Code generation is what I’m trying to avoid at all. But sometimes it can be useful, for example, to automate repeatable operations like creating ecs-system template or even ecs-startup behaviour.

New 20180722 version of ECS framework was released and time to break something again :)

Google docs sheets can be used as data source as universal replacement of ScriptableObject. Data can be downloaded as CSV and used as-is or be converted to JSON. I already did it as part of my huge library, but if you need only one small feature - its too complex project. So, I extracted google docs downloader to separate project.

Yes, IL2CPP still much slower than MONO in common cases. Can we decrease this performance penalty somehow?

Updates. They can be painful, but will help alot for moving forward in development. Time to break something in my ECS framework.

Reactive systems are systems that can detect changes on filters and process them. They are implemented at Entitas ECS framework too, but not at buitlin-unity ECS. Are they useful at all?

Mecanim in Unity - very powerful feature with automatic bones detection and retargeting introduced 4 years ago. What could go wrong?

Sometimes we have alot of data with complex processing logic and it takes alot of cpu time. Modern cpus have multiple cores and code execution can be splited between them for parallel data processing in same time. How we can get performance boost inside ECS framework systems without huge code changes?

Reactive systems in ECS can be very useful for processing custom reaction on component changes. But how we can be sure that filtered entities will be processed once and without performance penalties?

ECS framework works good enough, but it’s non-visual thing and works as background code. What if we want to get some visualization about data inside?