FP2: Fully In-Place Functional Programming [pdf]

This research is related to the Koka programming language: <a href="https:&#x2F;&#x2F;koka-lang.github.io&#x2F;koka&#x2F;doc&#x2F;" rel="nofollow noreferrer">https:&#x2F;&#x2F;koka-lang.github.io&#x2F;koka&#x2F;doc&#x2F;</a>

I was hoping it would reference roc-lang.org, which is actually doing in-place mutation for efficiency, but I didn’t see any mentions :-(

This is also used in Lean 4, which as it turns out is not just great for math proofs, but is also becoming a great general purpose programming language

Interesting article, I wonder if FIP procedures could be useful even in non functional oriented languages as a safer means of reusing allocations (in languages that support reference counting or ownership?).

Noob question: the Top type here is exactly the Node constructor of the stree type. It seems useful to declare a type that is “this ADT, but limited to these constructors”, but even Haskell seems to avoid it. Why?

I mean, `jq` does in-place mutations when the values being mutated have just one reference. With a lot of care one can really make that shine. `jq` has been doing this since day 1, years ago, though it&#x27;s had a few bugs in `reduce` that had to get fixed.

Looks like constant folding to me. What’s the difference exactly?