On 06/03/2016 02:24 PM, clipka wrote:
> Am 27.05.2016 um 02:11 schrieb William F Pokorny:
>
> It might be worth actually testing how the clockwise/counter-clockwise
> thing works out depending on the handedness of the coordinate system,
> rather than just speculating or relying on unconfirmed assumptions.
>

I assume you mean flipping the handedness of the camera to right and 
looking at all the test cases I created?

It is true I ran only one of the tests posted to povray github this 
way(1). It didn't change result with 3D handedness. It will take me some 
time to crank all the test cases with a right hand camera & review the 
results.

Is there some case you suspect will flip for the 2D curves that I might 
jump to first in an attempt to save some time? You're usually a few 
steps (as in gigantic, flying leaps) ahead of me... :-)

End of a long day, but I'm still thinking 2D is 2D & while given 
rotations and mirroring things might at first look different given a 
viewpoint, there is really only one normal planar 2D representation.

(1)- (bezier, lathe, 12 counter-clockwise loops varying point set 
rotations, straddling y=0 forming rings about y)

>
...
>> So, I guess the question is do we:
>>
...
>
> - Keep the "clockwise" and "counter-clockwise" wording, and add a note
> saying that these apply to POV-Ray's left-handed coordinate system,
> whereas with a right-handed coordinate system the direction is reverse?
>
> (Of course such a note should only be added if tests show that it is
> indeed the case ;))
>

I will check in on this thread tomorrow morning, before running what 
test cases I have with a right hand camera set up.

Bill P.

Post a reply to this message