Is Directus going to be re-written in Rust, and what does that mean in terms of migration and existing extensions? I read on this AMA that is the case. I am wondering if Directus is the right choose for us, but this makes me uncertain.
In the Directus email about this AMA, it mentions
> Our upcoming architectural rewrite in Rust
That is a big thing to just casually mention, I’ve been following the issue queue, discussions and discord and hadn’t seen anything about that!
Could you share more on what this is, and what kind of timelines are you looking at?
We’ve been a bit stealthy about it to make sure folks didn’t onboard toooo early (learned our lesson from Directus 8 → 9 and Nuxt 2 → 3 )
We have some big improvements in the pipeline that required a bit more scalability and resource management than we’re able to squeeze out of NodeJS, so we’re indeed jumping ship to Rust!
That sounds… scary. Why not just a better JS runtime? Deno, for example? Wouldn’t that save a lot of effort? I don’t think there are a ton of Rust developers. We’ll have to rewrite all of our custom extensions, no?
We’re hard at work preparing for the next decade of Directus, and that does indeed include refactoring our core internal logic to Rust for improved performance and scalability. JavaScript is a great language to develop in quickly, but its runtimes are quite resource intensive to scale. Whilst our internals are moving to Rust, that doesn’t necessarily mean all extensions have to as well. The amount of developers who are familiar with and skilled in Rust is but a tiny fraction of that of JS, so the plan is still to keep JS as a first class citizen target for extensions. It’s a bit too early to go into full detail on what that’ll look like as it’s still in active development, but hope to have some more fun news to share before the end of the year
You should also consider making dotnet a first-class citizen.
I would assume there’s a horde of corporate dotnet devs out there who would love to plug their controllers into a solid, manageable backend with an extendable, nice looking admin UI taking the burden of webdesign from them . It might not take much more than spinning up a dotnet runtime instead of JS, given the interfacing is done by gRPC.
I had some hope that Umbraco would fill this gap as a dotnet powered multipurpose backend, but its evolution seems to have slowed down over the course of the past years. So please consider giving the homeless a new home
Early access releases will always start a bit smaller, but it’s too early for me to tell with full confidence what that release timeline and process looks like
)
)
. It might not take much more than spinning up a dotnet runtime instead of JS, given the interfacing is done by gRPC.
