The 1MB blocks size is not a mess. If it was not the 1MB block size it would be something else later on.
This dispute is the natural result and challenge inherent in government by consensus. There is a reason that no human society has used government by consensus in the past. Its far easier to simply use force to subdue your opponents or get 51% of the voters to vote you power over your opponents.
Personally I would supports SegWit + 2Mb block fork provided the code was vetted and tested by the Bitcoin Core team and was supported by 95% of the available hashing power and all the major exchanges.
I believe Bitcoin Core's vision of scaling is ultimately the correct one but no one has convinced me that we cannot afford to throw the big block folks a bone in order to keep things moving along smoothly.
But I am not a technical expert and if someone can present a compelling case why SegWit + 2Mb would significantly damage the network or it becomes clear that it is impossible for SegWit + 2Mb to ever achieve consensus I would change my mind.
I still can't reconcile this ridiculous notion that everyone is arguing over 'SegWit with 1MB stab in the dark' or 'SegWit with 2MB stab in the dark' when both options are shortsighted, involve too much rigidity and are
guaranteed to require further hardforks in future. Why not just do
one hardfork allowing for a slow and gentle ease of pressure?
Both camps have clearly already decided on central planning to dictate or coerce whether it's going to be on-chain or off-chain scaling and we seemingly have a binary choice between the two stupid extremes. Neither wants to let the market choose freely and decide for itself how best to grow. I'd argue that both sides are spineless cowards in this regard.
Some parties that are have argued that ever increasing block sizes will lead to progressive centralization over time. Bigger block folks disagree.
Regardless of the underlying truth no matter what is decided we will face hard forks in the future. For bitcoin to grow
there are certain to be multiple necessary upgrades along the way some of which will require hard forks.
There is no avoiding this challenge by forcing though single automatic scaling algorithm and hoping for the best.
The fact that their is not widespread consensus currently for on-chain or off-chain scaling is yet another reason to support a compromise position along the lines of SegWit + 2Mb. It give both sides time to show what their solution offers without fundamentally committing us to either.