Most minimal static type system

What is the most minimal design for a static type system without slow compilation times, butt-ugly error messages or people saying "#lolnogenerics"?

Attached: maxresdefault.jpg (1280x720, 31K)

Other urls found in this thread:

elm-lang.org/
twitter.com/NSFWRedditImage

C

Background is a minimal scripting language that I might try to design and implement soon, btw.

I said most minimal, not most defective.

Attached: 1510495208148.jpg (659x1200, 435K)

ocaml?

You've got to decide what you want from the language. Maybe have a look at what golang has I seem to remember it was more reasonable.

Java-like classes and interfaces are pretty simple to implement, but you need methods and inheritance in your design. Parametric polymorphism if you don't want inheritance.
Templates with code generation feeded by simple function overloading is imperfect solution (2 hacks over an actual type system) that works surprisingly well.

Just add type keywords to Python and check that they match in runtime.
I mean, it'd still be a garbage language and implementation, but that's what you asked for.

Minimal doesn't have to mean shit.

>(You)
Maybe standard ML...
see

perl 5

Ducktyping.

He said without slow compilation times.

>

Attached: 1531460113529.jpg (700x963, 94K)

Typeless.

no u

Elm.
It's a very minimal language, and the type system is very simple, yet powerful enough.

elm-lang.org/

version 0.19, which will be released soon will have a very fast compiler, the current one is decent, but has some bugs that slow it down.

good hint.
Any specific version, though?

Nah, the current, 0.18 is fine and works great.

My point about the upcoming 0.19 release is just that it will improve the compile times, but that's really only an issue with a code base with 10000+ lines

nice self-portrait

Haskell's.

>slow compilation times

are you stuck in the 90s?

A badly designed type system can slow down compilation quite a bit. Not as much as headers, but enough for some languages that considered not doing type inference in generics for this reason.