1 2013-10-20 00:04:06 <jegs> omg latest bitcoin build failed
2 2013-10-20 00:04:25 <jegs> oh shit nevermind that weas in september
3 2013-10-20 00:04:43 <jegs> and only for BlueMatt
4 2013-10-20 00:04:44 <jegs> lol
5 2013-10-20 00:09:01 <sipa> just rebased it; checking whether it still builds
6 2013-10-20 00:26:34 <jegs> sipa: thanks! i was going to try that myself lol i appreciate it
7 2013-10-20 00:26:52 <jegs> what is the maximum length of a transaction id?
8 2013-10-20 00:27:24 <jegs> sorry i suck at using the wiki
9 2013-10-20 00:27:33 <jegs> lots of information but not the particular answer i'm looking for lol
10 2013-10-20 00:27:35 <sipa> ?
11 2013-10-20 00:27:42 <sipa> a transaction id is a 256-bit number
12 2013-10-20 00:27:55 <sipa> it's usually represented as a 64-character hex string
13 2013-10-20 00:28:07 <jegs> cool
14 2013-10-20 00:28:16 <jegs> i couldn't immediately ascertain that from https://en.bitcoin.it/wiki/Transactions lol
15 2013-10-20 00:28:28 <super3> my lastest pull request: https://github.com/bitcoin/bitcoin/pull/3106
16 2013-10-20 00:28:44 <super3> if there is anything I should include in there just let me know
17 2013-10-20 00:34:20 <super3> thanks sipa
18 2013-10-20 00:34:33 <sipa> yw
19 2013-10-20 00:35:11 <super3> one of these days i'm just going to put "SYN" in my commit message and see if anyone gets the joke.
20 2013-10-20 00:35:36 <sipa> i'm always sad that people use 'NACK' rather than 'NAK'
21 2013-10-20 00:36:39 <super3> heh
22 2013-10-20 00:37:00 <super3> i'm going to start using that then
23 2013-10-20 00:37:21 <super3> jegs: not your fault. the wiki is a mess.
24 2013-10-20 00:37:36 <sipa> the wiki is indeed very outdated and incomplete
25 2013-10-20 00:40:42 <jegs> ah cool, not just me then :p the wiki has a lot of good info on it though so no digs against whoever maintains it
26 2013-10-20 00:40:52 <jegs> sipa: + 80f2644...3d9c424 watchonly -> origin/watchonly (forced update)
27 2013-10-20 00:41:00 <jegs> first time i've seen "forced update" from git
28 2013-10-20 00:41:08 <jegs> what's that about?
29 2013-10-20 00:41:17 <sipa> ok
30 2013-10-20 00:41:37 <jegs> lol sorry hyper today just talking out loud
31 2013-10-20 00:41:40 <jegs> i'll ask google :p
32 2013-10-20 00:41:41 <sipa> so it happens because the current version of the remote branch is not a "descendant" from the previous version
33 2013-10-20 00:41:48 <jegs> ohh
34 2013-10-20 00:41:53 <super3> honestly i'm thinking of just building a website for Bitcoin docs
35 2013-10-20 00:41:57 <sipa> instead, it's a rewritten version based on newer history
36 2013-10-20 00:42:10 <sipa> super3: any reason for not just updating the wiki instead? :)
37 2013-10-20 00:42:40 <super3> organization of information, why not just build a reward/bounty system right in the website
38 2013-10-20 00:42:54 <super3> give people a reason to actually contribute
39 2013-10-20 00:43:03 <sipa> what would the reward be?
40 2013-10-20 00:43:14 <super3> probably micro bitcoins
41 2013-10-20 00:43:28 <sipa> oh, actual money
42 2013-10-20 00:43:31 <sipa> i doubt that will work
43 2013-10-20 00:44:06 <sipa> giving a monetary reward often makes people feel they're being paid for something rather than doing it voluntarily
44 2013-10-20 00:44:12 <super3> eh gamification works pretty well
45 2013-10-20 00:44:21 <super3> i'll give an example
46 2013-10-20 00:44:37 <sipa> so if it's not enough to be actually valuable, it can decrease participation
47 2013-10-20 00:44:56 <sipa> it's why you get a bag of presents when donating blood, rather than being paid
48 2013-10-20 00:45:55 <super3> i mean whenever i talk to a developer face to face about bitcoin docs we can pretty much agree that they need work
49 2013-10-20 00:46:11 <super3> im currently working on tutorials/documentation for bitcoin armory
50 2013-10-20 00:46:24 <super3> someone paid me to build it and offer it to the community
51 2013-10-20 00:46:37 <super3> i feel like it would work the same way for regular docs
52 2013-10-20 00:47:41 <tgs3> sipa: you need a Hayek's book? :)
53 2013-10-20 00:47:49 <super3> just throw out a 0.25 btc bounty on how to compile bitcoin on a certian special system, and someone has done that before could write it up easily
54 2013-10-20 00:48:22 <super3> i mean reddit karma points are worthess, but they are still one of the most trafficked sites on the net
55 2013-10-20 00:48:27 <tgs3> sipa: instrict/internal movitivation is good, but there are things like upkeep of family, rent, or being robbed by tax that one has to devote time to
56 2013-10-20 00:49:09 <tgs3> ideally you have both
57 2013-10-20 00:51:38 <sipa> super3: things change when the amounts are significant
58 2013-10-20 00:51:48 <sipa> 0.25 BTC != microbitcoins
59 2013-10-20 00:52:53 <sipa> super3: also, if it's specific questions, bitcoin.stackexchange.com works pretty well
60 2013-10-20 00:53:32 <sipa> but it's not very appropriate for reference material
61 2013-10-20 00:55:56 <sipa> tgs3: right, it's about when you _don't_ need the reward as an income
62 2013-10-20 00:56:38 <sipa> and "microbitcoins" certainly (at the current exchange rate) won't function as income for anyone i guess
63 2013-10-20 00:57:08 <jegs> sipa: configure is asking for libprotobuf after installing dependencies listed in the docs
64 2013-10-20 00:57:52 <sipa> so the documentation needs to be updated :)
65 2013-10-20 00:58:05 <sipa> qt in head now requires libprotobuf indeed
66 2013-10-20 00:58:46 <jegs> i see
67 2013-10-20 00:58:53 <jegs> who is in charge of build docs?
68 2013-10-20 00:59:02 <sipa> we're all volunteers
69 2013-10-20 00:59:10 <sipa> but i'm sure we'll fix things before releae
70 2013-10-20 00:59:13 <sipa> *release
71 2013-10-20 01:00:03 <jegs> i understand :) just wanted to do my part by communicating it to the right person
72 2013-10-20 01:00:19 <sipa> you're very welcome to fix it yourself
73 2013-10-20 01:00:40 <sipa> (or report an issue about it)
74 2013-10-20 01:01:16 <jegs> oh right, just clone on github then submit pull request right?
75 2013-10-20 01:01:28 <sipa> yes
76 2013-10-20 01:01:41 <sipa> you can even edit just on the github site, i believe
77 2013-10-20 01:01:45 <sipa> for single-file changes
78 2013-10-20 01:01:49 <sipa> though i've never tried that
79 2013-10-20 01:02:06 <jegs> cool shit
80 2013-10-20 01:25:11 <jegs> is this still true "The accounts code does not scale up to thousands of accounts with tens of thousands of transactions, because by-account (and by-account-by-time) indices are not implemented. So many operations (like computing an account balance) require accessing every wallet transaction."?
81 2013-10-20 01:28:40 <jegs> who here is working on multi-wallet?
82 2013-10-20 01:28:50 <CodeShark> ACTION
83 2013-10-20 01:29:19 <CodeShark> however, I've stopped working on the bitcoind multiwallet and instead have been working on an entirely new separate wallet app
84 2013-10-20 01:29:35 <Luke-Jr> â¦
85 2013-10-20 01:30:29 <CodeShark> I've long been voicing the opinion that wallet apps should not really be part of the verification/relay engine
86 2013-10-20 01:31:28 <Luke-Jr> nor should the GUI be part of the wallet app :p
87 2013-10-20 01:31:42 <jegs> yeah it's kind of ackward but i think that was done to get more bitcoin nodes out there right?
88 2013-10-20 01:32:09 <CodeShark> Luke-Jr: there are three basic use cases - GUI, CLI, and API
89 2013-10-20 01:32:54 <Luke-Jr> CodeShark: you forgot TUI
90 2013-10-20 01:33:01 <CodeShark> TUI?
91 2013-10-20 01:33:05 <Luke-Jr> curses
92 2013-10-20 01:33:13 <CodeShark> oh, lol
93 2013-10-20 01:33:21 <jgarzik> ACTION curses. I am so fucking tired of this stupid bitcointalk spam.
94 2013-10-20 01:33:22 <Luke-Jr> GUI/TUI/CLI can all be done on top of API
95 2013-10-20 01:33:30 <jgarzik> every time a freaking database gets hacked
96 2013-10-20 01:33:58 <jgarzik> don't these fools realize multiple mails per day will just decrease the spam effectiveness?
97 2013-10-20 01:34:17 <CodeShark> I have avoided bitcointalk for a while now - too much noise in there
98 2013-10-20 01:34:33 <sipa> haven't visited it in months
99 2013-10-20 01:34:44 <sipa> except when someone explicitly links to it
100 2013-10-20 01:35:10 <jgarzik> avoiding doesn't help in this case -- the database got hacked and posted
101 2013-10-20 01:35:27 <CodeShark> the bitcointalk db?
102 2013-10-20 01:35:29 <jgarzik> yes
103 2013-10-20 01:35:34 <sipa> ah, is that were all the weird spam comes from?
104 2013-10-20 01:35:36 <jgarzik> yes
105 2013-10-20 01:36:16 <sipa> multiple mails per day... from the same sender?
106 2013-10-20 01:36:30 <Luke-Jr> I think I get more from MtGox's old hack
107 2013-10-20 01:36:36 <Luke-Jr> tbh
108 2013-10-20 01:41:57 <jgarzik> sipa, same format, different senders
109 2013-10-20 01:42:42 <jegs> what does backupwallet do that cp doesnt?
110 2013-10-20 01:42:55 <jgarzik> sipa, some random appeal, sender always has a short name, the text is newly written each time, but each makes a pitch to send to bitcoin address XYZ
111 2013-10-20 01:43:10 <CodeShark> jegs: I believe just doing cp while a write operation is taking place can be dangerous
112 2013-10-20 01:43:12 <jgarzik> a lot of other similarity markers (I used to do email de-spamming... bleh. greylisting... bleh)
113 2013-10-20 01:43:35 <jgarzik> jegs, cp works fine, if bitcoind is turned off
114 2013-10-20 01:43:37 <super3> sipa: the point is not to provide an income for anyone, its more of a tip
115 2013-10-20 01:44:12 <sipa> super3: that's my point: if it's not an actually useful amount, it works counterproductive
116 2013-10-20 01:44:20 <jgarzik> Be careful about that guy named super3... he might be watching you with Google Glass ;p
117 2013-10-20 01:44:27 <super3> jgarzik: ha ha
118 2013-10-20 01:44:32 <jegs> i see thanks
119 2013-10-20 01:44:43 <sipa> though 0.25 BTC may be significant akready
120 2013-10-20 01:44:57 <jgarzik> jegs, backupwallet is a 'hot backup' or 'runtime backup'
121 2013-10-20 01:44:58 <super3> jgarzik: i'm too busy using it to watch the exchange prices
122 2013-10-20 01:45:42 <super3> ive got up 0.15 btc+ worth of tips from helping people out
123 2013-10-20 01:46:12 <super3> ive got 15 BTC for writing up documentation/tutorials for Bitcoin Armory
124 2013-10-20 01:46:44 <super3> so to say that people won't give out money for documentation is false
125 2013-10-20 01:46:53 <sipa> heh
126 2013-10-20 01:46:57 <sipa> i never said that
127 2013-10-20 01:47:08 <jgarzik> ACTION has even posted bounties for people to write docs
128 2013-10-20 01:47:12 <jgarzik> docs: quite important
129 2013-10-20 01:47:34 <super3> jgarzik: where are these bounties?
130 2013-10-20 01:47:49 <jgarzik> super3, no easy url to give :/ (cont.)
131 2013-10-20 01:47:49 <sipa> i said that a model where you pay small amounts for tasks works (for some value of small) worse than non-monetary rewards
132 2013-10-20 01:48:12 <sipa> as an incentive for doing the tasks
133 2013-10-20 01:48:35 <super3> sipa: well break down the tasks into something small that would warrant a reward
134 2013-10-20 01:49:13 <jgarzik> super3, gmaxwell and I have both said we would put up BTC for p2pool (gmax+me) user manual. "Push Here, Dummy" (PHD) manual that provides rock-simple instructions for setting up each component: bitcoind, p2pool, and your ASIC hardware (configured to p2pool).
135 2013-10-20 01:49:25 <sipa> all i'm saying is that it has to be a payment, not a tip-size payment that is promised
136 2013-10-20 01:49:43 <jgarzik> super3, then, I wanted to bounty more BTC to get that user manual translated into Chinese, maybe Russian too
137 2013-10-20 01:49:49 <sipa> tips are nice as an encouragement after the fact
138 2013-10-20 01:50:25 <sipa> super3: anyway, we're really not talking about the same thing anymore
139 2013-10-20 01:50:40 <super3> yeah lets move on
140 2013-10-20 01:50:45 <sipa> i just claimed that "microbitcoin" awards don't work :)
141 2013-10-20 01:50:46 <jgarzik> super3, the p2pool manual would have to cover all modern ASIC hardware: for each hardware { p2pool setup instructions }
142 2013-10-20 01:50:51 <jgarzik> super3, boring and droll, but useful
143 2013-10-20 01:51:27 <sipa> and the idea of having micropayment rewards for information was already tried by witcoin
144 2013-10-20 01:51:59 <CodeShark> it's highly inefficient if nothing else
145 2013-10-20 01:52:04 <jgarzik> Goal: decentralization++ when users try p2pool, rather than one of the Top Five Pools
146 2013-10-20 01:52:17 <super3> jgarzik: i just found some bounties, is the Chinese translation of Bitcoin wiki still valid?
147 2013-10-20 01:52:18 <CodeShark> much of this documentation requires someone who has a big picture grasp
148 2013-10-20 01:52:31 <CodeShark> not really possible to get someone like that if you just pay them pennies per paragraph
149 2013-10-20 01:52:59 <super3> i see it was posted way back in 2010
150 2013-10-20 01:53:35 <CodeShark> also, the cost of authorizing/managing so many tiny transactions can easily exceed the cost of the work itself
151 2013-10-20 01:53:51 <sipa> ?
152 2013-10-20 01:54:02 <sipa> oh, those aren't bitcoin transactions
153 2013-10-20 01:54:04 <super3> ok well you might have a point
154 2013-10-20 01:54:05 <CodeShark> well, someone has to review each contribution to see if it is payment-worthy
155 2013-10-20 01:54:19 <sipa> it was like stackoverflow
156 2013-10-20 01:54:29 <super3> well ok how if we threw out the micropayments, and just go with pure bounties
157 2013-10-20 01:54:39 <gmaxwell> technically it was like reddit but with an integrated micropayment system.
158 2013-10-20 01:54:45 <sipa> where tipping deducts your balance a bit, and being tipped increased it
159 2013-10-20 01:54:57 <gmaxwell> (I think it may have even been based on the reddit codebase)
160 2013-10-20 01:54:57 <sipa> right
161 2013-10-20 01:55:16 <super3> yeah +1 what CodeShark. bitcoin code is kinda dense
162 2013-10-20 01:55:17 <jgarzik> super3, probably not
163 2013-10-20 01:55:22 <jgarzik> *poof* baby bedtime
164 2013-10-20 01:55:35 <gmaxwell> it also cost very tiny amounts to post, and when people upvoted your posts you recieved some back. I thought it was pretty neat generally, but it didn't survive.
165 2013-10-20 01:55:58 <gmaxwell> (the cost of posting was supposted to be distributed to charities)
166 2013-10-20 01:56:17 <gmaxwell> (all tiny amounts like hundreths of a cent in USD equal value)
167 2013-10-20 01:56:38 <CodeShark> it's not publishing that costs money - it's reviewing the published content, editing it, compiling into a document with a consistent style, etc...
168 2013-10-20 01:56:38 <super3> well was its failure a problem with the actual idea, or just lack of people/marketing
169 2013-10-20 01:56:52 <gmaxwell> super3: a little of both, I think.
170 2013-10-20 01:57:14 <gmaxwell> I mean there was no real evidence that adding micropayments to a reddit thing actually made it work any better than reddit does.
171 2013-10-20 01:57:25 <CodeShark> the WWW made document publication essentially free
172 2013-10-20 01:57:36 <gmaxwell> CodeShark: yea, but eyeballs are valuable... thus spam.
173 2013-10-20 01:58:01 <gmaxwell> CodeShark: in theory "pay to post, get it paid back if people find your post useful" might help with that, but I think that hasn't been proven one way or another.
174 2013-10-20 01:58:43 <super3> well i found that you can get a lot of bodies to a website just offering crypto-cash, whether those people can be useful hasn't really be done right yet
175 2013-10-20 01:58:53 <CodeShark> but ultimately, for much of this stuff, we want to collect the documentation into a work that reads well and/or serves as a consistently-styled reference
176 2013-10-20 01:59:03 <gmaxwell> part of the problem, of course, is "I have something USEFUL to tell the world, and you want me to PAY to do so? screw you" unless the amounts involved are uselessly trivial.
177 2013-10-20 01:59:11 <super3> although you can say that reddit bitcoin-tip kinda does this to an extent
178 2013-10-20 01:59:26 <gmaxwell> reddit bitcoin tip is really poorly done.
179 2013-10-20 01:59:27 <CodeShark> you don't want to be searching through thousands of tiny posts to find things
180 2013-10-20 01:59:34 <super3> gmaxwell: how so?
181 2013-10-20 02:00:10 <super3> gmaxwell: it seems to move a decent amount of volume, and has exposes a good number of people to bitcoin
182 2013-10-20 02:00:12 <gmaxwell> super3: indivigual transactions for every tip, system costs more in transaction fees than it distributes in tips or darn near so. They also manage to pull random addresses for people from random public postings.
183 2013-10-20 02:00:27 <gmaxwell> The witcoin design was strictly superior, from a technical perspective.
184 2013-10-20 02:01:10 <super3> gmaxwell: well id say introducing someone to crypto-currency is more valuable than any txfee
185 2013-10-20 02:01:40 <gmaxwell> super3: not if the introdtion expirence is a bad one because you've recieved an amount so small you can't even uselfully pass it on to other people.
186 2013-10-20 02:01:40 <sipa> i disagree
187 2013-10-20 02:02:23 <sipa> i don't think bitcoin is ready for mass adoption
188 2013-10-20 02:02:27 <gmaxwell> (same problem those some of those clickfraud-faucet sites have... they pay people dust and then people are irritated when they can't actually spend it)
189 2013-10-20 02:02:27 <super3> gmaxwell: generalizing, not all the tips are small, and there is a limit on the smallest tip you can send
190 2013-10-20 02:02:46 <super3> gmaxwell: which is why we are working on trying to solve the microtransaction problem
191 2013-10-20 02:02:55 <sipa> how?
192 2013-10-20 02:03:07 <gmaxwell> super3: Whats to solve? witcoin had it solved for that context.
193 2013-10-20 02:03:15 <super3> Coinbase and Inputs.io can already do
194 2013-10-20 02:03:21 <super3> "microtransactions"
195 2013-10-20 02:03:33 <super3> not the real implementation like BitcoinJ, but hey it works
196 2013-10-20 02:03:47 <gmaxwell> micropayment channels are not applicable to this.
197 2013-10-20 02:05:07 <super3> Bitcoin is still very new to people, things like Bitcoin-tip are a novelty that help them understand the concept
198 2013-10-20 02:05:57 <gmaxwell> I don't think they're bad to have, but they way they are done on reddit results in a not very good user expirence. In any case, unexpected post-hoc tips are not the same thing as very tiny bounties.
199 2013-10-20 02:06:34 <super3> gmaxwell: any suggestions on how they could be done better?
200 2013-10-20 02:06:58 <gmaxwell> At least we seem to have observed that small bounties appear to be demotivating, they seem to have the opposite of the intended effect. Random unexpected tips don't seem to be harmful at least.
201 2013-10-20 02:07:22 <gmaxwell> super3: by using an actual micropayment system, preferrably one integrated into the site.
202 2013-10-20 02:08:20 <super3> that requires the blessing of the people running the site and a decent codebase
203 2013-10-20 02:08:30 <super3> sometimes you just have to work with what you got
204 2013-10-20 02:08:46 <sipa> it could be done externally just as well
205 2013-10-20 02:09:03 <sipa> it's just a slightly worse user experience if not well integrated
206 2013-10-20 02:09:20 <gmaxwell> integration does, yes... kinda, though you can use extensions to get integration even externally. Use of a micropayment system doesn't, however.
207 2013-10-20 02:09:44 <super3> id venture to say that Bitcoin doesn't really have the greatest user experience
208 2013-10-20 02:09:50 <super3> in general
209 2013-10-20 02:09:57 <sipa> it certainly doesn't
210 2013-10-20 02:10:19 <gmaxwell> And trying to use it for micropayments is somewhat doomed to have the worst of that expirence.
211 2013-10-20 02:10:54 <super3> i mean with time we will get there, but that time is not now
212 2013-10-20 02:10:57 <gmaxwell> Bitcointipbot or whatever could be replaced with a micropayment system. Support for displaying balances and such for that system could be added to reddit enhancement suite or a similar tipbot only extension.
213 2013-10-20 02:11:17 <gmaxwell> super3: or maybe not, if people keep using it poorly and expecting someone else to solve the problems maybe we won't get there.
214 2013-10-20 02:12:14 <super3> im working on a system to better integrate the micropayment hacks, and i'm sure many people are doing the same thing
215 2013-10-20 02:12:39 <gmaxwell> (esp since people are now frequently building services which are accutely hostile to viable micropayment systems existing. :( )
216 2013-10-20 02:12:57 <super3> i mean the micropayments problem has been around for a while, even before Bitcoin
217 2013-10-20 02:13:20 <gmaxwell> super3: sorry, I've seen a lot of people saying this stuff. But frankly none of them have produced a solution even as good as the witcoin one.
218 2013-10-20 02:13:52 <gmaxwell> super3: bitcoin potentially cuts through a lot of the problems that prevent solving it otherwise. E.g. witcoin didn't exist prior to bitcoin.
219 2013-10-20 02:14:40 <super3> yes so now that we have a good foundation we can go from there, which im optimistic about
220 2013-10-20 02:15:23 <super3> but its going to take a bit of time for people to work through it. the people that can solve these problems are just finding out about Bitcoin now
221 2013-10-20 02:15:49 <super3> so they still have to learn the core protocol, APIs, read the paper, before they can go off and do all this cool stuff
222 2013-10-20 02:16:38 <gmaxwell> No they don't.
223 2013-10-20 02:17:07 <gmaxwell> I mean, certantly better that they do. But not understanding bitcoin at all hasn't even stopped people from trying to write implementations of it. :P
224 2013-10-20 02:17:07 <super3> gmaxwell: ?
225 2013-10-20 02:17:14 <jegs> how do rejections work? like what does the JSON RPC API return?
226 2013-10-20 02:17:42 <jegs> i'm not knowledge enough to craft an invalid transaction so i can't test this myself :p
227 2013-10-20 02:17:42 <super3> jegs: like error messages?
228 2013-10-20 02:17:47 <jegs> yeah
229 2013-10-20 02:17:49 <jegs> error messages
230 2013-10-20 02:17:54 <jegs> or a sort of 'transaction status'
231 2013-10-20 02:18:05 <gmaxwell> And Bitcoin has existed for over 5 years now. We had viable microtransaction implementations in 2010. But people seem to reinvent the wheel or just not bother to use things pretty often.
232 2013-10-20 02:19:48 <super3> gmaxwell: but a majority of people didn't know it existed till this year, which i find quite sad for anyone in CS
233 2013-10-20 02:20:10 <gmaxwell> the majority of people in the world still don't know bitcoin exists and may never know it exists. :)
234 2013-10-20 02:21:02 <super3> gmaxwell: as long as our growth rate stays like this most of them will eventually
235 2013-10-20 02:22:33 <super3> jegs: trying sending some messed up transactions on the test-net
236 2013-10-20 02:23:10 <super3> jegs: might be worth playing with the coinbase or blockchain.info apis.
237 2013-10-20 02:23:28 <gmaxwell> jegs: rejections for what? it's easy to make a bad json rpc call.
238 2013-10-20 02:23:51 <gmaxwell> $bitcoind getrawtransaction jiffiepop
239 2013-10-20 02:23:59 <super3> heh
240 2013-10-20 02:24:43 <jegs> not like that
241 2013-10-20 02:24:56 <jegs> i mean someone trying to pay from an account with no money or something like that
242 2013-10-20 02:25:02 <jegs> the standard clients aren't going to let you do that
243 2013-10-20 02:25:16 <jegs> so testing this would be time consuming and difficult for me
244 2013-10-20 02:26:15 <super3> jegs: testing is always time consuming and difficult
245 2013-10-20 02:26:21 <super3> this is why you write the best unit tests you can
246 2013-10-20 02:26:38 <super3> http://json-rpc.org/wiki/specification might help a little bit
247 2013-10-20 02:26:45 <jegs> no i mean i would have to learn a lot of stuff in order to craft a raw transaction and have it broadcast
248 2013-10-20 02:27:12 <jegs> thanks i know the info is there just "ain't got no time for that shit" right now lol
249 2013-10-20 02:27:24 <gmaxwell> jegs: $ bitcoind sendfrom "random" 1GMaxweLLbo8mdXvnnC19Wt2wigiYUKgEB 10000000 < easy to test, unless you actually have 10 million coins, in which case you can easily hire me to help you more. :)
250 2013-10-20 02:27:25 <super3> jegs: maybe i could help more if i had actually help you
251 2013-10-20 02:27:30 <CodeShark> bitcoind will broadcast it for you
252 2013-10-20 02:27:42 <CodeShark> sendrawtransaction
253 2013-10-20 02:27:43 <gmaxwell> jegs: oh if you want to test a bad raw transaction, just manually twiddle the amount in the hex.
254 2013-10-20 02:27:43 <super3> oops
255 2013-10-20 02:28:07 <CodeShark> and decoderawtransaction is useful for debugging, testing as well
256 2013-10-20 02:28:37 <CodeShark> although the error messages could be a little more granular for a better tool
257 2013-10-20 02:29:03 <jegs> ok i'm sure i'll get to doing that at some point but for my immediate purposes... bad transactions simply aren't included in any 'block' so you can't lookup a transaction to see if it has a 'rejected' status right?
258 2013-10-20 02:29:05 <gmaxwell> $ bitcoind createrawtransaction '[{"txid":"9afba9a11cd3a315b4cf8cafa794d6f5f9cec4088202c8cf124963b8c0c6bccb","vout":0}]' '{"1GMaxweLLbo8mdXvnnC19Wt2wigiYUKgEB":1000000}'
259 2013-10-20 02:29:13 <gmaxwell> ^ tada, invalid raw transaction.
260 2013-10-20 02:29:21 <jegs> ok i'll put that in
261 2013-10-20 02:29:26 <gmaxwell> jegs: there is no such thing as "rejected" status.
262 2013-10-20 02:29:45 <gmaxwell> bad transactions are just instantly rejected and thereafter regarded as if the node had never heard of them.
263 2013-10-20 02:29:52 <CodeShark> if you look at the debug.log for bitcoind you can see whether it rejects your transactions when you sendrawtransaction
264 2013-10-20 02:29:58 <jegs> ok great you are confirming my understanding then
265 2013-10-20 02:30:16 <gmaxwell> CodeShark: you also get a return result from the sendrawtransaction command, but it's usually not very informative.
266 2013-10-20 02:30:38 <CodeShark> for some of the stuff I've done I've resorted to inserting granular tracers into bitcoind :p
267 2013-10-20 02:30:39 <super3> gmaxwell: looking at your proof of storage post, had an idea along those lines, but your concept is way better
268 2013-10-20 02:32:07 <sipa> super3: depending on what you mean by accounts, bitcoin *will* let you spend more from an account than it has (though the wallet needs to have enough funds)
269 2013-10-20 02:32:25 <CodeShark> the term "account" in its current use in bitcoind's wallet is a complete misnomer, IMHO
270 2013-10-20 02:32:33 <CodeShark> it should just be considered a label
271 2013-10-20 02:32:42 <CodeShark> transactions/addresses can be labeled
272 2013-10-20 02:32:46 <CodeShark> that's all
273 2013-10-20 02:33:14 <sipa> i guess it would be less confusing if label and accounts weren't confounded
274 2013-10-20 02:33:24 <CodeShark> the wallet still considers all the coins it has to be part of the same pile
275 2013-10-20 02:34:09 <sipa> but the accounts system as a whole is rarely useful
276 2013-10-20 02:34:20 <phantomcircuit> CodeShark, no it's an accounting account
277 2013-10-20 02:34:22 <phantomcircuit> it's not really labels
278 2013-10-20 02:34:24 <phantomcircuit> but it's not a very useful construct for anything serious since it's impossible to maintain backups
279 2013-10-20 02:34:47 <CodeShark> an accounting account would not let you have negative balances
280 2013-10-20 02:34:53 <CodeShark> you'd have to explicitly transfer funds
281 2013-10-20 02:35:14 <sipa> initially, accounts couldn't go negative
282 2013-10-20 02:35:28 <sipa> but that was impossible to actually enforce consistently
283 2013-10-20 02:35:43 <sipa> as transactions can get reverted
284 2013-10-20 02:35:45 <gmaxwell> super3: there were a lot of ideas along those lines that came first, but that one is both simple and should actually achieve the stated goal. (I dunno why it took so long to come up withâ¦)
285 2013-10-20 02:36:03 <CodeShark> the way bitcoind's wallet is typically used, the "accounting system" really doesn't make sense - it's really a single account, in which you can label transactions to indicate from whom they were received
286 2013-10-20 02:36:18 <sipa> that's not the account system
287 2013-10-20 02:36:35 <phantomcircuit> CodeShark, accounting accounts most certainly can go negative
288 2013-10-20 02:37:00 <sipa> the account system lets you mark addresses as "transactions to this credit account X"
289 2013-10-20 02:37:07 <phantomcircuit> actually the entire wallet can have a negative balance if you set the confirmation watermark to 0
290 2013-10-20 02:37:25 <sipa> and it lets you mark outgoing transactions as "deduct account Y"
291 2013-10-20 02:37:50 <CodeShark> in any case, the bitcoind wallet isn't designed to manage serious funds - it's only a single end-user personal wallet
292 2013-10-20 02:37:50 <sipa> the fact that these to-account tags on addresses overlap with labels is just confusimg
293 2013-10-20 02:38:10 <CodeShark> and in most such uses, what you care about is who sent you what and to whom did you send what
294 2013-10-20 02:38:14 <CodeShark> and it's a single account
295 2013-10-20 02:38:32 <CodeShark> if you truly want to manage multiple "accounts" using bitcoind you need multiple wallets
296 2013-10-20 02:38:43 <sipa> well the account system was designed for accounts on a shared wallet
297 2013-10-20 02:39:03 <super3> gmaxwell: how best could this be used to implement a distributed storage system?
298 2013-10-20 02:39:05 <sipa> if you want multiple wallets, you want multiple wallets - not accounts in the same wallet
299 2013-10-20 02:39:24 <sipa> and yes, for many uses multiple wallets is the right approach
300 2013-10-20 02:39:37 <CodeShark> sipa: in what I've been working on, an account consists of a script chain along with a particular signing policy
301 2013-10-20 02:39:37 <sipa> but not all
302 2013-10-20 02:40:15 <sipa> shared wallets make sense, they give you lower fees for example than keeping the coins separate
303 2013-10-20 02:40:21 <CodeShark> the signing is entirely separate, we have a separate abstraction called a "keychain"
304 2013-10-20 02:40:24 <gmaxwell> sipa: aren't you asleep?
305 2013-10-20 02:40:36 <sipa> gmaxwell: i believe i am
306 2013-10-20 02:40:41 <gmaxwell> Okay. Just checking.
307 2013-10-20 02:41:15 <CodeShark> the account specifies a set of scripts to observe/create new transactions for along with a signing policy which requires specific keychains for signing
308 2013-10-20 02:41:24 <MC1984> am i right in thinking that this accounts system is akin to email folders vs gmails 'labels'
309 2013-10-20 02:41:41 <sipa> CodeShark: ok, your account is what we call a wallet, i guess
310 2013-10-20 02:42:33 <gmaxwell> sipa: semi-shared wallets are perhaps a useful context. E.g. being able to select multiple wallets when making a transaction "You can spend from any, and potentially all of these"
311 2013-10-20 02:42:40 <super3> gmaxwell: also will this tree only have to be generate once?
312 2013-10-20 02:42:43 <gmaxwell> s/context/concept/
313 2013-10-20 02:42:56 <CodeShark> the term "wallet" is not really used in typical banking, though
314 2013-10-20 02:43:04 <gmaxwell> super3: the idea is to be once per-peer which you are proving storage to.
315 2013-10-20 02:43:19 <CodeShark> and in the case of bitcoind, the term "wallet" confuses accounts, signing policies, and keychains
316 2013-10-20 02:43:29 <sipa> CodeShark: s/wallet/bank/
317 2013-10-20 02:43:50 <sipa> your bank doesn't actually keep your bills and coins separately
318 2013-10-20 02:44:00 <sipa> it only remembers how many you own
319 2013-10-20 02:44:07 <sipa> that's an account
320 2013-10-20 02:44:16 <super3> gmaxwell: got it my main concern is that if you are storing a large number of files this way it might be a bit i/o intensive
321 2013-10-20 02:44:25 <sipa> i think it maps very closely to bitcoind's accounts :)
322 2013-10-20 02:44:29 <gmaxwell> yea, the idea of something that keeps coins seperate is non-existant in finance, it's only in bitcoin that people keep trying to impose it.
323 2013-10-20 02:44:40 <gmaxwell> super3: this isn't for storing files.
324 2013-10-20 02:45:06 <CodeShark> sipa, each bank account can have particular policies regarding its access
325 2013-10-20 02:45:16 <gmaxwell> super3: it's for proving that you're (temporarily) using disk space to prove that you're not filling up a network's capacity.
326 2013-10-20 02:45:42 <sipa> CodeShark: sure, and arguably we'd need that in accounts
327 2013-10-20 02:45:55 <CodeShark> in the case of bitcoin, if you want to have truly separate policies per account you can't really just lump in all the utxos
328 2013-10-20 02:46:17 <sipa> CodeShark: to continue the analogy, the bank owns all coins
329 2013-10-20 02:46:19 <CodeShark> although I do like gmaxwell's suggestion of policies that do allow it
330 2013-10-20 02:46:25 <super3> gmaxwell: i'm aware, trying to apply it to the storing files problem though
331 2013-10-20 02:46:32 <sipa> a wallet owns all coins
332 2013-10-20 02:46:43 <CodeShark> no, keychains own coins
333 2013-10-20 02:46:46 <sipa> the accounts are fpr different customers of the bank
334 2013-10-20 02:46:55 <sipa> well call it as you like
335 2013-10-20 02:47:03 <CodeShark> coin ownership is established by who owns the private keys, not who owns the list of scripts to watch
336 2013-10-20 02:47:16 <sipa> i own my coins on mtgox
337 2013-10-20 02:47:32 <sipa> though they have the keys
338 2013-10-20 02:47:36 <CodeShark> I'm talking at the cryptographic level
339 2013-10-20 02:47:47 <CodeShark> if you truly want to secure your own coins rather than trusting someone else
340 2013-10-20 02:47:56 <sipa> yes, and accounts have nothing to do with that level
341 2013-10-20 02:48:21 <CodeShark> well, they do if you want to retain separate policies on them
342 2013-10-20 02:48:21 <sipa> which isn't always the abstraction you want
343 2013-10-20 02:48:29 <super3> although i like the idea of a proof-of-storage coin
344 2013-10-20 02:48:30 <CodeShark> otherwise you still have to trust someone with all the keys
345 2013-10-20 02:48:33 <sipa> but that doesn't make it a useless one
346 2013-10-20 02:48:42 <sipa> of course
347 2013-10-20 02:48:55 <sipa> that's exactly what accounts are designed for
348 2013-10-20 02:49:04 <gmaxwell> CodeShark: I'm sad to hear that you've stopped working on multiwallet in bitcoin-qt. Finishing that would would be a real benefit to a lot of people in the short term.
349 2013-10-20 02:49:11 <CodeShark> one of the killer features of bitcoin is the ability to manage security at the cryptographic level rather than just trusting some authentication server somewhere
350 2013-10-20 02:49:18 <sipa> maintaining balances for different identities within one set of coims
351 2013-10-20 02:49:45 <sipa> CodeShark: sure, which is why muktiwallet would be awesome
352 2013-10-20 02:50:00 <gmaxwell> (and not just the users, we seem to be having increasing problem with people not running full nodes. A lack of multiwallet support is one of the causes, and thats going to get a lot worse when trezor comes out and there is just no way to use one w/ bitcoin-qt)
353 2013-10-20 02:50:37 <super3> gmaxwell: i think the main problem is just sync time
354 2013-10-20 02:50:58 <super3> people don't want to wait 2 days to use their client
355 2013-10-20 02:51:02 <gmaxwell> super3: this isn't an issue for _existing_ users.
356 2013-10-20 02:51:34 <sipa> CodeShark: just before you misunderstand, i'm not at all disagreeing with you about the possibilities of bitcoin, or what wallet priorities should be, or what abstractions to create
357 2013-10-20 02:51:57 <sipa> i'm only commenting on "accounts are silly, they should be X"
358 2013-10-20 02:51:58 <super3> gmaxwell: sure it is.
359 2013-10-20 02:52:19 <gmaxwell> super3: no it's not. It's a one time cost at install.
360 2013-10-20 02:52:39 <gmaxwell> (not to say that it doesn't need to be improved, and sipa has made tremendous progress there)
361 2013-10-20 02:52:45 <sipa> i think accounts are a halfway implemented feature, which is rarely what people wamt
362 2013-10-20 02:52:58 <sipa> but they offer exactly the abstraction they were intended to
363 2013-10-20 02:53:01 <CodeShark> I guess it's really pointless to argue over the bitcoind wallet account system - there are so many other aspects of it that are far worse :p
364 2013-10-20 02:53:51 <sipa> well, working on improving thimgs :)
365 2013-10-20 02:54:06 <CodeShark> indeed we are :)
366 2013-10-20 02:54:34 <super3> gmaxwell: for me its fine, but i don't recommend that new users run a full node because of it
367 2013-10-20 02:54:49 <super3> i agree that multiwallets should support it though
368 2013-10-20 02:55:00 <super3> would solve the problem with the least about of effort
369 2013-10-20 02:55:11 <gmaxwell> super3: not sure how you're disagreeing with my 'isn't an issue for _existing_ users'.
370 2013-10-20 02:55:12 <sipa> poke CodeShark!
371 2013-10-20 02:55:35 <CodeShark> in the future I foresee two types of users: typical, causal users who will most likely use SPV and corporate users who will likely run multiple verification/relay nodes as part of their IT infrastructure
372 2013-10-20 02:55:43 <MC1984> i just read how sipas headersfirst + libsecp got a sync down to 45 mins
373 2013-10-20 02:55:47 <MC1984> thats should help
374 2013-10-20 02:56:07 <super3> i have multiple computers, i wanted to do some dev work but had to stop because i had to wait for the blockchain to sync
375 2013-10-20 02:56:15 <sipa> MC1984: on a very good network connection, on a 12-core machine
376 2013-10-20 02:56:19 <gmaxwell> MC1984: well take care, the comparison there isn't two days.
377 2013-10-20 02:56:28 <CodeShark> I always keep at least two servers synched so I can test software from anywhere at any given time
378 2013-10-20 02:56:42 <gmaxwell> super3: Why?
379 2013-10-20 02:56:58 <CodeShark> but I'm certainly not a "typical, casual user"
380 2013-10-20 02:57:44 <super3> can't really send/in out coins if your wallet is still syncing with 2012 blocks
381 2013-10-20 02:58:03 <sipa> there are a few speedup tricks
382 2013-10-20 02:58:10 <sipa> like settimg high -dbcache
383 2013-10-20 02:58:17 <gmaxwell> super3: you can if your wallet has seen the coins. But why was your wallet syncing with 2012 blocks?
384 2013-10-20 02:58:21 <sipa> or -connect'img to a songle fast node
385 2013-10-20 02:58:32 <sipa> single
386 2013-10-20 02:58:58 <gmaxwell> Songle nodes are more harmonious.
387 2013-10-20 02:59:04 <phantomcircuit> lol
388 2013-10-20 03:00:11 <sipa> is 'songle' a word?
389 2013-10-20 03:00:16 <gmaxwell> super3: that shouldn't have been the case unless you are talking about a new node, or at least one that had not been run since 2012.
390 2013-10-20 03:00:29 <gmaxwell> sipa: sounds like something to do with singing.
391 2013-10-20 03:00:36 <sipa> right
392 2013-10-20 03:00:51 <super3> well yes i agree, but those are just workarounds the problem still exists
393 2013-10-20 03:00:53 <sipa> but then, single too :p
394 2013-10-20 03:01:08 <super3> people arn't running full nodes because they don't want to bother with a full sync
395 2013-10-20 03:01:09 <sipa> super3: you haven't answered
396 2013-10-20 03:01:35 <phantomcircuit> sipa, you have headers first working right?
397 2013-10-20 03:01:44 <super3> users including myself are lazy
398 2013-10-20 03:01:47 <sipa> phantomcircuit: somewhat, but i'm redoing it
399 2013-10-20 03:02:04 <gmaxwell> super3: okay, you're just talking in circles now. I give up.
400 2013-10-20 03:02:15 <sipa> super3: you still haven't answered why you were syncing with 2012 blocks
401 2013-10-20 03:02:43 <sipa> was it a new node?
402 2013-10-20 03:02:46 <super3> yeah
403 2013-10-20 03:03:05 <sipa> you can copy/sync from another
404 2013-10-20 03:03:13 <sipa> but yeah, it's not exact
405 2013-10-20 03:03:21 <super3> seems to go through the early blocks pretty quickly
406 2013-10-20 03:03:22 <sipa> ly user friendly
407 2013-10-20 03:03:37 <super3> maybe thats something that we can work on
408 2013-10-20 03:03:50 <gmaxwell> Again, this does nothing to counter the position I stated that the initial synctime is a one time burden which doesn't adversely impact most _existing_ users. E.g. it doesn't explain people who were using bitcoin-qt who switch off it.
409 2013-10-20 03:04:24 <gmaxwell> Which is fine, I don't care if you agree with that or not. I just wanted to make sure you hadn't run into a problem that I was unaware of.
410 2013-10-20 03:04:25 <super3> ah for that question there is no real incentive to keep it on
411 2013-10-20 03:04:52 <gmaxwell> super3: you don't have to keep it on, it catches up quite quickly unless you've left it without running it for months.
412 2013-10-20 03:05:00 <sipa> well i have heard people complaining about how slow it syncs even after just a few days/weeks
413 2013-10-20 03:05:21 <sipa> it may be fast, but it's not instant either
414 2013-10-20 03:05:54 <super3> usually its pretty bearable if its only a few days
415 2013-10-20 03:06:03 <gmaxwell> sipa: So far the complaints I'd seen about that were actually about the time when its displaying the splash screen.
416 2013-10-20 03:06:14 <sipa> ha!
417 2013-10-20 03:06:22 <super3> plus as long as there is coins in your wallet, you can probably still broadcast
418 2013-10-20 03:06:35 <sipa> just the block chain loading, can be really slow :)
419 2013-10-20 03:06:57 <gmaxwell> sipa: I've seen some users where it was taking a minute. Some kind of broken storage system. :(
420 2013-10-20 03:07:19 <sipa> yeah
421 2013-10-20 03:07:20 <super3> if people aren't using a program they turn it off, same thing for bitcoin. which would explain why they switch off qt.
422 2013-10-20 03:07:24 <gmaxwell> (as in, I got them to log timestamps because I wouldn't believe them!)
423 2013-10-20 03:07:39 <sipa> super3: oh, we're not talking about turning it off
424 2013-10-20 03:07:54 <sipa> we're talking about stoppimg to use it, amd move to another program
425 2013-10-20 03:08:13 <super3> ah i thought with the context you ment turn the program off
426 2013-10-20 03:08:26 <CodeShark> ideally, the sync would be part of a service which can always run in the background and can provide multiple levels of security, starting with headers-only sync and working its way up to full validation
427 2013-10-20 03:08:43 <sipa> CodeShark: that'd be awesome
428 2013-10-20 03:08:44 <super3> CodeShark: that would be nice
429 2013-10-20 03:09:02 <gmaxwell> CodeShark: s'not a new idea of course. Got a patch? :P
430 2013-10-20 03:10:15 <phantomcircuit> i should probably write a pull request to remove that 100ms sleep
431 2013-10-20 03:10:17 <phantomcircuit> but effort
432 2013-10-20 03:10:36 <sipa> yes, you should!
433 2013-10-20 03:10:42 <gmaxwell> phantomcircuit: we'll get even more very angry complaints about people who are unhappy that their node uses all their bandwidth. :P
434 2013-10-20 03:10:47 <phantomcircuit> sipa, yes but effort
435 2013-10-20 03:11:14 <MC1984> lol people install bitcoin
436 2013-10-20 03:11:20 <MC1984> "grrr why is it doing stuff!"
437 2013-10-20 03:11:21 <phantomcircuit> gmaxwell, lol
438 2013-10-20 03:11:31 <super3> MC1984: users can be like that sometimes...
439 2013-10-20 03:11:35 <gmaxwell> it should probably be removed, but also replaced with a ratelimiter that makes your node more latent if its gone over the ratelimiter. but ...
440 2013-10-20 03:11:41 <phantomcircuit> gmaxwell, for the most part the upnp client doesn't work right
441 2013-10-20 03:11:43 <gmaxwell> MC1984: "something is wrong, it's using my cpu!"
442 2013-10-20 03:11:46 <phantomcircuit> so not a big deal for most people
443 2013-10-20 03:11:47 <phantomcircuit> :/
444 2013-10-20 03:11:57 <MC1984> ive literally heard that many times
445 2013-10-20 03:12:03 <sipa> "it's in an infinite loop!"
446 2013-10-20 03:12:06 <gmaxwell> phantomcircuit: hm? it works surprisingly well.
447 2013-10-20 03:12:23 <gmaxwell> phantomcircuit: certantly well enough that we already get complaints! :P
448 2013-10-20 03:12:24 <phantomcircuit> gmaxwell, it randomly doesn't work with some routers
449 2013-10-20 03:12:25 <MC1984> it could do a better job of signalling exactly wht its doing
450 2013-10-20 03:12:29 <gmaxwell> phantomcircuit: yea sure.
451 2013-10-20 03:12:42 <phantomcircuit> like it doesn't work with my stupid comcast router