Funny you mentioned that, because we had a ton of requests on that and it has become now our first priority: we decided to address a larger audience which has waited too long and to produce a DVD on VEX, cos the info is sooo scarce…
And the first one will be on VEX in SOP context.
The first part will be inter-context VEX ( « context-independant »),
« workflow operators » ( how to visually use loops, conditionals into), using the inline VOP, using multiple inputs for deformers ( we will present the general « theory of deformers » where you use some geometry to deform, a rest geometry input which « captures » the geometry and a deformed geometry input, and we will show a lot of examples).
We will use fractal/noise functions ( like a special cookbook – understanding noise « octaves », different types of noise, loops for fractals etc..) for comprehensive detailed terrain generation (exactly like in Terragen using the strength of point attributes for displacement/color combined with some other SOPs ) , understanding object topologies to intelligently use VEX, coordinates for deformations, how to work with groups, point neighbours, etc.
The story will go on with VEX with shaders of course, depending on the requests/needs of our users.
Our policy at CmiVFX has always been to emphasize on DATA STRUCTURES.
I completely understand your request to know how shaders work obe with each other. Every serious programmer know that functions are nothing if you don't know ON what they operate, where/when they're executed, in other terms what is called a context.
The context must come first! ( even if this sounds unusual at a first time!)
IMHO data structure and dataflow is the backbone of everything. I maintain that there's no serious learning of 3d without a proper understanding of the underlying data structures and spending some time on them with pedagogical animation, diagrams etc will save you A LOT of time and free A LOT of brain space .
My motto is « Take time to build your shelves » first
![](/static/djangobb_forum/img/smilies/smile.png)
(in shaders, this means f.ex knowing the « adventures of color » from emission ( light) to the image ( COPs planes), through all the shaders).
The great thing is that shader communication can also be used in a more abstract way to provide general info transport ( you then « capture a lot of tricks » at once). This has already been used ( i think about the « magic lights » in the latest Siggraph renderman courses where lights carry « abstract information ») and this is something we intend to develop.