Quote from: Peter R on August 14, 2015, 04:19:02 PM

Quote from: NewLiberty on August 13, 2015, 08:36:02 PM

Quote from: Peter R on August 13, 2015, 03:17:41 PM

Quote from: Zangelbert Bingledack on August 13, 2015, 03:02:26 PM

At this point I'd say just find a way to put the forks on the market and let's arbitrage it out. I will submit if a fork cannot gain the market cap advantage, and I suspect the small-blockers will likewise if Core loses it. Money talks.

I had a strange idea recently: what if we don't even bother with BIP100, BIP101, etc., or trying to come to "consensus" in some formal way. What if, instead, we just make it very easy for node operators to adjust their block size limit. Imagine a drop down menu where you can select "1 MB, 2 MB, 4 MB, 8 MB, ." What would happen?

Personally, I'd just select some big block size limit, like 32 MB. This way, I'd be guaranteed to follow the longest proof of work chain, regardless of what the effective block size limit becomes. I'd expect many people to do the same thing. Eventually, it becomes obvious that the economic majority is supporting a larger limit, and a brave miner publishes a block that is 1.1 MB is size. We all witness that indeed that block got included into the longest proof of work chain, and then suddenly all miners are confident producing 1.1 MB blocks. Thus, the effective block size limit slowly creeps upwards, as this process is repeated over and over as demand for block space grows.

TL/DR: maybe we don't need a strict definition for the max block size limit.

You know that you can do this now, right? And always could.

The code is open source, you can (of course) just change it and compile.

We were discussing writing a BIP for the idea where the block size limit is removed from the consensus layer, and each client manually selects the max block size that they are willing to accept. If this idea truly makes sense and can gain some mindshare, then perhaps to go along with a BIP, we could fork Bitcoin Core to create a "reference implementation" for the new idea? Like Awemany said, it wouldn't require much coding.

I'd probably run it too, because by setting my personal block size limit high, I would always get dragged along with the longest proof-of-work chain. Who knows, it could actually catch on more widely too

We'd also need a name for this implementation.

You may also need to add a mechanism to alert the user when the network does fork away from their limit.