Original Post
Back when i first started dabbling with 3d graphics was a class in college and we made a Scene Graph Loader. Given a text file we would parse through it and build the scene. Our teacher told us it was bulky and would never be used in practice, but was a good way to learn about textures, lighting, transformations, even animation.
My question is that I am getting back into graphics programming in general and was wondering how you all structure your "engines". Mainly, I have a class for lights, and texture management from my scene graph loader and I was curious as to if you all encapsulate things so as lights and textures into objects or if you do a more C-style non-OOP implementation. I am used to OOP and dividing things up into objects and such, its how I learned. I would like to keep my Light class and Texture classes, but want to know if its feasible for things like speed of the final product. Any insight and thoughts on this matter or anything related to it would be greatly appreciated.