February 23, 2026
Smol code, colossal chaos
femtolisp: A lightweight, robust, scheme-like Lisp implementation
Tiny Lisp, big drama: devs torn between eye-rolls and awe over femtolisp
TLDR: femtolisp is a tiny, fast Lisp that prides itself on old-school features—and it secretly powers parts of Julia. The crowd is split between minimalism fanatics, bootstrapping nerds, and skeptics of “yet another Lisp,” with lively debate over how small and self-sufficient a language can be—and still matter.
A pint-sized coding language just kicked a hornet’s nest. femtolisp, a tiny “Scheme-like” Lisp that fits in about 150 KB, showed up flexing speed and old-school discipline—and its creator roasted copycat languages with a mic-drop: “If you’re going to waste everybody’s time with yet another lisp, at least do it right.” That spiky vibe set the tone. Then a bombshell from the comments: one user revealed that parts of the popular scientific language Julia actually run on femtolisp, accessible via a not-so-secret command. Cue the collective double-take and a chorus of “Wait, what?!”
From there, the thread split. Minimalism fans cheered the “small core, big power” ethos and compared it to the bootstrapping world of GNU Mes. Philosophers dove into a rabbit hole over the “smallest self-hosting Lisp,” with shout-outs to Paul Graham’s Bel. Meanwhile, skeptics rolled their eyes at “yet another Lisp,” but the Julia reveal blunted the snark. The author’s defense of “proper tail recursion” (a way to make loops efficient without blowing up memory) sparked debates about speed myths. And yes, the jokes flew: riffs on “everyone has their own Lisp (even pets),” “Smol Lisp, Big Mood,” and the eternal meme—“but can it run Doom?” The vibe: messy, nerdy, and wildly entertained.
Key Points
- •femtolisp is a compact, self-contained Scheme-like Lisp (~150 KB) with a bytecode compiler and VM; the compiler is written in femtolisp.
- •It offers extensive features including gensyms, backquote, exceptions, readable circular/shared structures, hash tables, UTF-8 IO/memory streams, and proper tail recursion.
- •The implementation is highly compatible with Scheme and includes some R6RS features, with a simple API and compacting garbage collector.
- •Many primitives are implemented in the language itself, yet the system claims speed among fast non-native-compiled Scheme implementations.
- •Design emphasizes concise, single-place implementations and demonstrates efficient proper tail calls in a pure s-expression interpreter (tiny/interpreter branches).