HI again. While investigating possible solution, for porting my project to ECS, I am learning, on what ECS can and can not do as of yet. My recent study, was based on the game design approach discussed in Unite Austin 2017 - Massive Battle in the Spellsouls Universe Discussion includes subject regarding tackling rendering and collisions, for thousands of entities. All nicely and highly applicable for their type of the project. They highlight, as the game entities (minions) behaves as particles, similar to the water, where when too close to each other, then they repeal their nearest neighbor. But such approach, is not the realistic collision behavior, for what for example rigid body does in classic programming methodology. And yet what I need. Hybrid ECS with RigidBodies In my project, I expect having many entities (building blocks) of possible some different convex shape. For OOP based programming, there are GameObjects. Each can have collision behavior, hence colliders and RigidBody, as well as rendering component. Since unity ECS do not support collisions yet, I was considering take a time and try write collision physics, or (leaning toward) stick to hybrid ECS approach for now, for tackling collisions. From my understanding, this is very achievable at this moment. And then let Hybrid ECS apply relative forces as required. I hope I got that last bit right. Pure ECS with Mesh Rendering Now the matter of rendering above objects. If my understanding is correct, this is least problematic, to get object rendered. By building pure ECS and assigning relevant meshes / textures to the bootstrap equivalent, or entity component, I should be able to render many of objects. However, I may expect, that some blocks have different shapes / textures. So I presume, is just matter of creating components groups / filters, which can store meshes, textures varieties. Pure and Hybrid ECS all together Now, I need just to just confirm, that coupling both pure and hybrid ECS, to control collides and render meshes respectively, is the right approach. I read from discussion, there is nothing against, with such approach. At least, should keeps code cleaner, and hopefully closer, to port into pure ECS in the future. Recap Questions The question is, Is the right way to build a systems of Hybrid and Pure ECS, to drive collision and rendering? Are there any additional potential traps, which may hinder such development? Or, how would you deal, or dealing with such matter? Cheers.