SBTB 2019: Paul Chiusano, Unison, and why the codebase of the future is a purely functional...

[@dazraf]

Love this!

[@micah9461]

oh gosh I'm totally sold

[@yan-amar]

If I understand correctly, this codebase manager could be used with any pure FP language with tweaks only to the storage format. In a way, Unison has this in common with Lisp that the code is stored as syntax tree. I would love to see such a tool generalized for arbitrary languages.

[@Karthik-yy6up]

What does he mean by "I'm just going to bind that dependency here (pieCrust)"? pieCrust doesn't seem to be in scope.

[@ksenialovi3716]

It's already defined, as we saw in the editor: .sbtb> find : Ingredients ... 12. pieCrust : Ingredients ...

[@TheSunscratch]

Just wondering how hash collisions are handled

[@MegalomaniaZ]

I'd guess they chose a key-space that makes this impossible.

[@malkdk]

There’s no need to handle collisions since they’re using a cryptographic hash function with a digest size of 512 bits.

[@malkdk]

For anyone else wondering how to change the implementation of a function (rather than just its name) see 20:40

[@davidprock904]

I'm working on an extremely different architecture like nothing currently existing. You have some commonality with what im doing, but not putting you down, mine is better, you should seriously contact me.

[@bibliusz777]

show me what u've got

[@okdoomer620]

I feel like this is kind of a big deal