Hacker News new | ask | show | jobs
by instig007 208 days ago
If you ever need this kind of stuff, you'll be better off building your own distributed interface by using plain regular GHC Haskell and https://haskell-distributed.github.io/
1 comments

`haskell-distributed` is awesome, but the reason Unison is a new language is exactly to avoid the limitations of such frameworks, namely that they _can't_ send arbitrary code and data around like Unison can.

You should totally write up a tutorial demoing the development of the same application using each. Folks would love that.

> but the reason Unison is a new language is exactly to avoid the limitations of such frameworks

by introducing another set of limitations that Unison docs don't state upfront, because it will affect their progression as the business.

> namely that they _can't_ send arbitrary code and data around like Unison can.

> You should totally write up a tutorial demoing the development of the same application using each. Folks would love that.

You should add "out of the box", and that "like Unison can" isn't the only way to implement a distributed runtime. This leads to another logical question: why should I "totally write up" the Unison way of things? At this point I'm not even sure that 99.999% of workloads need distributed runtimes of any kind, let alone the Unison one.