OK. Your idea is fine. It would solve some problems. It would create others. Maybe the net result would be better, maybe not.
In any case, it's not likely to happen without a herculean effort between a massive number of FOSS communities.
If you do decide to approach actual distro devs with this idea, expect it to be held to a critical examination far more strenuous than you've got going here. You'd be asking people to accept many constraints, take on a lot of extra work, and throw out a lot of things that work perfectly fine 90% of the time for a few dubious benefits; they aren't going to do that uncritically.