Again, I'm for an increase to the limit, and I think miners acting rationally would be too. They would do it because of their own self interest though, not because a poll of reddit users wanted them to.
Are you sure miners acting rationally wouldn't consider the block limit as 1 dimension of the potential way to increase transaction capacity? The real question isn't "how/when/to what do we increase the block size", it's "how do we increase transaction capacity in a way that maintains a healthy market for hashes". Anything that centralises towards the miners, or amongst the miners, is bad for the eco system as a whole. BIP 101 would create a trend toward both.
Miners create the blocks, so I'm not sure what decentralizing away from miners would even mean. If you mean that smaller miners are dependent on, or even supported by, 1MB blocks, I disagree. It may even be the opposite, smaller mining operations may have access to better bandwidth and speed vs remote industrial mining farms in rural china.
Economies of scale encourage, especially since the arrival of ASICs, the relative centralization of mining. This was not a surprise and was predicted from the very beginning. IMO maxblocksize has almost nothing to do with it.
What a great point.
The 5 largest miners in the world already told us they're trapped behinf the GFC with inferior connectivity speeds. And yes, we have to assume the relay network is factored into that statement.
It may turn out that a small miner could do the large block better connectivity attack Wiulle was going on about on reddit because of a direct connection to the relay network that causes the Chinese miners to choke.
