Flow’s OCaml to Rust Port

8 points by darichey


nemin

TypeScript ending up dominating the web is such a tragic catch-22.

On one hand, it makes sense for the lingua franca to be a conservative superset of JS, that doesn't really impose its own will on how code on the web should look like. It'd be the best if JS itself was endowed with various improvements instead and one wouldn't need to rely on 3rd party tools to provide basic stuff like types.

On the other hand, because TS dominates none of these other tools really ended up gaining any space and JS itself is moving at a snail's pace[0], leading to a really stagnant-feeling core that people try to desperately plug using an ecosystem that's held together by prayers.

Sloppification aside, Flow for instance seems to have some really damn handy features from its much better support for React, to the match expression (god, how many times I wished JS had that...), to nominal typing, etc. But outside Meta it's barely there and then there's all the other stuff people may not even have heard about, ReScript, Melange, Reason, etc.

And that ends up as a vicious cycle because, even if they have better ideas and more guarantees, why would you invest into these "lesser" languages, if TS is the gold standard? Which only leads to TS becoming even more disproportionately used and so on.

[0]: Not that this is necessarily a bad thing, rather this than another Date class situation, but it really feels like we haven't had any massive improvements since ES6...