Huang: Rust: A Critical Retrospective
Huang: Rust: A Critical Retrospective
Posted May 24, 2022 19:05 UTC (Tue) by khim (subscriber, #9252)In reply to: Huang: Rust: A Critical Retrospective by jmalcolm
Parent article: Huang: Rust: A Critical Retrospective
It's not that interesting, really. And while I wish them luck this would mean that this OS would forever stay a love letter to '90s user interfaces with a custom Unix-like core and wouldn't turn into the next Linux.
I understand the temptation to create a specialized language for an OS development, this may even be right thing to do for the academic project, but as fate of Oberon, Singularity and many other OSes shows this means you OS would be a purely academic/hobbyist world experience. Simply because you wouldn't have a robust compiler for everything-and-anything (which GCC/LLVM provide and which Rust, with it's popularity, may, theoretically, support).
Also: while Jakt certainly sounds interesting it looks like it would try to be another Swift-like language (kinda-sorta-safe-but-not-really) and not a replacement for Rust.
Simply from their small blog post where they boldly proclaim that they would have OOP and classes it's easy to see that they have almost given up to the templation of supporting implementation inheritance which means they don't understand what correctness even mean.
Rust fight against it tooth-and-nail because, simply put, implementation inheritance is SUID-level mistake: easy hack to solve some simple problems in simple cases which leads to horrible mess later.
Linus was always adamantly against such things and against opaque void-pointers everywhere… and for the same reason Rust abhors implementation inheritance.
It would be interesting to see what other errors would Jakt catch besides memory errors (Rust uses it's ownership-and-borrow approach to do many things safely, e.g. scoped threads).
I suspect it would be yet-another-failed-attempt-to-sway-C++-programmers. Unsuccessful because it would be too little return for too much pain.
