Quote from: kano on March 03, 2015, 12:17:41 AM

The bitcoin developers are effectively against p2pool since they are against micropayments.

Yes, I'm cherry picking the quote to respond to, and the rest I will get back with you on later, but this statement simply is not fair or true.

The smallest paid share in the p2pool share chain is currently (right now) 0.00108623 BTC ($0.27).

This hardly represents what most would consider a micro-payment, its over a quarter USD (not a couple pennies, or fractions there of).

Here is an (older) quote from a dev:

Quote

I think P2Pool makes mining fun again. -gmaxwell, Bitcoin Core Developer

And I would agree, it does.

TL;DR: Core dev's have nothing against, and in fact support decentralization and trust-less mining pools...

And my thought, is that serial micro-payments will actually reduce smaller payouts....

Lets say the smallest miner on the pool should be able to earn a single bit per block (0.00000100, 100 Satoshi))

The way I see them working:

Let say a miner starts on the pool with 1 TH/s.

Lets further assume that the payout is around .001 BTC per block found (for the sake of argument).

This miner opts-in to the serial micro-payment structure by setting up his multi-sig, a 2 of 2, 1 key signed by the pool on block generation, and 1 signed by him to spend.

If the miner does not spend the previous payouts they are included in the next block the pool finds, along with the .001 from the current block, effectively now a .002 payout (creating a potential double spend of an unspent multi-sig tx).

However since the first found blocks coins never got the second sig, the coins are not spent and are still available in a trust-less way.

This miner goes on to have a payout of .001 per block found, without ever signing the 2nd key for the payout, that only the miner has, and the payouts accrue over time, lets say for 10 blocks...

On the 10th block found the miner decides that .01 (10*.001) is enough to justify a payout and signs the 10th round of the .001 payout block transactions (that accrue), resulting in a payout of .01 (instead of 10 payouts of .001) to the address they specified when they set up the multi-sig and the accrued txs are all signed at once reducing tx fees.

This is grossly simplified, and I'm still wrapping my head around how to explain it, but I feel like I'm onto something.

If anyone gets it and thinks it might work, or has other holes to poke please do....

Edit: watch the video to get a feel for how serial multi-sig works:

Quote

If you watch, in his example, replace Alice with P2Pool and Bob with a miner...

https://www.youtube.com/watch?v=t3hJsFpPmXs&feature=youtu.be&t=34m30s