Mike Williams wrote:
> Wasn't it Stephen who wrote:
>> H. Karsten wrote:
>>> Well this would be maybe an option for the version 4.
>>>
>> LOL but seriously why do you want to scale things to zero? If I want 
>> to show something getting larger from nothing I initially translate it 
>> out of view then using a “step jump” translate it into view and scale 
>> it up from a small value.
> 
> If we're talking about zero in all three dimensions, then it would be 
> efficient to put all the objects inside if statements, and skip the 
> parsing and rendering if the scale is zero. Perhaps using a macro that 
> accepts an object and a scale factor if there are many such objects to 
> be handled.

Yes, that is a better way of doing it. I was thinking of my use of a 
modeller that could not utilise #if or #case statements.

> 
> However, the warning and the defaulting to 1.0 happen if only one of the 
> dimensions is scaled to zero. Perhaps Holger wants the objects to be 
> scaled infinitely thin, like part of a plane. With the current 
> behaviour, if you try to scale an object infinitely thin in one 
> direction, the parser thinks you've missed that parameter and scales it 
> to 1.0.
> 
> I can imagine some of the ray intersection solvers having big problems 
> with objects that really do have zero thickness, because they're trying 
> to find points where the ray goes from outside to inside something, and 
> that doesn't really happen if the object has no thickness. [Note that 
> thin objects like planes don't have that problem because they consider 
> half the universe to be inside.]
> 
> I can also imagine compound objects, like unions, suffering from 
> coincident surface artefacts wherever two components with different 
> textures overlap.
> 

You put what I was thinking into a well defined statement. I wrote in 
haste and now repent in leisure ;)
-- 

Best Regards,
	Stephen

Post a reply to this message