The xy2pol looks that still has problems with units:

Basile

Jean,

In your example:
The Magnitude (length) and The Angle (angle) in polar coordinates have different units. The rad is not a common unit. It is not correct.

In my example:
Look again please:


This was solved in recent upgrade by Andrey but apparently remains in the xy2pol.
http://en.smath.info/forum/yaf_postst7552_Units--Operation-cannot-be-performed.aspx

Basile

Better be in harmony wth the maths and Smath.

There is nothing to look gain, Smath is OK.
The Magnitude (length) and The Angle (angle) in polar coordinates have different units.
For sure they have different units: one is in "UnitGraph" the other is in "rad"

The rad is not a common unit. It is not correct.

YES: rad is correct, rad is the maths unit in trig approximations as well as in official tables.
If you want °: convert by appending ° in the Smath result

Jean,
1.
Please go previously and see the picture.
"rad" is the unit for both the magnitude and angle.
There is ONE placeholder for ONE unit.

The Magnitude (length) and The Angle (angle) in polar coordinates have different units.
For sure they have different units: one is in "UnitGraph" the other is in "rad"

The rad is not a common unit. It is not correct.

YES: rad is correct, [YOU JUST SAID THAT ONE IS LENGTH THE OTHER ANGLE!!!] rad is the maths unit in trig approximations as well as in official tables.
If you want °: convert by appending ° in the Smath result

2.
In this specific set of applications I have no interest in "numerical MathEMATICS". I use units. Thank you.

Basile

@Basile,

Just show what you are trying to describe verbally



Not that it would convince anyone who doesn't wish, though.

Mike,

The xy2pol does not work when units are included:

Do we agree?

When eventually it is working:
The units of the "Magnitude" and the units of the angle, (for the case of complex numbers here), will be different.

Note: I see you got W(att) how?
I see that you have a semicolon ( ; ) I have a comma (,).
Re(S);Im(S) versus Re(S),Im(S)

Basile

Something like that?



Notice that atan2 ins't in SMath, which is more correct than atan for this functions (see http://www.mathworks.com/help/matlab/ref/atan2.html for example).

Best regards.

Alvaro.

Wrote

Mike,

The xy2pol does not work when units are included:
Do we agree?

Yes we do

Цитата

When eventually it is working:
The units of the "Magnitude" and the units of the angle, (for the case of complex numbers here), will be different.

Yes, that's what I show in my screenshot, that uses ≡ instead of =, just to show the desired result: seeing is better than reading.

Цитата

Note: I see you got W(att) how?

Now I suppose it's apparent: Ctrl+= and manually make system with desired result

Цитата

I see that you have a semicolon ( ; ) I have a comma (,).
Re(S);Im(S) versus Re(S),Im(S)

Russian locale default settings

In this specific set of applications I have no interest in "numerical MathEMATICS". [Basile]

From A...Z, Smath is a numerical math tool as any other CAS and including your specific application.
xy2pol result is a vector of two components that need no unit as they are implicit. Extract each
component and append a dead unit decoration for clarity to yourself and the reader.

Wrote

Notice that atan2 ins't in SMath, which is more correct than atan for this functions (see http://www.mathworks.com/help/matlab/ref/atan2.html for example).

atan2.sm (35 КиБ) скачан 163 раз(а).

Hi Mike. I prefer this other, but have some round inconvenients:

Best regards.

Alvaro.

Файл не найден.Файл не найден.

Wrote

I guess that with the new try function samebody can make a good one atan2.

But with try, it won't differentiate again.

Wrote

Wrote

I guess that with the new try function samebody can make a good one atan2.

But with try, it won't differentiate again.

Oh, I think that this happen only to me :-). Not very sure that try don't accept diff.

So, this with numerical optimization must to be the best solution (until atan2 not be in SMath):

Best regards.

Alvaro.

It's a good option, and only ~30% slower.

Like this Alvaro: look for the Matlb code in the f(x) menu,
look for mwatan2 ... if you have in your version: then go

Jean



Maths atan2.sm (10 КиБ) скачан 160 раз(а).