A perfect circle has 360°?

[Joachim_L]

This is what I've been told decades ago. In the Affinityverse a perfect circle has 360,00001°

Test yourself with 6 decimal places for degrees.

[walt.farrell]

I've read elsewhere that Affinity does not have "true" circles. Perhaps this is another aspect of that?

[Solly]

Welcome to the world of computer representation of rational numbers. Rounding will always take place somewhere. Here you see a 0.00000278% error. This topic of numerical representation was an active topic of discussion in computer languages, programming, and scientific calculations in my student days in the early 1970's. The answer to the problem was always, "what amount of error no longer makes a difference in the result for practical purposes?" To think of it another way, a 360 degree circle is an ideal that is approximated to any degree of precision. The more like the ideal, the more time and expense involved. A more precise measuring instrument is more expensive. Very few of us have need of a micrometer for everyday use around the house, but someone wanting to track paper thickness would certainly have one. There's nothing to see here, move along.

 

[Gabe]

Hi all,

I moved this to feature requests as it's not a bug, but just a small rounding error 

[Ben]

This will be because all angles are represented internally as radians. We then convert them to degrees using a*180/pi.  That leads to a small amount of error in floating point precision.  The error is small enough not to matter in real terms.

 

In this example - the angle is not about the circle, but the arc start/end when presenting a pie shape.

 

As Walt also says - we use cubic bezier quadrant approximations for a circle/ellipse - so there is an element of error in the outline compared to a mathematically true circle.

[Alfred]

20 minutes ago, Ben said:

all angles are represented internally as radians

It’s been mentioned elsewhere that it would be useful to have an option for the user to work in radians instead of degrees. If that’s what you already use internally, it seems like a no-brainer to include such an option.

[NotMyFault]

4 hours ago, Alfred said:

It’s been mentioned elsewhere that it would be useful to have an option for the user to work in radians instead of degrees. If that’s what you already use internally, it seems like a no-brainer to include such an option.

The procedural texture filter allows to choose between these UOM. Unfortunately this doesn’t help here.