I do not think that this was done on purpose.

Some things get forgotten. Nobody is perfect.

I am not a big fan of AI because of all the slop but in this instance the response to "in maple proc statement: syntax to make a description" was useful

name := proc(parameter_list)
    description "your brief description here";
    local ...;
    global ...;
    # procedure body
end proc:

I also think that it would not harm to update the documentation.

@Carl Love 

I have removed listlist from the title. Is the type list(list) supposed to work in a type check? Here it does not

ToListOfLists:=proc(ex)`if`(nops(ex) = 1,ex,[op(0, ex), map(thisproc, [op(ex)])[]]) end proc:
ToListOfLists((a+b*c)^2);

         [^, [+, [*, b, c], a], 2]

type(%,list(list));

                  false

@acer 
I have understood now the usefullness of beeing type specific.

Lists, sets, sequences and vectors are expressions I had not in mind. I am missing here a term to name all expressions that are not of such types. I thought algebraic is suitable for that. 

I any case a reconstructor makes sense to check if the deconstructed list(list) is as intended.

The search in Maple help was not succesfull with ?[] and "?[]". I do not understand why I have to do to increace my search success rate. The search functionality needs an overhaul.

Anyway, I have all that I need to deal with scalar(?)/non-container(?) expressions. subsindets is my current favourite. 

Thank you for the rich feedback

@acer 

In the fun one-liner below nops(ex) = 1 can be replaced by ex::atomic 

ex:=(a+b*c)^2;
(proc(ex)`if`(nops(ex) = 1,ex,[op(0, ex), map(thisproc, [op(ex)])[]]) end proc)(ex);

to check whether an atomic leaf in the tree of operands has been reached. 
I realised in the meantime that nops(f(x))=1 which answers my question whether the two criteria can be used interchangeably.

What I could not work out is to look-up constructors like `?[]`

Is that, like the list constructor `[]` , documented?

I was looking for a solution that works on all algebraic expressions.
I also had in mind the reverse operation potentially manipulating only one operator before transforming everything back.

I stopped my trial when it came to choosing either checking for atomic or for nops(ex)=1. Better to ask.

By the way:

It is also possible to work with an anonymous proedure to shorten further.  Nothing for beginners.

Recursive code is also nothing for beginners. However I was also looking for a simple example which we see here and which I find very instructive. The absense of using data structures for storage of intermediate results makes the example particularly interesting to teach advanced concepts of compiter languages.

Unfortunately I cannot investigate all the answers in detail because I got fever in the meantime. Sunday fun is over...

I am not instructing but what you wrote reminded me allot of this introduction to a visual proach to understand calculus.

@acer 

Not(atomic) is perfectly fine. (I was not aware that there are bolean types)

Thanks

@acer 

I would have drawn the same conclusion ("A name is  a variable when there is a valued assigned to it") from maplemint's output.

With your comments, the matter is sufficiently clear now. Thank you

@sand15 

Does this mean that the way you presented is also valid for complex random variables?

(Maybe be the Statictics package represents a limitation to the real domain)

I was wondering why Maple Statistics package offers a kernel density plot ("advanced function") but not a cumulative distribution function/plot (basic functionality). There is only a CumulativeSumChart (which I find not very useful).

👍

@acer 

Interesting detail. What could be the purpose?

With Maple 2025.2 on Windows

I do not get the multiplication dots also when I replace Pi by x. 

Can you upload the worksheet using . This makes it easier for others to reproduce the error.

@nm 

First time I see this (maybe because I do not use Physics very often). Rerunning Download ex1_jan_18_2026.mw

gives a warning.

Maple does not restart propperly... a second restart is required.

However, removing timelimit from your code does not show the warning after restart.

Have we (unintended) come across two implementation weaknesses or is this a combined effect?

@nm 

Occasionally I encounter output of a command that depends on the use history of Maple (not necessarily the use of the command). The output is not wrong but different. In cases were  forget(command) is not effective I use map(forget, [anames()]) to the reset other remember tables of functions (which I do not know) that are called during the execution the command.

This works sometimes. However I am not sure whether really all remember tables that could influence Maple output are reset. The kernel for example could remember results that cannot be forgotten this way. Hence the question. My intention is not to permanently disable.

Ideally forget could offer an option that not only forgets the remember tables of a command but also all other remember  tables encountered during the execution of the command. This would be more specific with less computational effort than forgetting all.