-
there's a lot missing (function invocation/resolution, inline bytecode, specialization etc.) but the basic architecture is in place and working. as of now we're juuuust shy of 1K SLOC (993 SLOC) which is *very short*! stretch goal is to get the whole compiler under 5K SLOC
-
SLOC is kind of a silly thing to stress about/optimize for, but i do think it matters. you can read and internalize a 5K SLOC program in a weekend, but not a 50K SLOC one. and if the compiler is deliberately kept small then more is put into the language itself, which is best IMO
-
porting to (and learning) F# took about two weeks, which is quicker than i expected. the previous C# codebase was due for an overhaul and was already saddled with technical debt from incorrect assumptions and pivots. this codebase seems like it could be a final thing.

