I am a 1MB block supporter who feels 1MB is enough for many years to come. We have a network run on Raspberry Pi's and old laptops from the 90s, We need a stronger network and a stronger world wide web and a stronger Tor network; we need better mempool protocols to allow for rebroadcasting transactions with higher transaction fees and many other features like client side minimum fee requirement prediction but what we do not need is a larger block size. Bitcoin was never meant to be a free system, higher transaction fees are the only way to secure the network from government manipulation and control.
Hardforking now is impossible, consensus is everything and it will not be won by force, A hard fork now would mean fracturing the network into two distinct camps the 1MB faction and the others, it would be war. A war the 1MB camp will win.
There can only be one Bitcoin and one Bitcoin network and the new fork would be seen as nothing more then an alt coin bootstrapping itself with the Bitcoin Blockchain. When in doubt people will side with the 1MB faction and people know it, this is the only thing preventing a hardfork now.
Higher transaction fees mean more incentive for mining pools to process transactions and reinvest profits in the cutting edge hardware needed to stay ahead of the mining difficulty wave. The total cost of all transaction fees is more or less = to the amount of money spent on mining hardware, as that price falls so dose the price needed to manipulate the network. If miners are not constantly reinvesting profits to buy new hardware then governments will gain control simply by waiting for our mining hardware to become antiquated. We should only raise the block size after transaction fees grow to a point where they become prohibitive to the growth of the network that inflection point will be around 1 to 5 USD per transaction. Again Bitcoin was never meant to be free. Freedom is not free and nether is Bitcoin.
When transactions hit the 1MB limit nothing will happen other then through natural selection, the higher transaction fees will start culling out the useless tiny transactions (spam) of the network.
I'm very interesting to see how would a fee market become if the blocks become full at 1MB. It seems many people are afraid of seeing that happen, but what if it makes bitcoin more valuable?
Imagine that the mining fee per block becomes higher than 50 BTC, e.g. 0.0125btc per transaction for 4000 transactions per block at 1MB, and average transaction will be larger than 1 BTC to not feel the pain of fee. That is about 6 dollar per transaction at today's exchange rate, almost at the same level as today's international bank transfer fee, but still lower. So it won't impact two largest user group of bitcoin: Long term value store and international remittance
However, the result of 50 BTC fee per block is: In a 4 years' period, half of the coins will be collected by miners. Currently many people do not trust bitcoin because they doubt that it is a pump and dump scheme: Early adopters sitting on millions of coins, making lots of campaigns to merchants and users in an effort to dump them to late adopters and profit. However, if late adopters can become a miner and process transactions for early adopters and collect equal amount of coins like early adopters, then this monetary system is much better balanced and will not result in inequality long term wise. Because mining is almost free entry for anyone, this will attract more users, more than a low fee will do
In fact, this is already a concern when the next reward halving is coming, means over 75% of the coins have been mined. What would people do when a gold mine is depleted? Satoshi said that in future the block reward will be replaced by fees, but he did not mention how it can be done. If the block is always half-empty, then the fee per block will always be a small fraction of the block subsidy, which shrinks every 4 years