Quote from: gmaxwell on July 06, 2015, 06:49:46 PM

There is no requirement that mempools be in sync, -- in fact, they're not and the whole purpose of the blockchain is to synchronize nodes. The mempools of nodes with identical fee and filtering policies and whom are similarly positioned on the network will be similar, but any change in their policies will make them quite different.

Clean and synched mempools makes for a cleaner blockchain, else garbage in - garbage out. Most mempools are synched because node owners don't usually mess with tx policy. They accept the defaults. Pools like Eligius with very different policies are the outliers. IBLT will help by incentivising node owners to converge to the same policies.

Quote from: gmaxwell on July 06, 2015, 06:49:46 PM

IBLT doesn't currently exist, and other mechenisms like the relay network protocol don't care about mempool synchronization levels.

IBLT does exist as it has been prototyped by Kalle and Rusty. It is just nowhere near ready for a pull request. Since this block propagation efficiency was identified there could have been a lot of work done in Core Dev to advance it further (though I fully accept that other major advances like headers-first were in train and draw down finite resources). I recall that you had a tepid response summarizing the benefit of IBLT as a x2 improvement. Of course this is hugely dismissive because it ignores a very important factor in scaling systems: required information density per unit time. Blocks having to carry all the data in 1 second which earlier took 600 seconds is a bottleneck in the critical path.
It is the LN which doesn't exist yet and will arrive far too late to help with scaling when blocks are (nearer to) averaging 1MB.

Quote from: gmaxwell on July 06, 2015, 06:49:46 PM

the 1MB was either forward-looking, set too high, or only concenred about the peak (and assuming the average would be much lower) ... or a mixture of these cases.

So, in 2010 Satoshi was forward looking, when the 1MB was several orders of magnitude larger than block sizes.. Yet today we are no longer forward-looking or care about peak volumes, and get ready to fiddle while Rome burns. The 1MB is proving a magnet for spammers as every day the average block size creeps up and makes their job easier. A lot of people have vested interest in seeing Bitcoin crippled. We should not provide them an ever-widening attack vector.

Quote from: gmaxwell on July 06, 2015, 09:38:13 PM

To further make the point about mempools, here is what the mempool looks like on a node with mintxfee=0.0005 / minrelaytxfee=0.0005 set:

$ ~/bitcoin/src/bitcoin-cli getmempoolinfo { "size" : 301, "bytes" : 271464 }

That min fee at 0.0005 is 14 cents, and most users consider this to be way too high, especially if BTC goes back to $1000 and this becomes 50 cents. I kicked off a poll about tx fees and 55% of users don't want to pay more than 1 cent, 80% of users think 5 cents or less is enough of a fee.
https://bitcointalk.org/index.php?topic=827209.0

Maybe this is likely naive and unrealistic long-term, and a viable fees market (once the reward is lower) could push this up a little. Or is this another case where the majority of users are wrong yet again?

Peter made the point that Bitcoin is at a convergence of numerous disciplines, of which no-one is an expert in all. I suggest that while your technical knowledge is absolutely phenomenal, your grasp of the economic incentives in the global marketplace is much weaker.
While Cypherdoc might have had errors in judgment in the Hashfast matter (I know zero about this, and have zero interest in it), his knowledge of the financial marketplace is also phenomenal, and he correctly assesses how Bitcoin can be an economic force for good, empowering people trapped in dysfunctional 3rd world economies. He is right how Bitcoin has to scale and cheaply for users to maintain a virtuous feedback cycle of ecosystem growth, hashing power growth and SoV.
Lots of people will not pay fees of 14c per tx when cheaper alternatives like LTC are out there. I see the recent spike in it (disclaimer: I don't have any) as the market "pricing in" that BTC tx throughput is going to be artificially capped. Whle BTC tx throughput will always be capped by technology, we should not be capping it at some lower level in the misguided belief that this "helps".