20231223

I am not a burn-the-whole-system-down-and-start-over person.

I am an evenly-survey-the-available-tools-and-build-a-new-space-within-the-system, something-that-can-grow-into-a-rich-environment-of-its-own person.

(Note: The percentages below are incredibly figurative. Also none of this is correct. It might be useful, though.)

1. At the time they're built, systems serve 90% of their population well. 10% are not met well. A "complete" system can be configured -- with extra effort -- to serve the remaining 10%. It won't be awesome, but it'll be enough. That's a complete system. Incomplete systems tend to not get off the ground.

2. The composition of that poorly-served 10% will shift and bubble over time, with a varying set of sub-populations. Over time, one of those groups will catch momentum, and it'll start to push its way beyond that 10% line.

3. That it's-not-great-but-it-mostly-works population will eventually hit some critical threshold. The larger it gets (as a percentage of the whole), the more heat it'll generate as the system becomes less and less suited to the population. (Remember, it was built to serve the original 90% of the population well.) This is the time to start building something new, rooted in the original system, composed of its pieces, but using a composition that serves the new majority.

4. In time, the original system is left to history. The new one becomes self-hosting -- as in, an understanding of the original system is no longer required to operate in the new one. The new system can be defined and understood in its own terms. Goto 1.

It's a really good idea to design any system with this cycle in mind. It's not planned obsolescence; it's basically composting. Create systems that will degrade into useful substrate for what comes next. (Or: Always leave a door open. See also Kuhn Cycle.)