4 ms·
> If you're representing Wayland, this problem is your problem whether you want it or not. You're absolutely, completely, 100% correct. In this scenario it is
by Kalium 8y ago
> If you're representing Wayland, this problem is your problem whether you want it or not.
You're absolutely, completely, 100% correct. In this scenario it is the Wayland developer's problem - wanted or otherwise - that the user is incredibly frustrated. Like most frustrated people, the user chooses to blame what seems like a reasonable target who coincidentally isn't them. Such as Wayland!
This anger and frustration is something that Wayland devs unquestionably have to deal with. It's just perhaps possible that the choices made by NVidia driving this completely reasonable frustration might not be something Wayland devs have to own.
Again, you're completely right. Unhappy users are going to blame Wayland devs. There just might be some room for subtlety.
- eschaton 8y agoIt’s not just choices made by Nvidia, though, it’s equally due to choices made by Linux. Linux is trying to force its unique “all drivers are open source and in our tree” view of the world on everyone. Linux could stop doing this and start guaranteeing stable interfaces. If stable interfaces are available and Nvidia doesn’t use them, then it might be reasonable to call them “a bad actor.”
- chousuke 8y agoThat's disingenuous. Maintaining a stable driver API is far from free, and the project has done just fine without one since its inception. Nvidia is not playing by the rules, and there's an alternative vendor. Why spend extra effort just to help them? Assuming that a stable kernel API would just make things better ignores the potential downsides. I don't recall the name of the talk, but someone did comparative research into BSD kernel security and found that backwards compatibility was a significant source of bugs and that OpenBSD has benefited greatly from their aggressive removal of such code. There's no reason to doubt that this holds true for Linux as well