Comments by "Liam" (@1495978707) on "Brodie Robertson" channel.

  1. Single click is superior, you’ll never convince me otherwise. Opening files, renaming them, etc are more common than selecting only one item. If you’re selecting multiple files, there’s no difference between single and double click. The only time it matters is if you want to select a single file to copy/cut/paste, in which case, the extra press of ctrl is well worth the time save overall
    5
  2. Wtf do you need to only select a file for? If you want to select one file, same thing as selecting multiple files, either control+click or click and drag
    4
  3. Well this is one of the reasons that you use physical security at the end of the day. If you don’t want your computer to be reporting remotely, make it physically incapable of talking to a network. Then it can try to call home all it likes, but no remote access can be had
    4
  4. They count on it. They count on people just taking the abuse and groaning because it's just the way it is
    2
  5. Memory safety has nothing to do with leaks (forgetting to free memory), it has to do with accessing memory that you shouldn't be able to. Rust is memory safe because you have to be very explicit about how you interact with data, and it puts heavy restriction on how this can be done. To stop leaks, every part of the stack needs to remember to free memory, and people can fuck up at any level. This is why so many people just say screw it and take responsibility for their memory handling in C, even though it's harder. Its different when bugs happen because of you vs a giant team of people that you trusted
    2
  6. The best argument against Linux for this purpose is that it has a monolithic kernel, so if a driver craps out, it can bring down the entire system. But that’s true for windows and Mac too lol. Instead, GNU tools made for a microkernel is what we should work towards, and the open source community seems to have growing demand and interest for microkernels
    1
  7. The beautiful thing about linux is that it's out there, exists, and will forever, for anyone to freely use. And the really hard work has been done already to make usable operating systems out of it, by many independent teams. So even if every distro got shutdown somehow, or all the maintainers left or whatever other catastrophic scenario you can imagine, people can just pick up where they left off. When a proprietary organization goes under, the software dies with it. When they decide to stop supporting software, they can make it impossible to use or obtain, and if anyone wants to compete, they have to start from scratch. But that's not necessary with operating systems because there's so many open source ones
    1
  8. ​@gratuxAutomatic leak prevention basically requires deleting everything created within a scope when the scope ends. But what if you need to create things in a nested smaller scope and use them in the parent scope? Or another smaller scope? This kind of thing is why whole cloth OOP and factory type programming is so difficult to do right, it's full of cases where you allocate in one scope in order to construct an object, and then pass to another scope where more might be added, and so on. It may be very general and abstract, but it makes it mindnumbingly difficult to understand the code and find memory leaks, and it requires bypassing automatic protections that delete everything when a scope ends.
    1