Nim vs ATS

There are generics, templates, macros in Nim. They can allow you to write new DSL for your application, or avoid all boilerplate stuff.

Nim has built-in VM, which executes macros and some other code at compile time.
For example - you can check if you're on Windows, and Nim will generate code only for it.

The same code can be used for web, server, desktop and mobile.

Checks your code at compile time.

Imperative, OOP, functional programming in one language.

Nim has a lot of common with Python in terms of syntax.
Indentation-based syntax, for/while loops

You can use unicode names for variables, there is "unicode" module for operations with unicode

You only need to specify types in your procedures and objects - you don't need to specify type when you're creating a new variable (unless you're creating it without initialization)

With built-in "unittest" module you can create test with a very readable code

You don't need to deal with all those manual memory allocations, Nim can take care of it!
But also you can use another GC, or tweak it for you real-time application or a game

You can use Nim with any language that can be interfaced with C.
There's a tool which helps you to create new C and C++ bindings for Nim - c2nim
Also you can use Nim with Objective C or even JavaScript (if you're compiling for these backends)

Nim has "asyncdispatch" module, which allows you to write async applications.

writeLine(stdout, "hello") can be written as stdout.writeLine("hello")
proc add(a: int): int = a + 5 can be used like 6.add.echo or 6.add().echo()

ATS has a theorem-proving type system powerful enough to prove that its functions meet their specifications. This happens at compile time with no performance impact at runtime. This can be used to prove that an ATS program doesn't have bugs commonly found in C++ programs, like "this function never leaks memory" or "this program never attempts to divide by zero" or "this buffer never overflows" or to verify pointer arithmetic, etc.

The compiler (ATS/Postiats) is GPLv3.

The syntax is ML-like with the usual functional language features like pattern matching and tail-call optimization.

ATS works as a low-level systems language. ATS programs have performance and footprint comparable to programs written in C/C++.

Similar to Modula-3. This makes ATS a viable choice even for large-scale projects.

ATS can prove its concurrent programs have no deadlocks or race conditions.