Is it a bug in prime 2.0 with Jacobian??????

[ptc-4735478]

I was about to start a simple problem...and bam! Prime2.0 has its own math AGAIN! I am using Version M010 BTW...

See attached of the Jacobian! I am sure the jacobian matrix is not 3X1.. and the Jocobian determinent is not 0

.Or did they change the syntax for Jacobian ?

Oh my god...why would I trust prime 2.0 for any math?

It has just way too much problem....come one..it is not the first time...

if you have seen my older posts..you know this is not the first problem that I ran into with prime 2.0........................

[StuartBruff]

Hmm

henry leung wrote:

I was about to start a simple problem...and bam! Prime2.0 has its own math AGAIN! I am using Version M010 BTW...

See attached of the Jacobian! I am sure the jacobian matrix is not 3X1.. and the Jocobian determinent is not 0

.Or did they change the syntax for Jacobian ?

Oh my god...why would I trust prime 2.0 for any math?

It has just way too much problem....come one..it is not the first time...

if you have seen my older posts..you know this is not the first problem that I ran into with prime 2.0........................

It does look like something's different from M15.

I've only got Prime Express, which doesn't evaluate symbolics. Have you tried explicitly defining xi to be a 3-vector?

Stuart

[Werner_E]

Your worksheet did not work for me either.

I recreated it in MC15, converted it to Prime and now we have the correct results.

I could not spot a difference.

[AndyWesterman]

I went same route of re-creating in 15.

- looks like another candidate from the mathcad mad-house.

Problem is with the first ξ in the Jacobian.

Regards

Andy

[Werner_E · Accepted solution]

I found it, but I don't know what its called in the english version. The argument xi in the function X when you call jacob is formatted as "variable" whereas the other as "none". So to Prime its not the same variable.

WE

[AndyWesterman]

Of course. Keep on forgetting MP's little oddity there.

Nice to know why it fails

Thanks

Andy

[ptc-4735478]

PTC's introduction of labeling in prime is too troublesome....It was a good intention but often come out to be useless. I think they must add a reminder to user when ever the exact same 'name' has two different labelling. Math ad prime sometimes automatically choose different label for the same name and it is causing a headache to me.

thank WE very much for finding the glitch

[ptc-4735478]

Have a new question about the same worksheet.

This is more related to mathcad usage

Would like to solve x_greek[1, [2, [3, in terms of x[1, [2 and [3

or simply x_greek in terms of the component of x

I know how to do it without using index (I.E. using x_1 instead of x[1 )..

but if I do want to use it this way, what can I do? Because I do not want to type/copy and paste the equation again

Thanks

[Werner_E]

At first glance I thought the problem was the boolean equal sign but correcting it didn't solve the problem. Clueless!

[Werner_E]

OK, at second and third glance I'm not that clueless anymore. Find attached two possible workarounds. My third attempt to rewrite the equations in vectorform unfortunately failed - see attached.

P.S.: and yes, I realized soon after my last post that the boolean equal was OK anyway (the LHS should have been a row vector, though).

[ptc-4735478]

Thanks WE,

Actually I am aware of this work around.

What I wanted was to be able to symbolic solve individual elements of a vector in terms of the element of another vector given the relationship betweens the elements from the two vector; This will involve the use of index [i . I am not sure if this is possible in Mathcad prime or mathcad 15.

I guess for now I will just use different names for the individual elements of the two different vector, so that mathcad will interpret it as different variable instead of an element of a vector when using [ to differentiate them.

Thank you.

[Werner_E]

henry leung schrieb:

Thanks WE,

Actually I am aware of this work around.

What I wanted was to be able to symbolic solve individual elements of a vector in terms of the element of another vector given the relationship betweens the elements from the two vector; This will involve the use of index [i . I am not sure if this is possible in Mathcad prime or mathcad 15.

It doen't seem to be possible. Looks like mathcad is not aware of the dimensions of xi as xi never was used on the lhs of an assignment.

But at least with both workarounds you are able to avoid typing the equation a second time.

[Fred_Kohlhepp]

Not exactly sure this applies, but (in 15 at least) you can assign a symbolic rersult to a variable, then you can address elements of that variable.

[Werner_E]

Fred Kohlhepp schrieb:

Not exactly sure this applies,

Unfortunately not. The point is that a symbolic evaluation which works perfect in Mathcad 15 is broken in Prime.

All indices are matrix indices and nor vector x neither vector xi are defined before in any way.

In Prime the errormessage is about Mathcad cannot understand that syntax.

I have tried to convert the M15 sheet to Prime and at first glance it seemed to be OK, but thats not true as Prime does not evaluate on loading the workshet and that way we see the result which M15 had calculated. Recalculate in Prime gives the error.

I thought they would leave at least the symbolics as is but obviously PTC is enhancing symbolics the same way they were enhancing user interface and useability 😞

One thing I stumbled upon while dealing with that question was something different which doesn't work in M15 either. Why are the following two solve-expressions are seen so differently by Mathcad? Looks like in the equations matrices are expanded in their components while the solve variables are not!?

[Fred_Kohlhepp]

That's strange! In 14:

[Werner_E]

Thats strange, indeed!

I just checked again if the assignment to a variable XX would change the behaviour, but it doesn't. A*X evaluates as it should but solve gives the erreor and so XX is undefined later.

Obviously something has changed since M14. I am using Mathcad 15 M020.

Not sure if we should call it a bug but then, I would expect it to work the way you have shown in the screenshot.

[ptc-4735478]

Another question on the same exercise

So Mathcad 15 does not allow programming nested aray? Thank for any answer in advance.

[StuartBruff]

henry leung wrote:

Another question on the same exercise

So Mathcad 15 does not allow programming nested aray? Thank for any answer in advance.

[ptc-4935984]

es un aparato inteligente que te ayuda mas en tu mente

[ptc-4935984]

y valla que es de mucha utilidad y son muy buenos en el presio y en la calidad de aparato

[Werner_E]

So Mathcad 15 does not allow programming nested aray?

Sure it allows for nested arrays. I usually always initialize, even local variables which are not supposed to have any undesirable values from prior calculations. Preferrably one init too much rather than one too little as in your case.

In your case I would rather turn position into a function and omit the j-loop:

I have shortened the i-loop for better visability.

Werner