1 2012-01-10 00:01:10 <kam1l> ls
  2 2012-01-10 00:01:16 <kam1l> DERP where am I :P
  3 2012-01-10 00:01:42 <kam1l> luke-jr: you still around? What type of ram do you have on your dedi?
  4 2012-01-10 00:01:54 <luke-jr> BlueMatt: yt?
  5 2012-01-10 00:02:01 <luke-jr> kam1l: & who cares? O.o
  6 2012-01-10 00:02:18 <kam1l> I have 1 gb and just the pool and bitcoind gets it up to around 800mb
  7 2012-01-10 00:02:22 <kam1l> and the mysql ofc
  8 2012-01-10 00:02:50 <k9quaint> I have 16BG
  9 2012-01-10 00:02:53 <sipa> my vps has 1GiB of RAM and runs bitcoind just fine
 10 2012-01-10 00:02:55 <k9quaint> cuz chicks dig RAM
 11 2012-01-10 00:03:04 <kam1l> me too k9 on my PC :P
 12 2012-01-10 00:03:19 <sipa> i do limit the number of connections, though
 13 2012-01-10 00:03:23 <kam1l> ah
 14 2012-01-10 00:03:30 <gmaxwell> kam1l: I think you're reading virtual and not resident.
 15 2012-01-10 00:03:36 <k9quaint> sipa only lets the cool people mine
 16 2012-01-10 00:04:27 <da2ce7> lol
 17 2012-01-10 00:04:34 <k9quaint> words to live by
 18 2012-01-10 00:05:31 <sipa> ewww, naked pools
 19 2012-01-10 00:05:32 <da2ce7> *n.b. not a serious invitation as I don't own a pool yet.
 20 2012-01-10 00:06:09 <da2ce7> sipa: good point... I much prefer naked people before a naked pool.
 21 2012-01-10 00:06:21 <sipa> depends on certain properties of those people
 22 2012-01-10 00:06:28 <sipa> but that's maybe a bit off topic here
 23 2012-01-10 00:06:29 <k9quaint> the thought of naked bitcoin miners gives me pause
 24 2012-01-10 00:07:43 <k9quaint> wtf...paul ryan drops support for SOPA? did I fall through a wormhole or something?
 25 2012-01-10 00:08:55 <da2ce7> SOPA, makeing the interent safer for everyone! :)
 26 2012-01-10 00:09:14 <da2ce7> it gets rid of those peskey pirates and freespeech.
 27 2012-01-10 00:09:34 <k9quaint> no, it pretty much leaves the Somali pirates untouched
 28 2012-01-10 00:09:53 <k9quaint> piracy and the internet are completely discrete
 29 2012-01-10 00:10:08 <da2ce7> k9quaint: no I mean those so-called 'interent pirates'
 30 2012-01-10 00:10:24 <k9quaint> that term makes no sense
 31 2012-01-10 00:10:40 <sipa> piracy has nothing to do with copyright infringement
 32 2012-01-10 00:10:44 <k9quaint> indeed
 33 2012-01-10 00:10:49 <sipa> the latter does not involve stealing
 34 2012-01-10 00:10:54 <kam1l> what do you guys put your bitcoind max connections to?
 35 2012-01-10 00:10:56 <k9quaint> also true
 36 2012-01-10 00:10:57 <kam1l> do you leave it default?
 37 2012-01-10 00:11:08 <sipa> kam1l: i limit it to 40 on my bitcoind
 38 2012-01-10 00:11:13 <kam1l> kk
 39 2012-01-10 00:11:18 <sipa> it uses quite some buffer memory per connection
 40 2012-01-10 00:11:19 <k9quaint> sipa: you are dangerously well informed
 41 2012-01-10 00:11:46 <sipa> about?
 42 2012-01-10 00:11:51 <da2ce7> sipa: they are stealing the right of the copywrite owners to have exclusive controll over thier works.
 43 2012-01-10 00:12:02 <k9quaint> da2ce7: you cannot steal rights
 44 2012-01-10 00:12:08 <k9quaint> you can grant them, or deny them
 45 2012-01-10 00:12:15 <sipa> da2ce7: i don't think any legislation calls that stealing
 46 2012-01-10 00:12:54 <da2ce7> sipa: maybe they use the term 'copywrite infringement'
 47 2012-01-10 00:13:04 <sipa> yes, that's why i also use that term
 48 2012-01-10 00:13:59 <k9quaint> and its copyright
 49 2012-01-10 00:14:15 <sipa> oh, misread
 50 2012-01-10 00:14:17 <sipa> indeed
 51 2012-01-10 00:14:20 <k9quaint> copyread and copyleft
 52 2012-01-10 00:14:54 <da2ce7> I think that we are going to be soon moving to a while-list interent; insted of a black list.
 53 2012-01-10 00:14:58 <k9quaint> if SOPA passes, I will have to transfer assets to a corporate entity in the Netherlands
 54 2012-01-10 00:15:05 <k9quaint> this will take me about 2 hours
 55 2012-01-10 00:15:32 <k9quaint> then I will transfer the hosting to Europe, this will take several hours if things go well
 56 2012-01-10 00:15:45 <k9quaint> some poor IT guy will lose a tiny fraction of his job
 57 2012-01-10 00:15:53 <da2ce7> where only goverement approvded CA are alloud to make encripted connecttions.
 58 2012-01-10 00:16:22 <k9quaint> da2ce7: is english your first lang?
 59 2012-01-10 00:16:50 <k9quaint> (I need to set my parse filters accordingly)
 60 2012-01-10 00:17:01 <da2ce7> they will make a law where ISP's to drop all encripted data that isn't first signed by CA's that are goveremnt approved.
 61 2012-01-10 00:17:19 <da2ce7> 1st lang, but not very good at it.
 62 2012-01-10 00:17:28 <k9quaint> and I will buy satellite internet or move to a country that isn't owned by China
 63 2012-01-10 00:17:50 <k9quaint> the US has 300mil people, in the grand scheme of things, they matter less and less
 64 2012-01-10 00:18:56 <k9quaint> and by they, I mean all the fucktards who surround me but don't have the common decency to play in traffic on a busy street
 65 2012-01-10 00:20:00 <da2ce7> it shouldn't be too hard... that would mean that we could ban (and enforce) all encripted comminication dosn't come from valid CA issued certificates.
 66 2012-01-10 00:20:33 <luke-jr> da2ce7: NO
 67 2012-01-10 00:20:53 <da2ce7> luke-jr: have I missed something?
 68 2012-01-10 00:21:18 <luke-jr> da2ce7: CA model is flawed
 69 2012-01-10 00:21:25 <k9quaint> you can disguise encrypted communication
 70 2012-01-10 00:21:44 <k9quaint> and yes, CA model is an oxymoron
 71 2012-01-10 00:23:04 <da2ce7> k9quaint: what I'm proposeing is developing a proticol where to open a encripted channel; you must have a goverement approved certificate.  So if the goveremnt can list what CA's are trusted... we can make routers that check the hedder of the connection to see if it was made by a valid CA certificate.
 72 2012-01-10 00:23:12 <da2ce7> unless I'm missing something.
 73 2012-01-10 00:23:50 <k9quaint> encode an encrypted channel inside of an unencrypted one, now your protocol is useless
 74 2012-01-10 00:24:00 <sipa> ?
 75 2012-01-10 00:24:24 <da2ce7> k9quaint: well what happens when they ban all commnication that isn't encripted.
 76 2012-01-10 00:24:50 <k9quaint> I encrypt using their protocol, and embed my own protocol in the payload
 77 2012-01-10 00:24:58 <k9quaint> prove that I have not
 78 2012-01-10 00:25:21 <k9quaint> what I am transmitting is logs of radio static from the upper atmosphere
 79 2012-01-10 00:26:18 <sipa> without a pre-shared secret, and they knowing your protocol, they can mitm
 80 2012-01-10 00:26:39 <da2ce7> aka, facebook, google, youtube, skype, etc will all work... but p2p or bitcoin will not.
 81 2012-01-10 00:27:04 <cjdelisle> disguise your traffic as skype
 82 2012-01-10 00:27:17 <cjdelisle> which is not hard since skype is disguised as /dev/urandom
 83 2012-01-10 00:27:57 <da2ce7> cjdelisle: yeah... but with propper entroypy encoding, you can detect if you are passing voice or encripted data.
 84 2012-01-10 00:28:23 <da2ce7> I think that it wouldn't be too hard to add a server side skype pluggin that checks that you are not passing random data.
 85 2012-01-10 00:28:32 <gmaxwell> if you don't want high data rate you can hide things undetectably.
 86 2012-01-10 00:28:51 <gmaxwell> <3 <3 perturbed quantization stenography.
 87 2012-01-10 00:28:58 <gmaxwell> er .hahahha
 88 2012-01-10 00:29:14 <gmaxwell> steganography*
 89 2012-01-10 00:29:34 <da2ce7> gmaxwell: well technicaly you need two inderpendant channels, and both you send the data under the noise floor, and cancel out the background to get the data.
 90 2012-01-10 00:29:42 <cjdelisle> One silly idea I kind of like is to hide your handshake with a proof of work problem.
 91 2012-01-10 00:29:43 <da2ce7> so either channel inderpendantly has no data.
 92 2012-01-10 00:29:55 <sipa> da2ce7: data below the noise floor is still detectable
 93 2012-01-10 00:29:58 <gmaxwell> da2ce7: no no .. hah PQsteg solves that.. its sooo beautiful.
 94 2012-01-10 00:29:58 <k9quaint> gmaxwell and I have already preshared a secret
 95 2012-01-10 00:30:09 <gmaxwell> (well, the PQ part doesn't but the wet paper codes do)
 96 2012-01-10 00:30:17 <sipa> gmaxwell: link?
 97 2012-01-10 00:30:24 <cjdelisle> so you need to do like 16 bits of work to decrypt the handshake
 98 2012-01-10 00:30:34 <cjdelisle> which is not a big deal if you're expecting one
 99 2012-01-10 00:30:36 <gmaxwell> sipa: have papers access? http://dl.acm.org/citation.cfm?id=1022435
100 2012-01-10 00:30:36 <sipa> k9quaint: with a preshared secret, you can
101 2012-01-10 00:30:51 <cjdelisle> but if you're a DPI box handling gigabits of packets, not going to happen
102 2012-01-10 00:31:04 <gmaxwell> (if you don't have access, I have the paper someplace, just ask)
103 2012-01-10 00:31:08 <da2ce7> sipa: if it is below the noise floor how is it detectible, well assming you are missing random data (noise), with data (encripted data)?
104 2012-01-10 00:31:12 <k9quaint> sipa: ours is encoded in splatter patterns on each others backs!
105 2012-01-10 00:31:17 <gmaxwell> da2ce7: see paper.
106 2012-01-10 00:31:35 <sipa> da2ce7: e.g. deep space communication works using data below the noise floor
107 2012-01-10 00:31:57 <sipa> they use for example a 5000-bit pseudorandom pattern for a 1, and another one for a 0
108 2012-01-10 00:32:00 <gmaxwell> da2ce7: though an easy way of understanding it is this. I'm going to write you a message by sending you a picture. Should you attack today or not.
109 2012-01-10 00:32:18 <sipa> da2ce7: the energy per time for the data is way lower than the energe per time for the space noise
110 2012-01-10 00:32:20 <gmaxwell> da2ce7: you compute the parity of the message I send you, and if its 0 you don't attack.
111 2012-01-10 00:32:21 <sipa> at the receiver
112 2012-01-10 00:32:38 <gmaxwell> da2ce7: I'm not free to flip any bit I want to make the parity correct.
113 2012-01-10 00:33:08 <gmaxwell> da2ce7: wet paper codes extends this to multibit messages using general block linear error correcting codes.
114 2012-01-10 00:33:49 <gmaxwell> And the concept of perturbed quantization is that the sender does lossy compression of the cover message (e.g. jpeg) and chooses the bits to flip based on the values that are nearest the quantization thresholds.
115 2012-01-10 00:34:07 <cjdelisle> that is awesome
116 2012-01-10 00:34:18 <gmaxwell> so that the noise from the stego may be tens of dB below the quantization noise floor.
117 2012-01-10 00:35:09 <sipa> gmaxwell: got the paper
118 2012-01-10 00:35:20 <da2ce7> ok. I really have to go!! Gotta plane to catch!! have fun guys!
119 2012-01-10 00:35:32 <k9quaint> don't crash it into any of our buildings....
120 2012-01-10 00:35:36 <k9quaint> damn, he already left
121 2012-01-10 00:36:09 <gmaxwell> the only attack I'm aware of is that by using a more powerful analysis than the compresson algorithim you might be able to reconstruct a plausable image and find that there is more noise than expected from correct quantization.
122 2012-01-10 00:36:38 <gmaxwell> But even thats hard to distinguish from an R/D optimizing compressor. (which uses quantization threshold biasing in order to improve compression)
123 2012-01-10 00:37:31 <cjdelisle> they could compress the image more, or scramble the low bits a little bit and it would likely destroy your data
124 2012-01-10 00:37:55 <sipa> i guess you could do something like only bias the quantization if it is close enough to 0.5
125 2012-01-10 00:38:23 <gmaxwell> sipa: yes. Right, but for any definition of close enough an attacker with better statistics could find you.
126 2012-01-10 00:39:04 <gmaxwell> One thought I had is, of course you perform the same reconstruction yourself and only use bits where you are near threshold _and_ the reconstruction isn't that helpful. (mostly avoiding smooth areas)
127 2012-01-10 00:39:21 <gmaxwell> cjdelisle: yes, if they add any distortion the message is destroyed.
128 2012-01-10 00:39:57 <cjdelisle> yea if you use ECC to protect it you might be able to get something out of it but they can just distort it more
129 2012-01-10 00:39:57 <sipa> e.g. make 0-0.4 always 0, 0.4-0.6 0 or 1 based on whatever you're encoding, 0.6-1.0: always 1
130 2012-01-10 00:40:11 <sipa> anyway, indeed, it can always be beaten by better statistics
131 2012-01-10 00:40:19 <k9quaint> use ROT26, its uncrackable
132 2012-01-10 00:40:55 <gmaxwell> sipa: right you can be as tight as you want.. you can also take all the values in the image sort them by nearness to 0.5 and take the best N. where N depends on your message size.
133 2012-01-10 00:41:13 <gmaxwell> (at least, you can for jpeg.. other formats, not so much sadly)
134 2012-01-10 00:41:33 <sipa> depending on how little information you need to store, you can theoretically always beat any distortion
135 2012-01-10 00:41:44 <luke-jr> wow
136 2012-01-10 00:41:50 <kam1l> silly question
137 2012-01-10 00:41:50 <luke-jr> GitHub is ignoring a DMCA takedown
138 2012-01-10 00:41:51 <kam1l> bitcoind
139 2012-01-10 00:41:55 <gmaxwell> yes.
140 2012-01-10 00:41:57 <kam1l> does it have to be gen=1 to mine?
141 2012-01-10 00:42:03 <kam1l> like, through pool
142 2012-01-10 00:42:08 <sipa> if you want it to my itself, you need gen=1
143 2012-01-10 00:42:16 <gmaxwell> luke-jr: hm? they normally require you to file it formally. they did for diablo-d3 when someone put up a closed copy of his miner with a trojan.
144 2012-01-10 00:42:16 <sipa> in any other case, leave it off
145 2012-01-10 00:42:18 <kam1l> I don't want my server to be trying to generate
146 2012-01-10 00:42:28 <luke-jr> kam1l: mining on a pool doesn't require a Bitcoin client at all
147 2012-01-10 00:42:33 <kam1l> haven't gotten a block yet so I don't know if its properly connected
148 2012-01-10 00:42:49 <kam1l> its using RPC -> json connection between the pool and the client
149 2012-01-10 00:42:51 <luke-jr> gmaxwell: I followed the directions on their website.
150 2012-01-10 00:43:05 <luke-jr> gmaxwell: now they're claiming they didn't get it, after I had to revise it with an exact link to SC's repo
151 2012-01-10 00:44:53 <k9quaint> the irony is, if he reverts SC2 to the MIT license, it makes the altcoin stronger
152 2012-01-10 00:44:56 <gmaxwell> when I checked before they weren't even formally compliant with the dmca and so not elegable for the safe harbor in any case. (you _must_ have a designated agenet registered with the copyright office)
153 2012-01-10 00:45:12 <gmaxwell> luke-jr: oh you're making trouble for SC? meh..
154 2012-01-10 00:45:34 <k9quaint> at this point, I think any change to SC2 will improve it
155 2012-01-10 00:47:07 <k9quaint> if there was some way to partition the mining space, you could do lots of interesting things to defeat the 51% attack
156 2012-01-10 00:47:15 <k9quaint> oops, mt
157 2012-01-10 00:47:42 <gmaxwell> k9quaint: if you could do that you'd probably not need a block chain consensus at all. Just give one miner each one vote.
158 2012-01-10 00:47:55 <k9quaint> yes, that was essentially my point
159 2012-01-10 00:48:25 <gmaxwell> Most of the neat and effective solutions moot the chain entirely if they actually work you could just use them directly.
160 2012-01-10 00:48:37 <k9quaint> its the fact that the mining space is essentially random, and thus cannot be effectively partitioned
161 2012-01-10 00:49:10 <gmaxwell> no, it can be partitioned, just just that the partitions aren't useful because they don't follow trust boundaries.
162 2012-01-10 00:49:32 <gmaxwell> e.g. you can partition it but not in a way that puts the badguy into one box.
163 2012-01-10 00:49:36 <k9quaint> well, no partitions map usefully on top of human behavior
164 2012-01-10 00:49:42 <gmaxwell> right.
165 2012-01-10 00:49:55 <k9quaint> it makes it hard to be a racist, which is very vexing :(
166 2012-01-10 00:50:04 <gmaxwell> 0_o
167 2012-01-10 00:50:21 <k9quaint> fucking spaniards and their keepaway style of work cup play
168 2012-01-10 00:50:31 <k9quaint> burn them with fire
169 2012-01-10 00:50:39 <k9quaint> *world
170 2012-01-10 00:52:14 <k9quaint> on the other hand, Klinsman coaching the US team is very interesting
171 2012-01-10 00:53:09 <kam1l> what does increasing the number of connections for bitcoind do?
172 2012-01-10 00:53:11 <k9quaint> +n
173 2012-01-10 00:53:37 <k9quaint> past a certain threshold, not much
174 2012-01-10 00:53:40 <gmaxwell> kam1l: makes it use more memory. :)
175 2012-01-10 00:53:52 <kam1l> if I have alot of getworks, like tons
176 2012-01-10 00:53:55 <kam1l> but not many shares
177 2012-01-10 00:53:59 <kam1l> does that mean its bottlenecking?
178 2012-01-10 00:54:14 <gmaxwell> what poolserver software are you using?
179 2012-01-10 00:54:22 <kam1l> pushpool
180 2012-01-10 00:54:59 <gmaxwell> thats odd. I recommend switching to testnet to test your setup.
181 2012-01-10 00:55:11 <gmaxwell> (it has a difficulty of 35 so you'll find blocks must faster that way)
182 2012-01-10 00:55:19 <kam1l> kk
183 2012-01-10 00:55:21 <gmaxwell> kam1l: how much hash rate do you have?
184 2012-01-10 00:55:21 <kam1l> wil ldo
185 2012-01-10 00:57:07 <gmaxwell> I also strongly recommend you look at https://en.bitcoin.it/wiki/P2Pool  it's equal to or easier to maintain than a pushpool setup, and you some pooling benefits without the weaknesses. (p2pool is effectively solomining but with a zero trust agreement on shared payments)
186 2012-01-10 00:58:00 <k9quaint> my mind is not yet made up about p2pool
187 2012-01-10 00:58:08 <kam1l> saying like 85
188 2012-01-10 00:58:18 <kam1l> turne don testnest
189 2012-01-10 00:58:20 <kam1l> testnet*
190 2012-01-10 00:58:34 <kam1l> I think it should be alot more
191 2012-01-10 00:59:03 <gmaxwell> k9quaint: whats to not make up about it? It can't really get variance as low as centerlized pools... but there isn't much else negative you could say as far as I know.
192 2012-01-10 00:59:20 <k9quaint> gmaxwell: i haven't put a lot of thought into how to attack it
193 2012-01-10 00:59:28 <luke-jr> gmaxwell: I thought up a new reward scheme for p2pool&
194 2012-01-10 00:59:59 <gmaxwell> k9quaint: it's a blockchain, though somewhat stronger than most because it time locks against the bitcoin chain.
195 2012-01-10 01:00:02 <luke-jr> gmaxwell: when a block is found, instead of paying the last N people, pay the people with highest N shares
196 2012-01-10 01:00:23 <luke-jr> gmaxwell: (they then "lose" those shares&)
197 2012-01-10 01:00:52 <shadders> doesn't it have a much higher stale rate due to p2p latency?
198 2012-01-10 01:00:54 <luke-jr> additionally, continue to expire excessively old shares
199 2012-01-10 01:01:08 <luke-jr> shadders: I don't think so?
200 2012-01-10 01:01:17 <gmaxwell> if people were unusually luck and there was a big pile of high shares, then you'd be discouraged to join
201 2012-01-10 01:01:29 <gmaxwell> shadders: no. Not the kind of stale you mean.
202 2012-01-10 01:01:36 <luke-jr> gmaxwell: that's why you expire shares eventually
203 2012-01-10 01:01:49 <shadders> oh... I always assumed that was what was holding it back from mass uptake... lower mining efficiency...
204 2012-01-10 01:01:58 <gmaxwell> shadders: the bitcoin part has excellent latency because you have a bitcoin node right there.
205 2012-01-10 01:02:17 <luke-jr> gmaxwell: I presume p2pool's block chain has forks and merges?
206 2012-01-10 01:03:03 <gmaxwell> shadders: nah. You _do_ get high stales on the share chain because it has a 10 second mean but everyone gets high stales on it, and what counts is the relative performance.
207 2012-01-10 01:03:22 <shadders> it's been ages since I looked at it but I seem to recall thinking there was something in the design that would result in lower income than fast central pool...
208 2012-01-10 01:03:44 <shadders> gmaxwell: I see... so if everyone uses it there's no disadvatage
209 2012-01-10 01:03:46 <gmaxwell> The design changed from the original one.
210 2012-01-10 01:04:30 <shadders> but since it's competing with central pools performance is relative to those as well...
211 2012-01-10 01:04:32 <gmaxwell> shadders: really big p2pools have the problem of high share difficulty, but you solve that by just breaking up p2pools into smaller ones more like the sizes of current central pools.
212 2012-01-10 01:04:39 <gmaxwell> shadders: no, its not.
213 2012-01-10 01:05:00 <gmaxwell> shadders: the share chain is its own thing. Your share being stale for the share chain doesn't make it stale for bitcoin.
214 2012-01-10 01:05:21 <luke-jr> gmaxwell: fail
215 2012-01-10 01:05:26 <gmaxwell> So if you did find a block it would still happily be accepted, potentially better than a centeral pool because every p2pool user has a local node.
216 2012-01-10 01:05:46 <gmaxwell> 2012-01-09 21:03:06.817681 Pool stales: 8% Own: 9???1% Own efficiency: 99???1%
217 2012-01-10 01:06:13 <luke-jr> gmaxwell: that's because p2pool is small
218 2012-01-10 01:06:21 <gmaxwell> Thats my numbers on p2pool, so it's 98-100 with 95% confidence.
219 2012-01-10 01:06:21 <luke-jr> gmaxwell: it gets exponentially worse I think
220 2012-01-10 01:06:29 <shadders> hmm.. can't really argue when I don't know what I'm talking about...
221 2012-01-10 01:06:32 <luke-jr> gmaxwell: why can't p2pool allow blockchain merges?
222 2012-01-10 01:06:37 <gmaxwell> luke-jr: the share difficulty goes up linearly.
223 2012-01-10 01:07:24 <luke-jr> makomk: I'd really be interested in if my proposal solves CLC's problem btw
224 2012-01-10 01:07:33 <shadders> simple question then... why isn't it taking bulk market share from centralized pools?
225 2012-01-10 01:07:40 <k9quaint> the thing I like most about pools is I initiate the conversation and don't need to open ports to the scary scary world
226 2012-01-10 01:07:45 <gmaxwell> shadders: it's new and immature.
227 2012-01-10 01:07:48 <luke-jr> shadders: why isn't DeepBit dying?
228 2012-01-10 01:07:52 <gmaxwell> and it's growing like ganbusters.
229 2012-01-10 01:08:24 <shadders> it's hardly new... I remember hearing about when I first discovered bitcoin...
230 2012-01-10 01:08:38 <shadders> what's the hashrate total?  can it be measured?
231 2012-01-10 01:08:45 <gmaxwell> shadders: a week and a few days ago it had about 13GH.. it has 70GH now.
232 2012-01-10 01:08:54 <sipa> wow
233 2012-01-10 01:08:58 <BlueMatt> nice
234 2012-01-10 01:09:11 <gmaxwell> shadders: it didn't even process transactions until about a month or so ago.
235 2012-01-10 01:09:20 <gmaxwell> How could anyone really advocate it then?
236 2012-01-10 01:09:35 <luke-jr> does it support Windows?
237 2012-01-10 01:09:37 <luke-jr> easily?
238 2012-01-10 01:10:03 <gmaxwell> luke-jr: yes. But windows users are somewhat outside of the ideal market for it.. e.g. you need to run a full bitcoin node and start a program from the commandline.
239 2012-01-10 01:10:05 <shadders> I thought luke-jr didn't believe windows exists?
240 2012-01-10 01:10:22 <shadders> or that just java :p
241 2012-01-10 01:10:37 <luke-jr> shadders: shouldn't*
242 2012-01-10 01:11:12 <gmaxwell> BlueMatt: to be fair, something like 40GH of that is three miners right now.
243 2012-01-10 01:11:29 <k9quaint> the truth comes out!
244 2012-01-10 01:11:48 <k9quaint> I need to start a pitchfork and torch selling business online
245 2012-01-10 01:11:49 <BlueMatt> gmaxwell: oh...
246 2012-01-10 01:11:49 <gmaxwell> though the rest is smaller miners.. median of about 1.1 GH or so.
247 2012-01-10 01:11:57 <BlueMatt> well doesnt it take like 2 percent?
248 2012-01-10 01:12:07 <gmaxwell> No. Zero fee. You can optionally donate.
249 2012-01-10 01:12:16 <BlueMatt> oh, ok
250 2012-01-10 01:12:27 <gmaxwell> Though right now people are donating _to_ p2pool users.
251 2012-01-10 01:12:36 <BlueMatt> what?
252 2012-01-10 01:12:42 <luke-jr> gmaxwell: does p2pool currently share the txn fees, or does the block finder keep all?
253 2012-01-10 01:12:48 <k9quaint> to encourage decentralization
254 2012-01-10 01:12:56 <luke-jr> BlueMatt: there's a builtin sendmany-to-last-N-shares
255 2012-01-10 01:13:13 <gmaxwell> luke-jr: shares them .. technically it shares 99.5 percent of the total, with .5 going to the finder to discourage witholding.
256 2012-01-10 01:13:50 <luke-jr> gmaxwell: then why should I accept any transactions?
257 2012-01-10 01:14:02 <BlueMatt> luke-jr: so donating just means you dont post your address when you submit a share?
258 2012-01-10 01:14:20 <luke-jr> BlueMatt: no, I think it actually does a sendmany
259 2012-01-10 01:14:27 <gmaxwell> luke-jr: hm? you'll get paid for them, though its shared.. and its good for bitcoin. If it became a problem in the future it could require it.
260 2012-01-10 01:14:44 <luke-jr> but that WOULD be a good idea for a % donation
261 2012-01-10 01:14:50 <gmaxwell> BlueMatt: you can make a http get to the pool and it gives you a sendmany commandline basically.
262 2012-01-10 01:15:13 <gmaxwell> https://en.bitcoin.it/wiki/P2Pool#Donating_to_P2Pool_miners
263 2012-01-10 01:15:16 <BlueMatt> gmaxwell: oh, so people are mining specifically and sending money to the pool?
264 2012-01-10 01:15:27 <BlueMatt> or just sending coins
265 2012-01-10 01:16:07 <gmaxwell> just sending coins (well some of the people doing are p2pool users, and it seems some people are just bystanders who want to see it grow)
266 2012-01-10 01:16:46 <josephcp> from what i understand, that payout function has nothing to do with the way p2pool works though
267 2012-01-10 01:17:11 <gmaxwell> josephcp: What payout function?
268 2012-01-10 01:17:19 <josephcp> the donation
269 2012-01-10 01:17:31 <gmaxwell> josephcp: which donation?
270 2012-01-10 01:17:42 <josephcp> your link
271 2012-01-10 01:17:46 <gmaxwell> The 'donate to p2pool author' or the sendmany-to-the-miners?
272 2012-01-10 01:18:05 <luke-jr> gmaxwell: also, that reward system I suggested a few mins ago, p2pool could have a builtin "suggest old share expiracy change", with a vote in the p2pool chain ;)
273 2012-01-10 01:18:10 <gmaxwell> It does thats computed using the sharechain it's basically the current reward p2pool would pay out scaled to whatever amount you want to pay.
274 2012-01-10 01:18:14 <luke-jr> gmaxwell: or heck, builtin polls for ANYTHING :P
275 2012-01-10 01:18:38 <josephcp> oh yeah looks like it
276 2012-01-10 01:19:17 <gmaxwell> luke-jr: sounds like a cool idea. It's all in python you know. :)
277 2012-01-10 01:19:19 <luke-jr> gmaxwell: make all the p2pool config variables, and automate changing them by hashrate poll
278 2012-01-10 01:19:46 <luke-jr> gmaxwell: yes, but I'm with Eligius until it dies ;p
279 2012-01-10 01:19:52 <gmaxwell> luke-jr: what you should really do is vote on a bitcoin address which you delegate to choose for you. :)
280 2012-01-10 01:20:13 <luke-jr> gmaxwell: maybe.
281 2012-01-10 01:20:20 <luke-jr> yes, then you can instantly decide stuff
282 2012-01-10 01:20:22 <luke-jr> good idea
283 2012-01-10 01:20:38 <luke-jr> open the config, set your preferences, then leave it
284 2012-01-10 01:20:48 <luke-jr> the network adjusts based on the consensus of active configs
285 2012-01-10 01:20:54 <luke-jr> that'd be pretty cool
286 2012-01-10 01:21:15 <gmaxwell> Well, delegation would also be good so if I have 8 p2pool instances I don't need to configure all of them.
287 2012-01-10 01:21:35 <luke-jr> even better, if p2pool would segregate itself automatically when people disagree too much
288 2012-01-10 01:21:45 <gmaxwell> luke-jr: all these things are possible.
289 2012-01-10 01:21:48 <luke-jr> so miners can set "I prefer 50, but will tolerate down to 25 and up to 75"
290 2012-01-10 01:22:09 <luke-jr> and if the consensus would make it 24, you get split off into a new p2pool, and things recalculate based on that&
291 2012-01-10 01:22:26 <luke-jr> merging might be difficult tho
292 2012-01-10 01:23:20 <josephcp> merging wouldn't be that big of a deal if you can mine in "multiple p2pools" at a time ithink?
293 2012-01-10 01:23:43 <luke-jr> where is Gavin today? x.x
294 2012-01-10 01:23:46 <luke-jr> BlueMatt: ping
295 2012-01-10 01:24:11 <gmaxwell> josephcp: sure.
296 2012-01-10 01:37:21 <kam1l> if a pool is working (and it uses json-RPC with bitcoind)
297 2012-01-10 01:37:32 <kam1l> would the command bitcoind getinfo display the current hashrate of the server in total?
298 2012-01-10 01:37:36 <sipa> no
299 2012-01-10 01:37:40 <kam1l> didn't think so
300 2012-01-10 01:37:42 <sipa> it can't know
301 2012-01-10 01:38:15 <gmaxwell> iirc one of the http headers miners send includes their claimed hashrate. :) but it doesn't use it (I don't think anything does)
302 2012-01-10 01:38:58 <luke-jr> sipa: well, if it did share targets it could
303 2012-01-10 01:39:49 <kam1l> hmm, well I have a big fpga farm that my work uses
304 2012-01-10 01:39:55 <kam1l> and I have a ton of workers for each one
305 2012-01-10 01:40:04 <kam1l> but I am getting very few shares, yet loads of getworks
306 2012-01-10 01:40:14 <kam1l> back to the drawing board I guess
307 2012-01-10 01:42:01 <luke-jr> kam1l: solo mining doesn't use ANY shares
308 2012-01-10 01:42:08 <kam1l> not solo mining
309 2012-01-10 01:42:13 <kam1l> I have a pushpool pool server setup
310 2012-01-10 01:42:18 <gmaxwell> luke-jr: he's using pushpool.
311 2012-01-10 01:42:24 <gmaxwell> kam1l: what kinds of fpgas?
312 2012-01-10 01:42:26 <kam1l> I have it all running, no errors
313 2012-01-10 01:42:29 <kam1l> great question
314 2012-01-10 01:42:33 <kam1l> I don't know, just a co-op student
315 2012-01-10 01:42:44 <kam1l> this was supposed to be an easy setup :P
316 2012-01-10 01:42:52 <gmaxwell> How did you manage to get fpga mining working without knowing the FPGA type?
317 2012-01-10 01:43:01 <kam1l> I didn't setup the miners :P
318 2012-01-10 01:43:04 <gmaxwell> ah!
319 2012-01-10 01:43:08 <kam1l> someone who actually knows what hes doing did
320 2012-01-10 01:43:15 <kam1l> they just told me to get a cheap pool setup
321 2012-01-10 01:43:26 <gmaxwell> well, they might be broken... or just slower than you were expecting.
322 2012-01-10 01:43:31 <kam1l> well
323 2012-01-10 01:43:35 <kam1l> if by slower you mean like
324 2012-01-10 01:43:39 <kam1l> less than 100 mhash/s
325 2012-01-10 01:43:40 <gmaxwell> (by broken I mean running at too high a clock rate to work correctly)
326 2012-01-10 01:43:41 <kam1l> TOTAL
327 2012-01-10 01:43:49 <kam1l> could be
328 2012-01-10 01:43:54 <gmaxwell> kam1l: most inexpensive fpgas put out fairly little.
329 2012-01-10 01:43:55 <kam1l> pushpool gives no errors
330 2012-01-10 01:44:02 <kam1l> I get tons of getworks
331 2012-01-10 01:44:15 <kam1l> and like, 1 share a minute
332 2012-01-10 01:44:29 <gmaxwell> yes, thats suggestive of it just being slow. (or perhaps you don't have it all running?)
333 2012-01-10 01:45:07 <gmaxwell> in any case, at that rate you'd certantly be better off with p2pool or a traditional centeralized pool like eligius.
334 2012-01-10 01:45:26 <kam1l> I have alot of workers, making them by hand would be very tedious
335 2012-01-10 01:45:34 <kam1l> I assumed that each fpga would want its own worker
336 2012-01-10 01:46:49 <luke-jr> its own work, yes
337 2012-01-10 01:46:56 <luke-jr> but they can all be on the same pool worker
338 2012-01-10 01:47:07 <kam1l> really
339 2012-01-10 01:47:12 <kam1l> I assumed it would bottleneck
340 2012-01-10 01:47:32 <gmaxwell> it's own worker? meh no you can just point them all at the same address. At least on eligius, god knows if you'd trigger some anti-dos on some other pool but I doubt it.
341 2012-01-10 01:48:02 <gmaxwell> (eligius doesn't even have accounts, you just set the address you want paid as your username and put anything in the password field)
342 2012-01-10 01:51:49 <kam1l> well I guess I know where I'm going then
343 2012-01-10 02:05:18 <luke-jr> where is BlueMatt? -.-
344 2012-01-10 02:05:21 <luke-jr> and Gavin
345 2012-01-10 02:05:42 <luke-jr> tcatm: are you around if we get the builds finished tonight?
346 2012-01-10 02:08:53 <finway> p2pool is funny,  everybody's solo mining?
347 2012-01-10 02:09:53 <gmaxwell> finway: yes, basically solomining with a consensus payout, and attempted blocks with the payout spec count as shares to earn you a place in the consensus.
348 2012-01-10 02:10:37 <finway> What if somebody keep the bingo share?
349 2012-01-10 02:11:06 <luke-jr> finway: *why*?
350 2012-01-10 02:11:19 <finway> somebody find a block, and don't share the 50BTCs
351 2012-01-10 02:11:58 <gmaxwell> finway: same thing that prevent them on any pool.
352 2012-01-10 02:12:22 <gmaxwell> finway: if you're trying to only pay yourself your work won't credit you for the pool.
353 2012-01-10 02:12:34 <gmaxwell> you can't change it after the fact because the payout spec is part of the block.
354 2012-01-10 02:13:08 <finway> That's cool
355 2012-01-10 02:13:56 <finway> But i think since it's p2p, there's no OP, it's harder to prevent this kind of scam
356 2012-01-10 02:14:13 <gmaxwell> I think I am king of the moon.
357 2012-01-10 02:14:16 <gmaxwell> :)
358 2012-01-10 02:14:33 <finway> heh
359 2012-01-10 02:14:35 <gmaxwell> finway: there isn't any need for a 'pool op' to prevent that, it's just fundeimentally impossible.
360 2012-01-10 02:14:59 <gmaxwell> fundamentally too.
361 2012-01-10 02:15:21 <gmaxwell> And if it were possible, it would also work against central pools, and it's not like they're super able to stop bad behavior with proxies and such existing.
362 2012-01-10 02:15:34 <gmaxwell> p2pool doesn't have operators but it has developers and users.
363 2012-01-10 02:15:58 <luke-jr> finway: it works the same way Bitcoin itself deos
364 2012-01-10 02:16:00 <luke-jr> does*
365 2012-01-10 02:16:34 <luke-jr> finway: for every valid share, you earn a "point" in the merged-mine blockchain
366 2012-01-10 02:16:35 <finway> I'll try to understand it.
367 2012-01-10 02:16:48 <luke-jr> finway: valid shares need to include payments to all the people with points in that blockchain
368 2012-01-10 02:18:48 <finway> luke-jr: since you're an OP, i think i should stop worrying. :P
369 2012-01-10 02:19:16 <luke-jr> &
370 2012-01-10 02:20:18 <gmaxwell> finway: The real problems wit p2pool are that it take a bit more work to setup an run than just using a central pool, and that it can't achieve variance as low, or payout schemes as flexible, plus its less mature than other options.
371 2012-01-10 02:21:55 <finway> p2pool is the best thing happen to bitcoin in 2011
372 2012-01-10 02:23:28 <amiller> roconnor, i'd totally recommend taking a look at this set of course notes http://www.seas.upenn.edu/~cis500/current/sf/html/toc.html
373 2012-01-10 02:23:59 <roconnor> amiller: why is that?
374 2012-01-10 02:24:20 <amiller> it is a tour through verified programming using Coq basically
375 2012-01-10 02:24:38 <gmaxwell> **blinks**
376 2012-01-10 02:24:43 <amiller> i'm going through it for the second time basically
377 2012-01-10 02:24:48 <luke-jr> finway: let me add it to my list&
378 2012-01-10 02:25:16 <luke-jr> https://bitcointalk.org/index.php?topic=57325.0
379 2012-01-10 02:25:18 <amiller> it's very haskelly, so i think you'd probably get a lot out of it pretty easily... whereas it takes me a lot of difficult thought right now
380 2012-01-10 02:25:20 <luke-jr> there we go
381 2012-01-10 02:25:27 <roconnor> amiller: okay, but you gotta read my thesis then: http://r6.ca/Thesis/
382 2012-01-10 02:25:35 <gmaxwell> amiller: http://www.informatik.uni-trier.de/~ley/db/indices/a-tree/o/O=Connor:Russell.html
383 2012-01-10 02:26:10 <amiller> oh
384 2012-01-10 02:26:14 <amiller> quite.
385 2012-01-10 02:26:25 <copumpkin> roconnor: too many publications
386 2012-01-10 02:26:26 <gmaxwell> roconnor: I heard about this language you might like, it has a funny name. But it's just french for chicken so it's okay. :)
387 2012-01-10 02:26:35 <roconnor> amiller: ;)
388 2012-01-10 02:28:08 <roconnor> amiller: don't sweat it.  I appreciate it.
389 2012-01-10 02:28:16 <gmaxwell> hehe
390 2012-01-10 02:28:52 <finway> I can imagine the scenario:  lots of p2pool variations, the one support 32coins mergeming wins, cause to miners, more is better.
391 2012-01-10 02:28:58 <finway> Is that right?
392 2012-01-10 02:29:13 <gmaxwell> finway: what does wins mean?
393 2012-01-10 02:29:14 <finway> But the price was defined by NEEDs.
394 2012-01-10 02:29:14 <luke-jr> finway: no
395 2012-01-10 02:29:45 <finway> "win" mean get most mining power.
396 2012-01-10 02:30:13 <gmaxwell> finway: it seems that most miners don't care about merged mining, see also Deepbit.
397 2012-01-10 02:30:43 <finway> Yes, that's odd. I don't get it.
398 2012-01-10 02:30:54 <gmaxwell> finway: I would expect p2pools to bet geography / network topology scoped more than features in the future but we'll see.
399 2012-01-10 02:31:23 <gmaxwell> finway: btcguild lost ~no hashrate when it when from 0 fee to 5% fee (pps).
400 2012-01-10 02:31:49 <roconnor> I'm pretty impressed how well known Coq is.
401 2012-01-10 02:32:25 <luke-jr> gmaxwell: maybe when the CLC crap cools down, I should do an experiment and announce a 5% fee on Eligius
402 2012-01-10 02:32:30 <gmaxwell> finway: You may have mistaken Homo mygpumakesmemoniez with Homo economicus
403 2012-01-10 02:32:30 <luke-jr> gmaxwell: just for a week or smth
404 2012-01-10 02:32:57 <gmaxwell> luke-jr: use a signmessage thing to vote for the fee with a default of 5% if you don't get a majority for another option. :)
405 2012-01-10 02:33:02 <amiller> that's awesome, you've been making some fundamental contributions constructive mathematics since 2005
406 2012-01-10 02:33:22 <luke-jr> gmaxwell: I'm sure it'd end up 0 that way
407 2012-01-10 02:33:32 <roconnor> amiller: I don't know about fundamental, but I try.
408 2012-01-10 02:33:35 <luke-jr> gmaxwell: oh, you mean everyone votes 5% by default?
409 2012-01-10 02:33:39 <gmaxwell> luke-jr: yes.
410 2012-01-10 02:33:42 <luke-jr> hmm
411 2012-01-10 02:33:52 <luke-jr> gmaxwell: except right now, signmessage is only needed for NMC
412 2012-01-10 02:33:56 <luke-jr> ie, not many people
413 2012-01-10 02:34:32 <finway> I guess most miners are lazy, and scam coin does not worth changing... fee too... even nmc too...
414 2012-01-10 02:35:01 <gmaxwell> luke-jr: though I'd rather you left it alone until p2pool matures enough that you could just make it a p2pool frontend with enough fees to cover the increased stales from having high latency miners on p2pool. I think that would be a more worthwhile way to blow it up.
415 2012-01-10 02:35:09 <gmaxwell> (well, if indeed it blow it up)
416 2012-01-10 02:35:21 <gmaxwell> er blows
417 2012-01-10 02:36:08 <luke-jr> gmaxwell: ?
418 2012-01-10 02:36:16 <gmaxwell> roconnor: extractions from coq are very cool, hard to not remember that.
419 2012-01-10 02:36:40 <luke-jr> gmaxwell: remember, this is to promote Tonal. I don't plan to destroy Eligius.
420 2012-01-10 02:36:48 <roconnor> gmaxwell: did you see that talk about a PDF parser in Coq?
421 2012-01-10 02:37:05 <luke-jr> gmaxwell: I might let it die, but I won't actively try to destroy it.
422 2012-01-10 02:37:24 <gmaxwell> roconnor: no.. er.. 0_o as in all of pdf (including the postscript stack based virtual machine?)
423 2012-01-10 02:37:57 <gmaxwell> luke-jr: well I didn't mean that I'm more referring to bold changes which may have that outcome.
424 2012-01-10 02:38:28 <roconnor> gmaxwell: at first I didn't think it was a big deal; just a parser; but then he showed how crafted PDF files with loopy xrefs make your desktop file brower busyloop
425 2012-01-10 02:39:01 <roconnor> gmaxwell: his coq implemention provably terminates so such behaviour in some sense impossible.
426 2012-01-10 02:39:44 <gmaxwell> roconnor: a full pdf decoder has two Turing complete languages in it, plus a bunch of behavior that can use infinite cpu or memory.
427 2012-01-10 02:39:51 <gmaxwell> roconnor: thats pretty cool.
428 2012-01-10 02:40:26 <amiller> there's a really cool 'godels incompleteness' theorem in coq
429 2012-01-10 02:40:32 <amiller> it's on some pastebin, i've run into it a bunch of times
430 2012-01-10 02:40:39 <amiller> maybe i'll find out you wrote it, or at least recognize it
431 2012-01-10 02:40:44 <gmaxwell> amiller: the one written by roconnor ?
432 2012-01-10 02:40:53 <gmaxwell> hah
433 2012-01-10 02:40:59 <doublec> isn't that what is thesis was about?
434 2012-01-10 02:41:20 <amiller> it's the first of two main contributions in it!
435 2012-01-10 02:42:20 <amiller> er it was this: http://muaddibspace.blogspot.com/2009/10/cantors-diagonalization-proof-in-coq.html
436 2012-01-10 02:42:54 <amiller> never figured out who Muad`Dib was
437 2012-01-10 02:43:06 <roconnor> gmaxwell: http://www.reddit.com/r/dependent_types/comments/jm95w/dependent_types_demonstrated_with_a/
438 2012-01-10 02:43:16 <roconnor> gmaxwell: only the last few minutes are about the PDF parser
439 2012-01-10 02:43:25 <roconnor> gmaxwell: the rest of the talk is some sort of introduction to coq
440 2012-01-10 02:50:00 <gmaxwell> amiller: My understanding(/hope) is that roconnor is doing the Haskell implementation of bitcoin as design/precursor to to formally verfied on in coq.
441 2012-01-10 02:50:24 <amiller> this paper looks like it's the fun-size version of the second part of your thesis http://arxiv.org/pdf/1008.1213v2
442 2012-01-10 02:50:40 <amiller> also +1 for the CC clause in the document.
443 2012-01-10 02:51:30 <amiller> yeah that would make sense gmaxwell.
444 2012-01-10 02:51:59 <amiller> i doubt that there would be anything to do wtih bitcoin that's as interesting as constructive real analysis
445 2012-01-10 02:52:31 <amiller> unless you were going to try to formalize 'cryptography' which i've never heard of anyone attepmting
446 2012-01-10 02:53:09 <amiller> anyway um
447 2012-01-10 02:53:43 <amiller> i don't have enough practical experience with haskell (or coq for that matter) to know how valuable the haskell implementation is to get into coq
448 2012-01-10 02:55:02 <amiller> basically what i was considering two options, either 1) port your haskell code to coq, by implementing notations for the subset of haskell you mostly use
449 2012-01-10 02:55:33 <amiller> 2) implement bitcoin by making notations that let me line up with the layout in the bitcoin wiki
450 2012-01-10 02:56:09 <amiller> either way i think i'm being distracted too much by the 'scripting language' and not looking at the other aspects of bitcoin
451 2012-01-10 02:56:38 <gmaxwell> amiller: the documentation in the wiki is _far_ from complete.
452 2012-01-10 02:57:24 <amiller> yeah
453 2012-01-10 02:57:28 <lfm> read the source luke
454 2012-01-10 02:57:40 <luke-jr> no u
455 2012-01-10 02:57:48 <luke-jr> I just *rewrote* the source
456 2012-01-10 02:57:57 <luke-jr> an annoying portion of if anyhow
457 2012-01-10 02:58:16 <luke-jr> https://github.com/ckolivas/cgminer/pull/69/files
458 2012-01-10 02:59:10 <amiller> basically you would either have to be modeling either the satoshi implementation, or perhaps a haskell implementation (seems you've chosen haskell), or perhaps just one in coq to begin wtih (that's what i initially thought i would do)
459 2012-01-10 03:01:34 <amiller> so questions are 1. what's involved in moving from haskell to coq? is there a coq library (a repository of useful lemmas and notations) for working with haskell? 2. at a high level, what do you anticipate would be useful to prove about the specification?
460 2012-01-10 03:03:10 <k9quaint> never mucked about with Haskell
461 2012-01-10 03:13:12 <k9quaint> I wonder what happened to SAC to trigger so much anger
462 2012-01-10 03:13:24 <k9quaint> I suspect a pedobear was involved O.o
463 2012-01-10 03:14:34 <amiller> i guess another question i have roconnor is 3. have you thought of any generalizations of bitcoin that are perhaps simpler to model?
464 2012-01-10 03:14:48 <amiller> for example opentransactions has a much different architecture than bitcoin
465 2012-01-10 03:15:18 <amiller> it doesn't have any of the random proof of work stuff, it's not a decentralized currency with anonymous miners
466 2012-01-10 03:15:28 <amiller> instead it just has servers that sign for it
467 2012-01-10 03:16:17 <amiller> but the security properties you get are that if the server violates its specification (for example, lending its signature to a double-spend) then the public gets evidence that the spec was violated
468 2012-01-10 03:16:51 <luke-jr> k9quaint: presumably he was planning to make good money on the scam
469 2012-01-10 03:17:30 <k9quaint> I think he needs to show us on the doll where the bad man touched him
470 2012-01-10 03:39:04 <midnightmagic> if you move anything to coq, you can basically expect zero incoming patches from outsiders
471 2012-01-10 03:39:17 <luke-jr> lol
472 2012-01-10 03:39:26 <midnightmagic> :-)
473 2012-01-10 03:39:58 <gmaxwell> as opposed to the absolute flood of haskell contributors?
474 2012-01-10 03:40:24 <k9quaint> right it in lisp
475 2012-01-10 03:40:31 <amiller> i don't understand #haskell as a community
476 2012-01-10 03:40:32 <k9quaint> I will contribute a parenthesis or two
477 2012-01-10 03:40:33 <amiller> who are they?
478 2012-01-10 03:40:37 <amiller> i don't think i've ever met one
479 2012-01-10 03:40:46 <amiller> a haskell developer i mean
480 2012-01-10 03:41:06 <gmaxwell> amiller: big chunks of them are locked up in trading firms
481 2012-01-10 03:41:07 <amiller> i've met lisp people on the other hand
482 2012-01-10 03:41:21 <gmaxwell> they don't actually want haskell code, they just want the kind of people who find haskell attractive.
483 2012-01-10 03:41:25 <amiller> oh fuck, you know what that means
484 2012-01-10 03:41:30 <amiller> gotta liberate 'em
485 2012-01-10 03:41:30 <midnightmagic> gmaxwell: there are a lot of us who can build useful haskell patches. Even I can do that. :)
486 2012-01-10 03:41:33 <amiller> occupy haskell
487 2012-01-10 03:41:43 <amiller> occupy quants i suppose you mean
488 2012-01-10 03:41:57 <amiller> hmm
489 2012-01-10 03:42:00 <luke-jr> Perl is the obvious best choice.
490 2012-01-10 03:42:05 <gmaxwell> midnightmagic: will you fix my xmonad configuration for me then? If I ask for more free coding in #xmonad they're going to start catching on. ;)
491 2012-01-10 03:42:14 <luke-jr> & if you want a bunch of antisocial jerks contributing
492 2012-01-10 03:42:35 <luke-jr> mtve: btw, do you have work?
493 2012-01-10 03:42:45 <midnightmagic> gmaxwell: :)
494 2012-01-10 03:42:57 <midnightmagic> gmaxwell: Why are you writing in OCaml? :)
495 2012-01-10 03:43:08 <midnightmagic> ^are^aren't
496 2012-01-10 03:44:04 <amiller> what an odd project
497 2012-01-10 03:44:06 <amiller> xmonad is hillarious
498 2012-01-10 03:44:13 <gmaxwell> It's quite good software.
499 2012-01-10 03:44:39 <amiller> now i know what powers all the ridiculous multimonitor displays i typically associate with financial engineer silliness
500 2012-01-10 03:45:13 <midnightmagic> gmaxwell: Or just switch to dwm, it's easier to patch.
501 2012-01-10 03:46:00 <amiller> luke-jr, are you the only guy who mines nonstandard tx and weird blocks?
502 2012-01-10 03:46:06 <luke-jr> amiller: afaik
503 2012-01-10 03:46:21 <amiller> what's the point? i don't really think i understand what the significance of that is
504 2012-01-10 03:46:23 <midnightmagic> there used to be a group of miners doing that, who'd all agreed to connect directly to one another.
505 2012-01-10 03:46:27 <midnightmagic> who knows what happened to them
506 2012-01-10 03:47:03 <luke-jr> amiller: dunno, anti-"Gavin is the boss of Bitcoin"? :P
507 2012-01-10 03:47:18 <amiller> someone's gotta play the devil's advocate
508 2012-01-10 03:47:25 <luke-jr> I do that pretty well.
509 2012-01-10 03:47:28 <luke-jr> usually
510 2012-01-10 03:47:31 <amiller> better you i suppose than the russian mafia or whatever would fill the power gap.
511 2012-01-10 03:47:40 <gmaxwell> *cough*
512 2012-01-10 03:48:13 <amiller> lol
513 2012-01-10 03:48:23 <luke-jr> gmaxwell is part of the p2pool mafia
514 2012-01-10 03:48:35 <midnightmagic> forrest is russian?
515 2012-01-10 03:52:02 <gmaxwell> luke-jr: p2pool mafia is benevolent, it pays you to be more decentralized.
516 2012-01-10 03:52:17 <luke-jr> gmaxwell: suuuuure
517 2012-01-10 03:52:54 <gmaxwell> Or else.
518 2012-01-10 03:53:26 <amiller> carrots not sticks?
519 2012-01-10 03:53:30 <gmaxwell> (or else, & it doesn't)
520 2012-01-10 03:53:38 <onelineproof> How do I submit little bugs? Mostly just the make files I had to modify them to make them work properly... (bitcoin-qt.pro & makefile.unix)
521 2012-01-10 03:54:17 <nanotube> onelineproof: you can post pull requests on github....
522 2012-01-10 03:54:19 <luke-jr> onelineproof: you shouldn't need to modify them anymore
523 2012-01-10 03:54:58 <gmaxwell> onelineproof: if your modifications are general (e.g. won't just break it for someone else) create a pull request or just file an issue.  (though I'd bet they aren't& make isn't much of an automated build system)
524 2012-01-10 03:55:17 <onelineproof> Ok, Ill get to that sometime, I'm just making some custom builds, so I had to change some library/include directories
525 2012-01-10 03:55:34 <gmaxwell> onelineproof: on what system?
526 2012-01-10 03:56:22 <onelineproof> on ubuntu... I'm making my own package, which will soon be a custom distro
527 2012-01-10 04:07:05 <BlueMatt> luke-jr: linux: http://dl.dropbox.com/u/29653426/bitcoin-0.4.3rc1.tar.xz win32: http://dl.dropbox.com/u/29653426/bitcoin-0.4.3rc1-win32.tar.xz
528 2012-01-10 05:10:28 <heoa> Bug Report with Fresh Ubuntu and no earlier install with Bitcoin: http://pastie.org/3158574
529 2012-01-10 05:11:04 <heoa> I got bitcoin running as showed there but there is some prob with ~/.bitcoin -folder creation (I got it running after removing the dir)
530 2012-01-10 05:11:38 <heoa> (sorry totally newbie to bitcoin but there is the procedure what I did, you surely understand better it. Have fun!)
531 2012-01-10 05:17:22 <heoa> ..or wait, one thing I forgot: I may have done `$ sudo apt-get install bitcoind` before this. Did it intervene with the issue?
532 2012-01-10 05:18:20 <onelineproof> ya bitcoind probably created the first .bitcoin directory
533 2012-01-10 05:18:55 <onelineproof> and possibly dynamically linked to a version of libdb that is not compatible with the one bitcoin-qt uses
534 2012-01-10 05:19:05 <gmaxwell> onelineproof has it.
535 2012-01-10 05:19:46 <onelineproof> To be safe wallet.dat must be created with the same version of libdb
536 2012-01-10 05:20:05 <luke-jr> BlueMatt: O.o;;
537 2012-01-10 05:20:13 <gmaxwell> onelineproof: libdb only goes one direction
538 2012-01-10 05:20:20 <gmaxwell> you can upgrade but not downgrade.
539 2012-01-10 05:20:24 <luke-jr> BlueMatt: so is there a reason we can't rename the last ones, and that to 0.5.2 and 0.4.3? :P
540 2012-01-10 05:20:28 <onelineproof> o ok
541 2012-01-10 05:34:33 <BlueMatt> luke-jr: just remove the sigs and I guess you are fine...
542 2012-01-10 05:35:11 <luke-jr> BlueMatt: why remove the sigs? O.o
543 2012-01-10 05:35:17 <luke-jr> is the filename being signed?
544 2012-01-10 05:35:34 <BlueMatt> luke-jr: because they are at the top level dir and werent in previous versions
545 2012-01-10 05:35:41 <BlueMatt> (plus its not formatted for gitian as-is)
546 2012-01-10 05:35:47 <luke-jr> O.o?
547 2012-01-10 05:36:11 <BlueMatt> its close, but it cant be in a dir
548 2012-01-10 05:36:15 <BlueMatt> has to be zip
549 2012-01-10 05:36:20 <luke-jr> sigs being there = bug?
550 2012-01-10 05:36:21 <BlueMatt> and has to have at least 3 sigs
551 2012-01-10 05:36:28 <luke-jr> I have no idea what you're talking about.
552 2012-01-10 05:36:34 <BlueMatt> those are there to verify the files I gave youy
553 2012-01-10 05:37:01 <BlueMatt> just remove the sigs and make them the form of the previous uploads and you can remove the rc1s
554 2012-01-10 05:37:31 <luke-jr> ok&
555 2012-01-10 07:08:16 <barmstrong> hi all, could someone point me toward the code in the official client that checks if a transaction is "valid"?
556 2012-01-10 07:08:33 <barmstrong> for example I noticed this one has an extra op_nop in it http://blockexplorer.com/tx/5492a05f1edfbd29c525a3dbf45f654d0fc45a805ccd620d0a4dff47de63f90b
557 2012-01-10 07:08:42 <barmstrong> was not sure if this would still be considered valid
558 2012-01-10 07:08:49 <barmstrong> but wanted to see how the official client checks
559 2012-01-10 07:25:23 <luke-jr> barmstrong: obviously if it's in the block chain it's valid
560 2012-01-10 07:25:37 <luke-jr> barmstrong: and a decentralized system has no "official"
561 2012-01-10 07:26:39 <barmstrong> well there were other transactions that are no longer considered valid
562 2012-01-10 07:26:52 <barmstrong> but are in the main chain
563 2012-01-10 07:32:45 <heoa> Are there patches for autocompletion, getHashesToAddress -- on commandline?
564 2012-01-10 07:32:59 <heoa> getAddresses2Hashes
565 2012-01-10 08:31:16 <onelineproof> So if anyone wants to test my package: http://piratelinux.org/?page_id=2 . It's not focused on bitcoin but does contain some bitcoin functionality. Compatible with Ubuntu and Debian. ISO coming soon.
566 2012-01-10 08:42:52 <lfm> I have enuf virus already
567 2012-01-10 08:45:42 <cjdelisle> no such thing
568 2012-01-10 08:57:27 <cjdelisle> onelineproof: what exactly does it do?
569 2012-01-10 09:02:56 <cjdelisle> nvm interesting, probably would make a nice usb/fat32 thing
570 2012-01-10 09:03:19 <cjdelisle> syslinux 0wn-highschool-computer
571 2012-01-10 09:16:25 <onelineproof> cjdelisle: It's a project for the Pirate Party of Canada. The package installs all the modifications that I will implement into the Custom Distro.
572 2012-01-10 09:16:56 <cjdelisle> that's pretty cool
573 2012-01-10 09:17:39 <cjdelisle> you might find this interesting too: https://raw.github.com/cjdelisle/cjdns/master/rfcs/Whitepaper.txt
574 2012-01-10 09:17:43 <onelineproof> You don't have to trust my binaries. On one click with the package manager, all resulting binaries get compiled from source (except for the dependencies which normally come from trusted Ubuntu sources)
575 2012-01-10 09:18:10 <cjdelisle> it's not ready to be shipped to the world by any means but it's my idea for a better network.
576 2012-01-10 09:18:24 <cjdelisle> /nod
577 2012-01-10 09:19:33 <cjdelisle> I'm connected through cjdns to an irc network right now over 3 hops w/ one cjdns node on my computer, one in a vm, one on a server and one on the ircd
578 2012-01-10 09:19:59 <cjdelisle> it's in the "proof of consept" phase, it still makes some very bad routing decisions from time to time
579 2012-01-10 09:22:34 <onelineproof> cool. Is it like Tor with hidden services?
580 2012-01-10 09:23:02 <onelineproof> or like I2P?
581 2012-01-10 09:23:48 <cjdelisle> like i2p because there's no outproxy
582 2012-01-10 09:24:24 <cjdelisle> but like freenet in darknet mode because you need permission to connect to someone and connections are made manually through the configuration file.
583 2012-01-10 09:25:16 <cjdelisle> and unlike any of them because nodes try to route the fastest path instead of trying to bulster security by obfuscation
584 2012-01-10 09:34:13 <onelineproof> well ill give it a test some time. I really wanna find a network that is private and fast for file sharing, like torrents. I2P is too damn slow.
585 2012-01-10 09:34:37 <cjdelisle> if you're on efnet, drop by #cjdns
586 2012-01-10 09:34:59 <onelineproof> k ill add it to my chat list
587 2012-01-10 09:35:18 <cjdelisle> you have to get credentials to connect to someone's server and I hope you'll set your own vps or server up to be one of the backbone providers
588 2012-01-10 09:36:05 <onelineproof> ya once I get to know it...
589 2012-01-10 09:36:06 <cjdelisle> it's still in rapid development but the biggest thing in my mind is it needs to be fast and that means we need a lot of servers and few dsl/cable connections
590 2012-01-10 09:36:23 <cjdelisle> yea, once it becomes a little more stable too ;)
591 2012-01-10 12:57:04 <luke-jr> makomk: oh no, you might actually have to write code! what a nusance!
592 2012-01-10 12:59:25 <makomk> luke-jr: the problem isn't writing it, the problem is making sure that it behaves sensibly even when someone's trying to attack it, and for that matter figuring out how it should behave. This is rather harder to analyze than Bitcoin.
593 2012-01-10 13:00:11 <luke-jr> makomk: good thing I'm providing you with such an opportunity to test it
594 2012-01-10 13:01:13 <makomk> luke-jr: that can only prove that it doesn't have any flaws *that you managed to find and exploit*
595 2012-01-10 13:01:56 <UukGoblin> can you actually sign forum posts? won't the markup and line endings screw the signature up?
596 2012-01-10 13:02:26 <luke-jr> well, if you weren't so busy siding with the scammers, you could make it clear that you endorse trying to find flaws in it that might otherwise result in real actual theft
597 2012-01-10 13:02:41 <copumpkin> lol
598 2012-01-10 13:03:21 <makomk> luke-jr: I know there's a flaw in it that can result in real actual theft: the 51% attacker can still carry out a double spend. That's probably still unfixable.
599 2012-01-10 13:04:58 <luke-jr> makomk: you know, I seem to recall suggesting a theoretical solution the other day that would actually make 51% double spends more difficult&
600 2012-01-10 13:05:32 <makomk> luke-jr: oh? What side effects did this have?
601 2012-01-10 13:05:53 <UukGoblin> what's CLC?
602 2012-01-10 13:06:07 <luke-jr> UukGoblin: makomk's scamcoin
603 2012-01-10 13:06:20 <luke-jr> makomk: none I though of in the brief moment I thought about it
604 2012-01-10 13:06:41 <luke-jr> gmaxwell seemed to think it wouldn't be viable, but I couldn't understand his reason
605 2012-01-10 13:07:46 <luke-jr> so you don't need to go digging through logs, I'll summarize it again:
606 2012-01-10 13:08:20 <luke-jr> if two valid blocks are seen at height X within a 10 minute window, accept the one with the mathematically lower hash
607 2012-01-10 13:09:13 <makomk> That's, errrm, ...
608 2012-01-10 13:09:36 <luke-jr> it would certainly be more difficult, at least, for my blocks to always be mathematically lower than the people trying to compete
609 2012-01-10 13:10:14 <makomk> Yeah, I'm just not sure how you could enforce that in the face of chain forks which are more than a single block long.
610 2012-01-10 13:10:21 <makomk> Which they would be.
611 2012-01-10 13:10:26 <UukGoblin> fun.
612 2012-01-10 13:10:39 <luke-jr> makomk: just ignore the longer chain, if it breaks the rule
613 2012-01-10 13:11:00 <luke-jr> in ordinary cases, that'd probably mean a 2-block chain would lose to a 1-block chain occasionally
614 2012-01-10 13:11:16 <luke-jr> occasionally = rarer than current orphans
615 2012-01-10 13:11:29 <makomk> luke-jr: now what happens if I go back in history and create a block at height some time in the past with a lower hash than the existing block there.
616 2012-01-10 13:11:45 <luke-jr> makomk: that's why you have the requirement that it be received within a 10 minute window
617 2012-01-10 13:11:52 <makomk> Clients running at the time can obviously tell what happened, but any client that wasn't can't.
618 2012-01-10 13:12:02 <luke-jr> true
619 2012-01-10 13:12:29 <luke-jr> I'll leave you to ponder and solve that. ;)
620 2012-01-10 13:12:30 <luke-jr> bbl
621 2012-01-10 13:13:44 <makomk> I don't think it is solvable; it's basically the fundamental problem that prevents effective anti-double-spending measures.
622 2012-01-10 13:20:25 <UukGoblin> so what actually happened? do I understand correctly that luke-jr performed a 51% attack on a merged-mining chain by merged-mining from eligius?
623 2012-01-10 13:20:36 <luke-jr> UukGoblin: no.
624 2012-01-10 13:20:54 <makomk> UukGoblin: correct, if by "51% attack" you don't actually mean a double-spend.
625 2012-01-10 13:20:55 <UukGoblin> could I see some neutral information on the subject?
626 2012-01-10 13:21:03 <luke-jr> UukGoblin: but there's so much FUD, don't bother trying to understand&
627 2012-01-10 13:21:28 <luke-jr> UukGoblin: nobody is neutral. there's a few scammers and sympathizers, but most people are cool with it
628 2012-01-10 13:21:28 <makomk> UukGoblin: there should be a post by ArtForz on the forum confirming this.
629 2012-01-10 13:22:08 <UukGoblin> ok, then from both sides? :->
630 2012-01-10 13:23:14 <UukGoblin> luke-jr, perhaps you wrote something about the subject somewhere?
631 2012-01-10 13:23:21 <luke-jr> UukGoblin: I'm mining the heck out of it, using my own hardware. That's all that should matter.
632 2012-01-10 13:23:39 <luke-jr> UukGoblin: no, a complete writeup in all the details necessary to understand is IMO utterly impossible
633 2012-01-10 13:23:44 <luke-jr> UukGoblin: many people have tried, and failed
634 2012-01-10 13:23:47 <UukGoblin> luke-jr, ok
635 2012-01-10 13:24:00 <luke-jr> and 1-on-1 explaining takes way too long
636 2012-01-10 13:24:11 <luke-jr> (and even then, often fails)
637 2012-01-10 13:24:17 <UukGoblin> that'll just help keep the FUD up
638 2012-01-10 13:24:20 <makomk> luke-jr: come to think of it, is Eligius actually on your own hardware or is it MtGox-owned.
639 2012-01-10 13:24:49 <luke-jr> UukGoblin: much easier to make the point that it's nobody else's business
640 2012-01-10 13:25:21 <makomk> luke-jr's position is that because all the CLC-related part of mining happens in Eligius' pool software, it's "on his own hardware" and therefore not anyone's business.
641 2012-01-10 13:25:31 <UukGoblin> luke-jr, I'd disagree, if it really is offensive action you're taking
642 2012-01-10 13:26:50 <luke-jr> UukGoblin: nobody lost money, and it plays by the network rules; I just have more hashpower.
643 2012-01-10 13:27:17 <luke-jr> UukGoblin: in the case of SolidCoin, my 'attack' is 100% based on laws and legalities
644 2012-01-10 13:27:50 <luke-jr> (see https://github.com/solidcoin/solidcoin )
645 2012-01-10 13:27:59 <UukGoblin> luke-jr, ok a plain question then: could you have performed whatever you did on CLC if you didn't have eligius miners' hashpower?
646 2012-01-10 13:28:40 <Diablo-D3> are we still fucking discussing this?
647 2012-01-10 13:28:54 <Diablo-D3> UukGoblin: dude, you do realize 2 people got banned and 1 person close to it on the forums over this shit, right?
648 2012-01-10 13:28:56 <UukGoblin> Diablo-D3, I'm happy to move the question elsewhere
649 2012-01-10 13:28:58 <Diablo-D3> people are tired of hearing aobut it
650 2012-01-10 13:29:03 <Diablo-D3> UukGoblin: there is no proper forum.
651 2012-01-10 13:29:07 <Diablo-D3> basically, it boils down to this
652 2012-01-10 13:29:07 <luke-jr> &
653 2012-01-10 13:29:16 <Diablo-D3> ues, luke used eligius's hash power to merge mine CLC
654 2012-01-10 13:29:24 <Diablo-D3> no, it didnt effect miners and was not a violation of their trust
655 2012-01-10 13:29:31 <Diablo-D3> no, luke didnt make any BTC from it
656 2012-01-10 13:29:42 <Diablo-D3> yes, CLC is utterly fucked because the diff spiked to hell
657 2012-01-10 13:29:43 <UukGoblin> it doesn't belong to -dev indeed, but it was started here and I couldn't find any sensible info on google, so I kept asking here.
658 2012-01-10 13:29:55 <Diablo-D3> UukGoblin: so there, theres all the answers to your questions.
659 2012-01-10 13:30:02 <makomk> UukGoblin: that's usually how it does start, yeah.
660 2012-01-10 13:30:26 <luke-jr> sigh
661 2012-01-10 13:31:09 <luke-jr> Diablo-D3: don't start forcing me to spend hours going over this
662 2012-01-10 13:31:27 <UukGoblin> Diablo-D3, ok thanks for the heads up, I still have more questions though
663 2012-01-10 13:31:31 <luke-jr> in fact, I don't have hours
664 2012-01-10 13:32:15 <UukGoblin> oh great, #bitcoin-politics is +i. :-P
665 2012-01-10 13:32:17 <Diablo-D3> UukGoblin: CLC is claiming luke attacked them, luke is trolling them by going along with it
666 2012-01-10 13:32:22 <luke-jr> [09:23:21] <luke-jr> UukGoblin: I'm mining the heck out of it, using my own hardware. That's all that should matter.
667 2012-01-10 13:32:50 <Diablo-D3> UukGoblin: this would only be an issue if CLC had any actual real world value, but it doesnt
668 2012-01-10 13:32:55 <UukGoblin> luke-jr, are you using ONLY your own hardware?
669 2012-01-10 13:32:57 <Diablo-D3> no one is buying CLC
670 2012-01-10 13:33:08 <UukGoblin> I think you can't claim that ever on a p2p network
671 2012-01-10 13:33:09 <Diablo-D3> its CLC's fault for not having enough hash power to prevent the spike.
672 2012-01-10 13:33:38 <Diablo-D3> UukGoblin: basically, when luke joined, he was like 10 times more hash power than the rest of the CLC people combined
673 2012-01-10 13:33:43 <UukGoblin> Diablo-D3, no, I think it's a political issue
674 2012-01-10 13:34:00 <Diablo-D3> yes, it IS a political issue which directly becomes an offtopic and you will get banned for it issue
675 2012-01-10 13:34:00 <UukGoblin> Diablo-D3, it's not about money, but trust
676 2012-01-10 13:34:13 <UukGoblin> 143214 <@UukGoblin> oh great, #bitcoin-politics is +i. :-P
677 2012-01-10 13:34:17 <Diablo-D3> like I said, 2 people on the forums got banned for it, and 1 person came very close to being the third.
678 2012-01-10 13:34:17 <UukGoblin> where can I discuss it then?
679 2012-01-10 13:34:25 <Diablo-D3> there is no where you can discuss it
680 2012-01-10 13:34:36 <Diablo-D3> on the forums its completely off topic by word of theymos himself.
681 2012-01-10 13:34:56 <Diablo-D3> UukGoblin: you can discuss it in the clc irc channel if they have one
682 2012-01-10 13:35:04 <Diablo-D3> thats it
683 2012-01-10 13:35:09 <Diablo-D3> thats your only safe place
684 2012-01-10 13:35:50 <UukGoblin> Diablo-D3, ok, thanks for the info
685 2012-01-10 13:36:35 <Diablo-D3> btw, people claim its trust? no one trusts luke anyhow
686 2012-01-10 13:36:43 <Diablo-D3> this is bitcoin, we trust no one
687 2012-01-10 13:37:40 <UukGoblin> Diablo-D3, ok, stop doing OT here then.
688 2012-01-10 13:37:46 <UukGoblin> ;-)
689 2012-01-10 13:38:10 <Diablo-D3> Im just saying.
690 2012-01-10 13:38:18 <Diablo-D3> this has been hashed out over 9000 times
691 2012-01-10 13:38:23 <Diablo-D3> people are tired of hearing it
692 2012-01-10 13:38:32 <Diablo-D3> the CLC guys were annoying assholes, they got shut up.
693 2012-01-10 13:42:24 <dikidera> Just to ask
694 2012-01-10 13:42:40 <dikidera> Making new addresses starting with 3, does that mean the max possible addresses also increase?
695 2012-01-10 13:53:45 <gmaxwell> makomk: congrats on comming up with something that sounds like an interesting avenue of research.
696 2012-01-10 13:54:40 <gmaxwell> makomk: the merge stuff can't be as simple as you through out... e.g. consider if I produce a series is diff=1 1mb blocks parallel to your chain, do I manage to fill your disks?
697 2012-01-10 13:55:24 <CIA-100> libbitcoin: genjix * r214ba44a2fdd /include/bitcoin/utility/subscriber.hpp: subscriber: No need to copy regsistry stack http://tinyurl.com/88pr5qy
698 2012-01-10 13:55:25 <gmaxwell> oh you mentioned spam. yea. thats a problem but perhaps not insoluable.
699 2012-01-10 13:56:26 <gmaxwell> makomk: I've got another one of that ilk but it has problems of a similar class.
700 2012-01-10 13:58:40 <makomk> gmaxwell: yeah, it's interesting theoretically but not terribly practical even with the spam problem solved. Protects against luke-jr's attack but can't do anything useful against an attacker that's actually double-spending.
701 2012-01-10 13:59:14 <gmaxwell> makomk: I don't think that makes it worthless at all.
702 2012-01-10 14:00:09 <UukGoblin> so in theory, a malicious pool operator can defeat the idea of merged mining new chains?
703 2012-01-10 14:00:30 <gmaxwell> UukGoblin: no, not really. A better way to express this is
704 2012-01-10 14:00:56 <makomk> gmaxwell: it makes it worthless right now against anything except some theoretical honourable attacker that won't double-spend, because in the current proposal they can effectively nullify the protection and stop other transactions being confirmed again if they do.
705 2012-01-10 14:01:32 <gmaxwell> UukGoblin: that merged mining will not be successful if the net sentiment towards a new chain is neutral or somewhat negative, because merged mining is subject to extreme indifference.
706 2012-01-10 14:02:23 <gmaxwell> UukGoblin: slush, btcguild, or deepbit could indivigually undo the problem here, any of several possible pairs of smaller pools could, any tripplet of a larger set. They aren't however.
707 2012-01-10 14:02:57 <gmaxwell> makomk: someone can't reverse and respend anything but their own transactions.
708 2012-01-10 14:03:09 <UukGoblin> well, I hope it won't apply to my application...
709 2012-01-10 14:03:19 <gmaxwell> makomk: and an attacker that won't reverse and respend isn't theoretical.
710 2012-01-10 14:03:26 <UukGoblin> it probably won't though
711 2012-01-10 14:04:20 <gmaxwell> UukGoblin: oh boy. Whats this?
712 2012-01-10 14:04:52 <makomk> gmaxwell: that's not entirely true... an attacker can reverse (though not respend) other's transactions because I intentionally don't want to allow multiple conflicting transactions in the best chain.
713 2012-01-10 14:05:34 <gmaxwell> makomk: There can't be a reversal in that scheme without a respend.
714 2012-01-10 14:05:37 <UukGoblin> gmaxwell, well I was thinking of something nice and easy to get small amounts of data (checksums) into the bitcoin blockchain. That wouldn't actually create an altchain, no verification would take place, so reversing of stuff wouldn't make sense. People don't like my idea either, and hence the worry.
715 2012-01-10 14:06:11 <gmaxwell> (I switched to using the reverse/respend language because its more clear what a 'double spend' actually involves)
716 2012-01-10 14:06:37 <gmaxwell> UukGoblin: I'd like an altchain (but not a currency) for that. Stuffing data into the blockchain is a terrible idea for the reasons you probably already got blasted with.
717 2012-01-10 14:07:09 <gmaxwell> It doesn't even really have to be a chain so much as just tree grafts on merged mining.
718 2012-01-10 14:07:38 <makomk> gmaxwell: yes there can be. The chain merging currently requires both sides of the merge to be non-conflicting. If a 51% attacker introduces transactions into their opponents' version of the chain and a conflicting one into their own, the two can't be merged and their chain can grow to overwrite the non-malicious chain.
719 2012-01-10 14:07:38 <UukGoblin> gmaxwell, yeah, something standardized to make merkle trees of users' data to be hashed.
720 2012-01-10 14:07:51 <UukGoblin> and just put a root in coinbase's merged-mining area
721 2012-01-10 14:08:01 <UukGoblin> and actually store the data separately
722 2012-01-10 14:08:18 <makomk> The reason why I'm conceptually against merging chains with conflicting transactions is because it could probably be used to facilitate a double-spend.
723 2012-01-10 14:08:21 <gmaxwell> makomk: okay, so the non-attacker just has to take care to not conflict with the attackers chain. No different than normal mining.
724 2012-01-10 14:08:32 <UukGoblin> gmaxwell, but, frankly... if pruning was implemented in bitcoin, it'd be no different to just processing special transactions in bitcoin itself
725 2012-01-10 14:08:37 <gmaxwell> makomk: I wasn't suggesting that you merge chains with conflicting txn.
726 2012-01-10 14:08:56 <gmaxwell> makomk: I'm pointing out that there can't existing a conflicting txn without a respend.
727 2012-01-10 14:09:04 <makomk> gmaxwell: except they can't forsee the future, and tha attacker has no reason to follow this rule themself.
728 2012-01-10 14:09:37 <gmaxwell> UukGoblin: but it's not, and even then you can't prune until later because you need to see the txn to validate the history for yourself.
729 2012-01-10 14:10:33 <gmaxwell> makomk: Going to the simplest case: If I mine a block with only txn that I control, the attacker can not conflict those.
730 2012-01-10 14:11:10 <UukGoblin> well, keeping it separate is probably the best solution from political perspective. :->
731 2012-01-10 14:12:11 <gmaxwell> makomk: while an attack was happening, I guess.
732 2012-01-10 14:12:11 <makomk> gmaxwell: True. Which by itself would give everyone else an incentive not to mine anyone else's transactions.
733 2012-01-10 14:13:34 <makomk> Yeah. Except they'd have no way of knowing whether the attack was still going on, whether the attacker had stopped or was just waiting to spring their 51% chain on everyone.
734 2012-01-10 14:13:40 <gmaxwell> makomk: yea.
735 2012-01-10 14:15:34 <gmaxwell> makomk: the latecomers decision I think is what fundimentally kills that class of idea, perhaps.
736 2012-01-10 14:17:21 <makomk> Perhaps. Unlike some other solutions you should be able to guarantee the network won't fork persistently, but at the risk of really deep history rewrites...
737 2012-01-10 14:22:05 <UukGoblin> someone should really tell bitstamp (and rock) to fix their clocks. ;-)
738 2012-01-10 14:24:58 <Diablo-D3> lol
739 2012-01-10 14:25:02 <Diablo-D3> ntpd like a baws
740 2012-01-10 14:53:47 <ciscoftw> KNOCK #bitcoind pls send invite, wish to discuss bitcoind (rpc/server) behavior(s)
741 2012-01-10 14:54:00 <ciscoftw> gd it :(
742 2012-01-10 14:55:09 <gmaxwell> ciscoftw: there is no #bitcoind
743 2012-01-10 14:55:15 <gmaxwell> we discuss bitcoind in the regular channels.
744 2012-01-10 14:55:53 <ciscoftw> no? says invite only...
745 2012-01-10 14:56:00 <gmaxwell> Presumably to keep people from falling into a pit there.
746 2012-01-10 14:57:16 <ciscoftw> perhaps i should ask it here... my q is this, shouldnt't the bitcoind server report with +1 block to the current solved chain... for example if blockexplorer reports X as current block, bitcoind (./bitcoind getinfo) should report back with its block as x+1?
747 2012-01-10 15:04:55 <gmaxwell> No, iirc they should just agree.
748 2012-01-10 15:10:19 <ciscoftw> gmaxwell: i agree that the client is almost 100% of the time ahead of blockexplorer (by atleast 30secs anyway).. but i could have sworn bitcoind would stay +1 block ahead, and it a valid hash is created by a mining client it would broadcast to network?
749 2012-01-10 15:10:57 <gmaxwell> ciscoftw: thats not how getinfo works getinfo reports the position of the existing blocks not the hypothetical block your node may be working on.
750 2012-01-10 15:11:15 <ciscoftw> anybody have running bitcoind wanna check it out and report back? would be very helpful in understand correct behavoir
751 2012-01-10 15:11:23 <kinlo> the bitcoin client knows the state of the network
752 2012-01-10 15:11:24 <ciscoftw> do you have a bitcoind service running, mind checking for me pls?
753 2012-01-10 15:11:31 <gmaxwell> ciscoftw: I did check.
754 2012-01-10 15:11:34 <kinlo> if the bitcoin knows something the network doesn't, it must announce it or it will never get distributed
755 2012-01-10 15:12:00 <ciscoftw> gmaxwell: many thanx, and your service reports the current block aswell?
756 2012-01-10 15:12:28 <gmaxwell> Yes, my only lack of certanty there was I didn't know what blockexplorer was reporting.
757 2012-01-10 15:13:32 <ciscoftw> ...well thats like the entire premise of this... can you tell me what block your bitcoind is on this second
758 2012-01-10 15:13:49 <ciscoftw> blockexplore: 161571
759 2012-01-10 15:13:57 <kinlo> "blocks" : 161571,
760 2012-01-10 15:14:08 <kinlo> that will every client output
761 2012-01-10 15:14:11 <kinlo> should
762 2012-01-10 15:14:42 <ciscoftw> alirght :) thanx for helping me out fellas
763 2012-01-10 15:15:16 <gmaxwell> ciscoftw: I'm curious as to why you're asking.
764 2012-01-10 15:15:17 <ciscoftw> dont suppose your bitcoind reports "hashespersec"
765 2012-01-10 15:15:47 <kinlo> no, only those that mine do, but it would be stupid to mine with bitcoind
766 2012-01-10 15:15:50 <sipa> only for the built-in crippled miner
767 2012-01-10 15:16:08 <gmaxwell> ciscoftw: thats only reported with the (usless on bitcoin) integrated hashing.
768 2012-01-10 15:16:23 <kinlo> the build in miner is even not useable if you're on a test network with difficulty 1, too slow to do tests
769 2012-01-10 15:16:35 <ciscoftw> bitcoind has an integrated mining function :) awesome!!! so people dont mine via rpc?
770 2012-01-10 15:16:54 <kinlo> ciscoftw: nobody uses the internal function, it is just too slow
771 2012-01-10 15:16:54 <sipa> ciscoftw: satoshi intended everyone to be a miner
772 2012-01-10 15:16:58 <sipa> with the advent of gpu mining, it has become worthless
773 2012-01-10 15:17:32 <ciscoftw> gmaxwell: i stated messing with the bitcoin-dissector, just keeps opening a can or worms for me so to speak... hence these retarded q's about bitcoind
774 2012-01-10 15:18:00 <gmaxwell> ciscoftw: people mine using RPC miners external software that uses the getwork rpc call.
775 2012-01-10 15:18:16 <ciscoftw> yeah, its pretty crazy to look back at hashing power used to created genisus block ~8kH/sec iirc
776 2012-01-10 15:18:41 <kinlo> is the genesis block actually below any difficulty?
777 2012-01-10 15:19:14 <gmaxwell> ciscoftw: both computers and the cpuminer software have gotten a lot faster since then. :)
778 2012-01-10 15:19:18 <ciscoftw> sipa: yeah mining pools are kinda f'ing that up... part of the reason why i left 4monts ago
779 2012-01-10 15:19:19 <kinlo> I always assumed it would just be hardcoded to be accepted at any difficulty
780 2012-01-10 15:19:25 <gmaxwell> kinlo: sure. it's a valid block under the normal rules.
781 2012-01-10 15:19:52 <ciscoftw> i have less than 2000MH/s but i want to help the network, pool'd mining would actually make more sense for me, but like i said -wanna help the network
782 2012-01-10 15:19:52 <gmaxwell> kinlo: have to test to make sure the system works!
783 2012-01-10 15:19:54 <gmaxwell> ciscoftw: p2pool my friend, p2pool.
784 2012-01-10 15:19:56 <kinlo> so satoshi spent days creating a genesis block, while with modern gpu's we can do it in like a few seconds
785 2012-01-10 15:20:30 <gmaxwell> kinlo: Yes.
786 2012-01-10 15:20:33 <sipa> ;;bc,calcd 4000 1
787 2012-01-10 15:20:37 <gribble> The average time to generate a block at 4000 Khps, given the supplied difficulty of 1, is 17 minutes and 53 seconds
788 2012-01-10 15:20:40 <osearth> gmaxwell and ciscoftw are playing Pocket pool again
789 2012-01-10 15:20:47 <gmaxwell> ciscoftw: https://en.bitcoin.it/wiki/P2Pool
790 2012-01-10 15:20:47 <sipa> kinlo: he probably spend not more than half an hour to generate it
791 2012-01-10 15:20:51 <gmaxwell> sipa: nah, ever try the old bitcoin software? it was way slower than the current cpu miners.
792 2012-01-10 15:21:29 <kinlo> he had a good computer then :)
793 2012-01-10 15:21:57 <gmaxwell> I went and measured it a few months back. meh, I'd have to search the logs. but on contemporary hardware it was something like 40kh/s IIRC?
794 2012-01-10 15:21:57 <gribble> The average time to generate a block at 8 Khps, given the supplied difficulty of 1, is 6 days, 5 hours, 7 minutes, and 50 seconds
795 2012-01-10 15:21:57 <kinlo> ;;bc,calcd 8 1
796 2012-01-10 15:21:57 <sipa> that's what, a 386?
797 2012-01-10 15:21:59 <kinlo> stil a day :)
798 2012-01-10 15:22:01 <sipa> i think even in 2009 you easily got 1000 kbps
799 2012-01-10 15:22:10 <gmaxwell> I think 8 is the estimate you get between block 0 and 1.
800 2012-01-10 15:22:10 <sipa> kH/s
801 2012-01-10 15:22:44 <gmaxwell> sipa: not on the old openssl based code.
802 2012-01-10 15:22:46 <ciscoftw> ...so #bitcoind chan is no exisitant? not reserved for invite only? ...im not missing out on anything in there :)
803 2012-01-10 15:22:46 <kinlo> sipa: if you had optimized code
804 2012-01-10 15:23:08 <sipa> ciscoftw: this is the dev channel
805 2012-01-10 15:24:09 <sipa> the first 10 blocks are each less than 500s apart
806 2012-01-10 15:24:10 <sipa> which may mean pregeneration...
807 2012-01-10 15:24:37 <sipa> sorry, some are up to 800s
808 2012-01-10 15:25:12 <ciscoftw> how could you pregen a block? you need hash from pervious chain
809 2012-01-10 15:25:13 <gmaxwell> sipa: yea, now that you menion it.
810 2012-01-10 15:25:26 <sipa> ciscoftw: you can choose the time yourself
811 2012-01-10 15:25:34 <Joric> i used testnet-in-a-box for testing there's default difficulty 0.25
812 2012-01-10 15:26:01 <ciscoftw> assuming theres no tranactions too?
813 2012-01-10 15:26:01 <gmaxwell> Joric: thats .. kinda pointless.
814 2012-01-10 15:26:16 <sipa> ciscoftw: if satoshi was alone, all transaction would have been his own
815 2012-01-10 15:26:17 <gmaxwell> Joric: a lot of mining code can't really work below diff 1.
816 2012-01-10 15:26:30 <gmaxwell> Joric: I mean, it'll work, but it won't be any faster.
817 2012-01-10 15:48:45 <afmatt> Hello all - looking for some help getting mining setup on an Amazon GPU cluster similar to https://bitcointalk.org/index.php?topic=8405.0 - followed the steps in that thread but something isn't working right.
818 2012-01-10 15:57:06 <ciscoftw> afmatt: doesnt he conclude at end of thread that is was a waste of money? ...given it was a free 20dollar giftcard
819 2012-01-10 15:57:59 <afmatt> Yes he does - but I have access to a few of these "after hours" and want to make use of them :)
820 2012-01-10 15:58:42 <ciscoftw> good point, if its free, then i suppose why not...
821 2012-01-10 15:59:51 <ciscoftw> what specifically isnt working, his write up looks pretty good?