My Workflow

So I’ve been playing around with procedural generation for a while now and I wanted to share my workflow.

I’ve been programming for ages now, and i was mainly C++ programmer. Lately I’ve been moving to lower level/simpler (e.g. some call it C–) programming. This is somewhat reflected in the software I’m using/developing. And all of that experience and work is my project called “PixelDance”.

It started as a “what if shadertoy but with more flexible pipeline”. Now it grows as I need something (e.g. want more performance, lets add Transform feedback).

These are the guiding principles:

• Performance - many toys need huge computations thus GPU needs to help. I’ve chosen opengl.
• Easy to play with - there are things you want to change on the fly and see. That is why DearImgui is used for gui.
• Easy to reconfigure - some things need way different ‘pipeline’. That is why i’m using very performant and easy to integrate LuaJit. The language itself is very simple and binding is easy (when you get used to the stack).
• Easy to hack - the C style C++ “engine” allows quick addition of various things I need.
• Edit and see - inspired by Handmade hero and others (also called hot loading) the engine automatically reloads the script as soon as you save, thus change a value and see.
• OpenSource - I’m using CC-BY: Attribution as I don’t believe this is something benefits from being locked down.
• Save and remember - usually when saving an image I append the source to it so later I can load it and continue editing.

So I invite you to check it out here.