Entitas VS hecs

Github documentation for entitias looks like it is still good https://github.com/sschmid/Entitas-CSharp/wiki

For examples I usually turn to GitHub. It has numerous C# game engines. However, I find it easier to figure out implementations myself and use other engines only for API inspiration. There are also several C# ECS such as Entitas and DefaultECS that you can use if you do not want to create you own.

- Entitas

For Rust, I really like Bevy's, but it gets too much in the way. I'd probably use macroquad instead with something like hecs (I tried macroquad with Bevy ECS and didn't go well).

The question arises from seeing a plethora of projects using ECS: hecs , Bevy , specs, legion

My current design has struct Core that is basically "everything you need to save in a savefile", it has a hecs ECS (which needs a bit of boilerplate the hecs docs show you how to write to serialize it), and a bunch of simpler gamestate stuff like the discovered map positions, the current player etc. Everything is tree-like and serializes into a text file. Entity handles from hecs serve as "pseudo-pointers" that can represent cycle-like structures without running into endless cycles.

Out of the other Rust engines we've tried I think Macroquad was the most interesting option, and even though I've only made a few small projects in it worked extremely well and was what I'd expect from a game framework. While working in Bevy I felt like it was "writing fun Rust", but it wasn't really making a game. Macroquad on the other hand got immediately out of the way, and using it together with hecs was a painless experience where the whole time I felt like I was working on "the game" rather than "building systems that are invisible to the player".

The last crate worth mentioning is probably hecs for ECS, which we don't really use as heavily as some ECS fans would assume, but it made working around some problems between GDScript and Rust easier by storing things in ECS, passing around handles and querying ECS instead. Initially we did this with a global object and lots of state (which we still use for some things), but as the number of "things" grew it became easier to put it into ECS.

I use the awesome Vulkano for rendering and computation, and Rapier for simple physics. Contour is used for the initial shapes, but rapier forms the physics colliders from it. Hecs is used as well. And you might recognize Egui as gui :). I gotta say, I'm starting to be pretty happy with the rust ecosystem overall.

In the hecs documentation there is a section, Why Not ECS?. In it, the author states, "If your game will have few types of entities, consider a simpler architecture such as storing each type of entity in a separate plain Vec."

Usually, what I do is creating large, robust components in Rust. In my game, most of the logic lives inside the Factory node, which inside holds a full ECS (currently using https://github.com/Ralith/hecs) as well as other associated resources. This node takes care of holding the state and simulating all the machines in the factory and their interactions.

Generally, spatially-indexable data gets special treatment in games. (See Why not ECS for example, from the hecs ECS library.)

That impl basically says that Leftwalker is a System (or implements a System interface, to use a different parlance). Why exactly do Systems need lifetime is something better asked of the authors of the ECS library the tutorial uses. (personally I use hecs https://github.com/Ralith/hecs instead because I find it easier to use, no lifetime in sight :p)