1 2013-05-05 00:36:59 <cjd> hey guys, syncing up 8.x and I noticed a lot of "ProcessBlock: ORPHAN BLOCK, prev=0000" in the log. I would have thought such old orphans would be gone and forgotten.
  2 2013-05-05 00:37:30 <sipa> they're not orphans, just blocks downloaded in non-sequential order
  3 2013-05-05 00:37:37 <cjd> ahh gotchya
  4 2013-05-05 00:37:39 <cjd> thx
  5 2013-05-05 00:38:20 <sipa> note that the code uses orphan in the correct meaning of "without parent", and not the god-knows-why meaning of "block that is not part of the main chain"
  6 2013-05-05 00:39:19 <cjd> yeah, it makes perfect sense now. BTW: great job on the fast syncing
  7 2013-05-05 00:40:05 <The_Fly> http://s3.amazonaws.com/rapgenius/please-sir-can-i-have-some-moar.jpg
  8 2013-05-05 00:41:54 <sipa> yes
  9 2013-05-05 00:41:58 <sipa> :p
 10 2013-05-05 00:46:41 <The_Fly> woot, happy orphan :)
 11 2013-05-05 00:50:39 <cjd> so how bout that BIP for sending a client to an http URL to get 2 weeks of blocks? :)
 12 2013-05-05 00:51:06 <cjd> because right now my bottleneck is downloading
 13 2013-05-05 00:53:13 <sipa> there's a bootstrap.dat torrent
 14 2013-05-05 00:54:39 <amiller> sipa, what's the short hand for block-that-is-not-in-the-main-chain?
 15 2013-05-05 00:54:49 <The_Fly> cjd: and you could pass blocks/chainstate around your network pretty fast
 16 2013-05-05 00:54:51 <The_Fly> couldn't he?
 17 2013-05-05 00:55:30 <sipa> amiller: extinct block, disconnected block, stale block, side block, ...
 18 2013-05-05 00:56:39 <amiller> side block is maybe my favourite of those...
 19 2013-05-05 00:56:43 <cjd> The_Fly: just by watching my debug.log, it appears that network latency, out-of-order and other little issues are the main cause of slowness
 20 2013-05-05 00:57:13 <sipa> cjd: for initial sync, it's best to just connect to a single node (-connect=IP)
 21 2013-05-05 00:57:18 <sipa> and set dbcache high
 22 2013-05-05 00:57:45 <cjd> ahh right, db flushes
 23 2013-05-05 00:57:58 <The_Fly> torrenting the data in the client is an interesting idea
 24 2013-05-05 00:58:20 <The_Fly> but for the lazy
 25 2013-05-05 00:58:39 <The_Fly> or maybe better referred to as "efficient"
 26 2013-05-05 00:59:03 <cjd> well, I imagine eventually we will prune everything at the highest lockin and the amount to download will be less
 27 2013-05-05 01:00:13 <The_Fly> connecting to a single client sounds good, but user has to figure out the most local (if they can)
 28 2013-05-05 01:00:27 <cjd> could probably ship an exe with the "state" at the highest lockin so it only has to d/l blocks past that, unless you have an interest in historical data
 29 2013-05-05 01:00:38 <The_Fly> true
 30 2013-05-05 01:00:50 <sipa> cjd: means you have to trust the software distributor with data that cannot be validated
 31 2013-05-05 01:01:27 <cjd> well... you trust the distributor when you run the code
 32 2013-05-05 01:01:49 <sipa> well, it's open source and built deterministically
 33 2013-05-05 01:01:54 <shesek> are there any estimates anywhere on the number of machines running miners?
 34 2013-05-05 01:02:06 <sipa> you can verify the builds, and check the source code
 35 2013-05-05 01:02:17 <sipa> and even better, others can do the same
 36 2013-05-05 01:02:21 <cjd> yeah, the derivation of the state-at-lockin could be done with gitian as well IMO
 37 2013-05-05 01:02:32 <The_Fly> https://en.bitcoin.it/wiki/Checkpoint_Lockin
 38 2013-05-05 01:02:33 <sipa> that's an interesting idea
 39 2013-05-05 01:02:51 <gonffen> does anybody know off the top of their head roughly where incoming connection limits are established in the source?
 40 2013-05-05 01:03:00 <The_Fly> "PPCoin uses an ongoing "central checkpointing pulse" that isn't updated only on software updates."
 41 2013-05-05 01:03:09 <The_Fly> that sounds vulnerable
 42 2013-05-05 01:03:49 <cjd> gonffen: IIRC it's a flag when you start the client
 43 2013-05-05 01:04:02 <cjd> but remember to turn up your file descriptor count before playing
 44 2013-05-05 01:04:17 <gonffen> I didn't run it with any flags...
 45 2013-05-05 01:04:37 <gonffen> I'm actually concerned about litecoin-qt
 46 2013-05-05 01:04:43 <gonffen> for some reason I have 60 connections
 47 2013-05-05 01:04:46 <gonffen> it was up to 63 for a bit
 48 2013-05-05 01:05:01 <gonffen> seems like a horrible idea
 49 2013-05-05 01:05:27 <sipa> git head automatically raises file descriptors if necessary and possible
 50 2013-05-05 01:05:57 <cjd> -maxconnections
 51 2013-05-05 01:07:37 <cjd> ERROR: ProcessBlock() : already have block 129442 00000000000002ce894fe7d9c421fab742ef656ca753fe95c338fdd3bd0d69da
 52 2013-05-05 01:07:37 <cjd> Misbehaving: 58.47.122.218:8333 (0 -> 0)
 53 2013-05-05 01:07:46 <gonffen> I should probably not look through code this evening anyways...
 54 2013-05-05 01:07:49 <cjd> ^^ Ask everyone for blocks, complain about duplicates ?
 55 2013-05-05 01:09:29 <sipa> cjd: bingo
 56 2013-05-05 01:09:53 <ProfMac_> I have a transaction in block 234567 :-)
 57 2013-05-05 01:11:52 <cjd> never attribute to malice what can adaquately be explained by lazieness :)
 58 2013-05-05 01:14:59 <cjd> has there been any discussion of a protobuf protocol? it would be annoying in the short term maintaining 2 of everything but in the long term you could dump serialize.h and alt client devs would have a much easier time.
 59 2013-05-05 01:15:20 <sipa> cjd: unfortunately, not possible
 60 2013-05-05 01:15:40 <sipa> as the serializations are used for hashes
 61 2013-05-05 01:16:10 <sipa> (well, it would be possible as long as you have a serialize-only version of the old code, just to compute hashes)
 62 2013-05-05 01:16:41 <cjd> ahh right
 63 2013-05-05 01:16:51 <cjd> so the transaction format is forever
 64 2013-05-05 01:17:07 <cjd> block header format too... so that kinda kills it
 65 2013-05-05 01:20:44 <sipa> 
ACTION has a bitcoind that doesn't need OpenSSL/EC!

 66 2013-05-05 01:20:46 <sipa> 
ACTION sleeps

 67 2013-05-05 01:22:02 <cjd> ooo
 68 2013-05-05 01:22:06 <cjd> I saw your repo
 69 2013-05-05 01:22:09 <freewil> what version was getbloctemplate added to the api
 70 2013-05-05 01:24:31 <cjd> hah
 71 2013-05-05 01:24:39 <cjd> you could compact bootstrap.dat
 72 2013-05-05 01:26:05 <cjd> so you have just one block with one coinbase transaction which pays out btc to every address which has btc
 73 2013-05-05 01:26:11 <cjd> basically constant folding for the chain
 74 2013-05-05 01:26:39 <sipa> eh no
 75 2013-05-05 01:26:45 <cjd> as long as the clients are ok with a single block which violates every rule ever...
 76 2013-05-05 01:26:45 <sipa> that would be unvalidatable
 77 2013-05-05 01:26:56 <sipa> that defeats the purpose
 78 2013-05-05 01:27:21 <cjd> well, I'm assuming the client has received this compacted bootstrap from a trusted source
 79 2013-05-05 01:27:26 <sipa> if you're going to trust someone to give you something unverifiable, just use a lightweight client
 80 2013-05-05 01:27:40 <sipa> which doesn't need a trusted database in the first place
 81 2013-05-05 01:28:08 <sipa> or if you have a really trusted source (like yourself), just copy the datadir
 82 2013-05-05 01:28:16 <cjd> annoying thing with light clients for me is the fact that they can't discover that they have been paid by looking at the chain
 83 2013-05-05 01:28:27 <sipa> huh?
 84 2013-05-05 01:28:42 <cjd> ok, guess the one I was looking at is lighter than others
 85 2013-05-05 01:28:55 <sipa> i mean SPV nodes like Multibit and Bitcoin Wallet for Android
 86 2013-05-05 01:29:05 <cjd> I was looking at picocoin which apparently needs to get a payment broadcast to it to know it was paid
 87 2013-05-05 01:29:10 <cjd> sorry
 88 2013-05-05 01:29:51 <cjd> this non-openssl ECC thing, fast?
 89 2013-05-05 01:29:57 <sipa> very
 90 2013-05-05 01:30:09 <cjd> using djb's weird trick?
 91 2013-05-05 01:30:18 <sipa> which one?
 92 2013-05-05 01:30:34 <cjd> heh
 93 2013-05-05 01:30:53 <sipa> if you mean using 5 52-bit limbs instead of 4 64-bit ones to represent a 256-bit field element: yes
 94 2013-05-05 01:31:17 <cjd> there's this strange tree like block of code in his multiplier
 95 2013-05-05 01:32:05 <cjd> https://github.com/cjdelisle/cnacl/blob/master/crypto_scalarmult/curve25519/donna_c64/smult.c#L392
 96 2013-05-05 01:32:47 <sipa> eh, i have something similar yes
 97 2013-05-05 01:32:48 <cjd> IIRC justmoon was using the same thing in his bitcoinjs code
 98 2013-05-05 01:32:59 <sipa> for inverses and square roots
 99 2013-05-05 01:33:06 <sipa> that's just a precomputed exponentiation ladder
100 2013-05-05 01:33:17 <cjd> ok :)
101 2013-05-05 01:33:28 <cjd> I never really got that deep into that stuff
102 2013-05-05 01:34:03 <sipa> https://github.com/sipa/secp256k1/blob/master/src/impl/field.h#L90
103 2013-05-05 01:36:34 <cjd> neat
104 2013-05-05 01:37:11 <sipa> for raw signature verifications it's around 6.5x faster than OpenSSL
105 2013-05-05 01:37:24 <sipa> on x86_64, when using the assembly code
106 2013-05-05 01:38:26 <sipa> anyway, now i really need sleep
107 2013-05-05 01:38:52 <cjd> ok gn
108 2013-05-05 01:38:58 <cjd> awesome stuff
109 2013-05-05 02:11:36 <FlyingLeap> litecoin
110 2013-05-05 02:11:47 <FlyingLeap> going to overtake bitcoin?
111 2013-05-05 02:19:53 <cjd> FlyingLeap: my opinion is this: There are 2 types of people holding bitcoin, the ones who are speculating and the ones who are buying it to make a political statement ("I don't trust the bank"), LTC offers an interesting place to speculate but what political statement does it make?
112 2013-05-05 02:20:55 <cjd> And if deep down it doesn't promise us a better life then the only investors in it will be speculators and it will rise but when there is a problem and it crashes, it will not recover.
113 2013-05-05 02:21:29 <FlyingLeap> cjd, litecoin doesn't also make the same political statement?
114 2013-05-05 02:22:07 <FlyingLeap> if it's mostly speculators, sounds shady
115 2013-05-05 02:22:12 <cjd> so then it is essentially the same as bitcoin but with fewer places to spend it?
116 2013-05-05 02:22:30 <cjd> why should I bother with it (assuming "get rich quick" is not an option)
117 2013-05-05 02:22:55 <FlyingLeap> don't know. do they have useful developments that get imported into bitcoin?
118 2013-05-05 02:23:14 <cjd> no, the only difference is the mining algorithm
119 2013-05-05 02:23:40 <cjd> which was supposed to only work on CPUs but it was cracked so now GPUs work
120 2013-05-05 02:23:42 <FlyingLeap> scrypt is a good hash algorithm though isn't it.better than bcrypt.
121 2013-05-05 02:23:45 <cjd> so it's effectively the same
122 2013-05-05 02:24:21 <FlyingLeap> that's news to me. scrypt cracked?
123 2013-05-05 02:24:42 <cjd> "cracked" has a different meaning to cryptographers
124 2013-05-05 02:24:52 <FlyingLeap> i see what you mean.
125 2013-05-05 02:25:00 <cjd> someone found an easier (but still hard) way to get a solution
126 2013-05-05 02:25:05 <cjd> and now GPUs can hash
127 2013-05-05 02:25:35 <FlyingLeap> sounds too much like a clone with no real advantage then
128 2013-05-05 02:25:42 <cjd> +1
129 2013-05-05 02:25:52 <cjd> I actually described half of this problem in 2010, I somehow neglected to spot the other half which is what made it totally broken
130 2013-05-05 02:26:07 <cjd> (for memory hardness, it's still ok for wasting CPU/GPU cycles)
131 2013-05-05 02:26:36 <shesek> https://en.bitcoin.it/wiki/Weaknesses#Attacker_has_a_lot_of_computing_power "The attacker can't ... Reverse other people's transactions". How so? couldn't he remove the block with the transaction he wishes to remove, and create a new a, longer, block chain without that transaction?
132 2013-05-05 02:26:48 <FlyingLeap> memory hardness is what was originally assumed to be such that gpus would not work for scrypt?
133 2013-05-05 02:26:59 <cjd> FlyingLeap: correct
134 2013-05-05 02:27:32 <cjd> shesek: he could make them disappear but it is not useful to him unless he is collaborating with the other people
135 2013-05-05 02:28:39 <shesek> what does that sentence from the wiki mean than?
136 2013-05-05 02:28:59 <cjd> FlyingLeap: We need new currencies, we need currencies which are backed by stocks and bonds in our local communities and backed by our town governments and by gold and silver and all sorts of things, we need money which makes our lives better
137 2013-05-05 02:29:07 <cjd> we don't need more speculation and useless clones
138 2013-05-05 02:30:23 <shesek> cjd, the advantage of GPUs over CPUs is still much more minimal than what's going on with bitcoin
139 2013-05-05 02:30:31 <shesek> its not nearly the same
140 2013-05-05 02:32:24 <cjd> When Lolcust created TBX which eventually evolved through a series of forks into LTC, he intended it to be a currency with builtin money laundry.
141 2013-05-05 02:32:39 <cjd> This is something, not something I agree with but it is a real change
142 2013-05-05 02:33:38 <rs0> AIUI GPU's don't have much of an advantage over CPUs when it comes to scrypt. Scrypt ASICs however have a tremendous advantage
143 2013-05-05 02:34:07 <rs0> in fact i think Luke-Jr basically considers it a structural weakness in the Litecoin design
144 2013-05-05 02:34:50 <shesek> built in money laundry? how so?
145 2013-05-05 02:35:10 <cjd> the technique he tried to use didn't work
146 2013-05-05 02:39:08 <shesek> what you mean by money laundry? making it impossible to trace the coin origins by following the transactions chain?
147 2013-05-05 02:39:47 <cjd> yeah
148 2013-05-05 02:39:58 <shesek> how did he try to do that?
149 2013-05-05 02:40:19 <cjd> google tennenbrix
150 2013-05-05 02:46:31 <cjd> http://image.bayimg.com/cc32968d52b31233fa45bc27cc9f502cdcd8366c.jpg
151 2013-05-05 03:14:24 <freewil> how come `walletlock` doesn't show up in `bitcoind help | grep walletlock`
152 2013-05-05 03:21:32 <freewil> hmm
153 2013-05-05 03:22:14 <freewil> so if your wallet isnt encrypted then `help` doesnt show commands applicable to an encrypted wallet only
154 2013-05-05 03:22:26 <freewil> thats somewhat annoying
155 2013-05-05 03:54:45 <sserrano44> hi
156 2013-05-05 03:56:39 <sserrano44> I'm having a weird issue with a transaction, when I broadcast it is rejected instantly by bitcoind
157 2013-05-05 03:56:52 <Luke-Jr> sserrano44: broadcast it how?
158 2013-05-05 03:57:11 <sserrano44> but if I decoded all parameters look good
159 2013-05-05 03:57:21 <sserrano44> bitcoind sendrawtransaction ...
160 2013-05-05 03:57:39 <sserrano44> testing it with signtransaction returns completed
161 2013-05-05 03:57:46 <sserrano44> I guess all the parts are there
162 2013-05-05 03:58:05 <sserrano44> this is the transaction: 0100000001c58579976b9be9e18ab0b6d27375f27546a47d144ae5b2d27d0c6fb9a01fad4d010000008b483045022100e6cf1b0002217eca92506477be690c1b43b429260d2482dd99c20a96c4a270110220798f0082751f8f4d4a8812a02f145311e438184d1f20e749b8970766708efa91014104f7a84351c1341e504c5e8b6eba6a99262d62fc4b27433132a0bf760da9ec9d7bd1338181685a88d51ede5cae13611a1212d3f5708c9fead0c8ed0546abaf44a8ffffffff02c011dd00000000001976a9148a49ba9ff
163 2013-05-05 03:58:41 <sserrano44> bitcoind returns error: {"code":-22,"message":"TX rejected"}
164 2013-05-05 04:00:13 <sserrano44> any ideas?
165 2013-05-05 04:01:16 <lianj> sserrano44: pastebin it. irc cut off part of the tx
166 2013-05-05 04:03:28 <freewil> Luke-Jr, do you think it makes sense to hide the wallet encryption-related commands from `bitcoind help` if your wallet isn't currently encrypted?
167 2013-05-05 04:03:40 <Luke-Jr> freewil: debatable
168 2013-05-05 04:03:51 <Luke-Jr> I could probably argue both positions
169 2013-05-05 04:04:05 <freewil> seems like a bad idea
170 2013-05-05 04:04:06 <jchp> hiding functions is usually more confusing
171 2013-05-05 04:04:37 <freewil> yeah i was going to write a script to make sure my Node.js module has all the commands in it
172 2013-05-05 04:04:46 <freewil> then i ran into this issue of why it was hiding some commands
173 2013-05-05 04:04:46 <Luke-Jr> well, so is showing functions that effectively don't exist
174 2013-05-05 04:05:13 <Luke-Jr> it's not like "encrypted wallet" is a state, it's kinda encompassing of the client
175 2013-05-05 04:05:19 <jchp> yes, but that's a known unknown, you know some may fail. hiding functions could be an unknown-unknown feature for people
176 2013-05-05 04:05:47 <freewil> lol jchp agreed
177 2013-05-05 04:06:38 <freewil> imagine if `man` tried to encompass state into the output
178 2013-05-05 04:07:24 <Luke-Jr> freewil: ?
179 2013-05-05 04:07:38 <freewil> it seems like help is a basic "documenting" feature
180 2013-05-05 04:07:56 <Luke-Jr> freewil: wallet encryption isn't a state
181 2013-05-05 04:07:58 <freewil> and so it's output shouldnt depend on your current state (whether you have an encrypted wallet currently or not)
182 2013-05-05 04:08:14 <freewil> what would you call it
183 2013-05-05 04:08:18 <Luke-Jr> it's like the difference between 0.4 and 0.8
184 2013-05-05 04:08:44 <freewil> you can still run an encrypted wallet in 0.8
185 2013-05-05 04:08:50 <Luke-Jr> irrelevant
186 2013-05-05 04:08:58 <Luke-Jr> 0.4's help doesn't have commands only in 0.8
187 2013-05-05 04:09:16 <freewil> thats a terrible analogy
188 2013-05-05 04:09:29 <freewil> unencrypted wallet is still the default in 0.8
189 2013-05-05 04:09:45 <freewil> if you just start bitcoind and don't already have a wallet that is what you get
190 2013-05-05 04:10:13 <freewil> so say a new user is using `bitcoind help` to study the commands
191 2013-05-05 04:10:31 <freewil> see what their options are with working on the api
192 2013-05-05 04:10:54 <freewil> they have unknown unknowns because they cant see all the commands because of their current configuration
193 2013-05-05 04:11:52 <sserrano44> lianj: this is the decoded transaction http://pastebin.com/GHF7Gcyp
194 2013-05-05 04:15:02 <Luke-Jr> freewil: bitcoind isn't for users anyway, it's for developers
195 2013-05-05 04:15:28 <freewil> right, so in this case the user is the developer
196 2013-05-05 04:23:13 <freewil> oh well, let your voice be heard https://github.com/bitcoin/bitcoin/pull/2619
197 2013-05-05 04:26:51 <freewil> Luke-Jr, wouldn't you consider whether the wallet is encrypted a type of configuration or "state"
198 2013-05-05 04:27:23 <freewil> you could encrypt or unencrypt at runtime
199 2013-05-05 04:32:49 <Luke-Jr> freewil: you cannot unencrypt
200 2013-05-05 04:33:23 <Luke-Jr> the state is locked/unlocked (for an encrypted wallet)
201 2013-05-05 04:33:27 <Luke-Jr> and no commands are hidden for that
202 2013-05-05 04:35:45 <freewil> i see
203 2013-05-05 04:42:59 <cjd> is there an agreed upon standard way to write a transaction which pays noone ?
204 2013-05-05 04:43:03 <cjd> which is prunable
205 2013-05-05 04:43:42 <jgarzik> cjd: an output with OP_RETURN is an example of something that is provably prunable
206 2013-05-05 04:46:28 <cjd> has any progress been made on the proposal to make that IsStandard() ?
207 2013-05-05 04:46:50 <jgarzik> cjd: the recent data spamming rather reduced motivation...
208 2013-05-05 04:48:18 <cjd> hrm, chicken-and-egg problem... ofc the spammers will never want to use such a thing since they want people to carry around their data
209 2013-05-05 04:59:25 <jeremias> hmm, I am upgrading bitcoind from 0.7.0 to 0.8.1, it claims corrupt wallet, first gives a warning and then an error
210 2013-05-05 04:59:57 <Luke-Jr> cjd: what is the purpose of a transaction which pays noone?
211 2013-05-05 05:02:48 <jeremias> do you know what is the order of operations, if the bitcoind fails with corrupt wallet, has it altered the blockchain files?
212 2013-05-05 05:02:56 <jeremias> eg. can I try downgrading to 0.7.0?
213 2013-05-05 05:12:29 <Luke-Jr> jeremias: makes more sense to downgrade to 0.7.3rc3
214 2013-05-05 05:14:46 <cjd> I need notarization of some data, not much and not often, I want any SPV node with the btc header chain to be able to verify it and I don't want to have to buy some alt coins to do it.
215 2013-05-05 05:15:24 <cjd> I'm happy to pay for inclusion and I'm happy to see it pruned
216 2013-05-05 05:16:28 <Luke-Jr> cjd: see ChronoBit and/or Namecoin
217 2013-05-05 05:16:40 <SomeoneWeird> lol.
218 2013-05-05 05:16:52 <cjd> namecoin is out because altcoin
219 2013-05-05 05:16:54 <SomeoneWeird> if you don't need the coins at the other end just op_return it
220 2013-05-05 05:17:04 <Luke-Jr> cjd: altcoin is what you need. Bitcoin is not for your spam :p
221 2013-05-05 05:17:22 <amiller> the strangest limitation of bitcoin is how difficult it is to find a way to pay for extra features you want in it!
222 2013-05-05 05:17:51 <Luke-Jr> cjd: pruning is not complete; full nodes *still* need to store it forever
223 2013-05-05 05:18:36 <cjd> I want to collaborate and find a solution which makes everyone happy but it is easier for me to include fake standard transactions which can't be pruned than it is to fool around with alt chains.
224 2013-05-05 05:19:27 <gmaxwell> cjd: wait. ""I don't want to have to buy some alt coins to do it" ... if I just give you enough namecoin to stamp a million things will you give up on trying to bloat bitcoin with it?
225 2013-05-05 05:20:38 <gmaxwell> cjd: you should be using something like chronibit, except that everyone who says that timestamping is important to them ??? at least so far??? seems to be full of it, since they won't bother to do even the modest work to run something with O(1) scaling which has already been written.
226 2013-05-05 05:20:49 <cjd> The nonstandard op_return is unfortunate because it forces legitimate users into the same group as people who want to force everyone to carry around their data
227 2013-05-05 05:21:41 <gmaxwell> cjd: outputs are not a good place for time stamping in any case, even with purning they still burden the network with storage proportional to usage.
228 2013-05-05 05:21:55 <Luke-Jr> gmaxwell: otoh, ChronoBit is still pool-specific :/
229 2013-05-05 05:21:57 <gmaxwell> cjd: vs coinbase timestampers like chronobit which are O(1).
230 2013-05-05 05:22:03 <gmaxwell> Luke-Jr: so fix it.
231 2013-05-05 05:22:35 <gmaxwell> Luke-Jr: I think the chronobit proofs themselves actually don't care and would work find in a regular coinbase.
232 2013-05-05 05:22:53 <cjd> what was the depth of a cronobit hash chain? I seem to recall it being long
233 2013-05-05 05:22:57 <Luke-Jr> gmaxwell: but there's no p2p protocol to communicate them to other pools
234 2013-05-05 05:23:20 <gmaxwell> Luke-Jr: any reasonably sizable pool is getting multiply blocks a day.
235 2013-05-05 05:23:28 <Luke-Jr> relevance?
236 2013-05-05 05:23:36 <gmaxwell> cjd: it's a couple kilobytes typically, depends on p2pools relative hashrate.
237 2013-05-05 05:23:53 <cjd> hrm
238 2013-05-05 05:23:58 <gmaxwell> Luke-Jr: just submit to one, you'll get timestamped??? I agree there should be some work sharing mechenism, but for lots of things its not important.
239 2013-05-05 05:25:10 <gmaxwell> cjd: it could be made much smaller if p2pool had more than just a prev backreference. But see also my whining about people not actually caring about this application.
240 2013-05-05 05:25:36 <Luke-Jr> gmaxwell: actually, the length could be helpful for more accuracy ;)
241 2013-05-05 05:25:48 <Luke-Jr> gmaxwell: you can prove X amount of work was done before it hit the blockchain
242 2013-05-05 05:25:57 <Luke-Jr> and even the timestamp at any given step
243 2013-05-05 05:27:36 <gmaxwell> yes, it effectively gives you reduced security increased precision timestamps.
244 2013-05-05 05:28:07 <gmaxwell> but multiple backreferences allowing you to skip the middle ones don't preclude you keeping them for proving the more precise time.
245 2013-05-05 05:30:04 <cjd> this is a weird situation since like paying $0.25 for a download or something is considered valid but paying $0.25 for a timestamp is abuse
246 2013-05-05 05:30:18 <cjd> (paying the miner)
247 2013-05-05 05:31:14 <cjd> I suppose one can hope that the download site cleans up their payments so there are not dust payments all lying around...
248 2013-05-05 05:33:18 <gmaxwell> cjd: using bitcoin as a currency is something that all bitcoin users have consented to, throwing non-currency usage onto them is something they haven't consented to.  Perhaps in some cases its generally harmless, ??? I'd think timestamping is certantly among the least harmful of the non-currency uses. But timestamping has a tidy O(1) solution which is obviously in everyone's best interest to use, but it requires a little more effort to ...
249 2013-05-05 05:33:24 <gmaxwell> ... setup.
250 2013-05-05 05:34:11 <cjd> There is some validity to the "what they signed up for" argument
251 2013-05-05 05:34:39 <gmaxwell> Of course, widespread use of the poorly scaling time would remove incentive to do the additional work to setup the O(1) kind, which is why it's more important to yell at you for it then it might be otherwise. :P
252 2013-05-05 05:35:19 <cjd> well.. the proofs also have to be quite small
253 2013-05-05 05:35:36 <cjd> Has the dust outputs PR been included or did the colored coin people keep it out?
254 2013-05-05 05:35:55 <cjd> Because dust outputs strike me as a much bigger threat overall
255 2013-05-05 05:37:40 <gmaxwell> afaik, colored coin went nowhere there, because it has a tidy answer "create outputs with values large enough to be worth spending"
256 2013-05-05 05:38:05 <cjd> Yeah, I got that after about 2 minutes of reading the thread
257 2013-05-05 05:38:32 <cjd> sadly there doesn't exist any implementation which uses deterministic satoshi tracking which IMO is the only sane way to implement it
258 2013-05-05 05:39:48 <gmaxwell> cjd: so, if you want more compact proofs than chronobit can currenty give you??? make a tool that fits into the merged mining infrastructure, and at least one multiple-block-a-day pool will run it, probably at least two.
259 2013-05-05 05:40:23 <Luke-Jr> cjd: it's abuse because the nodes storing the blockchain have agreed to the storage for currency/trade purposes, not to timestamp your data
260 2013-05-05 05:40:44 <Luke-Jr> cjd: you may be paying the miner, but that does not compensate all the nodes storing it, nor do they have the opt-out ability
261 2013-05-05 05:41:00 <SomeoneWeird> ABUSE CJD
262 2013-05-05 05:41:03 <SomeoneWeird> :P
263 2013-05-05 05:41:04 <Luke-Jr> oops, guess gmaxwell already said that
264 2013-05-05 05:41:28 <cjd> Well a generic validator would be able to root in any transaction, mine or the coinbase
265 2013-05-05 05:41:29 <gmaxwell> If the timestamping were really costless it wouldn't be much of an issue, but every timestamp done in the lame transaction way increases the blockchain size forever. :(
266 2013-05-05 05:42:25 <Luke-Jr> cjd: merged mining enables that, yes
267 2013-05-05 05:42:33 <cjd> well.. long term the only way I see this thing scaling is by making the payer provide the part of the chain containing his coins
268 2013-05-05 05:43:27 <Luke-Jr> btw, I did throw this template together https://en.bitcoin.it/wiki/BIP_MergedMiningScript
269 2013-05-05 05:43:42 <cjd> so then it occupies a few bits of storage to flag an output as unspent
270 2013-05-05 05:43:51 <gmaxwell> cjd: uh, that doesn't actually prevent the chain from growing.
271 2013-05-05 05:44:09 <cjd> yeah, it just prevents anybody from having to store it
272 2013-05-05 05:44:20 <gmaxwell> No, it does not.
273 2013-05-05 05:44:21 <gmaxwell> cjd: nodes still need to inspect the historic chain to convince themselves that no one has inflated the currency in the past.
274 2013-05-05 05:44:22 <Luke-Jr> cjd: nodes need to store the chain
275 2013-05-05 05:44:41 <gmaxwell> And to inspect it they need to be able to obtain it.
276 2013-05-05 05:45:29 <gmaxwell> Which means that _someone_ has to store it.
277 2013-05-05 05:45:34 <cjd> well if everybody throws everything away after 5 years, you have a 5 year period where any funny business would cause someone's code to break
278 2013-05-05 05:46:11 <cjd> yeah, someone will always want to store the whole history of the universe, but it should not be a requirement
279 2013-05-05 05:46:16 <gmaxwell> cjd: there are already plenty of people willing to believe that satoshi or whomever has a billion bitcoins, the fact that each node independantly verifies that this isn't so is fundimental to the trust in the system.
280 2013-05-05 05:46:46 <gmaxwell> cjd: it's not a requirement for everyone to store everything??? but it's a requirement under our security model for it to be stored and available.
281 2013-05-05 05:47:35 <gmaxwell> cjd: and regardless, it doesn't change the fact that there is a tidy O(1) solution to your problem which requires a small amount of additional upfront development work.
282 2013-05-05 05:47:36 <cjd> plenty of people willing to believe <-- we both know that proving them wrong won't even slow them down ;)
283 2013-05-05 05:47:52 <gmaxwell> cjd: It actually has.
284 2013-05-05 05:48:10 <gmaxwell> (well at the very least, redirected them to more factually based paranoia!)
285 2013-05-05 05:48:16 <cjd> heh
286 2013-05-05 05:49:07 <gmaxwell> cjd: How's cjdns going in any case?  Also?????did you ever write up what you want the timestamps for?
287 2013-05-05 05:49:08 <sydna> oddly enough, people *don't* want to store history. that's why there's groups like archive.org and ArchiveTeam running around the joint
288 2013-05-05 05:49:42 <cjd> cjdns is kind of just doing it's thing, I have not made many changes to it lately
289 2013-05-05 05:49:45 <SomeoneWeird> cjdns imploded
290 2013-05-05 05:49:46 <cjd> it basically works
291 2013-05-05 05:49:48 <SomeoneWeird> jks
292 2013-05-05 05:50:07 <Luke-Jr> sydna: I think it's more for people too ignorant to realize they want to store history, not that people have made a conscious decision not to
293 2013-05-05 05:50:24 <cjd> the network is a bit fragile and slow to converge because I need to rewrite a bunch of things
294 2013-05-05 05:50:42 <cjd> still putting that off since in it's current form it's reasonably free of bugs
295 2013-05-05 05:50:50 <cjd> and it took a very long time to get there
296 2013-05-05 05:51:12 <sydna> Luke-Jr: I've approached people with offers to personally archive dead forums and websites that look in danger of disappearing. on more than one occasion I've had people nuke the web server to stop people from archiving the contents.
297 2013-05-05 05:51:56 <sydna> Luke-Jr: there does seem to be some resistance that people have towards the archiving of internet history, or history in general.
298 2013-05-05 05:52:15 <Luke-Jr> weird
299 2013-05-05 05:52:21 <gmaxwell> sydna: part of the problem is that old things look new online, the pages don't yellow.
300 2013-05-05 05:53:17 <sydna> gmaxwell: I hadn't considered that. it's fairly easy to tell old websites apart though, all the outgoing links will be broken
301 2013-05-05 05:53:42 <cjd> The motivation for this timestamping business is a semi-secret project, I don't talk much about it because when I do, people come out of the woodwork to argue with me about design details.
302 2013-05-05 05:53:55 <gmaxwell> I feel very fortunate that most of my mispent online youth happened before the WWW and what was said on BBSes, fidonet, compuserve, and early IRC largely went down the memory hole. ... I've seen people's reputations torn up by things they wrote on the internet a decade prior as kids.
303 2013-05-05 05:54:25 <gmaxwell> cjd: ah. well. just make it so complicated that most people give up trying to understand it.
304 2013-05-05 05:54:28 <gmaxwell> :P
305 2013-05-05 05:54:36 <cjd> :D
306 2013-05-05 05:54:37 <sydna> I understand that. there's a difference between public forums and IRC logs though
307 2013-05-05 05:55:23 <gmaxwell> sydna: somewhat... but when people write in a forum they're writing as part of a dialog with the existing people there that share substantial context and culture with them and also have the ability to engage in two way discussions to clarify misunderstandings.
308 2013-05-05 05:55:36 <gmaxwell> So while public forums and irc are different, they're not entirely different.
309 2013-05-05 05:55:52 <cjd> The Internet makes the whole world nakid, eventually society will catch up with the understanding that we are all human, it just sucks to be the first person to lose their pants.
310 2013-05-05 05:56:09 <gmaxwell> (I'm not opposed to archiving either, but I've actually seen the long memory of the internet hurt people severely, and so I think I see why some might be!)
311 2013-05-05 05:56:35 <sydna> gmaxwell: is decidedly less formal than a forum though. I'm fairly sure that your demeanour is slightly different on here than when you post in the mailing lists or bitcointalk.org
312 2013-05-05 05:56:44 <sydna> gmaxwell: I completely understand your stance though
313 2013-05-05 05:56:52 <jeremias> is it safe to run pywallet while bitcoind is running?
314 2013-05-05 05:57:27 <gmaxwell> cjd: a good bit of is is hygiene and expectations. It's less bad to lose your pants when you know you were going to be without them.
315 2013-05-05 05:57:38 <cjd> /nod
316 2013-05-05 05:57:44 <gmaxwell> People are generally not ready for how pantsless the internet makes them??? its surprising.
317 2013-05-05 05:57:50 <gmaxwell> jeremias: No.
318 2013-05-05 05:57:51 <cjd> yeap
319 2013-05-05 05:58:14 <sydna> people are generally not ready for how much an alias exposes them
320 2013-05-05 05:59:25 <gmaxwell> sydna: a lot of people think they're far more hidden than they are??? I often think it's better to just participate under your legal name and know that you're not hidden. At least your expectations match reality.
321 2013-05-05 05:59:40 <cjd> +1
322 2013-05-05 06:00:37 <cjd> Also helps remind one's self to avoid saying things which might be taken the wrong way
323 2013-05-05 06:01:24 <sydna> gmaxwell: the way I think of pseudonyms varies on a day to day basis. I suspect that I would sorely regret using my true name for many things online. not because any of them are particularly bad, just that I would look back on myself and wince
324 2013-05-05 06:01:38 <sydna> *on a day to day basis changes
325 2013-05-05 06:01:41 <gmaxwell> Thats quite hard sometimes, particularly where you're communicating in forums (like IRC) where there isn't an expectation of editorial review before posting.
326 2013-05-05 06:01:56 <Luke-Jr> sydna: what is a true name?
327 2013-05-05 06:02:51 <sydna> Luke-Jr: my full name, at least, one of them
328 2013-05-05 06:03:08 <Luke-Jr> sydna: I'm trying to get philosophical here, I mean in general :P
329 2013-05-05 06:03:32 <sydna> oh, sorry
330 2013-05-05 06:03:45 <kauzu> good morning
331 2013-05-05 06:04:32 <kauzu> how can i calculate the priority of a tx ? i mean automaticaly... there are way to much inputs for do it manualy
332 2013-05-05 06:05:38 <sydna> Luke-Jr: I'm actually not sure. have you read the Earthsea Trilogy?
333 2013-05-05 06:05:44 <Luke-Jr> sydna: no
334 2013-05-05 06:07:34 <sydna> Luke-Jr: my quote won't make any sense then. just relates to a guy who always uses a pseudonym for his entire life, and only one person in the word uses his true given name. something I somewhat aspire to.
335 2013-05-05 06:08:31 <sydna> Luke-Jr: unfortunately outside the world of magic and dragons, everybody wants my real ID. my bank, my rental agency, mt gox. everything is contaminated with my real identity.
336 2013-05-05 06:08:47 <Luke-Jr> sydna: my point is that a name is how you are referred to in writing/speech.
337 2013-05-05 06:10:56 <kauzu> how can i calculate the priority of a tx ? i mean automaticaly... there are way to much inputs for do it manualy
338 2013-05-05 06:12:37 <Luke-Jr> kauzu: why
339 2013-05-05 06:12:54 <sydna> Luke-Jr: ideally for me that would be a pseudonym in real life as well as here, but there's always legal documents contaminating that. I might call myself sydna to the people at the coffee shop, but the lady at the rental place can still see my real identity on the documents.
340 2013-05-05 06:13:26 <Luke-Jr> sydna: move to the USA, where you have a constitutional right to use whatever name you want
341 2013-05-05 06:13:31 <kauzu> luke-jr: why not
342 2013-05-05 06:13:45 <Luke-Jr> kauzu: because it doesn't have a universal definition
343 2013-05-05 06:14:05 <Luke-Jr> kauzu: different clients and miners can calculate prioirities however they like
344 2013-05-05 06:14:42 <kauzu> luke-jr: then for the reference implementation
345 2013-05-05 06:14:49 <sydna> Luke-Jr: I wish. I suspect their other laws might be less friendly to me.
346 2013-05-05 06:14:55 <Luke-Jr> kauzu: it still varies between versions
347 2013-05-05 06:15:07 <Luke-Jr> sydna: every country has laws that suck.
348 2013-05-05 06:15:18 <Luke-Jr> if you find a perfect one, let me know ;P
349 2013-05-05 06:15:40 <sydna> Australia is getting less and less perfect as time goes on
350 2013-05-05 06:15:55 <duSn> law in massachusettes -> you have to walk 100ft in front of car with a lantern
351 2013-05-05 06:16:21 <Luke-Jr> I don't like Australia at all.
352 2013-05-05 06:17:23 <sydna> any particular reason? I'm always curious about our reputation
353 2013-05-05 06:19:35 <SomeoneWeird> AUStrAAAAAALIA
354 2013-05-05 06:55:35 <Kwanf6> how does a client find other peers? Is there some "Tracker" sites?
355 2013-05-05 06:58:26 <jaakkos> Kwanf6: https://en.bitcoin.it/wiki/Network#Bootstrapping
356 2013-05-05 06:59:31 <gmaxwell> Kwanf6: do you mean initially on first start or steady state?
357 2013-05-05 06:59:54 <gmaxwell> if you mean initially, thats bootstrapping??? after that it hears about peers from other peers.
358 2013-05-05 07:02:25 <cjd> is TxIn->previous_output->hash reversed?
359 2013-05-05 07:03:41 <alex_fun> heya
360 2013-05-05 07:04:18 <alex_fun> if chain starts at low diff and checkpoints are added say after every 150th block
361 2013-05-05 07:04:41 <alex_fun> does it make it immune to 51% chain ovewrite
362 2013-05-05 07:05:09 <alex_fun> maybe in future some coins clients are going to auto push checkpoints updates? :)
363 2013-05-05 07:05:36 <gmaxwell> alex_fun: it makes it a centeralized system, as other than "trust me" there is no unambigious decenteralized way to "checkpoint" that frequently ??? so yes, sure??? but why bother with a blockchain? pretty inefficient way to have a centeralized system.
364 2013-05-05 07:06:03 <gmaxwell> Perhaps you should consider using paypal instead?
365 2013-05-05 07:06:18 <alex_fun> interesting point
366 2013-05-05 07:07:11 <cjd> What I find more interesting is the prospect of issuing a currency using colored coins
367 2013-05-05 07:07:20 <alex_fun> can block chain issue checkpoints itself somehow?
368 2013-05-05 07:07:47 <gmaxwell> alex_fun: every block is the blockchain checkpointing itself
369 2013-05-05 07:08:12 <cjd> because it can be backed by businesses in the local community and these guys here keep the infrastructure working :)
370 2013-05-05 07:08:56 <Luke-Jr> gmaxwell: btw, what if we got rid of the block size limit, and replaced it with a p2pool rule that every sub-block could add 15 kB to the block data but was required to include all previous transactions? ;)
371 2013-05-05 07:09:09 <Luke-Jr> gmaxwell: and blocks are only valid if they complied with the p2pool rule
372 2013-05-05 07:09:17 <Luke-Jr> (not necessarily p2pool itself, but something related)
373 2013-05-05 07:10:05 <alex_fun> Luke-Jr:  how did u `stoped` that new old coin? technically speaking? you did not proccess blocks
374 2013-05-05 07:10:14 <alex_fun> looks interesting
375 2013-05-05 07:10:26 <Luke-Jr> alex_fun: I just raised the transaction fee to some huge amount
376 2013-05-05 07:10:31 <gmaxwell> alex_fun: the whole idea of "immune to 51%" is basically misguided. Unless you've found a better way to prevent sybils than computing/energy consumption thats another way of saying "allow a minority to control the state of the system". There are many ways to allow a minority of something to control the state, but if you're interested in making something secure and decenteralized most of them are not very interesting.
377 2013-05-05 07:11:17 <cjd> xD
378 2013-05-05 07:11:51 <gmaxwell> Luke-Jr: dunno. too tired to think
379 2013-05-05 07:12:06 <alex_fun> Luke-Jr:  as far as I know tx fees go into mempool them via mining into reward, some miners could simple keep processing blocks at theirs low fees however how would that impact target say 1 block per 10 min?
380 2013-05-05 07:12:46 <Luke-Jr> alex_fun: dunno, too tired to parse that
381 2013-05-05 07:12:48 <Luke-Jr> <.<
382 2013-05-05 07:12:54 <alex_fun> lol
383 2013-05-05 07:13:01 <alex_fun> I give u 200 bbq coins :D
384 2013-05-05 07:13:04 <alex_fun> haha
385 2013-05-05 07:13:07 <alex_fun> kidding
386 2013-05-05 07:15:33 <alex_fun> I mean people sent coins - go to every node mem pool - nodes who accept low tx fees say 30% from 100%. Their rate of blocks production is 30%. 70% is not mining just sits on client. so 30% can still proccess all transactions.
387 2013-05-05 07:16:09 <Luke-Jr> >50% is 100%
388 2013-05-05 07:20:41 <alex_fun> becouse it becomes auto selected best chain up where blocks can be double spent up till  latest checkpoint?
389 2013-05-05 07:21:53 <alex_fun> ok so 51% becomes longest chain, it still includes variety of miners who can set low fees individually or what am I missing?:)
390 2013-05-05 07:26:15 <Luke-Jr> alex_fun: 51% can ignore other miners
391 2013-05-05 07:27:12 <r0sc0e> anyone can tell me why in my appdata directory there is a directory called devcoin, with a wallet.dat, blkindex.dat ect..
392 2013-05-05 07:27:30 <r0sc0e> thought these files are in the bitcoin-appdata directory?
393 2013-05-05 07:27:35 <alex_fun> Luke-Jr: oo becomes its longest chain so unless stuff accepted by those nodes blocks proccesses at low text fee would not be accepted by clients
394 2013-05-05 07:28:04 <alex_fun> since their block hash and merkle hash would be diff from 51% chain
395 2013-05-05 07:28:04 <Luke-Jr> r0sc0e: because you ran a scamcoin called 'devcoin'?
396 2013-05-05 07:28:05 <alex_fun> right? :)
397 2013-05-05 07:28:32 <alex_fun> 
ACTION 51% lukes tonal coin :P

398 2013-05-05 07:28:33 <alex_fun> haha
399 2013-05-05 07:36:02 <alex_fun> ok thanks for explanation anyways
400 2013-05-05 07:36:03 <alex_fun> )
401 2013-05-05 07:56:43 <r0sc0e> whats the correct way to save my wallet on bitcoin-qt 0.8.1? could anyone explain it to me please?
402 2013-05-05 08:08:41 <bitanarchy> hi
403 2013-05-05 08:09:15 <bitanarchy> why is there a list of bitcoin addresses in every the bitcoin-qt wallet?
404 2013-05-05 08:34:38 <takeyourhatoff> How far away are we until the HD wallets spec is finalised?