You're not making your case very well, and claiming it to be a matter of education isn't helping. You can't just say someone is wrong without actually articulating why and expect to sound smart. Yes you can attempt to build operating systems in a manner that approaches the way you'd use to build an executable. But you're chasing after a moving target and there's not that much benefit in putting in the work, and the reason why is the exact thesis of the article.
What you lack isn't education, it's experience. Experience to determine what is a good use of time and what doesn't matter. Experience to determine how to approach implementing ideas like reproducibility without shooting yourself in the foot, as the author did. Experience you gain by screwing it up and learning the lesson. It sounds to me like you screw stuff up and don't learn the lesson, leaving you to develop hard-headed idealistic positions like this one that you rely on in lieu of said experience.
> You can't just say someone is wrong without actually articulating
I can, I just did it.
> But you're chasing after a moving target
Please educate yourself. Read about bootstrap, hermeticity and reproducibility. Nix has a byte-to-byte reproducible core and that is very important and valuable.
What you lack isn't education, it's experience. Experience to determine what is a good use of time and what doesn't matter. Experience to determine how to approach implementing ideas like reproducibility without shooting yourself in the foot, as the author did. Experience you gain by screwing it up and learning the lesson. It sounds to me like you screw stuff up and don't learn the lesson, leaving you to develop hard-headed idealistic positions like this one that you rely on in lieu of said experience.