Well, a big problem here might be engineering thinking, as opposed to design thinking. If you view the problem as, "I have to enumerate all the [Linux on Deskt

Well, a big problem here might be engineering thinking, as opposed to design thinking. If you view the problem as, "I have to enumerate all the [Linux on Desktop] problems, and if I enumerate enough problems and fix all the ones I enumerated in the right order, then I've solved [Linux on Desktop]," then "Main Linux problems on the desktop" is, by virtue of having done exactly that, and not being the first or last iteration of that, part of the problem! Not that I begrudge Linus Torvalds for saying things sort of like, Linux is evolved, not designed. It just shows that one of the big holes it has is its use on the desktop, and the major Linux designed product (Android) isn't really libre, and that Linus is a brilliant guy but he doesn't have answers to everything nor claims to. Meanwhile take a look at elementaryOS, which I think has a lot of opinions you'd never find in an evolved or engineering-rational platform (like their own UI programming language). I think if they had the resources of a giant corporation they could make a meaningful impact on the desktop market. I'd say it's very similar to a debate I heard from head of an architecture school attached to a university better known for its engineering. "When we looked at expanding the campus, decisions were made in terms of parking spaces per square foot, and whichever had the most parking spaces per square foot is where we would build the building." It's not that he's going around arguing his ambitious and less efficient designs are better, just because he's coming in with qualitative or emotional impacts incalculable by an enumerative cost-benefit analysis. He's just saying enumerating all the considerations, and then solving, is a really reductive way of thinking about things.