@chkat 

If I interprete correctly, Maple kernels (mserver) terminate external processes but the GUI does not do that.

The behaviour might be OS dependend. You use Linux. Correct?

Could you upload demo code that works under Windows?

By "closing the Maple window" you mean crtl-F4 or crtl-shift-F4 or close tab (right click on tab) or exiting the Maple seesion (i.e. the GUI, Alt-F4).

@one man 

It is possible to simulate rolling without slipage with MapleSim. Sometimes unreasonable high values for parameters of the contact have to be used. The problem is that no slip detector is available in MapleSim. Only in rare cases slip can be detected. This make simulation nice to look at but limits the interpretation. Slip equals wear and that is of interest.

Here it took me a while to reproduce an ideal rolling case.

I had no time for a code review. You say:Then we look at where its current location on the trajectory should be

Does this mean that the trajectory is given? By that, do you mean the trajectory of the contact point?

@Rouben Rostamian  

Then it looks like as if the output is not properly generated.

Thank you for the explanations. I would love to see Maple supporting quaternions.

This is a nice animation that raises a question: Are there initinal conditions that reverse the direction of travel of the small torus inside the large one?

Too bad that MapleSim does not allow a torus inside a torus contact. Simulating this with Maple is tempting but would take allot of time. Maybe there are other trustwothy alternatives that could be employed to find such solutions.

@Andiguys 

The problem is complex. In those cases answers might be obtained by restructuring expressions. The expression P in Simple_T can be converted to a factored expression: P=t1*t2*t3 .
These factors can be investigated separately and you might be able to extract meaningful relations. Here is one for t1 telling us conditions when t1 is positive and that apparently t1 cannot be negative. Now you have to combine this result with solutions for t2 and t3 and identify realtions for the parameters. In the best case you can derive form these realtions ranges for parameters (using solve).  There are quite some combinations to be investigated.
I now have to work on a less intersting task but I hope this helps to advance. It's not exculded that this problem can be boiled down to something meaningfull.

Download Simple_T_rply.mw

Interrupt works in screen reader.

Looks to me that the interrupt does get through to the kernel.
To interrup: Click on the bug icon an then quit and you are back to normal.

@nm

When my laptop runs on battery it switches to a balanced mode which limits tasks running full steam on the processor and the graphics card. This leads to Maple running slower. I assume that in the case we are discussig simplify was interupted by timelimit before this dll is called

What I see is that the dll is still active: the offset adress pointer (yellow) is still changing its adress.

Whatever simplify has asked the dll to do, it seems to be in an infinite loop. No new memory is allocated which is often the case in other instances where Maple "hangs" but effectivley an algorithm piles up result after result without comming to an end.

If timelimit would accept an instruction limit or processor cycles instead of a real time limit, a termination would be more deterministic.

My hypothesis is that Maple cannot termiate (all?) third party dlls. mserver.exe (including timelimit) is waiting for the dll to terminate. All functions in the stack of mserver wait for the above functions to terminate.

This case should be debugged.

@nm 

I had the same question. Maybe the sqrt expressions causes the call.

By the way:

I wanted to have a look at the stack again and restarted Maple.

This time the system ran in optimized battery mode. Simplify was interrupted before the call to GMP. If timelimit would work on CPU time this would not have happend. Maybe an option you could profit from and easier to implement than a true realtime kernel. I guess...

It hang's here:

If the page ?GMP is relevant then it could be that the hang is in a third party library and timelimit cannot/ does not terminate the call to this library.
Note: No memory piling up is visible which I see often when Maple becomes unresponsive.

Something must be special about "e".

@acer 

I used the form in the past but did not get feedback. I find the form not appropriate in cases where it is unclear whether a command performs as intended or where I do not want/need a software change. In those cases initial, informal clarification is helpfull.

For the future development of the Physics package: Where can/should observations (not necessarily bugs) made with the package be reported? 

Does it always make sense to contact Maplesoft support in any case?

@xavier

It's not working on my installation.

It would be good if someone else could try it on Windows (10).

How much time does it take to generate a sodoku?

Seconds, minutes, hours? I stopped after 2 hours

@ecterrab 

It's a long time ago that I had this lecture on tensors, which mainly dealt with curvlinear coordinate systems and deformations of them (structural mechanics). So I might not have used propper terms.

What I had in mind with expressions like "V M V" are physical quantity equations (literally translated - that's how I remember) which are tensorial equations meaning that they are valid independend of coordinate systems.

If possible I work with equations, do algebra to get symbolic results and only at the end plug in parameters for numerical results. These equation are most of the time scalar equations. For the parameters I use equations (no assignements) and eval/subs as well, which does not anihilate the symbolic results.

When browsing through physics tensors I was looking for examples that do the same with tensor equations. I.e. start with a tensor equation and then switch perspective between, for example, an innertial reference system and a rotating reference frame. This switching is similar to entering different sets of parameters but not anihilating the tensor equations.

Kind of stripped down tensor calculus in flat spacetime for engineers where different point of views are desirable or large deformations play role. The first part can be done with matrix and vectors provided the user knows the transformations (and does not confound back and forth transformation matricees). The second part needs tensors.

My dream was that Maple provides the constituing equations in coordinates of a desired reference frame with a push of a button.

Take this example of a rotating observer that neither sits in the innertial frame nor on the rotating body: (The Bizarre Behavior of Rotating Bodies. Doing here the physics right is quite a task.

This is maybe asked too much or easily possible with a subset of the Tensor package.

Please consider my question not as a feature request. My question was simply triggered by your explanations and the prospect that the Physics package could mabe lay a foundation for a larger user base. 

The package is huge achivement. I feel that the gap between what physist are taught (need to know and feel comfortable with; and what the tensor package provides) and what engineering will finally use is too large that non-physicist will have a look at the package. Some engineering examples might change this. As you said, Phycis,Tensors is more a compulsive reference manual.

@xavier

Do you mean not compatible with Maple 2015 or with Maple 2025?

I have tried

and [123,432] for rand number
and [1,1] number of grids

as you have indicated.

Maple 2025 and 2021 are evaluating but not finishing.

Can you give an example that works on your installation?