Am 26.07.2013 11:32, schrieb Shay:
> "clipka" <ano### [at] anonymousorg> wrote in message
> news:51f1eefb@news.povray.org...
>
>>> That train was impressive. Used prisms*, so, imo, a whole other animal.
>>
>> No, not really. There, too, it was all about cutting stuff away and
>> adding stuff back in.
>
> Similar techniques and an equal challenge, for sure. But, if there's a
> distinction between a "csg scene" and a "hand-coded scene", I see the
> primitive primitives as the logical place to make that distinction.
> Primitive primitives are the Neoplasticism of hand-coding.

Would you consider boxes as primitive primitives for that matter? 
Because I see not much of a conceptual difference between those and 
(linear) prisms: Both can be replaced by a compound of planes.

I'd also like to note that the power of (linear) prisms does not lie in 
being a shortcut in terms of CSG (after all it would be comparatively 
easy to throw together a macro that takes an array of vertices and 
generates a prism constructed from planes), but in being much faster to 
render.

As for the terms "csg scene" and "hand-coded scene", here's how I'd 
define them:

- CSG scene: Any scene that defines all its complex shapes as unions, 
merges, intersections and/or differences of less complex solid shapes 
(which in turn might also be defined this way); after all, that is 
exactly what "CSG" means: "Constructive Solid Geometry". (In the strict 
sense this would even allow for meshes, provided they have an 
inside_vector, but I'd be ok with still disallowing them.)

- Hand-coded scene: Any scene that was created using nothing more than a 
text editor and POV-Ray itself.

The two terms are perfectly orthogonal: The shapes in a CSG scene can be 
both hand-coded and/or created with a CSG modeling tool, and a 
hand-coded scene can have both CSG shapes and/or non-CSG shapes.

Post a reply to this message