I fully agree with the purpose of Bitcoin, it's gold that you can cypher, backup, teleport for a network fee that will remain way much lower than transfering physical gold.
It's also eaisier to store and protect.
I agree that having a self correcting guess that auto adjust to follow parameters is switching the issue and risky.
So what, we chose an arbitrary size ? (let's suck my finger and show the sky, for me it's a fixed 5mb)
We chose a block filling treshold, and when blocks are full at x% for x time, we go to the agreed size ?
Because I agree with RoadStress when he say we should think of it now (but we (readers here) are)
But I agree with Davout that say you don't fix something not broken (I would just prefer to fix it when orange light turns on)
We all wish for this, but alas you cannot build the pyramids like this. The very building of them is going to decide many things in your civilization, but you cannot argue about your royal cubit throughout the process. You might use SOME past data, but eventually because of the nature and difficulty (and uncertainty) of future
consensus the tradeoff is exactly what you brush off. What is interesting and perhaps even useful is that the proposed 20mb is nearly the same size as the royal cubit in inches. I don't think this is unrelated in fact, but it would take me a long time to show, but it could allow it to become the "popular" choice.
But we need to be clear on this exact difficultly, that seemingly it is a fallacy to suggest we can deal with this in the future and/or extrapolate future data to make a better decision, OR set up a market adjusting parameter. We don't have these possibilities, and THAT is why this dialogue has arisen.
It's go time. (not fixing it now risks the possibility of a 1mb size forever)