I don’t disagree that’s true, but I think partly that’s because we have historically had close to zero idea what we’re doing. We’re seeing languages with better design successfully supplanting dominant languages in their space (eg Rust/C, TS/JS)
Comments
Log in with your Bluesky account to leave a comment
TS is an interesting example because all the people coming from C++/Java/C#/... complaining about JS not having types would never suggest a type system like the one TS adopted (gradual). They couldn't see the value of dynamic type checking and how to impose types w/o restricting it too much.
Yeah. I mean I am sympathetic. I didn't know _much_ of what I was doing getting started with Rust either. And most PLs predate even what little consolidated teachable knowledge I had access to! The history is one of people stumbling into the task and piecing something together that kinda works ish.
Insofar as Rust seems better or "more" intentionally designed, I think this is a consequence of (a) hiring people with more training after it got going and (b) having tons of well funded lead time to iterate / explore the user-needs surface area while nobody made very real demands of it.
(And I'd also debate the idea that there's even _that_ big a set of consolidated and widely-agreed-upon knowledge to learn. A lot of PL knowledge is still quite disputed, fragmented, niche or under-explored. The focus lately is on trying to consolidate type system knowledge. Good but insufficient.)
Comments