I would really like to see types on the beam. Right now the one that I’m keeping the closest eye on is Hamler (https://github.com/hamler-lang/hamler). I know the other project is Gleam, but I’m indifferent to the syntax. It must be pretty hard since projects like Alpaca, and Elchemy have been abandoned.

For me it’s not about checking types, it’s really about reasoning about the code. When I put my Elixir projects down and come back 3 months later, when I navigate to the function that needs changing, it involves tracing every call site to infer what types are getting passed in. I loose interest in having to Hindley/Milner all my types when the computer can do it better than me To be fair typespecs are great but it’s so easy for them to get out of sync and doesn’t help when you pass a function to another function and need to know the type signature (at runtime).

I am fully with you on the sentiment – but I am trying very hard to detach myself from syntax tastes. I have found so many times during my career that syntax barely matters. If the language constructs and the runtime are good I can swallow almost any syntax (well, maybe not COBOL but who knows).

Very true.

Personally I quite like the ML-style syntax, speak Haskell, Elm, and the like. At the same time I’m not really fond of the C++ style syntax with deeply nested :: namespaces with lots of shortened names, like std or Buf (seriously, you couldn’t call if Buffer)?

But the latter also applies to Rust and I cannot deny that Rust is a very thoughtfully designed language with a lot of merits (I’m actually learning it at the moment).

So yeah, as I see it good language design is nearly orthogonal to choice of syntax.