Another symbolic Integral returns wrong result

[ScottMilligan]

Running M030, I found that MathCad returns an inconsistent symbolic result (see attached file), depending on which of two equivalent integrand definitions is used. Comments are welcomed.

Scott

[ptc-1368288]

On 11/28/2009 6:58:50 PM, Astroverted wrote:
>Running M030, I found that
>MathCad returns an
>inconsistent symbolic result
>(see attached file), depending
>on which of two equivalent
>integrand definitions is used.
>Comments are welcomed.
>
>Scott
>_____________________________

Comments:

1. Plug in the Mathematica web integrator.
2. Check with Maple.
Note that if you "Save as" 11 or lower, more collabs will read and might eventually get the right answer. A better idea is to attach an image so that collabs in all Mathcad/Maple versions will read.

jmG

[ScottMilligan]

Jean:

Here it is in MathCad 11. Also, I confirm Tom's observation regarding the indefinite form.

Scott

[ptc-1368288]

On 11/28/2009 8:03:39 PM, Astroverted wrote:
>Jean:
>
>Here it is in MathCad 11.
>Also, I confirm Tom's
>observation regarding the
>indefinite form.
>
>Scott
__________________________

You seem confusing the integral function with the cumulative integral. As you can observe, Mathcad/Maple is all correct vs the book. A common misunderstanding that many user do/would commit with any CAS. Your Mathcad 14 should agree. If not: MuPad is not correct. Posting the work sheet results in this simple verdict serving well again the collab community.



Jean

[ScottMilligan]

Jean:



My point wasn't that an iterated sum of the definite integral and the value of the definite integral are not the same, but rather that MathCad returns different symolic results for the definite integral vs. MAPLE 13. Perhaps more telling, MathCad also returns different results comparing between its own symbolic evaluation and numeric evaluation of the same integral, as I hope the attached sheet demonstrates.



Scott

[ptc-1368288]

>On 11/28/2009 8:03:39 PM, Astroverted
>wrote:
>>Jean:
>>
>>Here it is in MathCad 11.
>>Also, I confirm Tom's
>>observation regarding the
>>indefinite form.
>>
>>Scott
___________________________

No idea what Tom has confirmed, because it seems you have a problem with Mathcad 14 [MuPad], then not concerning your integral. Whether right/wrong in MuPad, you may be able to put it right following the guidance attached. Incorrect means: not printing correctly in the mind. The mind result goes on the paper or work sheet if you wish, what goes on the paper if invalid is then "wrong". You seem to be wrong, maybe as well Tom. Assume MuPad and PTC were incorrect in the first place, it could very well be that they are now "wrong".

Please, do not hesitate to come back before this coming saturday. I will be back January, 15 2010. BTW, Mathcad has an "iterated product operator". but you said "iterated Sum". Whether it has a mathematical meaning and use, don't know, never heard of "iterated Sum".

Jean

[TomGutman]

Strange. It calculates the indefinite integral correctly, so the problem seems to be in the application of the limits.
__________________
� � � � Tom Gutman

[ptc-1368288]

On 11/28/2009 7:32:09 PM, Tom_Gutman wrote:
>Strange. It calculates the
>indefinite integral correctly,
>so the problem seems to be in
>the application of the limits.
>__________________
>� � � � Tom Gutman
_______________________________

Apply the same method I have suggested to Rich.

jmG

[TomGutman]

>>Apply the same method I have suggested to Rich.<<

Whatever for? Once again, you seem to have completely missed the point of the post. I don't happen to have Mathematica, nor Maple, nor is the correct value of the integral in question. This is purely a MC14 issue, which somehow gets the value wrong. If you don't have MC14 to test it with, you have no useful comment on the issue.
__________________
� � � � Tom Gutman

[ptc-1368288]

On 11/28/2009 9:28:19 PM, Tom_Gutman wrote:
...
>This is purely a MC14 issue, which
>somehow gets the value wrong.
...
__________________
>� � � � Tom Gutman
====================================

Then, one more issue with MuPad, but no surprise. Though I have some doubt with such a simple pattern recognition. I can't verdict in the absence of the MuPad result and whether it is correctly assigned or not.

jmG

[AlvaroDíaz]

On 11/28/2009 7:32:09 PM, Tom_Gutman wrote:
>Strange. ...

MuPad reproved applying Barrow's rule. It's a dangerous bug.

Regards. Alvaro.

[mzeftel]

I logged it.

Mona

[rmix22-disabled]

On 12/3/2009 2:16:34 PM, MonaZ wrote:

>I logged it.

and according to to PTC-Support its already fixed in M040 - no date could be told. I ran into the same bug with a slighty different integral.

Does anybody know when M040 is scheduled. I thought the next release would be MC15 in June or July!?

Hope that some other integral errors will be fixed too, like the missing summand in INT(x*ln(x),x)

RMix

[IRstuff]

I think M030 and above are only available to maintenance customers.


TTFN,
Eden

[rmix22-disabled]

On 3/20/2010 6:02:22 PM, eden_mei wrote:
>I think M030 and above are
>only available to maintenance
>customers.

Thats true but I am maintainance customer and I am running M030 (because M035 does not offer any advantage apart from Windchill but has that nasty memory bug).

It's a shame that PTC does not provide a bugfix to _all_ customers _immediately_. Producing wrong results is such a serious bug for a math-software that I would consider a quick reaction mandatory.

According to PTC-Support the bug _is_ already fixed, they just don't offer it yet.

Simply waiting until the next bigger maintainance release is rolled out and then only offering the bugfix to customers who paid twice ... - does not sound like a company one could recommend in good conscience.
I remember PTC selling MC14-M011 (this version was full of bugs) to single users at a time when release M020 already was rolled out and insisted their new customers to pay a second time (for maintainace) to get the bugfixed M020. Finally and far too late M020 was then released for all customers. To stay with the Topic: I think the integral-bug this thread is about was introduced with M020 (or maybe even prior when the switched from F000 to M010 or M011). It does not affect the oldest MC14 version.

Maintainance is a nice thing to have and PTC-Support is usually quick and helpful. But the fix of such serious bugs should not be dependent on the customer having a paid maintainance contract.

Just my 2 cents.

My version is under maintainance anyway and I will get M040 whenever PTC decides to let it free. It seems nobody has an idea when this will be (that being my original question) and I can only hope that PTC will handle it like M020 and offer it to all of us.

RMix