--- Log opened Wed Nov 06 00:00:26 2013
08:25 < adam3us> can you put a locktime inside a script address? or is the locktime an attribute of the transaction only
08:26 < sipa> it's a transaction attribute
08:26 < adam3us> sipa: eg can i pay you, in such a way that I cant later (after confirmation) spend the inputs; and you can only onwards spend after  locktime
08:27 < sipa> yes, that's what locktime does
08:28 < sipa> it prevents putting the transaction in a block until a certain timestamp or block height has passed
08:28 < adam3us> sipa: well the first transaction must be confirmed, but i suppose eg the script requires two signatures, and i give the recipient one signature that is on a transaction serlaiziation with a  locktime
08:28 < adam3us> sipa: i guess that is how you would do it
09:58 < amiller> i'm starting to feel more concerned about the selfish mining attack
09:58 < amiller> first of all i think the thing to focus on is th 1/3 limit
09:58 < amiller> since everything else really gives the attacker too much control over latenc
09:59 < amiller> as an impossibility result you'd want to say that we're screwed even with a weak adversary
09:59 < amiller> so here's the thing, previously we thought there was no direct way to profit disproportionately even with a 51% attack
10:00 < amiller> the thing is, this closes the gap between 1/3 and 1/2 because once you hit 1/3, you can gain disproportionate profit
10:00 < amiller> you don't even have to commit fully to the selfish mining!
10:00 < amiller> if you are above 1/3, you can withold blocks for a little bit, some of the time, and still get the benefit
11:00 < petertodd> amiller: announce/commit sacrifices to fees are an excellent example where withholding blocks for just a little bit can be worth it
11:01 < petertodd> amiller: basically, the sacrifice unlocks at some block height, so anything you do to keep the rest of the network at least a block behind can be very profitable
12:54 < amiller> it's hard for me to understand how trickle could actually slow tx propagation much
12:54 < amiller> every inv has a 1/4 chance of getting sent, every 100 ms or so
12:55 < amiller> since that's how often each peer is processed
12:55 < amiller> the chance of a tx not getting sent after a couple seconds is really low
--- Log closed Thu Nov 07 00:00:33 2013