Gavin Andresen - 2012-09-06 22:12:27

@s{quotedtext} @s{quotedtext} Good point. I assume that miners will choose to mine the version of a transaction with the highest fee-per-kilobyte, since that will give them the best profit, but actually changing the code to implement that policy has been controversial when I've brought it up before.

To fight transaction spam, I think the relaying logic will need to get smarter, too.  A large, expensive-to-verify double-spend should be way down on the "stuff that should be relayed when there is enough bandwidth" list.

@s{quotedtext} @s{quotedtext}
That's a different issue, and a new feature. I think the best way to implement that feature is "child pays for parent" (see https://github.com/bitcoin/bitcoin/pull/1647 for a proposed implementation), and then the user can broadcast a high-fee pay-to-self child transaction to get the parent accepted into a block.