PLOT ERROR

[ssato]

I have a quetion.

I attatched the file.

I do not understand the error on last plot of file.

I think "n" is no problem.

Best Rgs

[Werner_E]

---

I do not understand the error on last plot of file.

I think "n" is no problem.

---

I agree and as I don't see any rational reason for this error I would call it a strange Prime bug.

Prime seems not to like the fact, that only every fourth value is non-zero? Not sure what exactly causes that bug.

But I can confirm that it should work and it also does work OK in Mathcad 15

I found two possible workarounds to do it in Prime (not satisfactory but at least you get your plots):

1) change the factor 2 of the cosine to something close to 2

2) change the step width of your range n to 1,1.5 .. 50

[Fred_Kohlhepp]

Interesting!  In Prime 3.0:

Going backwards?

[Werner_E]

Fred, your function v differs from the original function. The 2 is only a factor of the cosine.

[Fred_Kohlhepp]

Sorry, needed more coffee!

Still:

[Werner_E]

Yeah! Lack of coffee really can have bad side effects 😉

But nonetheless we see that your Prime 3.0 plot is quite happy with that expression.

So the bug seems to be newly introduced in Prime 4.0 (guess it works still OK in P31, too).

Maybe the only left-over of the plot enhancements we were promised that Prime 4 will include.

[Werner_E]

Here is another workaround using a different, but equivalent function:

[Werner_E]

Here are two other workarounds using the round function

I also noticed, that if we change the range from n:=1,2...50 to n:=1,2...7 Prime has no problem. But as soon as we go up to 8 or more - error. Thats crazy.

[DJF]

Don't we just need n to be a vector (not a range)(add the = sign) and vectorize the equation?  

[Fred_Kohlhepp]

There are two primary uses for range variables:

1. As indices for vectors, and
2. As plotting independent variables.

If making n a vector fixes this problem in Prime 4, then it's still a bug, and an aggravating one.

[Werner_E]

Sure, this is a valid workaround, too. Given that we don't need n as range later.
Point for me is that it should work as initially done and due to a, obviously freshly (as Fred showed) introduced, bug it doesn't.

I might add a third use to Freds list - ranges are also needed in programs for for-loops.

[ssato]

Thank you. everyone.

I will attach another file.
If "a" is an odd number in the last graph, an error will occur only with graph creation.
If "a" is an even number, no error occurs.

In the end, is this a bug in Prim 4.0?

If you are a bug, where should I contact?

[Werner_E]

It sure IS a bug and obviously a new one as Fred head shown that in Prime 3.0 it works OK

PTC is not very interested in bug reports.

In fact you have to pay for the privilege(!) to send a bug report - only if you are under maintenance you may open a support ticket an report that bug.

[LucMeekes]

And then it may take a while to get fixed...

Success!
Luc

[Fred_Kohlhepp]

---

@LucMeekes wrote:

And then it may take a while to get fixed...

---

My stopwatch doesn't carry intervals long enough!

Anybody have a "stop calendar"?  😉

[LucMeekes]

As a matter of fact, I have a small clock that I can set for a (one) date and time in the future. If it reaches that point in time it will sound the alarm; it was initially set for 2000-01-01 00:00 and would show how many days until the set time. I don't have one that counts up, but they're not too difficult to make on a web site.

Come to think of it:

Would be fun to have a down counter to the predicted date of https://community.ptc.com/t5/PTC-Mathcad-Questions/Mathcad-Prime-5-0-some-new-updates/m-p/489877#M174257 as a reply to that post.

Luc

[Werner_E]

---

@Fred_Kohlhepp wrote:

---

@LucMeekes wrote:

And then it may take a while to get fixed...

---

My stopwatch doesn't carry intervals long enough!

Anybody have a "stop calendar"?  😉

---

How's a calendar called which covers more than just one year?

[ssato · Accepted solution]

Everyone, I understood well.......