1 2011-11-07 04:22:18 <coderrr> AlexWaters, sup
  2 2011-11-07 06:11:59 <ThomasV> gmaxwell: http://s3.ecdsa.org/seed.png
  3 2011-11-07 06:12:57 <Diablo-D3> also
  4 2011-11-07 06:13:01 <Diablo-D3> SIGMA GARLIC
  5 2011-11-07 06:13:19 <Diablo-D3> sigma garlic has a very high ramirez value
  6 2011-11-07 06:13:35 <ThomasV> yes, definitely a correlation
  7 2011-11-07 06:14:13 <ThomasV> in another trial I had "frank" and "sinatra" in the same sequence
  8 2011-11-07 06:14:56 <Diablo-D3> :D
  9 2011-11-07 06:17:15 <ThomasV> I decided not to use it for the screenshot :)
 10 2011-11-07 06:21:54 <gmaxwell> ThomasV: Pretty great. It's perhaps not clear that you can remember either the hex or the mnemonic.
 11 2011-11-07 06:26:18 <ThomasV> well, I initially used the term 'corresponding' instead of 'equivalent'
 12 2011-11-07 06:26:30 <ThomasV> but I agree, it's not obvious
 13 2011-11-07 06:26:31 <gmaxwell> If you don't have one already, you might want to add an extra mnemonic word which is simply the sum of all the others (mod whatever the size of the wordlist) as a check digit.. which will enable you to detect a screwed up one as well as recover it.
 14 2011-11-07 06:27:12 <ThomasV> an error correcting word code :-)
 15 2011-11-07 06:27:14 <gmaxwell> I see you have 128 bits of hex there.. but 12 words
 16 2011-11-07 06:27:26 <ThomasV> yes, I don't use pgp
 17 2011-11-07 06:27:38 <ThomasV> I infringe a patent instead :-)
 18 2011-11-07 06:27:39 <Diablo-D3> what, you odnt have enough words?
 19 2011-11-07 06:27:53 <gmaxwell> so thats ... 10.6666 bits per word?
 20 2011-11-07 06:27:54 <ThomasV> google 'mnemonic encoding'
 21 2011-11-07 06:28:26 <ThomasV> it's a small dict of 1626 words
 22 2011-11-07 06:28:28 <gmaxwell> ah 32 bits into three words.
 23 2011-11-07 06:29:40 <ThomasV> it can be argued that it's a trivial patent
 24 2011-11-07 06:30:19 <gmaxwell> Know what the patent number is?
 25 2011-11-07 06:30:27 <ThomasV> just google 'mnemonic encoding'
 26 2011-11-07 06:30:40 <gmaxwell> (e.g. if what they patented was mostly the list of words then it would be simple and prudent to change it)
 27 2011-11-07 06:31:13 <ThomasV> http://www.patentstorm.us/patents/5892470/description.html
 28 2011-11-07 06:31:53 <gmaxwell> Well, claim one can be easily avoided in the implemention. (haven't read the other ones yet)
 29 2011-11-07 06:32:54 <ThomasV> the list of words is here: https://github.com/singpolyma/mnemonicode
 30 2011-11-07 06:33:01 <gmaxwell> Three (and the computer-medium form of it) is the nasty one.
 31 2011-11-07 06:33:33 <gmaxwell> Claim one and two can be avoided by just changing how you compute it (e.g. using repeated subtraction in the encoder instead of division)
 32 2011-11-07 06:34:35 <Diablo-D3> wat
 33 2011-11-07 06:34:37 <gmaxwell> For three I'd prefer to add a tiny amount of inefficiency by adding one word to the list.
 34 2011-11-07 06:35:05 <gmaxwell> (there may be other ways to avoid the claim, but making hte numberbase non-identical is so easy and harmless)
 35 2011-11-07 06:36:12 <coderrr> i wonder if you could make it easier to remember by using short phrases instead of words
 36 2011-11-07 06:36:12 <gmaxwell> (also because the inefficiency slays the other independant claims too e.g. a "unique representation"
 37 2011-11-07 06:36:28 <gmaxwell> coderrr: you lose entropy that way and make it bigger...
 38 2011-11-07 06:36:31 <ThomasV> but I think the patent claims the method on arbitrary dictionaries. which would include the pgp word list, it's just another base
 39 2011-11-07 06:37:19 <coderrr> gmaxwell, how do you lose entropy? if you use a phrase in place of a word, they're both just a unique ID
 40 2011-11-07 06:38:01 <coderrr> and if you have 100k phrases versus 1k words, you need less phrases and maybe end up with similar # of total words ?
 41 2011-11-07 06:38:09 <coderrr> but yea maybe im wrong, hadnt thot it through
 42 2011-11-07 06:38:09 <gmaxwell> ThomasV: nah, it doesn't. claim two requires multiplication in the decoding, one requires division in the encoding, three requires there is a plurality of words which combine to form the base of the number you're encoding.
 43 2011-11-07 06:38:18 <gmaxwell> (e.g. in PGP words you encode one number into one word)
 44 2011-11-07 06:38:38 <gmaxwell> I suspect that couldn't patent one word one number due to prior art, but I didn't look at the prior art they cited.
 45 2011-11-07 06:39:13 <ThomasV> gmaxwell: I am not sure if it is worth adding an error correcting code. this is not a situation where you want to send the key to someone over an unreliable channel, and where the other party can ask you to send it again if it's not correct
 46 2011-11-07 06:39:31 <ThomasV> if it's not correct, you are just screwed
 47 2011-11-07 06:40:50 <ThomasV> the wallet recovery procedure will stop because it will see that there are no coins at this address
 48 2011-11-07 06:41:07 <gmaxwell> ThomasV: Perhaps, though adding a tiny amount of additional length isn't costly. And you'll know for sure it was a typo.
 49 2011-11-07 06:41:57 <gmaxwell> you could even just make it a single word.. take all other words, convert them to numbers 0-1625 and take mod 1625 of the result. Thats your last word.
 50 2011-11-07 06:43:47 <ThomasV> it makes it 13 words instead of 12... that's the number of the devil
 51 2011-11-07 06:44:26 <gmaxwell> you could just display a number at the end then ... 12 words and a number? heh
 52 2011-11-07 06:47:51 <ThomasV> ok, bbl
 53 2011-11-07 07:18:13 <GoodAD> how many block now?
 54 2011-11-07 07:18:30 <GoodAD> 150600?
 55 2011-11-07 07:33:13 <min0r> how come it takes only a few mins for bitcoin 0.4 to download blocks 0-140,000... but after 140,000... it takes forever?
 56 2011-11-07 07:34:38 <gmaxwell> min0r: because there is fairly little activity early on.
 57 2011-11-07 07:34:44 <gmaxwell> min0r: and lots more later.
 58 2011-11-07 07:34:59 <gmaxwell> And most of the time is spent validating not downloading (though more is downloaded too)
 59 2011-11-07 07:35:08 <min0r> heh, that makes sense...
 60 2011-11-07 07:35:27 <GoodAD> how many block now 151112?
 61 2011-11-07 07:35:56 <gmaxwell> GoodAD: you can look at sites like https://blockexplorer.com/ to see. Its never to far behind.
 62 2011-11-07 07:36:00 <gmaxwell> ;;bc,blocks
 63 2011-11-07 07:36:01 <gribble> 152219
 64 2011-11-07 07:36:10 <min0r> gmaxwell, on macosx, bitcoin is taking about 35% of CPU, and im seeing tons of disk activity... this normal?
 65 2011-11-07 07:36:13 <GoodAD> merci
 66 2011-11-07 07:36:26 <gmaxwell> min0r: it is during initial syncup. Once it's in sync then it should be fairly little.
 67 2011-11-07 07:36:48 <gmaxwell> min0r: all that disk activity is being spend finding old transactions which the newly recieved blocks are spending.
 68 2011-11-07 07:37:03 <gmaxwell> (and writing the new data to disk synchronously)
 69 2011-11-07 07:39:01 <min0r> gmaxwell, ive heard talk of eventually hashing the whole blockchain up to a certain block when it becomes too big... u think that will happen?
 70 2011-11-07 07:40:03 <gmaxwell> No. I mean it might be an option. But any kind of summery requires that you trust the party that gave you the summary.
 71 2011-11-07 07:40:25 <gmaxwell> Of course, there is pruning which you can do to reduce storage, but it doesnt reduce the initial syncup time.
 72 2011-11-07 07:40:39 <min0r> also, would it ever be possible to speed up block generation if all clients agreed to such a thing? or is that part of the code impossible to update?
 73 2011-11-07 07:40:46 <gmaxwell> min0r: I expect that when lite clients of varrious kinds become mature most people who would care about syncup time would use those.
 74 2011-11-07 07:41:04 <gmaxwell> min0r: You could replace bitcoin with paypal if all clients update.
 75 2011-11-07 07:41:21 <gmaxwell> But speeding up block generation isn't even desirable.
 76 2011-11-07 07:41:40 <min0r> how about from a brick-and-mortar POS angle?
 77 2011-11-07 07:41:42 <gmaxwell> (well, it's desired by people who don't understand the tradeoffs of the system)
 78 2011-11-07 07:41:56 <min0r> i.e. restaurants, etc
 79 2011-11-07 07:42:05 <gmaxwell> min0r: speeding up the chain wouldn't make POS much better, and there are many other solutions for POS.
 80 2011-11-07 07:42:51 <min0r> yah , i think ive seen some 3rd parties that guarantee 0 confirmation transactions for a fee somewhere...
 81 2011-11-07 07:42:53 <gmaxwell> I posed a list with several here: https://bitcointalk.org/index.php?topic=28565.msg359948#msg359948
 82 2011-11-07 07:43:04 <gmaxwell> But I could probably add three or four more now.
 83 2011-11-07 07:43:07 <min0r> shorter block generation = less secure?
 84 2011-11-07 07:44:11 <min0r> btw, what do you think of litecoin (using scrypt instead of SHA?)
 85 2011-11-07 07:44:33 <gmaxwell> min0r: even at 2 minutes it would be unacceptable for most POS... and yes, shorter times reduce security and increase disk usage (even for lite clients, which must know the headers). Also, short times risk creating perverse incentives like encouraging miners not to process txn because they can't process many in the time between blocks.
 86 2011-11-07 07:45:05 <min0r> i sort of wish bitcoin did that from the get-go... it seems like its harder to gather a monopoly on mining if you must be CPU based
 87 2011-11-07 07:45:09 <JFK911> 2 minutes?  it used to take DAYS to process credit card transactions.
 88 2011-11-07 07:45:14 <JFK911> bring back the imprinter for bitcoins!
 89 2011-11-07 07:45:23 <gmaxwell> min0r: I had advocated scrypt for bitcoin once a long time back and was talked out of it: giving a monopoly to botnet operators is a bad thing considering how easily bought they are.
 90 2011-11-07 07:45:24 <min0r> whereas something with a few mil can just make an ASIC farm and snag 50% of BTC
 91 2011-11-07 07:45:50 <gmaxwell> min0r: scrypt as used by those chains (e.g. litecoin) does _NOT_ prevent that.
 92 2011-11-07 07:46:42 <min0r> gmaxwell, hrmm, good point... tho there are now botnets/worms that control GPUs
 93 2011-11-07 07:46:51 <gmaxwell> Thats my other concern about litecoin. I'm concerned that some funny business went on with the scrypt parameter selection. The memory used by it is unusually small, _maybe_ enough to frustrate gpu implementations (though Diablo-D3 doesn't even think that) but it's not too big to prevent fast FPGA implementations.
 94 2011-11-07 07:47:22 <min0r> really? then whats the point of using it? lol
 95 2011-11-07 07:47:38 <gmaxwell> If it were really intended to be CPU only it should be using tens of megabytes per hashing instance.. it's using more like 128 kilobits.
 96 2011-11-07 07:48:02 <gmaxwell> It's small enough that you should be able to implement it without external memory of the FPGAs people are using for mining.
 97 2011-11-07 07:48:22 <min0r> hrmm, good to know, thanks for the info...
 98 2011-11-07 07:48:48 <Diablo-D3> gmaxwell: you're correct, I dont think its enough to matter
 99 2011-11-07 07:49:16 <Diablo-D3> the worst case is it limits how many runs I can do per workgroup or such
100 2011-11-07 07:49:18 <gmaxwell> I'm not sure whats up, I sent art an email asking about it since he's apparently the source of the scrypt they're all using, and he also did the first bitcoin mining fpgas so he should be fully aware of the dynamics there.  Maybe I'm missing something.
101 2011-11-07 07:52:06 <min0r> IMO these alternate blockchains are pointless.... the only way there would be a serious competitor to BTC would be if someone comes up with a chain that can't be 'gamed' by a small group of miners
102 2011-11-07 07:52:29 <Diablo-D3> min0r: that'd be impossible
103 2011-11-07 07:52:37 <Diablo-D3> because you're defining the problemw rong
104 2011-11-07 07:52:43 <Diablo-D3> 1 hash == 1 attempt, period
105 2011-11-07 07:52:46 <Diablo-D3> you cant outgame that.
106 2011-11-07 07:52:50 <gmaxwell> Or rather, if it were known to be possible we'd have that instead of bitcoin.
107 2011-11-07 07:53:16 <Diablo-D3> if you start just giving yourself more hashpower, you are NOT winning
108 2011-11-07 07:53:19 <Diablo-D3> hashpower is expensive
109 2011-11-07 07:53:30 <Diablo-D3> and the more you have, the more time and money you have to spend to maintain it
110 2011-11-07 07:53:38 <gmaxwell> People have been trying to solve this problem hard since the 80s.. and they had access to all the tools we do today. Many people believed that sybil proof trustless distributed systems were _impossible_ until bitcoin proved that wrong.
111 2011-11-07 07:53:42 <Diablo-D3> and btc diff ends up fucking you in the end as well
112 2011-11-07 07:54:34 <min0r> good point...
113 2011-11-07 07:55:16 <min0r> so, what do you guys think the main barriers are to seeing mainstream BTC adoption?  what has to happen for more retailers to use it (which, of course is critical to its long-term success)
114 2011-11-07 07:55:35 <min0r> price stability?
115 2011-11-07 07:55:54 <Diablo-D3> min0r: generally? no market.
116 2011-11-07 07:56:03 <Diablo-D3> when I can btc amazon and newegg, then its mainstream
117 2011-11-07 07:56:06 <Diablo-D3> and not a moment before that.
118 2011-11-07 07:56:21 <gmaxwell> Maturity in general. Price stability isn't really an issue on its own since if we had a big enough enviroment you could have traders that sell exchange contracts.
119 2011-11-07 07:58:22 <min0r> diablo: whats going to make amazon and newegg adopt BTC? how do we get there? and whats stopping everyone from adopting BTC now?
120 2011-11-07 07:59:18 <min0r> gmaxwell, if you had to put a % chance on seeing newegg accept BTC within 5 years, what would you guess at?
121 2011-11-07 07:59:31 <gmaxwell> I couldn't guess.
122 2011-11-07 07:59:57 <gmaxwell> I also don't know how critial that is... I think that if someone offered a debt card that used bitcoin it wouldn't matter so much if newegg didn't.
123 2011-11-07 08:00:09 <gmaxwell> And that would allow the bitcoin userbase to grow until newegg does.
124 2011-11-07 08:00:10 <Diablo-D3> min0r: its a chicken and egg problem
125 2011-11-07 08:00:19 <Diablo-D3> and you'll have to solve both simulaniously
126 2011-11-07 08:00:31 <min0r> gmaxwell, thats a good idea...
127 2011-11-07 08:00:43 <min0r> is anyone doing that?
128 2011-11-07 08:01:05 <min0r> ive seen people sell those prepaid mastercards, but at insane premiums -- 20%+
129 2011-11-07 08:01:12 <gmaxwell> mtgox talked about doing it  ... well there is one, but the fee is kinda crazy... like 10% or something.
130 2011-11-07 08:01:35 <gmaxwell> I dunno what fee would make it sensible.. perhaps 1-2%
131 2011-11-07 08:03:30 <min0r> gox would be the perfect place to do it
132 2011-11-07 08:07:58 <min0r> btw, any theories on what happened to satoshi?  cashed out a million coins at $30 and living on a beach somwhere? ;)
133 2011-11-07 08:09:30 <Diablo-D3> he never existed to begin with
134 2011-11-07 08:09:56 <gjs278> he is from pokemon
135 2011-11-07 08:10:05 <gjs278> he spent his gains on playing cards
136 2011-11-07 08:56:41 <coderrr> gmaxwell, heh, "zbekistan had official denial them will possible architecture estruction and blade debris case represented"
137 2011-11-07 08:57:11 <coderrr> that was my two word phrase encoding of the same # thomasv did
138 2011-11-07 09:06:27 <epscy> min0r: there needs to be a compelling use for bitcoin, something which makes people want to use BTC rather than government backed money
139 2011-11-07 09:07:20 <epscy> this is why i think it will have a hard time taking off in the west
140 2011-11-07 09:07:55 <epscy> but there are many places in the world where people don't have access to basic banking
141 2011-11-07 09:08:32 <epscy> with computers getting cheaper (see raspberry pi) bitcoin has shot in the developing world
142 2011-11-07 09:09:51 <sipa> i think bitcoii has one indisputable advantage to traditional money -  at leat potdntially: complex transactions aka contracts
143 2011-11-07 09:09:58 <sipa> bitcoin
144 2011-11-07 09:10:55 <sipa> at least potentially
145 2011-11-07 09:11:23 <lianj> epscy: we have yet to see the raspberry pi ;)
146 2011-11-07 09:21:36 <coderrr> gmaxwell, not sure if its better or not https://gist.github.com/3a9e65bb5d8509abef7d
147 2011-11-07 09:22:56 <epscy> lianj: december apparently, though yes they did originally say november
148 2011-11-07 09:23:32 <epscy> sipa: i am worried that will just lead to bizzare money gambling schemes like we see on wall st
149 2011-11-07 09:24:11 <sipa> human nature...
150 2011-11-07 09:24:39 <epscy> meet the new boss, same as the old boss
151 2011-11-07 09:32:50 <UukGoblin> epscy, slight difference here is that end-users don't have to give their money to banks at all
152 2011-11-07 09:32:54 <UukGoblin> so banks have no money to gamble with
153 2011-11-07 09:33:47 <sipa> eventually, they will
154 2011-11-07 09:34:09 <epscy> UukGoblin: well in the west investment banking subsidises retail banking, without it we would be paying a monthly fee for current accounts and a fee everytime we use an ATM
155 2011-11-07 09:34:53 <epscy> it would be entirely possible to have a safe well capitalised retail bank, but consumers would consider it too expensive
156 2011-11-07 09:35:21 <UukGoblin> or useless
157 2011-11-07 09:35:24 <UukGoblin> like they should
158 2011-11-07 09:35:40 <epscy> i can see bitcoin "investment" companies becoming popular too, why hold your money when it could be working for you?
159 2011-11-07 09:36:04 <epscy> but yes it is nice that bitcoin ultimately gives the choice to the end user
160 2011-11-07 09:37:05 <UukGoblin> one reason to hold it yourself is that the economy doesn't get screwed up by large high risk trades
161 2011-11-07 09:38:30 <epscy> i would like to think that bitcoin would promote fiscal responsibility but in practice people don't make smarter financial decisions just because a commodity is deflationary
162 2011-11-07 09:39:11 <epscy> i don't think bitcoin is immune from the kind of financial crisis we have seen in the last 3 years with government backed money
163 2011-11-07 09:40:31 <UukGoblin> :-<
164 2011-11-07 09:41:11 <sipa> it's not, but for different reason
165 2011-11-07 09:41:16 <epscy> there are actually good reasons to discourage large holdings, economies that do well tend to have money moving about the most
166 2011-11-07 09:41:37 <epscy> it is really about the quality of the investment
167 2011-11-07 09:42:23 <UukGoblin> imho, problems with money shouldn't affect world's economies
168 2011-11-07 09:42:36 <UukGoblin> everyone should be able to get their own food and water
169 2011-11-07 09:43:01 <UukGoblin> but that's way outside the scope of #-dev ;-P
170 2011-11-07 09:43:49 <epscy> yeah, take it to #zeitgeist ;)
171 2011-11-07 09:44:39 <UukGoblin> CBA with their silly password policy
172 2011-11-07 09:44:55 <UukGoblin> and no, they're wrong too
173 2011-11-07 09:44:57 <epscy> loi, it exists?
174 2011-11-07 09:46:25 <UukGoblin> sure
175 2011-11-07 09:46:36 <UukGoblin> they used to have their own IRC server and stuff
176 2011-11-07 09:46:50 <sipa> cba?
177 2011-11-07 09:46:55 <epscy> i can see myself trolling them when i am drunk
178 2011-11-07 09:47:03 <UukGoblin> can't be a*ed
179 2011-11-07 09:47:10 <UukGoblin> epscy, lol
180 2011-11-07 09:53:26 <Eliel> bitcoin doesn't fix the main problem with money but it does fix many of the ones that have been built on money.
181 2011-11-07 09:54:49 <epscy> UukGoblin: http://pastebin.com/3c6y75xf not sure who was trolling who there
182 2011-11-07 09:55:33 <Eliel> :D
183 2011-11-07 09:56:17 <jeremias> Eliel: what have you been up to lately, still developing offline wallet? :)
184 2011-11-07 09:56:52 <Eliel> jeremias: unfortunately, haven't found too much time for it. Also, someone else kind of implemented it already.
185 2011-11-07 09:57:27 <jeremias> yep
186 2011-11-07 09:57:42 <Eliel> not exactly the same thing but close enough.
187 2011-11-07 09:58:23 <UukGoblin> epscy, lol
188 2011-11-07 11:51:03 <makomk> gmaxwell: I'm not convinced that implementing scrypt on FPGAs would be efficient even if it does fit on the larger ones...
189 2011-11-07 11:53:22 <cjdelisle> the whole point of scrypt is that is would be inefficient on fpga/asic/gpu
190 2011-11-07 11:53:44 <cjdelisle> and by inefficient I mean not much more efficient than a plain old processor.
191 2011-11-07 11:55:14 <makomk> Yeah, I know; gmaxwell doesn't seem to think that the implementation in litecoin etc achieves this though.
192 2011-11-07 12:02:59 <cjdelisle> for each "thread" you need 128k of fast memory, the die space adds up quick
193 2011-11-07 12:04:20 <Ycros> botnets, om nom nom
194 2011-11-07 12:16:40 <cjdelisle> Ycros: yes botnets are very very good at mining litecoin, that was understood and accepted as a cost of developing TBX (the first scrypt coin).
195 2011-11-07 14:19:16 <CIA-34> bitcoin: Gavin Andresen master * ra5f07cb / (3 files in 3 dirs): Static link on Linux. And better document release process on OSX. - http://git.io/g_Yg4g
196 2011-11-07 14:44:02 <BlueMattBot> Project Bitcoin build #85: STILL FAILING in 21 min: http://jenkins.bluematt.me/job/Bitcoin/85/
197 2011-11-07 15:02:16 <BlueMatt> gavinandresen: reason for link against QtCore/QtGui libs is that the qt-dev package on ubuntu doesnt have static packages...
198 2011-11-07 15:02:26 <BlueMatt> and I didnt see them when I briefly looked through synaptic
199 2011-11-07 15:02:43 <gavinandresen> BlueMatt: I was just writing an email about that....
200 2011-11-07 15:03:13 <gavinandresen> BlueMatt: so what's the right answer that will get us a 0.5 release out in the next couple of days?
201 2011-11-07 15:03:18 <BlueMatt> Id say requiring the installation of qt4 libs isnt unreasonable
202 2011-11-07 15:03:25 <gavinandresen> I agree
203 2011-11-07 15:03:47 <gavinandresen> Change the readme ...    what should the readme say?
204 2011-11-07 15:03:48 <Eliel> that might infact be preferable anyway.
205 2011-11-07 15:03:55 <gavinandresen> apt-get install qt4-dev   ?
206 2011-11-07 15:04:13 <Eliel> that's instructions for compiling?
207 2011-11-07 15:04:21 <gavinandresen> No, binary linux distribution
208 2011-11-07 15:04:35 <BlueMatt> I think they are there by default...
209 2011-11-07 15:05:06 <BlueMatt> (and I dont see qt-core packages installed on my system...)
210 2011-11-07 15:06:05 <sipa> i'm building git head bitcoin-qt on oneiric now
211 2011-11-07 15:06:09 <BlueMatt> libQtGui.so.4 is in libqtgui4
212 2011-11-07 15:06:13 <sipa> let's see which libs it links against
213 2011-11-07 15:06:28 <BlueMatt> and libQtCore.so.4 is in libqtcore4
214 2011-11-07 15:06:34 <BlueMatt> so those two are required
215 2011-11-07 15:07:03 <BlueMatt> and that should be it afaict
216 2011-11-07 15:08:12 <Eliel> I'd vote for including the installation command in the documentation. It'll do nothing if they're there and will be helpful even for people on other distros that don't have them preinstalled.
217 2011-11-07 15:11:36 <sipa> libaudio2 libboost-filesystem1.42.0 libboost-program-options1.42.0 libboost-system1.42.0 libboost-thread1.42.0 libc6 libdb5.1++ libexpat1 libffi6 libfontconfig1 libfreetype6 libgcc1 libglib2.0-0 libice6 libpcre3 libpng12-0 libqtcore4 libqtgui4 libsm6 libssl1.0.0 libstdc++6 libuuid1 libx11-6 libxau6 libxcb1 libxdmcp6 libxext6 libxi6 libxrender1 libxt6 zlib1g
218 2011-11-07 15:11:53 <sipa> that are that packages that provide libraries my built bitcoin-qt links against
219 2011-11-07 15:12:13 <sipa> i suppose most of them are dependencies of libqtgui4 anyway
220 2011-11-07 15:12:28 <BlueMatt> many of those are very explicitly deps of default install package ubuntu-desktop
221 2011-11-07 15:12:43 <gavinandresen> sipa: I committed a change this morning so you can qmake RELEASE=1 and static link to most of those
222 2011-11-07 15:14:30 <BlueMatt> RELEASE is a poor name imho, it is release for generic binaries, for distro packages, you would never use it
223 2011-11-07 15:14:41 <BlueMatt> (not that it really matters...)
224 2011-11-07 15:14:42 <sipa> agree
225 2011-11-07 15:15:07 <gavinandresen> happy to change it to something better... how about qmake BRUCE=1
226 2011-11-07 15:15:21 <BlueMatt> I prefer curtis
227 2011-11-07 15:15:29 <sipa> i vote qmake ANSWER=42
228 2011-11-07 15:15:45 <Eliel> how about just plain STATIC=1?
229 2011-11-07 15:15:52 <BlueMatt> na, thats too obvious
230 2011-11-07 15:16:04 <gavinandresen> RELEASE=1 also does some mac osx-specific stuff
231 2011-11-07 15:16:26 <BlueMatt> meh, whatever
232 2011-11-07 15:17:08 <Guest21618> heh
233 2011-11-07 15:17:12 <Guest21618> grr
234 2011-11-07 15:17:25 <jgarzik> that's better
235 2011-11-07 15:18:25 <sipa> libaudio2 libc6 libexpat1 libffi6 libfontconfig1 libfreetype6 libgcc1 libglib2.0-0 libice6 libpcre3 libpng12-0 libqtcore4 libqtgui4 libsm6 libstdc++6 libuuid1 libx11-6 libxau6 libxcb1 libxdmcp6 libxext6 libxi6 libxrender1 libxt6 zlib1g
236 2011-11-07 15:18:30 <sipa> those are left now
237 2011-11-07 15:18:35 <sipa> with RELEASE=1
238 2011-11-07 15:18:48 <gavinandresen> sipa: that looks about right.
239 2011-11-07 15:18:50 <BlueMatt> "libpng" grrr
240 2011-11-07 15:19:15 <BlueMatt> where is libpng used?
241 2011-11-07 15:19:23 <BlueMatt> (it should be removed if possible)
242 2011-11-07 15:19:35 <BlueMatt> (had problems with different distros and libpng deps in the past)
243 2011-11-07 15:20:03 <gavinandresen> src/qt/res/icons  has a bunch of png files
244 2011-11-07 15:20:13 <BlueMatt> m, well that would do it I suppose...
245 2011-11-07 15:21:02 <gavinandresen> So: what's the magic apt-get install  line that will make sure all of those dso's are installed?  Anybody?
246 2011-11-07 15:21:16 <gavinandresen> (I'm itching to update doc/README and tag a rc3)
247 2011-11-07 15:21:20 <sipa> libQtGui.so.4.7.4 depends on libpng12.so.0
248 2011-11-07 15:21:27 <BlueMatt> sudo apt-get install libqtcore4 libqtgui4
249 2011-11-07 15:21:32 <BlueMatt> should do it by itself Id assume
250 2011-11-07 15:21:45 <sipa> anyone know how to find all dependencies of an ubuntu package, recursively?
251 2011-11-07 15:22:00 <BlueMatt> probably some dpkg flag somewhere...
252 2011-11-07 15:23:04 <BlueMatt> apt-rdepends
253 2011-11-07 15:23:07 <BlueMatt> sipa: ^
254 2011-11-07 15:23:28 <sipa> don't have that
255 2011-11-07 15:23:36 <sipa> but that sounds like reverse dependencies?
256 2011-11-07 15:23:43 <BlueMatt> yep
257 2011-11-07 15:23:43 <sipa> apt-cache dotty seems to give what i need
258 2011-11-07 15:23:52 <gavinandresen> I just happen to have a fresh ubuntu server, I'll see what I get...
259 2011-11-07 15:23:53 <BlueMatt> m, well whatever
260 2011-11-07 15:28:02 <gavinandresen> Looks like I didn't get expat, pcre3, png, zlib, and a couple of others on (how do I tell what version of Ubuntu I'm running?)
261 2011-11-07 15:28:10 <sipa> lsb_release -a
262 2011-11-07 15:28:11 <BlueMatt> cat /etc/issue
263 2011-11-07 15:28:34 <gavinandresen> ... on 10.10 'maverick' server
264 2011-11-07 15:28:48 <sipa> ldd bitcoin-qt | fgrep / | cut -d '/' -f 2- | cut -d ' ' -f 1 | while read L; do dpkg -S "/$L"; done | cut -d ':' -f 1 | sort | uniq >/tmp/need.txt
265 2011-11-07 15:28:58 <sipa> (to get all packages bitcoin-qt needs)
266 2011-11-07 15:29:05 <BlueMatt> does bitcoin-qt run directly after installing the libqt packages?
267 2011-11-07 15:29:14 <gavinandresen> Installing bitcoin-qt on the server is dumb anyway, so maybe we don't care...
268 2011-11-07 15:29:22 <sipa> apt-cache dotty libqtgui4 libqtcore4 | fgrep '"' | fgrep ' -> ' | cut -d ' ' -f 1 | cut -d '"' -f 2 | cut -d ':' -f 1 | sort | uniq >/tmp/deps.txt
269 2011-11-07 15:29:28 <BlueMatt> on default ubuntu install, it should pretty much just run
270 2011-11-07 15:29:38 <sipa> (to get a list of all packages libqtgui4 and libqtcore4 depend on)
271 2011-11-07 15:29:43 <Eliel> gavinandresen: someone might want to run it with remote X.
272 2011-11-07 15:29:52 <Eliel> that is, ssh -X
273 2011-11-07 15:30:11 <BlueMatt> mmm, very true
274 2011-11-07 15:30:14 <gavinandresen> Eliel: true...
275 2011-11-07 15:31:35 <makomk> BlueMatt: you might want --as-needed btw.
276 2011-11-07 15:31:58 <BlueMatt> for?
277 2011-11-07 15:32:30 <makomk> bitcoin-qt, to avoid unnecesary dependencies on libraries you don't use directly.
278 2011-11-07 15:32:49 <BlueMatt> re: what?
279 2011-11-07 15:33:00 <gavinandresen> Is --as-needed a linker flag?
280 2011-11-07 15:33:06 <BlueMatt> yea
281 2011-11-07 15:33:07 <makomk> libpng in particular.
282 2011-11-07 15:33:25 <BlueMatt> libpng is not linked against bitcoin, its against qt (now that I think about it)
283 2011-11-07 15:33:36 <BlueMatt> so its not a problem (as both qt and png are system-installed)
284 2011-11-07 15:33:46 <BlueMatt> its when you static qt that it would be a problem
285 2011-11-07 15:34:48 <BlueMatt> it would, however, remove dep on libpthread libm and libgcc_s
286 2011-11-07 15:34:58 <BlueMatt> .so.0, .so.6 and .so.1 respectively
287 2011-11-07 15:35:22 <BlueMatt> not that that really matters, but...
288 2011-11-07 15:36:02 <makomk> You may find that without --as-needed you're linking against libpng directly in the dynamic case even though you don't need to, due to... quirks of library dependency handling on Linux. Dunno.
289 2011-11-07 15:36:51 <BlueMatt> I kinda doubt we are, though I suppose its theoretically possible...
290 2011-11-07 15:37:06 <makomk> Looks like it is actually doing the right thing.
291 2011-11-07 15:39:24 <gavinandresen> I'm afk for lunch for a while. Somebody please figure out the right "sudo apt-get install ..." line for the doc/README and submit a pull (or just get sipa to push)
292 2011-11-07 15:41:07 <sipa> gavinandresen: you installed libqtgui4 on maverick, and you didn't have libpng afterwards?
293 2011-11-07 15:42:21 <sipa> it seems to depend on it: http://packages.ubuntu.com/maverick/libqtgui4
294 2011-11-07 15:49:30 <Eliel> sipa: my guess is that the missing ones were already installed.
295 2011-11-07 15:49:39 <sipa> ?
296 2011-11-07 15:51:32 <Eliel> that is, I think gavin just checked which packages got installed along with qt4 as dependencies but didn't check if they were already installed.
297 2011-11-07 16:00:23 <gavinandresen> back.  Hmm?  No, I looked for /usr/lib/*png* and see nothing...
298 2011-11-07 16:01:00 <gavinandresen> is it named something other than libpng.so.#  or put someplace else?
299 2011-11-07 16:01:11 <sipa> /usr/lib/x86_64-linux-gnu/libpng12.so
300 2011-11-07 16:01:12 <sipa> here
301 2011-11-07 16:01:34 <gavinandresen> no x86_64 directory
302 2011-11-07 16:01:48 <sipa> did you check /lib ?
303 2011-11-07 16:01:50 <nanotube> find /usr/lib -name libpng* maybe?
304 2011-11-07 16:02:00 <sipa> dpkg -L libpng12-0
305 2011-11-07 16:02:14 <gavinandresen> sipa: didn't check lib... there it is /lib/libpn
306 2011-11-07 16:02:26 <nanotube> yea it's in /lib
307 2011-11-07 16:02:32 <gavinandresen> of COURSE it is....
308 2011-11-07 16:02:47 <sipa> sure, decoding png is such an essential system function
309 2011-11-07 16:02:53 <gavinandresen> sigh
310 2011-11-07 16:03:05 <nanotube> heh
311 2011-11-07 16:03:50 <sipa> gavinandresen: can you run apt-cache dotty libqtgui4 libqtcore4 | fgrep '"' | fgrep ' -> ' | cut -d ' ' -f 1 | cut -d '"' -f 2 | cut -d ':' -f 1 | sort | uniq >/tmp/deps.txt ?
312 2011-11-07 16:04:16 <sipa> if that's a superset of the dependencies of what we build, we're fine with just demanding the installation of libqtgui4
313 2011-11-07 16:04:18 <gavinandresen> sipa:  if I could figure out how to copy and paste into VirtualBox, yeah....
314 2011-11-07 16:05:33 <BlueMatt> should be able to if youve installed guest tools
315 2011-11-07 16:07:11 <gavinandresen> sipa:  that got me a 1,148-line deps.txt file (assuming I didn't make a typo... still can't figure out how to get copy&past with OSX VirtualBox working)
316 2011-11-07 16:07:43 <sipa> gavinandresen: quite possible, it's the recursive list of everything that will be installed if you install libqtgui4
317 2011-11-07 16:08:12 <gavinandresen> good enough for me-- I'm going to update the readme to suggest installing the two qt libraries, then tag rc3
318 2011-11-07 16:10:13 <nanotube> gavinandresen: i remember i had to install the 'guest additions' on the virtualbox client os in order to be able to copy/paste into a windows vbox vm...
319 2011-11-07 16:10:14 <sipa> actually, just libqtgui4 suffices
320 2011-11-07 16:10:18 <nanotube> maybe something similar in your case?
321 2011-11-07 16:10:22 <sipa> as that depends on libqtcore4
322 2011-11-07 16:10:29 <BlueMatt> gavinandresen: ^
323 2011-11-07 16:10:40 <gavinandresen> ok
324 2011-11-07 16:12:23 <CIA-34> bitcoin: Gavin Andresen v0.5.0rc3 * rd6245ca / doc/README : Mention Qt4 runtime dependencies - http://git.io/GYX4wg
325 2011-11-07 16:14:30 <gavinandresen> Ok, rc3 tagged.  Time to bin/gbuild....
326 2011-11-07 17:01:42 <gavinandresen> Lol xkcd today:  http://xkcd.com/974/
327 2011-11-07 17:02:36 <nanotube> hehe
328 2011-11-07 17:04:25 <sipa> gavinandresen: in the IT world, everything always happens automatically, but nothing just works :)
329 2011-11-07 17:05:59 <gavinandresen> sipa:  I use a couple of IT products that "just work" -- my iPhone (not all the apps Just Work, of course) and Chrome (ditto for its plugins)
330 2011-11-07 17:07:24 <sipa> right, not exactly what i mean - it's a not entirely accurate translation of a joke in dutch
331 2011-11-07 17:08:02 <gavinandresen> sipa:  ah... yeah, some humor was lost in translation
332 2011-11-07 17:08:13 <BlueMatt> humor always is...
333 2011-11-07 17:08:35 <gavinandresen> Can somebody sanity test bitcoin-0.5.0rc3-linux.tar.gz for me :  https://sourceforge.net/projects/bitcoin/files/Bitcoin/bitcoin-0.5.0/test/
334 2011-11-07 17:09:03 <gavinandresen> And bitcoin-0.5.0rc3-macosx.dmg
335 2011-11-07 17:09:15 <gavinandresen> (windows builds soon, I'm compiling now)
336 2011-11-07 17:09:42 <sipa> i'll do gitian builds when i get home
337 2011-11-07 17:10:42 <luke-jr> gavinandresen: where's the source releases?
338 2011-11-07 17:11:37 <gavinandresen> luke-jr: I think this will work:   https://github.com/bitcoin/bitcoin/tarball/v0.5.0rc3
339 2011-11-07 17:11:47 <luke-jr> lame
340 2011-11-07 17:11:54 <gavinandresen> (if it doesn't, let me know)
341 2011-11-07 18:45:06 <shurdeek> Does anyone know if the Atheros from bitcointalk.org forums is the same Atheros as the one from bitcoin wiki?
342 2011-11-07 18:45:24 <makomk> Heads up: merged mining's support for multiple chains is actually flawed (see: https://bitcointalk.org/index.php?topic=51069.0), it looks like the easiest solution for chains to take is to add their own block hashes to the Bitcoin coinbase.
343 2011-11-07 18:55:06 <luke-jr> makomk: nonsense
344 2011-11-07 19:01:03 <makomk> luke-jr: the alternative involves every new auxiliary chain co-ordinating with every other aux chain to make sure their IDs are as different as possible.
345 2011-11-07 19:01:47 <luke-jr> makomk: nope
346 2011-11-07 19:01:57 <makomk> Got a better solution then?
347 2011-11-07 19:02:06 <luke-jr> makomk: the alternative just means other aux chains use a sane cheat-prevention
348 2011-11-07 19:02:08 <Eliel> makomk: luke's way would be to ignore altchains (except namecoin) I think.
349 2011-11-07 19:02:12 <luke-jr> like not letting two blocks in order have the same proof
350 2011-11-07 19:02:35 <luke-jr> namecoin's algorithm was never intended for other aux chains in the first place
351 2011-11-07 19:04:18 <makomk> luke-jr: what happens if I'm a 51% attacker and mine a proof of with two versions of the same-numbered aux chain block, one sending coin A to address B and one sending it to address C, then confirm it with more blocks that use the same POW to confirm both?
352 2011-11-07 19:05:09 <makomk> It seems to me that doing it your way would make double-spending (and triple, quadruple, ...) free once you'd got enough hashpower.
353 2011-11-07 19:05:18 <luke-jr> hmm
354 2011-11-07 19:05:30 <denisx> luke-jr: I ported pushpoold to libevent2 and now this big memleak is gone
355 2011-11-07 19:05:41 <luke-jr> makomk: ok, you have a good point there
356 2011-11-07 19:05:55 <luke-jr> makomk: but there's still no reason a solution needs to drop namecoin support
357 2011-11-07 19:06:00 <luke-jr> denisx: i c
358 2011-11-07 19:08:12 <DrHaribo> makomk: I think you found a problem that hasn't really been thought of before :)
359 2011-11-07 19:08:14 <Eliel> or... when a miner sees two possible continuation forks in the blockchain, they could just avoid the risk of mining the wrong one and mine them both...
360 2011-11-07 19:08:51 <Eliel> although, I'm not sure if that'd actually be a bad thing.
361 2011-11-07 19:59:48 <Mqrius> Heya! Is there a way to dump a raw transaction? Either from the client, or from a site like blockexplorer, or from pywallet or something...
362 2011-11-07 20:00:20 <Mqrius> (If I can do it without compiling the client with modifications, that'd be great. Compiling in windows is a pain :( )
363 2011-11-07 20:07:02 <Mqrius> No?
364 2011-11-07 20:09:15 <Edward_Black> Mqrius I think I saw something like that, but don't quote me on that. let me trawl my x-coin favorites folder...
365 2011-11-07 20:10:17 <Lolcust> Maybe pywallet would do ? https://bitcointalk.org/index.php?topic=34028.0
366 2011-11-07 20:13:16 <Edward_Black> Yeah, i think pywallet can dump tx, among other things
367 2011-11-07 20:13:16 <Mqrius> Lolcust: Pywallet doesn't output it raw. It has some things that look raw-ish, but I'm not sure what they are (tx_k and tx_v)
368 2011-11-07 20:14:21 <Mqrius> Unless tx_v is actually the raw transaction, but I'm not sure. For one, it asks for both tx_k and tx_v to import a transaction
369 2011-11-07 20:14:48 <Lolcust> hm, let me double-check
370 2011-11-07 20:23:50 <ThomasV> why is bitcoin-info down?
371 2011-11-07 20:24:45 <Mqrius> lolcat: any idea?
372 2011-11-07 20:24:53 <Lolcust> Well, Mqrius, it appears that both txk and txv are the raw data the client soft stuffs in the wallet
373 2011-11-07 20:24:56 <Lolcust> source https://bitcointalk.org/index.php?topic=35469.msg440599#msg440599
374 2011-11-07 20:25:11 <Lolcust> both are needed to "manually" import a single transaction
375 2011-11-07 20:26:05 <Lolcust> and in fact they are the _only_ thing needed for transaction import, so, from that I reckon they are exactly what "raw transaction dump" is
376 2011-11-07 20:26:25 <Mqrius> Lolcust: I'm looking for the hex string that's broadcasted on the network
377 2011-11-07 20:26:35 <Mqrius> I'm building a javascript tool that generates a transaction, but I want to check it against a transaction already made, to see if there's any difference
378 2011-11-07 20:27:18 <Lolcust> I am no coder, but wouldn't it be easier to just broadcast it to testnet and see if it sticks, Mqrius ?
379 2011-11-07 20:27:38 <Lolcust> And methinks you would need wireshark to intercept the string it broadcasts
380 2011-11-07 20:28:08 <Lolcust> AFAIK that is the best way to grab what it is sending to the net as opposed to what it is writing to the wallet
381 2011-11-07 20:28:20 <Mqrius> Hmm, I could try that
382 2011-11-07 20:28:30 <Mqrius> I hope it's not too convoluted :P
383 2011-11-07 20:28:40 <sipa> Mqrius: what would you need, ideally?
384 2011-11-07 20:29:00 <Lolcust> Sending to testnet is piece o cake Mqrius :) Wireshark - not so much
385 2011-11-07 20:29:18 <sipa> just hex dump of raw transactions, import+export?
386 2011-11-07 20:29:30 <Mqrius> sipa: Any transaction hex string from an address I have the private key of, so I can reconstruct it in my custom tool to check if it matches
387 2011-11-07 20:30:08 <sipa> i'll write you a patch that exposes it as an rpc?
388 2011-11-07 20:30:29 <Mqrius> sipa: Ideally, without having to compile. I always have a load of trouble with that on windows >_<
389 2011-11-07 20:30:33 <Mqrius> Sorry I'm difficult :P
390 2011-11-07 20:30:51 <sipa> bah, windows
391 2011-11-07 20:30:52 <Lolcust> Mqrius, I can build modified bitcoins for windows =)
392 2011-11-07 20:31:32 <Eliel> Lolcust: do you feel like you understand OP_EVAL now?
393 2011-11-07 20:31:34 <Mqrius> Lolcust: I don't have much knowledge about testnet, how to set it up/get coins for it/what's different in transactions, etc. Figuring that out would cost me a decent amount of time
394 2011-11-07 20:31:48 <ThomasV> Mqrius: I have a python script that creates transactions
395 2011-11-07 20:32:10 <Mqrius> ThomasV: That would also work. Have you checked that it works?
396 2011-11-07 20:32:27 <Lolcust> Eliel saying that I "understand" it would be a mite of an overstatement, but I am starting to like the idea a lot
397 2011-11-07 20:32:44 <ThomasV> Mqrius: yes it works, it is part of my new client
398 2011-11-07 20:33:15 <Lolcust> I am still not partial to the "use same address format for simple things like sending auntie some coins, and mind-bending stuff like exotic escrow setups" philosophy
399 2011-11-07 20:33:19 <Mqrius> ThomasV: Can I get it somewhere? (Also, client in python... Nice. I like coding in python :))
400 2011-11-07 20:33:23 <ThomasV> Mqrius: http://ecdsa.org/electrum
401 2011-11-07 20:33:48 <ThomasV> the code is short
402 2011-11-07 20:33:52 <Guest62385> Hi, I'm testing bitcoin 0.5 rc3 and on tab transaction on the field to filter address, it's case senistive. Why?
403 2011-11-07 20:34:02 <Guest62385> its a bug?
404 2011-11-07 20:34:17 <sipa> addresses are case sensitive...
405 2011-11-07 20:34:52 <Eliel> Lolcust: I see no problem with it. The receiver made the address anyway, they're supposed to know how to use it :)
406 2011-11-07 20:35:06 <Guest62385> but if I edit the address to be for example: Leonardo and I type leonardo, I don't see any result
407 2011-11-07 20:35:30 <Lolcust> Eliel, maybe you expect them to read the manual, too ? =)
408 2011-11-07 20:35:40 <Lolcust> or Cthulhu forbid, wiki ?
409 2011-11-07 20:35:51 <Guest62385> I don't think the filter search should be case sensitive
410 2011-11-07 20:35:54 <Eliel> Guest62385: not a good idea to get used to handling addresses without paying attention to case.
411 2011-11-07 20:36:22 <Guest62385> to edited address
412 2011-11-07 20:36:57 <Eliel> Lolcust: who would need to read manual? user or the coder who's creating a wallet program? In the latter case, yes. In the former, I don't see any particular need, their wallet software ought to take care of things.
413 2011-11-07 20:37:19 <Eliel> Lolcust: unless, it's a sophisticated wallet, in that case it puts some expectations on the user too.
414 2011-11-07 20:38:17 <Guest62385> Why Bitcoin (tested on bitcoin 0.5 rc3) take so much time to exit?
415 2011-11-07 20:38:38 <Guest62385> to be closed
416 2011-11-07 20:38:45 <Lolcust> Eliel well, escrow/multisign and whatever super-duper stuff will be implemented with OP_EVAL later are "advanced wallet functionality"
417 2011-11-07 20:39:27 <Eliel> Lolcust: well, yes, if you use them, you need to understand them, of course.
418 2011-11-07 20:39:44 <Eliel> unless the wallet software can do it for you.
419 2011-11-07 20:40:04 <Lolcust> User getting exposed to it without being explicitly and repeatedly told that it is advanced stuff for big boys and girls is a nice way to get a bunch of butthurt "rich but not cow-pewter savvy" people whining across the internets
420 2011-11-07 20:40:42 <Eliel> Lolcust: you know, that's not criticism for OP_EVAL but rather the existence of complex transactions.
421 2011-11-07 20:41:42 <Lolcust> Eliel, basically, my position boils down to "it won't be a horrible burden to have a separate addresstype as catch-all for Mind Bendy Advanced Transactions, to make it very very explicit that there are "basic" send coins to auntie transactions and Complicated Ones that REQUIRE reading the manual to avoid trouble"
422 2011-11-07 20:42:47 <Lolcust> It's not a criticism of OP_EVAL as a code construct, just criticism of particualar non-crucial design facet of it which as far as I can tell is fairly trivial to do the "other way"
423 2011-11-07 20:43:03 <Eliel> Lolcust: An escrow kind of transaction requires more communication to build the transaction in the first place than just getting the address.
424 2011-11-07 20:43:09 <gavinandresen> Lolcust: huh?  Why would complex transactions require Aunt Ellie to read the manual?  All the complication should be taken care of by the software she's using
425 2011-11-07 20:44:33 <ThomasV> who is Aunt Ellie?
426 2011-11-07 20:44:53 <gavinandresen> Aunt Ellie is your doesn't-know-how-to-copy-and-paste relative....
427 2011-11-07 20:44:57 <Lolcust> Well, auntie's better understand how escrow works so that she doesn't end up in a situation when there is one party too many or something. And escrow is quite frankly not the last exotic transaction we will see, so it seems prudent, to me, to compartmentalize the relatively complex stuff to a separate "advanced" addresstype
428 2011-11-07 20:45:09 <ThomasV> it's my mom, not my aunt
429 2011-11-07 20:45:22 <gavinandresen> What's your mom's name?
430 2011-11-07 20:45:31 <Lolcust> Also, I'm kinda scared of theymos's implication that in the future we will have >256 exotic transaction types
431 2011-11-07 20:46:03 <ThomasV> gavinandresen: "password"
432 2011-11-07 20:46:20 <Eliel> Lolcust: even if there are >256 exotic transaction types, most users don't need to worry about them.
433 2011-11-07 20:46:36 <Eliel> just don't use them to receive coins.
434 2011-11-07 20:46:55 <gavinandresen> Lolcust: well, I'm proposing two bitcoin address types; the one we have now, and a new one that does the arbitrarily-complicated-op-eval stuff.
435 2011-11-07 20:47:19 <lolcat> Mqrius: About what?
436 2011-11-07 20:47:21 <Lolcust> gavinandresen ah, sorry, I must have misunderstood then.
437 2011-11-07 20:47:37 <Lolcust> I thought "classic" addresses are getting phased out post op-eval
438 2011-11-07 20:47:43 <gavinandresen> Lolcust: nope
439 2011-11-07 20:47:43 <Mqrius> lolcat: Sorry, wrong tab
440 2011-11-07 20:48:09 <Eliel> Lolcust: I'd guess they'll eventually get phased out. But that could take a long time.
441 2011-11-07 20:48:09 <Lolcust> gavinandresen, that's cool, I'm fine with two - address approach
442 2011-11-07 20:48:10 <lolcat> What is wrong with old versions?
443 2011-11-07 20:48:27 <Lolcust> nothing lolcat, they are gonna be fine )
444 2011-11-07 20:49:00 <lolcat> 0.3.21, is that a good version?
445 2011-11-07 20:49:03 <gavinandresen> I think the old address format will be around essentially forever, they'll just become more rare as people use signatures-on-multiple-devices-required addresses more and more
446 2011-11-07 20:49:27 <sipa> lolcat: why do you keep using such an old version?
447 2011-11-07 20:49:43 <lolcat> sipa: It is not in my respitories and compiling is a pain
448 2011-11-07 20:50:00 <sipa> which distribution/os?
449 2011-11-07 20:50:10 <Eliel> Lolcust: there's binaries even for linux.
450 2011-11-07 20:50:17 <Eliel> oops, meant lolcat :)
451 2011-11-07 20:50:36 <Lolcust> gavinandresen, well, as long as old addresses are around for just sending coins to auntie, we're gonna be fine I guess. Mind a stupid question, Gavin?
452 2011-11-07 20:50:50 <lolcat> sipa: Debian Squeeze I belive
453 2011-11-07 20:50:57 <gavinandresen> Lolcust: shoot.  I reserve the right to give a stupid answer, though
454 2011-11-07 20:50:59 <ThomasV> oh, blockchain.info is up again
455 2011-11-07 20:51:02 <sipa> auntie will probably be using a web-based bitcoin service
456 2011-11-07 20:51:18 <lolcat> What is these new adresses I keep hearing about?
457 2011-11-07 20:51:26 <sipa> that sevice will expose a url or bitcoin address for payments, that she can give to her neighbour for paying back  those eggs
458 2011-11-07 20:51:34 <gavinandresen> ... or an app on her iPhone.  (that you installed for her, because installing apps is too complicated....)
459 2011-11-07 20:51:41 <sipa> she doesn't know nor care what that address represents
460 2011-11-07 20:51:54 <Lolcust> lolcat , it's new kind of powerfull eldritch mathemagics
461 2011-11-07 20:51:59 <Eliel> Lolcust: nothing requires OP_EVAL transactions to be complex though.
462 2011-11-07 20:52:01 <sipa> whether it is a template from a complex txou script, or a simple refernece to a public key
463 2011-11-07 20:52:16 <Lolcust> Eliel well, simple transactions do fine without
464 2011-11-07 20:52:19 <lolcat> Lolcust: What was wrong with the old format?
465 2011-11-07 20:52:24 <gavinandresen> lolcat:  see https://en.bitcoin.it/wiki/BIP_0013 and https://en.bitcoin.it/wiki/BIP_0012
466 2011-11-07 20:52:44 <Eliel> Lolcust: true, but still, it's likely that once OP_EVAL is out there, any new transactions will use that, simple or not.
467 2011-11-07 20:53:18 <Lolcust> Eliel not sure I understand you )
468 2011-11-07 20:53:26 <sipa> i see no reason for that
469 2011-11-07 20:53:46 <gavinandresen> Eliel: it is easier to use the existing address type for single-address sends... (that's the way I coded it)
470 2011-11-07 20:54:05 <Eliel> Lolcust: ah, I mistyped, meant new transactions types.
471 2011-11-07 20:54:18 <Lolcust> gavinandresen well, I was curious if my personal favorite custom transaction (your hidden spend thing) might eventually see a comeback, and how extensive and dangerous and overhaul it would imply
472 2011-11-07 20:54:43 <gavinandresen> Lolcust: OP_EVAL is a lot like the hidden spend thing.  Nobody knows what the heck you're doing until you spend the coins you got.
473 2011-11-07 20:55:01 <ThomasV> gmaxwell: you were serious about that patent circumvention strategy?
474 2011-11-07 20:57:02 <lolcat> gavinandresen: Is it invented?
475 2011-11-07 20:57:06 <Lolcust> Now I wonder if there is a way to stuff some cunning way to confuse and befuddle coin-history analysis into the op-eval, especially in the context of my poor little Tenebrix and its eventually-upcoming laundr...Strong Decorrelator service  ^__^
476 2011-11-07 20:57:50 <gavinandresen> lolcat: is what invented?  OP_EVAL?  Yes, I implemented it a couple weeks ago.
477 2011-11-07 20:58:41 <gavinandresen> Lolcust: I don't think it helps, because All Is Revealed when you spend the coins.  And all has to be revealed, because everybody has to be able to validate your transaction.
478 2011-11-07 20:59:12 <lolcat> gavinandresen: So if I upgrade I have to use opeval?
479 2011-11-07 20:59:24 <gavinandresen> lolcat: no
480 2011-11-07 20:59:33 <Eliel> Lolcust: you could perhaps put together a transaction type for strong decorrelation that works in a p2p fashion. Design the transaction and require signatures from every participant for it to be valid.
481 2011-11-07 20:59:59 <sipa> lolcat: it's just a new kind of transactions
482 2011-11-07 21:00:05 <sipa> the old ones keep working fine
483 2011-11-07 21:00:06 <gavinandresen> Lolcust:  multisig might help with the "I don't trust the mixer not to run off with my coins" problem, though.
484 2011-11-07 21:00:42 <lolcat> sipa: Like a new kind of bitcoins?
485 2011-11-07 21:00:49 <sipa> no
486 2011-11-07 21:01:00 <sipa> just more things you can do with them
487 2011-11-07 21:01:10 <lolcat> I can buy relestate with them=
488 2011-11-07 21:01:14 <lolcat> ?
489 2011-11-07 21:01:24 <sipa> that's completely independent
490 2011-11-07 21:01:27 <Eliel> lolcat: if someone will sell you some.
491 2011-11-07 21:01:33 <Lolcust> gavinandresen well, yes, though you have to agree that it would be silly for a mixer with several millions as "clean-ish" coin buffer to run off with, say, a single million. But it would definitely improve BTC mixers in terms of trust
492 2011-11-07 21:01:36 <Eliel> but that doesn't really have to do with the tech :)
493 2011-11-07 21:01:44 <lolcat> sipa: then what can I do now that I couldn't before?
494 2011-11-07 21:02:16 <lolcat> Lolcust: what is a bitcoin mixer?
495 2011-11-07 21:02:34 <sipa> lolcat: it will allow you to create addresses that correspond to more than a single key
496 2011-11-07 21:02:46 <Lolcust> lolcat very long story, let me find you a wiki link to get you started
497 2011-11-07 21:02:54 <lolcat> sipa: Won't that give me twice the amount of bitcoins?
498 2011-11-07 21:02:58 <sipa> lolcat: no
499 2011-11-07 21:03:09 <sipa> please, read how bitcoin works
500 2011-11-07 21:03:28 <Lolcust> lolcat: https://en.bitcoin.it/wiki/Mixing_service
501 2011-11-07 21:03:51 <helo> so the address sent to will be the hash of two (or more) concatenated public keys?
502 2011-11-07 21:04:10 <sipa> helo: no, it will be the hash of a script that will do the checking
503 2011-11-07 21:04:19 <sipa> which will contains the public keys
504 2011-11-07 21:04:21 <helo> ahhh right
505 2011-11-07 21:04:35 <lolcat> Lolcust: Oo, so I can use it to make terrorist moeny traceless?
506 2011-11-07 21:04:42 <sipa> no
507 2011-11-07 21:04:46 <lolcat> sipa: I know how bitcoins work, I had a ot of them
508 2011-11-07 21:05:01 <Lolcust> lolcat not exactly, it's ... complicated
509 2011-11-07 21:05:05 <helo> i'm glad i found the OP_EVAL BIP, as it forced me to finally comprehend the Script wiki
510 2011-11-07 21:05:41 <sipa> lolcat: if you want to understand how OP_EVAL works, or what it allows you to do, you'll need to understand how bitcoin scripts works
511 2011-11-07 21:05:56 <sipa> otherwise, there are just some extra use cases one can least that it will enable
512 2011-11-07 21:06:29 <helo> lolcat: understanding https://en.bitcoin.it/wiki/Script is the next level of bitcoin understanding after mining, i think...
513 2011-11-07 21:06:35 <lolcat> What do I need bitcoin scripts for?
514 2011-11-07 21:06:48 <sipa> what do you need bitcoin for?
515 2011-11-07 21:06:51 <Lolcust> they are, essentially, the brain of the system
516 2011-11-07 21:07:07 <helo> lolcat: do you know why e-wallet services are dangerous to use?
517 2011-11-07 21:07:27 <lolcat> helo: Because they get hacked and the owners migth be evil
518 2011-11-07 21:07:39 <sipa> right
519 2011-11-07 21:07:45 <helo> lolcat: right... they have full access to everyone's bitcoin
520 2011-11-07 21:07:48 <sipa> now, what if you could prevent them from doing so?
521 2011-11-07 21:08:06 <lolcat> sipa: By having my own e-wallet service or own client!
522 2011-11-07 21:08:07 <sipa> allow them to handle your coins, but you still need to give permission for everything they do with it
523 2011-11-07 21:08:38 <lolcat> Sounds like a lot of extra effort
524 2011-11-07 21:08:54 <lolcat> Why not have the client on your server?
525 2011-11-07 21:09:09 <sipa> because not every human has a server to run it on
526 2011-11-07 21:09:20 <helo> lolcat: you can be sure that someone would have to compromise both the bitcoin wallet service and your local wallet to have access to your bitcoin
527 2011-11-07 21:09:29 <gavinandresen> ... and we're not all security-expert sysadmins who are 100% confident we will never ever be hacked.
528 2011-11-07 21:09:34 <sipa> plus running it yourself involves a lot a responsabilities many people don't want to take, e.g. server crash -> lose coins
529 2011-11-07 21:09:52 <sipa> wallet services are much better placed to deal with that for all but the 0.01% most tech-savvy ysers
530 2011-11-07 21:09:55 <gavinandresen> In fact, if you are 100% confident you'll never be hacked, then either you keep your bitcoins on paper or you're delusional.
531 2011-11-07 21:10:01 <helo> with OP_EVAL, a server crash would still lead to losing coins though, right?
532 2011-11-07 21:10:05 <lolcat> sipa: No, if the server crash you just reboot
533 2011-11-07 21:10:08 <jrmithdobbs> gavinandresen: mine are on paper
534 2011-11-07 21:10:09 <jrmithdobbs> lol
535 2011-11-07 21:10:14 <sipa> lolcat: what about hard drive crash?
536 2011-11-07 21:10:22 <lolcat> sipa: I use raid 0 or 1
537 2011-11-07 21:10:31 <sipa> what if your server room burns down?
538 2011-11-07 21:10:39 <jrmithdobbs> offsite backups
539 2011-11-07 21:10:44 <helo> so you could just be a little less paranoid about where you store your private key backups?
540 2011-11-07 21:10:58 <sipa> what if there is a bug in the software, so that the wallet file, including all its replicas are corrupt?
541 2011-11-07 21:11:02 <Lolcust> Well, so far, none of my crashes resulted in coinloss. Forgetting to back up that obscure windoze folder did
542 2011-11-07 21:11:05 <lolcat> sipa: It would be filled with co2 and the air would be sucked out. It is a professional datacenter
543 2011-11-07 21:11:28 <makomk> gavinandresen: that reminds me, suppose I'm a wallet service that wanted to use 2-of-2 multi-signature transactions to provide instant transactions by refusing to sign a double-spend. How practical would this be?
544 2011-11-07 21:11:42 <sipa> lolcat: right, if you're willing to do all that effort (have decent hosting, know about the risks, make backups, ...), you're very well placed to manage your own coins
545 2011-11-07 21:11:49 <sipa> lolcat: that's what bitcoin was intended for
546 2011-11-07 21:11:59 <sipa> but it's not what everyone wants to do
547 2011-11-07 21:12:12 <lolcat> sipa: Yeah, if you give money to a child (or someone stupid) they will loose them, be it cash or digital.
548 2011-11-07 21:12:43 <Lolcust> makomk seems practical on first glance, as long as you trust the escrowfolk
549 2011-11-07 21:12:58 <Lolcust> or in this case, Trust Provider
550 2011-11-07 21:13:17 <gavinandresen> makomk: that should be the same as a Wallet Protection Service...
551 2011-11-07 21:13:39 <helo> with OP_EVAL, you can have three private keys, only one of which you need to get your bitcoin, and a fourth private key that is held by the e-wallet service that is alaways required?
552 2011-11-07 21:13:48 <makomk> Except, of course, that a wallet protection service doesn't necessarily have to track transactions on their end.
553 2011-11-07 21:13:49 <sipa> for example
554 2011-11-07 21:13:54 <gavinandresen> ... and I expect the wallet protection and "green" address functions will get bundled together.
555 2011-11-07 21:15:14 <gavinandresen> helo: almost.  I'm proposing at most n-of-3 keys to start (I'm worried about people spamming the block chain just because they can)
556 2011-11-07 21:16:05 <helo> but one of two could be optional, with the 3rd mandatory, or all 3 mandatory, or any one of the three, etc, depending on the script?
557 2011-11-07 21:17:36 <gavinandresen> helo: BIP 11 proposes n-of-m, m < 3, transactions.  You can't do "a and b  OR  c and d", but you could do "any 2 of 3"  (and it'd be trivial to allow "any 2 of 4", which is what you want, if we allowed m == 4)
558 2011-11-07 21:17:37 <Guest62385> Why Bitcoin (tested on bitcoin 0.5 rc3) take so much time to exit?
559 2011-11-07 21:18:02 <helo> i see
560 2011-11-07 21:18:37 <phantomcircuit> Guest62385, it's waiting for things to write to disk
561 2011-11-07 21:18:42 <gmaxwell> ThomasV: Yes, If you care about it I'd want to do a more careful review before pronouncing it done but I was being serious.
562 2011-11-07 21:18:43 <gavinandresen> Guest62385: it tries to be extra careful cleaning up database handles, making sure transactions are all processed, etc.
563 2011-11-07 21:18:55 <gavinandresen> Guest62385: ... although I'm sure it could be optimized to shutdown faster.
564 2011-11-07 21:19:04 <Guest62385> ok
565 2011-11-07 21:19:11 <Guest62385> thanks for the answer
566 2011-11-07 21:19:31 <sipa> Mqrius: my 'rawtx' branch on github has a getrawtransaction RPC call
567 2011-11-07 21:20:02 <gavinandresen> sipa: have you looked at BIP 10?
568 2011-11-07 21:20:02 <ThomasV> gmaxwell: I do care about it. but I do not know what protection it offers, and against what
569 2011-11-07 21:20:46 <Eliel> gmaxwell: patent protection? got a link for more details somewhere?
570 2011-11-07 21:21:25 <ThomasV> Eliel: it's about circumventing a software patent
571 2011-11-07 21:21:32 <sipa> which one?
572 2011-11-07 21:21:36 <gavinandresen> sipa:   https://gist.github.com/1321518     Proposal for Standardized, Multi-Signature Transaction Execution
573 2011-11-07 21:21:47 <ThomasV> http://www.patentstorm.us/patents/5892470/description.html
574 2011-11-07 21:22:04 <gmaxwell> Eliel: he had a specific patent, and I suggested some changes to the implementation which would moot it.
575 2011-11-07 21:22:21 <sipa> gavinandresen: reading now
576 2011-11-07 21:23:01 <makomk> ThomasV: how the hell did that get granted?
577 2011-11-07 21:23:17 <ThomasV> makomk: lol
578 2011-11-07 21:23:29 <ThomasV> because people are dumb
579 2011-11-07 21:23:41 <gmaxwell> ThomasV: http://www.pattools.com/claimchart.html < feed it to that, then mark all the dependant claims (the ones that say 'a method of claim x' or the like) and ignore them. Then for each dependant claim explain why you're not doing _exactly_ what this claim says.
580 2011-11-07 21:23:45 <Lolcust> gavinandresen actually, blockchain spam should be managed by txfees. If Bob's wonderfull huge 4 of 14 multisign is too huge, he will pay an exuberant fee for it, no ?
581 2011-11-07 21:23:49 <gmaxwell> makomk: because it's really quite narrow.
582 2011-11-07 21:24:07 <gmaxwell> ThomasV: (and change the software to make your reason true, of course)
583 2011-11-07 21:25:06 <makomk> Claim 3 seems broad to me.
584 2011-11-07 21:25:17 <ThomasV> gmaxwell: "then for each dependent claim" <--- independent ?
585 2011-11-07 21:25:35 <gmaxwell> ThomasV: ha yea.
586 2011-11-07 21:25:51 <gmaxwell> makomk: "radix that is the same as the count of the selected plurality of words"  < _same_ I recommended adding a tiny amount of redundancy in the encoding.
587 2011-11-07 21:26:10 <gmaxwell> (e.g. make the radix of the grop of words larger by adding one more word to the dictionary)
588 2011-11-07 21:26:36 <makomk> In fact, it seems broad enough to cover the PGP word list, which suggests the examiner screwed up.
589 2011-11-07 21:27:12 <gmaxwell> makomk: nah, it doesn't cover pgp words because "selecting a plurality of words" must represent a digit. And pgp words are arguably 1:1, at least thats how they'd understand it.
590 2011-11-07 21:27:37 <makomk> Aha - the PGP word list has even and odd word lists for different byte positions or something.
591 2011-11-07 21:27:52 <gmaxwell> The invention here mostly seems to be e.g. how do you represent gf 2^n words using a non-power of two wordlist.
592 2011-11-07 21:28:36 <makomk> gmaxwell: huh, "radix that is the same as the count of the selected plurality of words" suggests one word per digit surely?
593 2011-11-07 21:29:11 <ThomasV> gmaxwell: what will it bring to me in practice? I mean, is it common for free software projects to adopt such a strategy? is the threat real?
594 2011-11-07 21:29:16 <gavinandresen> Lolcust:  yes, but I worry because transaction fees are broken right now-- clients and miners really need more flexibility to let fees go where the market decides, instead of us guessing what the right fees are.
595 2011-11-07 21:29:20 <gmaxwell> makomk: thats not how I read it. I'm reading it as there are a plurality of words, and they form digits. A single word is not a plurality.
596 2011-11-07 21:30:03 <makomk> Huh?
597 2011-11-07 21:30:04 <gmaxwell> ThomasV: It's probably not very real, but you're aware of it. Its risk you can easily mitigate if not eliminate with a half hours work.
598 2011-11-07 21:30:43 <Lolcust> gavinandresen I am mightily skeptical about market-driven txfees, though I agree that a faster, better way (compared to client version update) to update "mandatory" fees that play the antispam role is needed
599 2011-11-07 21:30:52 <ThomasV> gmaxwell: but then my code will look funny :-)
600 2011-11-07 21:31:03 <gmaxwell> makomk: pgp words uses single words to represent digits of base 2^8. In the scheme they're discussing you use 3 words to represent a single digit in base 2^32.
601 2011-11-07 21:31:30 <Lolcust> The problem with market-driven fee adjust, is that neither miners nor "simple users" care much about blockchain being spammed with a pile of megahuge transactions
602 2011-11-07 21:31:48 <Lolcust> Tragedy of the commons, and all that jazz Gavin :-(
603 2011-11-07 21:32:01 <sipa> nobody knows how that will evolve
604 2011-11-07 21:32:08 <gavinandresen> Lolcust: space in blocks isn't a commons, so I'm not worried about that...
605 2011-11-07 21:32:17 <gmaxwell> ThomasV: the more important thing to do is the changes you can't make later. e.g. you can change exactly how its calculated, but if you want to make the list slightly redundant you can't do that later.
606 2011-11-07 21:32:20 <gavinandresen> (miners "own" that space)
607 2011-11-07 21:32:46 <makomk> The trouble is that some of the claims don't seem to require that two-stage process of breaking the number down into digits and subdigits.
608 2011-11-07 21:33:43 <gmaxwell> makomk: yes, though if you read the rest of the text it seems to direct the interpertation. I strongly suspect if you check the file wrapper you'll see that the applicant clarified that outright (unless the office was totally asleep)
609 2011-11-07 21:34:06 <gmaxwell> makomk: If the holder were to push for that interpertation I'm pretty sure that would fail to lots and lots of prior art including SKEY.
610 2011-11-07 21:34:11 <Lolcust> gavinandresen, are you sure supermajority of miners (or rather, pool-ops) realize their ownership and its implications ? Because it seems to me, some  of those folks are just damn happy that, as Edward_Black puts it, "WITH BTC THEIR COMPUTER SHITS TEH DOLLERS!"
611 2011-11-07 21:34:52 <gmaxwell> makomk: IIRC the patent was an older one, before 2005 or so the file wrappers aren't online so you'll have to spend about $150 to get a copy of it. :(
612 2011-11-07 21:34:59 <Lolcust> I would not count on "market rationality", markets are neither rational nor perfectly efficient.
613 2011-11-07 21:35:37 <gmaxwell> (and for one of my projects I wouldn't bother pulling the wrapper on a patent as simple as this)
614 2011-11-07 21:36:11 <Lolcust> gavinandresen , IMHO, some strongly enforced baseline fee, just sufficient to keep folks from crapping into the blockchain for sheer griefer joy or out of petty opportunistic malice, is a necessity. Which seems to be exactly what we have now
615 2011-11-07 21:37:04 <Lolcust> And to me current fees seem quite bitey enough to keep people from making truck-sized 6-out-of-20 multisigns when they will get this opportunity
616 2011-11-07 21:37:21 <gavinandresen> Lolcust: agreed, we have to be careful.
617 2011-11-07 21:37:47 <sipa> but the maintainers of the software should not be the people who decide what the fees are
618 2011-11-07 21:38:43 <Lolcust> well, I am all for letting miners charge above baseline fee. Now, bypassing the minimum transaction fees imposed on huge transactions is another story
619 2011-11-07 21:39:19 <gavinandresen> I see a three step process:  1.  Make most clients not care about block-chain-size (do the headers-only-semi-lightweight-node-thing.  2. Un-hardcode the fees in both client and miner, but have reasonable guesses for defaults (that most people will probably use).  3. Get more liberal about transaction size
620 2011-11-07 21:40:10 <makomk> gmaxwell: from the rather short list of references cited, I'm guessing the patent office screwed up, but I'm no expert.
621 2011-11-07 21:40:17 <CIA-34> bitcoin: Daniel Folkinshteyn * rec60191eff37 supybot-bitcoin-marketmonitor/OTCWebsite/index.php: add coinabul ad
622 2011-11-07 21:40:27 <gavinandresen> I think most miners would be happy to include a LOT more free transactions per block, for example...
623 2011-11-07 21:40:38 <gmaxwell> makomk: that happens less often than most people think. Usually when it looks like that there is something in the file wrapper that makes it clear.
624 2011-11-07 21:40:44 <helo> when miners are spending 100x on GPU as they are storage, limiting the block size wouldn't really be a concern right?
625 2011-11-07 21:40:56 <Lolcust> gavinandresen I am worried that at this point the 4. is : observe as one ultra-liberal pool drags gargantual transactions into the blockchain for a tiny fee, to everyone's chagrin.
626 2011-11-07 21:41:03 <ThomasV> <gmaxwell> ThomasV: http://www.pattools.com/claimchart.html < feed it to that, then mark all the dependant claims (the ones that say 'a method of claim x' or the like) and ignore them. Then for each dependant claim explain why you're not doing _exactly_ what this claim says.
627 2011-11-07 21:41:14 <sipa> Lolcust: that may indeed be a possible future
628 2011-11-07 21:41:17 <ThomasV> gmaxwell: "explain" in my code ?
629 2011-11-07 21:41:30 <gmaxwell> makomk: e.g. the applicant has adopted the weird and highly limiting definition of some word. They get rejected and then they clarify " no no radix means only numbers than end in three!"
630 2011-11-07 21:41:50 <sipa> that the largest mining consortium says "we'll decide what we put in the chain, can't follow the load it generates to maintain a full node? stop mining yourself, and leave it to us"
631 2011-11-07 21:41:52 <Lolcust> sipa also, a possible way for pool cartels (which of course do not exist ^__^ plox don't kill me guys) to wage war
632 2011-11-07 21:41:55 <gavinandresen> Lolcust: yeah.... I've got some ideas on maximum block size rules that might fit in, too...
633 2011-11-07 21:42:16 <sipa> the decision to change the block size will probably be one of the hardest to make