Resizing matrices - again!

[RadovanOmorjan]

More than few times I tried to explain that(in some cases) Prime Beta,Alpha is resizing simultaneously more than one matrices although only a single matrix is selected for resizing. I attached an AVI file (the quality sacrified to the size, sorry).



More than few times I was told that this could not be reproduced. Am I doing anything wrong here?

[mzeftel]

Radovan,

I am using the same Alpha software and tried to duplicate your steps about 6 times and the second matrix does not resize. Can anyone else report seeing the same problem? I will give your video to our quality assurance group. I don't know what's different about your system versus mine that would cause that.

Thanks for posting the video.

Mona

[mzeftel]

Our QA department was able to duplicate this, but only sometimes. Can you list the exact steps, to help them narrow it down.

Thanks,

Mona

[RadovanOmorjan]

Here it is. I must say that I do not see any unusul step here.

1. Run Prime
2. Type A:identity(50)[ENTER]
3. Type A=
4. Move the matrix A down with the mouse
5. On the right type A*A^-1[SPACE][SPACE]=
6. Click anywhere outside the result
7. Place the cursor on the right lower corner on the matrix A
8. Click and Drag it in order to resize it
9. And here we are: both matrices will resize

Sometimes you can put below the third matrice, say, again A= , or A*A^-1[SPACE][SPACE]= and click on the empty space. All the three matrices could be resized at the same time by resizing either of them

Regards,
Radovan

[IRstuff]

I've tried repeatedly to duplicate this without success.

MP1A running on W7 Q8200 PC

TTFN,
Eden

[Ninetrees]

I can't duplicate in W XP Pro SP3.

Rich
http://www.downeastengineering.com/

[mzeftel]

We've narrowed down the issue to only happening when you click in a certain area in the worksheet. Now that we can duplicate it, our developers can fix the problem.

Thanks for all the efforts to help us track this down.

Mona

[Ninetrees]

Wish that I knew where that spot is!!

In the attached image, you can see that I was working with the matrix Ch. I was shrinking it vertically. Then I noticed that all 3 matrices were shrinking at the same time. Then M' went nuts. Changed the color of Ch here and in the plot to black. Turned the plot color gray. The doc scrolled up and down by itself for several minutes, then stopped.

Then, too, the attached M' file works in M14, but not here. 2 CPUs @ 85%, but never ends...

Rich
http://www.downeastengineering.com/

[IRstuff]

Huh, your file seems to have no problem in my installtion of MP1a.


TTFN,
Eden

[Ninetrees]

Sooo...that means that you get real values for a, mu, sigma, and the chi squared stats? How long does it run? And what do you get for results?

It has /never/ run to completion for me, even after I let it run for many minutes, and it should run in a few seconds, tops.

I wonder if I should re-install M'? I'm running Win XP Pro SP3 on a laptop w 2 GB RAM. Java 6 Update 17.

Rich
http://www.downeastengineering.com/

[IRstuff]

Your function seems to work OK by itself; PhotopeakStats takes about 0.7 seconds running under W7-32b running on a 4GB Q8200 system.

I get a [161x1] 0.958 1.205 [3x1]

The [160x] I guess runs from 1465 to 1625 or so
The [3x1] returns 3731 1544 37.916

The assignment to the vector never finishes, though, but it appears to be a problem with the nested array that's assigned. If you replace that with a single variable, it completes. I don't do programming, so I don't know if that's a bug or not.

TTFN,
Eden

[Ninetrees]

Thanks for the detailed reply. The time and the values are acceptable. As far as the nested array goes, I think that is a bug. OTOH, I replaced the nested array in PhotopeakStats with a const = 1, and a variable name replaces the array in the call to PhotopeakStats at the bottom of p 2, yet all I see in the lower RH corner of the call to PhotopeakStats on p 2 is a slowly spinning yellow circle, never any sort of answer. I've tried disabling and re-enabling calculations, and ctrl-F5 to recalc the entire file, but no change. I submitted it as a bug report...

Rich
http://www.downeastengineering.com/

[IRstuff]

When I replace the LHS nested array with a single variable, it seems to work.

Attached is the modified file.

TTFN,
Eden

[Ninetrees]

Thanks for posting that. I actually did ALMOST exactly what you did. The only difference was that I DIDN'T post the standalone call to PhotopeakStats at first. When I did that, the program executed, and my time is similar to yours. When I then included the assignment, as you did, including the variable "blah," my copy did not run to completion. When I replaced all the variables in the assignment matrix (LHS) as [a1 a2 a3 a4], it DID run. Then I put back the variables, one at a time, until I discovered that if I include Ch in the LHS matrix, the program fails. In your copy, running on my machine, it does not.

I suspect that there is some confusion with labeling. In M14, if I select a constant, the tool bar indicates that; if I select a variable, the too bar indicates that. I don't see this in M'.

I deleted and re-entered the variable Ch, but no change.

So I'll send off both copies to PTC and let them
see what they find...

Rich
http://www.downeastengineering.com/