I am using version 0.87.3728. I performed the calculation below:


As you can see the answer should be simply -0.4063804, but Smath added an unwarranted, very small imaginary number to the answer.

Hello Will,

I tried this on 0.87 stable (0.87.3692)

Could you double check it. This looks to me Ok, although quite different from your result.
I again forgot to bring 0.87 Beta with me, sorry. I'll check this out tomorow.

Regards,
Radovan

Wrote

Hello Will,

I tried this on 0.87 stable (0.87.3692)

Could you double check it. This looks to me Ok, although quite different from your result.
I again forgot to bring 0.87 Beta with me, sorry. I'll check this out tomorow.

Regards,
Radovan

I downloaded and installed the stable version (0.87) and it also worked fine for me.

So this must be a bug in 0.87 Beta.

Thanks for the help.

You are welcome ,

You might be right. I mentioned some odd behavior wnen using trigonometric fuctions with units arguments in the first Beta. Andrey fixed that. It might remain some unfinished work on the same issue.

Regards,
Radovan

Hello Will,

Here is the 0.87_3728 version of your example made on my computer.

It is again different than yours but closer to mine previously made on 0.87 Stable.
I think that the imaginary part could be freely neglected here (less than 10^(-15). If the SMath would have had the system of representing results with the given precision, this imaginary part would have not been presented here. Therefore, I suggest you to use just the real part of the results.

On the other hand I am not sure why our results are so different?

Regards,
Radovan

Wrote

Wrote

Hello Will,

I tried this on 0.87 stable (0.87.3692)

Could you double check it. This looks to me Ok, although quite different from your result.
I again forgot to bring 0.87 Beta with me, sorry. I'll check this out tomorow.

Regards,
Radovan

I downloaded and installed the stable version (0.87) and it also worked fine for me.

So this must be a bug in 0.87 Beta.

Thanks for the help.

Note that you had the angle mode set to degrees and omorr has his set to radians (0.87 seems to return a real result in either mode)

CB

Oh yes, thank you captainblack. I just forgot to check that out . It did not come to my mind, sorry.

Regards,
Radovan

Wrote

Hello Will,
If the SMath would have had the system of representing results with the given precision, this imaginary part would have not been presented here. Therefore, I suggest you to use just the real part of the results.

Radovan,

As CB said, I was performing the calculation in degrees, but, as you have demonstrated, the inclusion of an imaginary component occurs regardless of whether degrees or radians are used. In this case, I did end up using the real component. However, since I use Smath to clearly document calculations, the results produced by Smath in this case are unacceptable since it introduces unnecessary clutter to what should be a straightforward calculation. So, I would most certainly add this problem to the "to be fixed" list for 0.87 Beta.

Sincerely,
Will

Hello Will,

I suppose that SMath should have Zero and Complex treshold in order to avoid these situations. This way would overcome the mentioned problems.

Regards,
Radovan

Wrote

I suppose that SMath should have Zero and Complex treshold in order to avoid these situations. This way would overcome the mentioned problems.

Agree with you. Will implement this.