Hacker News Reader

Comment by anothername12 | Hacker News Reader

A lot of people write Lisp interpreters, post it to HN and call it a day, but I think compilation is the far more interesting task for a Lisp implementation.

SBCL is great to poke around in and the decompiler’s great. Also check out ABCL, ECL for what compilation to JVM bytecode and C looks like.

View on HN

This is a bit Haskell oriented, but the "Implementation of Functional Programming Languages" by SPJ is one of my all time favorite books.

https://simon.peytonjones.org/slpj-book-1987/

These compilers were pulling heroic data structure optimization stunts in 1985, that no modern production compiler for major compiled languages can replicate in 2024.

Why? Is it mostly because of Haskell's purity (side-effect-freeness), so optimizations can change expressions more freely and safely?

That, and the uniform s-expr/gadt representation of data structures instead of going wild with clever pointers.

I am not sure I understand what you mean in the context of Haskell.

https://wiki.haskell.org/GADTs_for_dummies

Huh – I didn't know he'd left MSR for Epic Games.

He did. He’s working on a new language called Verse now:

https://dev.epicgames.com/documentation/en-us/uefn/verse-lan...

Huh interesting.

The language seems sort of complicated for the domain (gameplay scripting).

The scheme compiler in Paradigms of Artificial Intelligence Programming by Peter Norvig is a great introduction to compiling Lisp. I was particularly impressed by it only taking a change in one line to go from regular recursion to handling tail call optimization.

Reply on news.ycombinator.com