Thoughts on Zig

2024-09-30

In continuation with my previous post about the new systems programming languages, I mentioned I was considering Zig as a potential complementary systems language to our main one, which is Go(lang). Well, for the past month or so, in my spare time, I tried writing something more substantial in it to understand the language more. For some time now, I’ve been “itching” to write something similar to Hashicorp’s memberlist library, but in a lower-level language for performance, smaller footprint and minimal network load. Now, I’ve used memberlist before, and it is a superb piece of code, but I wanted something that supports a consistent leader across the whole fleet. It’s a requirement to a system I plan on building in the near future (more on this in a later post). My top choices were C, Rust, and Zig, and as I said, I took a liking to Zig due to its promised simplicity, so I wrote it in Zig. The project is called zgroup, and you can check it out on GitHub if you’re interested. It’s still similar to memberlist but with the added capability of electing a leader across the whole group. It uses both the SWIM Protocol, which memberlist uses, and Raft’s leader election algorithm.

Anyway, so what do I think of Zig? I think it has a lot of promise. It still feels incomplete, which it is, especially the standard library. Which is expected of course for a < v1.0 software. But coming from C, and Go, I think Zig is the sweet spot for me between C++ and Rust. But I don’t think I will be introducing it to Alphaus anytime soon though; maybe when it’s tagged v1.0+, which I think is still years from now.