@nm 

I have submitted a software change request. Maple does not seem to recognise in this case that numerator and denominator contain polynoms with same primitive parts (but of different powers). Primitive parts, if I understand correctly, are fundamental in algebra and can be used to normailize expression. This probably also explains why evala sometimes is helpfull at simplifing expressions. 

Only if combine is used with my original assumptions, sqrt(2) is extracted and a subsequent simplify is effective with assumptions on beta. Intuetively I would not have expected that combine can extract sqrt(2).

I stopped trials with your assumption -1 < x which indicates that the result is valid for a larger domain. How to get there with Maple?

Output[61] without using a combine command is of course reason enough that Maplesoft has close look.

221298  
228066 

@acer @dharr

Thank you

Maybe an alternative:

In the view tab on the ribbon is an option where mouse scrolling worked on my Windows 11 installation. 

Using it involves quite some clicks and mouse movements...

The problem is that is can get "polutet" with Help Pages.

@acer 

Too bad that your utilities have dissapeared. They might have adressed a niche topic but there will allways be users striving for simplicity (in code and results) beyond looking for pragmatic solutions.

(Dumb and boring) AI is not a help here and I do not see an (affordable) AI agend testing "all" combinations of Maple commands with brute force anytime soon. 

I keep it with Leonadro Da Vinci who supposedly said:

Simplicity is the ultimate sophistication.

@acer 

I am almost sure that I tried this but maybe I forgot to add the assumptions. 

It would be nice to have a routine that does the trial and error to find a solution. (Like int with the option 

method=_RETURNVERBOSE). I think @nm once showed here a routine that tried many "simplification" ways. 

Anyway, I am glad that I have asked. The command works for the elliptic integral to my full satisfaction.
Thank you.

PS.: I did not expect Maple 2026 to be any better. Is it?

@dharr 
Interesting and unexpected. I could not dedcuce this use from the help page of evala.
Thanks.

@acer 

Thank you

`&#128077`;

or

👍(not printing in text mode)

Good initiative. I am a bit worried that a new generation of engineers and managers trust AI too much.

I noticed in the ladder example provided on ExaktAI that units are not well handled (yet).

When ExaktAI (the underlying LLMs) can handle this, a list of 100 "typical" engineering problems would be interesting to benchmark.

A calcualtion annecdote came to my mind when reading your post: I learned from a retired optics specialist that, before the advent of computers, 3 independent human calculations where required before design and manufacturing of lens systems could start. Work was mentally so exhausting that the human caculators had a 6 hours working day... That was less than 70 years ago.

Does Maple "chokes" with the uploaded example or does it require special settings? What do you exactly mean by choking?

Example_Generate_Models-2025.2.mw

@wingho

I assume that with "literature" you mean the Maple help system.

I had a look at pdsolve documentation. I do not see a pdsolve option that reports of solver settings (including the type of solver). This would be useful when it comes to modifying default parameters and solvers.

I also tried infolevel[pdsolve] := 5, but this does not give more information (this works for dsolve).

Have you consulted: ?pdsolve,numeric,education/p>

This seems to offer more options (solvers and parameters) to solve your pde.

About the physical problem:

Can I assume that u(x,0) is equivalent to the compression of a rod at t=0?

If so, the solutions make quite sense to me. The relaxation propagates with the speed of sound. It reminds me of a falling slinky.

Remarks: The initial compression with 25% is quite high which could require modeling that accounts for large deflections (non-linear effects).

@dharr 

Compared to my original post, your puzzle can be reproduced. It is probably a regression.

Unfortuneately we cannot probe what the user interface sends to the kernel and what the kernel receives.

Edit: I seems that MaplePrimes has different challenges interpreting 2D-Math. Better open the attached worksheet

Download ditto_2-D_Math.mw

@acer

I edited my reply.

I don't think it's a bug. But I think I did not understand what you were demonstrating with the uneval example with respect to equation labels. I could not make use of it.

I am only guessing that the user interface attaches equation labels to output from the kernel and replaces equations labels with whatever is attached before sending this to the kernel. If I may say so.

Why the parameter modifier ::uneval is not making form the argument sol[1] a 'sol[1]' is something I do not understand since I do not know the evaluation rules when a procedure is called. This is probably descirbed somewhere in the documentation. The only difference between the arguments is that sol[1] is a name assigned to an expression and the second argument is an "anonymous" expression. In this example c is not evaluated

c:=2;
f := proc(x::uneval) x^2 end proc:
f(c)
                             c := 2

                                2
                               c 

So why is sol[1] replaced by the expression to which sol[1] is assinged to?

@acer

Maybe I am wrong. I though that

sends only an indexed name for sol[1] and the whole output of equation label (2) with the index 1 to the kernel. If I copy the command and paste it here (or in 1D-Math input) I get exactly that.

ExpressionTools:-Compare(sol[1], (y(x) = RootOf(-Int(1/(2*cos(_a) + RootOf(-Int(1/(2*cos(_a) + _Z)^(1/2), _a = 0 .. y__IC) + x__IC + c__2))^(1/2), _a = 0 .. _Z) + x + c__2), y(x) = RootOf(Int(1/(2*cos(_a) + RootOf(Int(1/(2*cos(_a) + _Z)^(1/2), _a = 0 .. y__IC) + x__IC + c__2))^(1/2), _a = 0 .. _Z) + x + c__2))[1])

When the kernel evaluates the above, the second argument is taken as is and first
edit: the first argument 
is taken from the kernels memory wihtout any further evaluation.

Can the communication between the kernel and the UI be recorded?

@acer 

Thank you.

My takeway finding is that the kernel did not know the output I was refering to.

When names are used, those surprises won't happen. By habit I use euqation labels. This comes from text book style documenation where variable names and assignemt operators could distract or confuse readers that do not know Maple.