I like the distributed farmers idea. There would be lots of gameplay ramifications for eco efficiency and vulnerability to raids.
One outstanding point is the slot ("trees" as mentioned by @DesertRose) priority per farmers. Instead of some script that would place each additional farmer at the best possible slot depending on dropsite and farm orientation, slot prioritization should be done based off the path the farmer takes to access the farm when tasked to farm on it (whichever slot is closest).
I have a sanction idea that satisfies both sides of the argument but it is a bit difficult to implement technically:
Every time JC says a racist comment, his units increase in cost by 20%.
When he insults another player, his units loose 20% of attack damage.
When he trolls, his units loose 10% max HP.
The penalty changes depending on the severity of the offence.
That's not necessarily true. It's definitely a discussion we can have.
Your other points are 100% fair. My main reasons for starting DE.
Collaboration, especially online collaboration, can be very frustrating at times. But what you can do (which you've already done and which I've done) is use your mod for experimentation, etc., then parse out bits and pieces you think would serve well the base game and make PRs.
It's a bit more complicated then that however. It's not just about good will.
There are things that I can do in a mod that won't be accepted in vanilla, like dropping support for minuscule screens.
The time spent on developing anything for vanilla is of a large order of magnitude (often even two) longer then for a mod, due to requirements, process, need of communications..
Reviewer availability is scarce and mismatch with your own motivation and availability creates a even bigger gap between making something in a mod and for vanilla.
Some ideas are very often initiated in the mod as a totally buggy POC, and refined over time since you can immediately play test it (or dropped). Which is something you cannot really do with vanilla.