I think the situation is more nuanced than that.

Of course, the F-35 program was an incredibly expensive mess (litterally the most expensive weapon program of all time), because of conflicting specs, data leaks, political infighting, cost overruns which are the stuff of legend, etc… At some moments, there were certainly reasons to think the whole program would collapse on itself like wet tissue paper.

But there are operational F35 now. 900+ as of 2023, which is 4 time more than the rest of Gen 5 fighters combined. And performance-wise, it is good, especially on the stealth & avionics parts. On the other side, the J-20 is largely unproven (probably a decent design, but not as good), and the Su-57 is a bunch of glorified prototypes.

Now sure, cost is high, maintenance is time-consuming, availability somewhat below target, but it's not particularly surprising for high-performance equipment. It may fall short of the ambition of the program on the cost part, but by itself it's a dangerous and fully operational fighter.

While I respect the degree of technical expertise you can get on SO, I am very much convinced this is a failed experiment which will slowly collapse on itself. And I think that because the organization of the site is completely at odds with it's professed goal.

SO is not really a forum, because time is supposed to be irrelevant : a 12 years old answer takes precedence over a question asked now, failing to consider that, maybe, the context of this question has changed. Its ridiculous, especially in the context of software engineering.

On the other hand, it's not a library either : a library is a collection of book, which are relatively self-contained knowledge systems. And in practice, the SO answers are not self-contained. They merely answer a specific point, with no guaranteed coherence from one to another, so a beginner cannot use them to build a broad understanding of a subject.

To take the analogy of the Cathedral and the Bazar, I am under the impression that SO members are trying to build a cathedral out of the stone sold by hundred of bazaar people, and refuse to see that fact that all stones all having different sizes and dimensions is maybe kind of a problem when you build a cathedral.

I think a lot of people are misunderstanding what Proton actually brought to Linux gaming.

I had been running Linux exclusively for some moths in 2013-2014, and trying to get games to work on Linux felt like this : Wine is likely able to run it if you can found the right configuration, but good luck with that. I think the only game I managed to run without issues was Civilization 4, so I rolled back on Windows some time later.

Of course, Valve contributed to Wine, and projects like dxvk and others are major achievements (if a team effort), but that’s not their main contribution. Valve understood that gamers may be somewhat more tech-litterate than other people, but that making games work on Linux should be easy. And that’s what Proton was made for.

Nowadays, most games I buy on Steam work out of the box. I sometimes forget to check protondb before buying a games, and I rarely had an issue. Even if in 2018 you had to tinker a bit, you rarely needed more than to choose the correct Proton version (big up to Glorious Eggroll).

I think it’s symptomatic of the situation of the Linux Desktop : technically, it’s where it needs to be. But there is still a gap in accessibility and easiness. Tinkering is nice, but you should not have to do it to have something that works.