Crystal 2026 Predictions

Community
1

The time has come to peer into the crystal ball and tell me your predictions for 2026. Do you think Crystal is going to pivot into a functional language? You think a major startup will adopt Crystal? Is this the year of the Crystal LSP? Or maybe you’ve got a project in the works that will come out this year. Here’s your chance to let us all know what you think.

3 Likes
2

Let’s start with going over mine for last year:

  1. Better language tooling.
    • Tree sitter is in a pretty stable state
    • ameba-la is now owned by the crystal-ameba org
    • active work is being put into ameba to improve it by myself and primarily @Sija
    • we’ve started a working group (to be announced soon) to help coordinate everything required for getting tooling into a better state
  2. Semantic analysis in ameba
    • Not a lot of progress made on this front as I’ve had to split my attention and energy between several things. Not sure if we want this to be part of ameba or not
  3. Type-safe equality
  4. More structured macro system
    • No progress on this front
  5. Rewrite crystaldoc.info from the ground up to make it faster / more reliable and provide more functionality.
    • Waiting until we have better tooling to test out any new LSPs. What exists now is ugly but it works
  6. Fault-tolerant stdlib parser + semantic passes
    • A lot of ideas on this front, including just reusing the tree-sitter for language server things
    • Going to be one area of focus for the working group

Now for the coming year:

  1. The working group has had an initial meeting, but that’s only a starting point. Want to have more regular meetings with delegating tasks and working on things in parallel
  2. Simplify larimar as much as possible. It’s just a language server framework at this point, would be good to toss out stuff that’s not needed anymore and make it as lean as possible.
  3. tree-sitter to 100%. Only a few edge cases left
  4. markd gfm support to 100% and a new release
  5. Would be really nice to find a way to do partial semantic analysis that’s significantly faster, only requires going through methods once, and can be done incrementally. This would go a long way towards helping tooling
6 Likes
3

There are pretty high chances that Crystal will replace Rust as a low-level system language, and I hope that popularity will start to grow.

PS: I will make an announcement soon (some people on Discord are already familiar with what I will introduce), so stay tuned.

8 Likes
4

Over the past two years, Crystal has made two significant advancements: Windows support and parallel execution. I believe Crystal will continue its steady growth this year as well.

1 Like
5

As for me, I have some projects in the works. I’ve also had some thoughts around writing or making videos about crystal.

I think there’s going to be more problems this year around libraries falling out of date. Very few projects have a team to keep up with them and a lot of devs have moved on over time. I only hope we can give people a reason to come back

4 Likes
6

When I write Crystal, I often think: “I don’t want dependencies. I want full control over the libraries I use. I want to build my own pieces and make my own castle.” The Crystal community has people from many different backgrounds around the world, but from what I see, many Crystalist eventually start thinking this way. I’m not sure if this is bad. But I feel like something could be changed and would be better, but I can’t explain it well yet.

1 Like