1. 32
  1.  

  2. 4

    I really like how this both acknowledges the incredible work Egorov did while also justifying why it doesn’t invalidate what you did, and vice versa.

    The opportunities for algorithmic improvements that he noticed helped speed up the Rust and WebAssembly implementation another 3x over what we had previously seen.

    This is also really nice and something a lot of people don’t often take into account: it’s not a choice of optimize or switch tech. You can do both.

    1. 3

      I really like how this both acknowledges the incredible work Egorov did while also justifying why it doesn’t invalidate what you did, and vice versa.

      Yes. Exemplary and polite way to discuss language comparisons. I bookmarked it for that as much as the Javascript profiling advice.

    2. 1

      I think there is a place for “struct for JS”, probably a DSL that transpiles to JS using typed array. I think LLJS was one such effort? I am not sure why it failed.

      1. 2

        I’m not certain of this but I think my recollection is that LLJS predated typed arrays (or at least widespread support for them?) and maybe a few other of the building blocks it would’ve needed to be really good.

        It helped inspire wasm to at least some extent so it didn’t fail in my estimation.