I didnt mean it too seriously
Wondering how Mesa and Gallium3D in AROS would compare with TinyGL/Goa in MorphOS.
As I see it, TinyGL might be faster (at least in some cases) but Gallium3D is way more complete.
Don't have any experience with Gallium3D, but if I had to mention 2 major things that Gallium3D has (and TinyGL doesn't), they would be:
1)VBO (Vertex Buffer Object) support.
Instead of sending mesh data over (relatively slow) memory bus each time drawn, the data is uploaded to gfx ram (whenever it's initialized / changed) and drawn from there. According to my tests, that can even achieve 5x speedup. Basically it "adds a few functions around what you have with vertex arrays", so it's (relatively) easy to #ifdef for tinygl compatibility.
Should be relatively easy to implement in TinyGL too, at least compared to:
2)Shaders.
With shaders the coder can have (almost) total control what happens with vertices when given to OpenGL, and how the pixels will get their final color. Includes a C/C++-like language that's compiled when the program is ran (so the same binary will work on different gfx cards). This is totally different way to code for OpenGL, when compared to fixed function pipeline (which TinyGL uses)
OpenGL ES2 doesn't have any fixed function pipeline (Which ES1 was based on), and the entire fixed function pipeline was also deprecated in OpenGL 3.0 (and removed from 3.1, current version being 4.2).
While the fixed function pipeline will likely never be removed from "desktop implementations", more and more code will be written using shaders instead, as that's the "current" way of doing things.
In addition, TinyGL is also missing several "minor" features, such as color index mode and stipple support (both of which are almost totally useless, and also deprecated) - occassionally you might find some OpenGL game that uses some exotic functionality that TinyGL is missing, or just isn't working as expected.