Quote from: sickpig on November 18, 2014, 11:23:53 PM

Quote from: Peter R on November 15, 2014, 08:39:31 PM

Quote from: NewLiberty on November 15, 2014, 07:56:32 PM

I am not a fan of Gavin's current Max Blocksize proposal.

I've been following your scalability proposal in the Technical Discussion section. You want to see the max block size adjusted based on blockchain feedback (closed loop), whereas Gavin proposes an open-loop solution. One risk with Gavin's solution is that it assumes historical growth rates for internet bandwidth will continue moving forward in time. One risk with your proposal is that it introduces another feedback term into what is surely a nonlinear system (right now the only feedback term is Bitcoin's difficulty adjustment).

By the way, has anyone heard from DeathAndTaxes? I'd love to hear his view on both sidechains and scalability.

me too.

I've just checked his profile and his last login was on October 7th, last post on 29 of August...

D&T is struggling with banks with his exchange type company, not surprising that he is pretty busy.

On the MBS proposal topic.

The solutions I suggest for the scalability issue based on the Max_blocksize is more a hybrid of the two.

There are at least two problems with Gavin's proposal that are resolved very simply with mine.
1) What if the open-loop solution is too big and knocks lots of nodes off the network increasing centralization.
2) How do we know when the next time we need some central money planning decision authority to suggest a new target MAX_BLOCKSIZE?

Simply stated: We use a factor of the recent blocksize with adjustment restrictions to prevent burst abuse conditions (TX spam, etc) -the closed loop, and use an upper limit of the exponential solution -the open loop.
When they converge, then we know that it is time for whomever is the new consensus builder to propose something different (or to recognize that there is some other problem needing attention).

What this does (among other things) is it avoids the potential condition where we have solved scalability in other ways (such as side chains or something else) and typical block sizes are still in the 1-2MB range (decades from now) and the network of that time is more or less built around this expectation, lots of hobbyists, and we have very good resilience.
Then suddenly we get a bunch of 1GB blocks (because it is under the limit of the protocol) but folks aren't expecting it and we lose 7/8 of the nodes then running to this bandwidth attack.
Our children will look back at us and think we were idiots for not preventing this sort of thing when we had the chance.

Worse case scenario, it is only as bad as Gavin's proposal.
Best case scenario, we have a more effective backstop to prevent protocol abuse and encourage node running.