On Sun, May 07, 2023 at 11:31:04AM +0200, Страхиња Радић wrote: > but the arguments presented in it leave me unconvinced.
The "maneuverings" argument in specific was entirely misdiagnosed. Even in his own example (redhat trying to remove /etc) you can clearly see it has nothing to do with GPL and everything to do with _adoption_. Let's imagine a world where the linux ecosystem (including the GNU tools) were permissively licensed. And now redhat wants to remove `/etc`. Of course, if redhat was the only distro that did this, most applications wouldn't adopt their policy - leaving them with an insane amount of packages which will be broken on their distro - making this plan infeasible. According to that article's logic: This kind of "hijacking" and political maneuverings never happen with a permissive license like the BSD or MIT licenses. in this imaginary universe, redhat would just go, "Hmm, we didn't get what we wanted, but since linux is BSD licensed, we'll just give up." Does that sound realistic? Of course not. They would've done exactly what they're doing right now - trying to push their policies unto every (or at least most) distros via the virus that is systemd so that their policies gain _adoption_ among applications/libraries. Linux being GPL/BSD makes zero difference when you need adoption from a massive amount of third party applications. If you want a real world example, just look at wayland-protocol, which is MIT licensed. According to that articles logic, Valve can just fork it and add their custom-protocol and implement it in their compositor. But if Valve's compositor is the only compositor that implement's a protocol, then most applications won't follow/adopt it and thus the protocol will be useless. So the reality of the situation is that Valve is still "maneuvering" and trying to get what they want into the upstream wayland-protocol so that it gains adoption. The MIT license made zero difference. - NRK