pow [index (iterate (mul 0) (+1))]

…**… pow

:test ((+2) ** (+3) =? (+8)) (true)

Functional programming language based on pure de Bruijn indexed lambda calculus.

Lambdas all the way down.
No primitive functions.

> (+5)
[[[[2 (2 (1 3))]]]]
> 'a'
[[[1 (0 (0 (0 (0 (1 (1 (0 2)))))))]]]
> add
[[(([([[1 0 [[0]]]] ((((0 [[(((0...

> :time factorial (+42)
0.01 seconds

Efficient reduction using abstract machines and higher-order encodings.

Substantial standard library.
Source

> ∏ (+1) → (+3) | ++‣
> number! <$> (lines "42\n25")
> sum (take (+3) (repeat (+4)))
> (+10b) ⋀! (+12b)

$ echo "main [0]" > echo.bruijn
$ bruijn -b echo.bruijn > echo
$ wc -c echo
2 echo
$ echo "hello world!" | bruijn -e echo
hello world!

Compilation to Tromp's binary lambda calculus.
Support for byte and ASCII encoding.

Meta-programming
and self-interpretation.

> length `factorial
> !(swap `((+2u) (+3u)))
> lhs (blc→meta "010000100000110")

Why?

By having a very small core (the reducer), bruijn is safe, consistent, and (potentially) proven to be correct!
Since it doesn't have builtin functions, bruijn is independent of hardware internals and could easily be run on almost any architecture.
Compiled binary lambda calculus is incredibly expressive and tiny. Read the articles by Justine and Tromp.
Exploring different encodings of data as function abstractions is really fascinating.
Naming parameters of functions is annoying. De Bruijn indices are a universal reference independent of the function and can actually help readability!
Really, just for fun.

bruijn