yorickpeterse's recent activity

  1. Comment on Asynchronous IO: the next billion-dollar mistake? in ~comp

    yorickpeterse
    Link Parent
    It's true that time and effort has been spent in optimizing threads, and they're certainly in a better state than they were say 20 years ago. But my very thought is "what if we'd done a better...

    It's true that time and effort has been spent in optimizing threads, and they're certainly in a better state than they were say 20 years ago. But my very thought is "what if we'd done a better job", followed by "what if we can/could do better?".

    That's the point ultimately: we are where we are today, but I'm curious where we could've been if we took a different path that focused on optimizing threads further, rather than take the path that involves asynchronous IO.

  3. Comment on What programming/technical projects have you been working on? in ~comp

    yorickpeterse
    Link Parent
    You're right, that should be "data races", so I'll fix that :)

    You're right, that should be "data races", so I'll fix that :)

    1 vote

  4. Comment on What programming/technical projects have you been working on? in ~comp

    yorickpeterse
    Link
    For Inko I started working on type specialization/monomorphization, albeit slightly differently: instead of specializing across types, I'll be specializing over ownership/value types, sometimes...

    For Inko I started working on type specialization/monomorphization, albeit slightly differently: instead of specializing across types, I'll be specializing over ownership/value types, sometimes called "kinds" or "shapes". Basically if a value is passed around differently on the stack or has different logic surrounding creating/disposing of aliases, that gets specialized over.

    The reason for this is that most objects are addressed through pointers, and this will remain the case due to the type/ownership system requiring this. In such a setup, specializing using types makes less sense, as you'd end up with e.g. 15 Array specializations that all use the same memory layout, and possibly the same machine code (or at least very similar code). Some more details are found here.

    Now the tricky thing is that I have a rough idea of how to implement this, but I'm at this stage where I haven't fully convinced myself of that yet, so I'm going around in circles a bit. This usually takes a week or so to settle in, but it's a frustrating experience during that time, as it feels I'm not making any progress.

    4 votes

  5. Comment on Octopuses sleep—and possibly dream—just like humans in ~science

    yorickpeterse
    Link Parent
    I believe this has been experimented with, and IIRC the conclusion was that if an octopus doesn't mate, they generally do live a bit longer and just go about their business, but the difference...

    I believe this has been experimented with, and IIRC the conclusion was that if an octopus doesn't mate, they generally do live a bit longer and just go about their business, but the difference isn't that big.