I have had great success in generating 2D graphs using OpenSCAD, and just wanted to share. https://www.youtube.com/watch?v=ThDIWKD-4LY
I am still working on 3D software rendering from the ground up; my goal is to not be limited by what APIs such as OpenGL have exposed, nor be subjected to algorithm anomalies that hardware vendors such as Nvidia have introduced; the world understands graphics pipelines well enough, and floating point operations in CPU have reached speeds that are high enough, and computers come equipped with processors that have a number of cores enough, for it to be possible to now "reinvent the wheel" by creating a 3D graphics API that exposes more of the algorithmic details to do fancy things, and exposes more hooks to do custom pixel/zbuffer manipulations e.g. described by Kirsch with CSG rendering. I'm just not very happy with OpenGL or with how the hardware-accelerated rendering hardware has been. The tricks they have up their sleeves is marketing and parallelism, that's pretty much it.
Also over the next few years I intend to work on a full CSG library, but steps have to be taken carefully and this is going slowly. CSG is very similar to pixel choosing in 3D because in both cases you need to pay meticulous attention to accumulation errors and rounding strategies.
Cheers, Nerius
Sent with Proton Mail secure email.
I have had great success in generating 2D graphs using OpenSCAD, and just wanted to share. https://www.youtube.com/watch?v=ThDIWKD-4LY
I am still working on 3D software rendering from the ground up; my goal is to not be limited by what APIs such as OpenGL have exposed, nor be subjected to algorithm anomalies that hardware vendors such as Nvidia have introduced; the world understands graphics pipelines well enough, and floating point operations in CPU have reached speeds that are high enough, and computers come equipped with processors that have a number of cores enough, for it to be possible to now "reinvent the wheel" by creating a 3D graphics API that exposes more of the algorithmic details to do fancy things, and exposes more hooks to do custom pixel/zbuffer manipulations e.g. described by Kirsch with CSG rendering. I'm just not very happy with OpenGL or with how the hardware-accelerated rendering hardware has been. The tricks they have up their sleeves is marketing and parallelism, that's pretty much it.
Also over the next few years I intend to work on a full CSG library, but steps have to be taken carefully and this is going slowly. CSG is very similar to pixel choosing in 3D because in both cases you need to pay meticulous attention to accumulation errors and rounding strategies.
Cheers, Nerius
Sent with [Proton Mail](https://proton.me/) secure email.