad - Сообщения
#181 Опубликовано: 26.10.2024 20:25:46
Coming soon...
The ability to explicitly specify the placement of block elements will be added.
It looks strange, maybe I'll add some numbering: #1, #2, #3, ...
The ability to explicitly specify the placement of block elements will be added.
It looks strange, maybe I'll add some numbering: #1, #2, #3, ...
Russia ☭ forever, Viacheslav N. Mezentsev
#182 Опубликовано: 27.10.2024 09:34:21
... It looks strange, maybe I'll add some numbering: #1, #2, #3, ...
Hi. That's the Mathematica default notation for pure function's arguments: https://reference.wolfram.com/language/howto/WorkWithPureFunctions.html
Best regards.
Alvaro.
1 пользователям понравился этот пост
Вячеслав Мезенцев 27.10.2024 13:26:00
#183 Опубликовано: 13.11.2024 03:47:53
Plugin updated.
Changes:
- fixed image handling functions (READRGB(),...);
- the appearance of brackets has been changed;
- added the ability to explicitly specify an insert for MathcadBlock (##, #1, #2, ...).
mathcadblock-explicit-paste.sm
Russia ☭ forever, Viacheslav N. Mezentsev
1 пользователям понравился этот пост
sergio 13.11.2024 16:08:00
#184 Опубликовано: 13.11.2024 16:42:40
Expanding the idea of numbering, we can add syntax:
## - take all as sys|mat|line
#i#j - take from i to j
##j - take from 1 to j
#i# - take from i to end
and
#i#js - sys(#i,..,#j)
#i#jm - mat(#i,..,#j)
#i#jl - line(#i,..,#j)
#i#jfun - fun(#i,..,#j)
or
### - insert as parameters: foo(x,y,###) -> foo(x,y,#1,#2,...)
##3# - insert as parameters: foo(x,y,##3#) -> foo(x,y,#1,#2,#3)
## - take all as sys|mat|line
#i#j - take from i to j
##j - take from 1 to j
#i# - take from i to end
and
#i#js - sys(#i,..,#j)
#i#jm - mat(#i,..,#j)
#i#jl - line(#i,..,#j)
#i#jfun - fun(#i,..,#j)
or
### - insert as parameters: foo(x,y,###) -> foo(x,y,#1,#2,...)
##3# - insert as parameters: foo(x,y,##3#) -> foo(x,y,#1,#2,#3)
Russia ☭ forever, Viacheslav N. Mezentsev
#185 Опубликовано: 13.11.2024 23:20:14
I encountered a plot problem with the list version of rkfixed(). Probably this also applies to other ODE-integrators with embedded function assignment.
In the native 2D Plot:
- When plotting a solution function along with a matrix of marker points, then the points won't display
- Yet, when using lspline() directly, both objects plot without problems.
In X-Y Plot:
- When the independent variable has the same name as the solution function (e.g. x(x), then x() is not treated as a function but as the expression x (similar with v(v) or any other name).
- A workaround is to use a variable name different from the function name. Yet this is strange, because in SMath x() and x usually are independent objects.
plotproblem.sm
![2024-11-13 17_08_58-SMath Solver - [plotproblem.sm_].png](/ru-RU/files/Download/da36GC/2024-11-13-17_08_58-SMath-Solver---[plotproblem.sm_].png)
In the native 2D Plot:
- When plotting a solution function along with a matrix of marker points, then the points won't display
- Yet, when using lspline() directly, both objects plot without problems.
In X-Y Plot:
- When the independent variable has the same name as the solution function (e.g. x(x), then x() is not treated as a function but as the expression x (similar with v(v) or any other name).
- A workaround is to use a variable name different from the function name. Yet this is strange, because in SMath x() and x usually are independent objects.
plotproblem.sm
Martin KraskaPre-configured portable distribution of SMath Studio: https://en.smath.info/wiki/SMath%20with%20Plugins.ashx
1 пользователям понравился этот пост
Вячеслав Мезенцев 13.11.2024 23:47:00
#186 Опубликовано: 14.11.2024 01:45:31
A little step further? (And I say little because I'm not the one who has to improve the code) Take #arg as function parameter?
Best regards.
Alvaro.
Best regards.
Alvaro.
1 пользователям понравился этот пост
Вячеслав Мезенцев 14.11.2024 01:54:00
#187 Опубликовано: 15.11.2024 03:34:05
Plugin updated.
Changes:
- MathcadBlock: the syntax for explicit insertion has been extended.
Russia ☭ forever, Viacheslav N. Mezentsev
1 пользователям понравился этот пост
Alvaro Diaz Falconi 16.11.2024 08:51:00
#188 Опубликовано: 15.11.2024 05:14:39
Russia ☭ forever, Viacheslav N. Mezentsev
1 пользователям понравился этот пост
Alvaro Diaz Falconi 16.11.2024 08:51:00
#189 Опубликовано: 15.11.2024 16:11:09
Visualization of transformations and matrix norms using implicit plots and automatic scaling of the X-Y plots.
Matrixnormen 2D.sm
Matrixnormen 2D.sm
Martin KraskaPre-configured portable distribution of SMath Studio: https://en.smath.info/wiki/SMath%20with%20Plugins.ashx
3 пользователям понравился этот пост
#190 Опубликовано: 12.12.2024 04:09:39
I want to add a set of simple examples to the Mathcad block context menu that demonstrate how it works. Can you suggest such examples?
Similar to the X-Y Plot component. In this case, you will be able to fill in not only the input values, but also the description. You can insert arbitrary expressions inside the component.
Imagine that you can formalize the solution to a simple problem. What would you suggest as educational examples?
Similar to the X-Y Plot component. In this case, you will be able to fill in not only the input values, but also the description. You can insert arbitrary expressions inside the component.
Imagine that you can formalize the solution to a simple problem. What would you suggest as educational examples?
Russia ☭ forever, Viacheslav N. Mezentsev
1 пользователям понравился этот пост
Martin Kraska 13.12.2024 20:06:00
#191 Опубликовано: 13.12.2024 17:18:46
Imagine that you can formalize the solution to a simple problem. What would you suggest as educational examples?
I suggest a handbook like Martin's one, with examples where the user can reuse the code (Code reusability is the technical expression for when you don't want to say that you are going to copy and paste). Or like Mathcad quicksheets. This is my idea of a clean use of the block for solving high school problems.
BasicHandbook-nsol.pdf
BasicHandbook-nsol.sm
Best regards.
Alvaro.
5 пользователям понравился этот пост
Вячеслав Мезенцев 13.12.2024 23:41:00, francesco rapuano 13.12.2024 18:43:00, NDTM Amarasekera 13.12.2024 19:02:00, Martin Kraska 13.12.2024 20:06:00, sergio 13.12.2024 21:15:00
#192 Опубликовано: 13.12.2024 20:06:09
I'd like to have such things like nsol in SMath core.
How about embedded assignment, similar to FindRoot? I don't like poking the values out of S. I see that this might reduce flexibility, e.g. in the case of parametrized initial guess.
I tried to demonstrate a generic approach, i.e. you get the variables assigned without having to issue any specific command except for the generic content of the placeholder. It is just a try and fails with single variables but you get the idea.
Of course, one would like to have a solve block without the cryptic generic expression in the placeholder.
Perhaps, such a block could be created from the insert menu with pre-allocated content of the placeholder and with setting it to hidden.
![2024-12-13 13_47_32-SMath Solver - [BasicHandbook-nsol.sm_].png](/ru-RU/files/Download/Rkunpm/2024-12-13-13_47_32-SMath-Solver---[BasicHandbook-nsol.sm_].png)
BasicHandbook-nsol_Kr.sm
How about embedded assignment, similar to FindRoot? I don't like poking the values out of S. I see that this might reduce flexibility, e.g. in the case of parametrized initial guess.
I tried to demonstrate a generic approach, i.e. you get the variables assigned without having to issue any specific command except for the generic content of the placeholder. It is just a try and fails with single variables but you get the idea.
Of course, one would like to have a solve block without the cryptic generic expression in the placeholder.
Perhaps, such a block could be created from the insert menu with pre-allocated content of the placeholder and with setting it to hidden.
BasicHandbook-nsol_Kr.sm
Martin KraskaPre-configured portable distribution of SMath Studio: https://en.smath.info/wiki/SMath%20with%20Plugins.ashx
1 пользователям понравился этот пост
Alvaro Diaz Falconi 13.12.2024 22:00:00
#193 Опубликовано: 13.12.2024 22:21:36
Hi Martin. I put "0" as argument because there are no more args. The second could be a vector of options, like this:
or the solved variables. Also, I use "S" because this error, which I imagine that could be corrected, but don't know how hard it could be:
Best regards.
Alvaro.
PD: Notice that x# : transpose(str2num(x#))*diag(U#) or x# : diag(U#)*str2num(x#) if you want to return the answer as column vector, like normal people, returns x# as scalar if length(x#) = 1.
or the solved variables. Also, I use "S" because this error, which I imagine that could be corrected, but don't know how hard it could be:
Best regards.
Alvaro.
PD: Notice that x# : transpose(str2num(x#))*diag(U#) or x# : diag(U#)*str2num(x#) if you want to return the answer as column vector, like normal people, returns x# as scalar if length(x#) = 1.
#194 Опубликовано: 14.12.2024 07:36:10
... How about embedded assignment, similar to FindRoot? I don't like poking the values out of S. I see that this might reduce flexibility, e.g. in the case of parametrized initial guess
Hi Martin. I think the opposite: the implicit assignment of the names of the variables sought to a solution makes its call very rigid, since many times there is no direct relationship between the variables of the equation and the names of the variables that are intended to be assigned to the solution, as is precisely demonstrated in this case
Another variant among many would be to find a limiting speed and call the result v.lim, even though the variable of the equation is "v".
Best regards.
Alvaro.
#195 Опубликовано: 31.12.2024 00:07:50
Plugin updated
Changes:
- Mathcad Block: restored drawing of curly bracket (SS >= 1.3.*).
Russia ☭ forever, Viacheslav N. Mezentsev
4 пользователям понравился этот пост
Oscar Campo 01.01.2025 06:23:00, sergio 31.12.2024 01:47:00, churichuro 31.12.2024 02:12:00, Alvaro Diaz Falconi 01.01.2025 13:20:00
#196 Опубликовано: 20.08.2025 23:31:03
Plugin updated.
Changes:
- CreateMesh() and CreateSpace() added;
- cspline() added;
- lspline() and pspline() fixed.
Russia ☭ forever, Viacheslav N. Mezentsev
#197 Опубликовано: 21.08.2025 13:22:30
Plugin updated.
Changes:
- find() function added.
MathcadToolbox-find-handbook.sm
MathcadToolbox-find-handbook.pdf
There are some problems with obtaining the values themselves. Their nature is still unclear. This is the first trial implementation.
Отредактировано 21.08.2025 13:23:20
Russia ☭ forever, Viacheslav N. Mezentsev
#198 Опубликовано: 21.08.2025 14:13:10
This effect is called a "classic case of circular dependency"
Never use the same variables for initial approximations and to store results - AI advised me.
Never use the same variables for initial approximations and to store results - AI advised me.
Russia ☭ forever, Viacheslav N. Mezentsev
1 пользователям понравился этот пост
Martin Kraska 21.08.2025 20:59:57
#199 Опубликовано: 21.08.2025 17:11:48
The problem is in the delayed evaluation. It is nice if you want to evaluate the solution for multiple configurations with S just storing the task instead of the solution.
It may well be fine to have this behaviour by design, but then we should not get an error message (undefined Ix).
So my proposal is
- if the solution can be found numerically with the data given at processing of the block, then store the data and not the task description.
- if the task has undefined variables other than the solution variables, then just store the task description (as it is done now but don't complain.
If you enforce intermediate evaluation, then there are no restrictions on the variable names any more. The recursivity is broken.
BTW, I'd propose to allow for scalar arguments for the single equation/single variable case.
In case you provide solvers with range specification, my ideas would be to use Davides ternary operators a
If no initial guesses or ranges are given, try to apply a linear solver.
MathcadToolbox-find-handbook_Kr.sm
1-2-Maßeinheiten.sm
It may well be fine to have this behaviour by design, but then we should not get an error message (undefined Ix).
So my proposal is
- if the solution can be found numerically with the data given at processing of the block, then store the data and not the task description.
- if the task has undefined variables other than the solution variables, then just store the task description (as it is done now but don't complain.
If you enforce intermediate evaluation, then there are no restrictions on the variable names any more. The recursivity is broken.
BTW, I'd propose to allow for scalar arguments for the single equation/single variable case.
In case you provide solvers with range specification, my ideas would be to use Davides ternary operators a
If no initial guesses or ranges are given, try to apply a linear solver.
MathcadToolbox-find-handbook_Kr.sm
1-2-Maßeinheiten.sm
Отредактировано 21.08.2025 17:12:37
Martin KraskaPre-configured portable distribution of SMath Studio: https://en.smath.info/wiki/SMath%20with%20Plugins.ashx
#200 Опубликовано: 21.08.2025 18:43:51
Plugin updated.
Changes:
Refactors the `find` function to be more powerful and intuitive, aligning its behavior with user expectations from similar math software.
- The set of variables to be solved is now implicitly determined from all initial guesses (`≈`) provided in the input system. This removes the need to explicitly declare all variables.
- The function signature is now more flexible. Users can either omit the variable list to get all results, or specify which variables they want returned.
- Output variables can be specified either as a flat list (`find(sys, x, y)`) or as a `sys`/`mat` block (`find(sys, sys(x,y))`).
- Introduces `_` as a wildcard argument (`find(sys, _)` or `MathcadBlock -> find(_)`). This provides an elegant way to explicitly request all solved variables, which is particularly useful for integration with `MathcadBlock`.
- The output is now filtered to return only the subset of variables requested by the user, in the specified order.
Available nonlinear solvers: FixedPoint, LevenbergMarquardt, NewtonRaphson, Broyden.
Default solver: "LevenbergMarquardt"
Understanding the Calculation Behavior. A Circular Dependency (Gemini).pdf
Отредактировано 21.08.2025 19:02:15
Russia ☭ forever, Viacheslav N. Mezentsev
2 пользователям понравился этот пост
- Новые сообщения
- Нет новых сообщений