On 4-10-2018 19:00, Kenneth wrote:
> Thomas de Groot <tho### [at] degrootorg> wrote:
> 
>>
>> After testing, I think you will get more reliable results using
>> Reorient_Trans(y,Norm) instead. Otherwise the arrow text has a tendency
>> to rotate around the original y axis.
>>
> 
> [off-topic]
> That little suggestion of yours has actually cleared up a long-standing mystery
> for me. Having assumed that Point_At_Trans and Reorient_Trans produced the same
> results, I had always used the former, for normals found on the surface of a
> sphere (it just seemed easier)-- with the result being that my placed object
> there rotated 90-degrees every now and then, relative to the surface's own
> normal-- specifically, when the traced point crossed the border of
> triangular-shaped 'quadrants' on the sphere surface. (It also occurs on the dips
> and undulations of a height_field.)
> 
> I thought that this was due to a natural ambiguity of the found normal. But
> Reorient_Trans doesn't perform that un-wanted rotation, as I see now. Excellent!
> 
> 

Thanks for this little insight into working processes. From my side, I 
have always used Reorient_Trans instead of Point_At_Trans, probably for 
the reasons you mention, through the advice of somebody else, like 
usually happily happens in these ng's. The descriptions in the docs are 
the shortest possible, but it tells nonetheless the difference in 
concept from which one can understand the behaviour, with a bit of luck. 
;-)

Maybe this is a little bit clearer:
http://www.f-lohmueller.de/pov_tut/trans/trans_470e.htm

-- 
Thomas

Post a reply to this message