[Bf-committers] How about adding boost uBLAS librarytoblender/extern?

[Yves Poissant]

> Simply put, while I like GI algorithms and new shading
> frameworks/concepts the "legacy systems (as defined in you preamble) are
> not as legacy & outdated as you make out. Renderman & it's clones are
> all based around these "legacy" concept and are pretty much the standard
> against which GI frameworks / algorithms are tested... and then
> generally integrated into in some fashion.

Yes. Exactly. Starting with "Cars", Pixar have been integrating more and 
more modern approaches to rendering. But I would hesitate comparing 
Renderman, a shading language, with a closed function renderer such as all 
legacy renderer are. It is very easy adding new functionality in a shading 
language. Even adding completely new paradigms. With a shading language, you 
don't have to use any legacy functionalities if you don't want. You can even 
create your own parallel system.

> It's nice to
> talk about the next generation of rendering algorithms and how Blender
> can be a part of said revolution, but replacing the scanline / shaders
> approach is (in my opinion) madness.

I hear you. Actually, I hear that argument a lot.

> but focusing their
> rendering architecture on the scanline / shader concept (like most
> professional pipelines I have used) is the way to go.

In the end, I don't think it is your opinion or mine that will change 
anything. The end user will decide. There are plenty of alternative 
renderers available. More and more Blender users are trying them and 
adopting them. And they are just following the trend. Looking at the 
"Exibition" section of 3D World no 112 and 113, I counted 11 out of 17 
renders that used VRay, or MentalRay, or even Maxwell. More than half. And 
that proportion is shifting more and more as months passes.

Time will tell.

Yves