"tom" ha scritto 
> 
> Has someone looked in the sphere_sweep problem that was previously reported
> in :
> 
> http://news.povray.org/web.44779e5a358a67bc5c947f990%40news.povray.org
> 
> The last version of povray that correctly rendered the file mentioned
> in the bug report was compiled v3.5 on Linux.
> 
> 

The problem is that the intersection between a planar speresweep and a ray 
orthogonal to the plane is ill-conditioned. The equation is a polynomial
that is a perfect square:

    p(d) = p2(d) * p2(d)

meaning that it is always non negative and zero only on the surface.

And every root is multiple and sturm's sequences should be special cased,
not mentioning that bisection is pratically useless...

I have an implementation fixing these and other problems, just waiting
for the sources to become open source.

Massimo

Post a reply to this message