| 00:45:27 | d34th_: | d34th_ is now known as d34th |
| 02:27:29 | roidster: | roidster is now known as Guest73433 |
| 02:30:24 | Guest73433: | Guest73433 is now known as zzyzx |
| 07:25:35 | BlueMatt_: | BlueMatt_ is now known as BlueMatt |
| 09:31:26 | petertodd: | gmawell: I suspect that's a general class of problems, all boiling down to the fact that EE's think in terms of "average behavior" and probabilities; a general counter measure might be to use cryptographic cipher-like tools to randomize actual behavior in a way (hopefully) unpredictable to the attacker |
| 09:32:20 | petertodd: | gmaxwell: "scramble" is the term I'm looking for |
| 09:35:12 | gmaxwell: | yea, not just unpredictable to an attacker but to make certian pathological cases unlikely. E.g. self synchronizing line encodings that lose sync if they send too many zeros in a row. |
| 09:35:56 | gmaxwell: | xor with a self synchronizing stream cipher and now a really common and obvious bitpattern (all zeros) becomes just another one in a zillion never going to see it random sequences. |
| 10:54:43 | Pasha: | Pasha is now known as Cory |
| 13:49:18 | gmaxwell: | There is a 'cloud mining' operation advertising on the forum with "No heat and radiation". |
| 13:50:37 | Luke-Jr: | gmaxwell: wow, nice reply on the PoW post |
| 13:58:16 | gmaxwell: | I kinda hit him with a bus there, which I was trying to avoid... Unfortunately there is no nice thing I can point people at "read this background and you'll be less likely to make posts where I have 6 point bulleted arguments about all the ways your proposal can't work as is". I don't want to be discouraging... but I thought being more complete would be more educational. |
| 14:00:19 | sipa: | * sipa slaps gmaxwell a bit around with a large bus |
| 14:00:44 | jgarzik: | haha! |
| 14:00:47 | jgarzik: | here's a fun design: |
| 14:02:06 | jgarzik: | merged mining without the coinbase action. Run a side chain by putting block headers in OP_RETURN transaction. Rest of chain is OOB. Consensus rule: only one block header per bitcoin block is valid, the one with the highest fees. |
| 14:02:35 | jgarzik: | miners have an incentive to self-mine such transactions. |
| 14:02:55 | sipa: | why would you have a consensus rule to limit to one side-chain? |
| 14:03:28 | gmaxwell: | well you'd do one block per sidechain per bitcoin block, I think. No? |
| 14:03:28 | jgarzik: | I wouldn't. I was unclear. only one block header per side chain per bitcoin block must be valid. |
| 14:03:36 | jgarzik: | correct. |
| 14:04:32 | gmaxwell: | Indeed, that sounds fun. Though probably unfortunately prone to thundering hurds.. though I suppose thats part of the fun. |
| 14:05:13 | jgarzik: | thundering herds would quickly burn through miner fees, unless they are themselves a miner. |
| 14:05:29 | jgarzik: | (unless there is another thundering herd to which is being referred) |
| 14:06:36 | jgarzik: | very miner-centric |
| 14:06:54 | gmaxwell: | Thats all I mean, that you might get a bunch of people throwing away fees to try to get a block unsuccessfully. But yes, you probably wouldn't bother unless you were a miner. |
| 14:07:27 | gmaxwell: | If all sidechain blocks were required to make it into the bitcoin chain, however, you couldn't have a sidechain with an interblock gap lower than bitcoin's— which is something that a more loosely coupled design can do fine. |
| 14:08:43 | jgarzik: | agreed. that can be part of the design.. in the BitPay offices, we discussed a chain design with multiple levels of difficulty. lower levels of difficulty between MMORPG clients with sub-second blocks, and higher levels of difficulty for gaming servers that lock in the game timeline every 10 minutes or so. |
| 14:09:08 | jgarzik: | you could apply the same model. transactions in the bitcoin chain are the "higher level difficulty" blocks in that example. |
| 14:09:34 | jgarzik: | higher levels of difficulty can rewrite the chain over lower levels of difficulty, if they choose. |
| 14:10:00 | jgarzik: | (enables gaming to an extent... always trade-offs) |
| 14:10:05 | jgarzik: | no pun intended |
| 14:10:55 | jgarzik: | Anyway, it would be fun to do a merge-mined namecoin side chain in this design, maybe. That works just fine with 10 minute blocks IMO, so no need for complications. |
| 14:59:38 | fanquake: | fanquake has left #bitcoin-wizards |
| 19:56:29 | copumpkin: | copumpkin is now known as puritypunisher |
| 20:01:10 | nsh: | nsh has left #bitcoin-wizards |
| 20:02:55 | puritypunisher: | puritypunisher is now known as copumpkin |
| 23:05:44 | [Tristan]: | [Tristan] is now known as Guest9070 |