Conversation
compilers are bourgeois, jits and interpreters are proletariat. i will not be taking questions
1
0
0
@fiore @kopper i mean technically so is java just in this case the vm is a stack machine and the bytecode is... peculiar
0
0
1
@coolbean @fiore @kopper the interpreter is the compiler. Can obv also do jit
1
0
1
@snacks @fiore @kopper no

you compile words to threaded code which is then executed by an interpreter which handles entering and exiting words through the return stack
1
0
0
@coolbean @fiore @kopper implementation specific even in ans
2
0
0
@coolbean @fiore @kopper iirc gforth compiles to actual machine code. flk def does
1
0
0
@snacks @fiore @kopper the majority of implementation (direct and indirect threaded code) use this and even subroutine threaded code works similarly but leveraging actual processor instructions in place of the interpreter (at the potential slight expense of memory since subroutine instructions arent as compact as execution tokens)
0
0
0
@snacks @fiore @kopper geforth does not use subroutine threaded code, it does actually have a really convoluted interpreter that uses a mix of both direct and indirect threaded code

this is one of the things i dont like about it and why id like to hack around on pforth instead but to be perfectly honest ocaml or sml might be the better choice in the long run
1
0
1
@coolbean @fiore @kopper yeah, gforth is an abomination, pforths main disadvantage is gonna be that it's slow as balls in exchange for being very hackable and easy to understand.
I welcome any exploration into languages that arenct plagued by trying to be like c and lose the weird shit that comes with it's syntax
1
0
1
@snacks @fiore @kopper have you looked into ocaml or standard ml before? they're like about on equal ground with lisp and forth for me in the realm of "best programming languages"
1
0
1
@coolbean @fiore @kopper i've heard about ocaml before but never looked too deep into it. Always just makes me hear haskell's siren call again
1
0
1
@snacks @fiore @kopper the main thing the ml family has going over haskell is that actually decent compilers exist

i also just happen to prefer the syntax
1
0
0
@coolbean @fiore @kopper wdym decent compiler? Haskell seems to be faster than ocaml, optimizing it just requires arcane knowledge compared to anything else because of immutability and lazy eval
1
0
0
@snacks @fiore @kopper both ocaml and sml can do lazy eval as well

im mostly just throwing shade at ghc though. actual hell software
1
0
1
@coolbean @fiore @kopper haskell devs also have webdev brain with everything having a billion dependencies
1
0
0
@snacks @fiore @kopper im afraid ocaml also has package manageritis. opam is like the thing pushing me away the most. i want to try sml instead because of that. smlnj and mlton just dont have a centralised package manager poisoning the ecosystem
0
0
1