1 2012-01-17 00:00:01 <diki> it causes bitcoin to crash
  2 2012-01-17 00:01:17 <diki> sl
  3 2012-01-17 00:01:19 <diki> l
  4 2012-01-17 00:01:22 <diki> Damn...
  5 2012-01-17 00:01:30 <diki> Pressed enter instead of letters
  6 2012-01-17 00:01:44 <diki> Anywho, is there a way to disable the log file?
  7 2012-01-17 00:02:29 <riush> diki -printtoconsole disables it i think
  8 2012-01-17 00:06:48 <gmaxwell> Hm. Would it be unreasonable to use an additional 200 MiB ram during initial syncup if it made the syncup fast?
  9 2012-01-17 00:08:04 <gmaxwell> It looks like it wouldn't be _too_ hard to replace CTxDB with a class that basically maintains an in memory hash table, for all the txn and block indexes, and then just flushes it all out at once when you tell it to.
 10 2012-01-17 00:09:00 <gmaxwell> I believe that change alone would make us no longer IO bound during syncup.
 11 2012-01-17 00:10:37 <gmaxwell> Hm. considering that our defaults allow 125 connections .. we're already allowing memory usage much higher than that, we could even derate maxconnections based on the size of the index buffer.
 12 2012-01-17 00:12:06 <BlueMatt> gmaxwell: if the os cache is doing its job, that will provide 0 read advantage and only speed up writes, which could be accomplished by batch-writing blocks
 13 2012-01-17 00:12:31 <BlueMatt> s/accomplished/partially accomplished/
 14 2012-01-17 00:12:56 <BlueMatt> and if the os cache is too overloaded to do its job, then stealing more memory would be even worse
 15 2012-01-17 00:13:02 <sipa> gmaxwell: can you verify that my fix_ports branch fixes things?
 16 2012-01-17 00:13:07 <gmaxwell> It's not caching I'm concerned about.
 17 2012-01-17 00:13:11 <sipa> (i can't test right now myself)
 18 2012-01-17 00:13:12 <gmaxwell> sipa: in a minute, will do.
 19 2012-01-17 00:14:21 <gmaxwell> BlueMatt: We're currently causing something like 23 gigbytes of writes, in millions of write operations, most synchronous and randomly ordered. And it's not the writing of the blocks themselves that matters, it's the index updates.
 20 2012-01-17 00:15:00 <gmaxwell> The index updates take multiple writes each because of the index stuff. And at least as the code is currently written, you can't just delay this, because it's used for the connect inputs for the newly recieved stuff.
 21 2012-01-17 00:15:49 <BlueMatt> (a part of batch block writing has to be in-memory CTxDB)
 22 2012-01-17 00:16:12 <BlueMatt> but Im saying do that and let CTxDB be updated as you go instead of all at once at the end
 23 2012-01-17 00:16:42 <gmaxwell> Okay, I'm suggesting that in-memory CTxDB is basically enough.  But yes, flusing incrementally is probably fine and removes my memory question.
 24 2012-01-17 00:17:11 <sipa> flushing incrementally requires a completely different way of handling connectinputs
 25 2012-01-17 00:17:42 <BlueMatt> sort of, it just has to do checking against a to-be-flushed buffer as well as on-disk
 26 2012-01-17 00:17:55 <sipa> indeed
 27 2012-01-17 00:18:03 <gmaxwell> damnit is there a way to get a git format patch of a commit on github?
 28 2012-01-17 00:18:43 <gmaxwell> I don't think the reads are the problem, as BlueMatt says OS cache to the rescue.
 29 2012-01-17 00:19:02 <sipa> gmaxwell: https://github.com/sipa/bitcoin/commit/c981d768b3cb0cf0879b3d70d3b548692fff0882.patch
 30 2012-01-17 00:23:53 <gmaxwell> sipa: connect works now.
 31 2012-01-17 00:25:49 <CIA-76> bitcoin: Pieter Wuille master * rc981d76 / (5 files): Fix handling of default ports - http://git.io/rmfL9Q https://github.com/bitcoin/bitcoin/commit/c981d768b3cb0cf0879b3d70d3b548692fff0882
 32 2012-01-17 00:29:14 <gruez> gmaxwell: i got datadir to work
 33 2012-01-17 00:29:16 <gruez> it wasn't broken
 34 2012-01-17 00:29:44 <gruez> instead of
 35 2012-01-17 00:29:54 <gruez> bitcoin-qt -datadir PATH
 36 2012-01-17 00:29:57 <gruez> it should be
 37 2012-01-17 00:30:02 <gruez> bitcoin-qt -datadir=PATH
 38 2012-01-17 00:30:04 <gruez> :p
 39 2012-01-17 00:39:35 <BlueMattBot> Project Bitcoin build #177: FAILURE in 12 min: http://jenkins.bluematt.me/job/Bitcoin/177/
 40 2012-01-17 00:39:36 <BlueMattBot> sipa: Fix handling of default ports
 41 2012-01-17 00:46:52 <gmaxwell> gruez: hmph, I thought both were supposted to work. Thought my logs show I did give you a -datadir= version at some point. :)  Glad you got it working.
 42 2012-01-17 00:46:58 <gmaxwell> gruez: is bitcoin fast in the ramdisk?
 43 2012-01-17 00:47:49 <gruez> gmaxwell: testing now
 44 2012-01-17 01:42:15 <gruez> gmaxwell: can't tell
 45 2012-01-17 01:42:26 <gruez> got 81% in less than 1 hour though
 46 2012-01-17 01:43:47 <gruez> for some reason, even with ramdisk, the cpu gets used for a while
 47 2012-01-17 01:43:51 <gruez> then it idles
 48 2012-01-17 01:43:56 <gruez> then it gets used
 49 2012-01-17 01:44:50 <gruez> http://i.imgur.com/LjLbY.png
 50 2012-01-17 01:56:41 <etotheipi_> is there a way to disable the wallet encryption in Satoshi client 0.5.0?
 51 2012-01-17 01:56:48 <etotheipi_> or any version
 52 2012-01-17 01:58:44 <gruez> etotheipi_: don't know
 53 2012-01-17 01:58:50 <gruez> never set a password
 54 2012-01-17 02:01:22 <luke-jr> etotheipi_: no :/
 55 2012-01-17 02:01:55 <gmaxwell> etotheipi_: what do you mean by disable?
 56 2012-01-17 02:02:05 <etotheipi_> I meant, remove the passphrase to be unencrypted
 57 2012-01-17 02:02:14 <gmaxwell> gruez: what version are you running?
 58 2012-01-17 02:02:38 <gruez> gmaxwell: the dropbox version
 59 2012-01-17 02:02:48 <gmaxwell> Okay 0.5.2rc.
 60 2012-01-17 02:02:52 <gruez> http://dl.dropbox.com/u/29653426/bitcoin-0.5.2rc1-win32.tar.bz2
 61 2012-01-17 02:03:21 <etotheipi_> gah, how on earth am I supposed to get my private keys out of my wallet??
 62 2012-01-17 02:04:05 <gmaxwell> etotheipi_: (1) why do you want to do that, and (2) dumpprivkey
 63 2012-01-17 02:04:20 <etotheipi_> I am going to import all of them into Armory
 64 2012-01-17 02:04:42 <gmaxwell> grab the branch from sipa that has dumpwallet.
 65 2012-01-17 02:05:07 <etotheipi_> ahh... excellent
 66 2012-01-17 02:05:12 <lianj> "Armory - The most advanced Bitcoin Client in existence! " sounds ... hm .. optimistic
 67 2012-01-17 02:05:44 <etotheipi_> lianj, did you see the features?
 68 2012-01-17 02:05:53 <lianj> looks nice though
 69 2012-01-17 02:05:56 <lianj> etotheipi_: yea
 70 2012-01-17 02:06:25 <etotheipi_> I still desperately need some more people to help me test it... but I also need to work on the build system
 71 2012-01-17 02:06:35 <etotheipi_> Makefiles and such, are not my forte
 72 2012-01-17 02:07:42 <lianj> etotheipi_: ah, you built it. didnt notice until now
 73 2012-01-17 02:08:05 <etotheipi_> with the exception of zero-conf transactions... I haven't found any way to break it yet :)
 74 2012-01-17 02:08:30 <etotheipi_> all the features seem to work exactly as I'd hoped... but the lack of zero-conf is really killing the experience (I'm workign on it now)
 75 2012-01-17 02:08:49 <lianj> didnt look like many files, but i forgot that pythonists like to have huge files :P
 76 2012-01-17 02:09:04 <etotheipi_> C++ is done "properly"
 77 2012-01-17 02:09:16 <etotheipi_> but yes, I like single, large python files
 78 2012-01-17 02:09:53 <gruez> what's with all the coding conventions for bitcoin?
 79 2012-01-17 02:10:11 <etotheipi_> gruez, what do you mean by that?
 80 2012-01-17 02:10:17 <gruez> like prefixing variable names with their type, sorta like hungerian notation
 81 2012-01-17 02:10:33 <gruez> which is supposedly bad according to this https://blogs.msdn.com/b/nicholg/archive/2006/04/13/576274.aspx
 82 2012-01-17 02:11:12 <josephcp> pretty sure satoshi did it
 83 2012-01-17 02:12:09 <lianj> etotheipi_: single-large files and no tests - hello python :D .. but from the feature list it looks great, congrats! keep up with the implemenation changes
 84 2012-01-17 02:12:20 <etotheipi_> what do yo umean , no tests?
 85 2012-01-17 02:12:24 <etotheipi_> I have 2500 lines of unit tests
 86 2012-01-17 02:12:55 <lianj> oh yes, sorry missed to find it because its a single file :)
 87 2012-01-17 02:13:14 <etotheipi_> there 2/3 of it are python tests:  unittest.py
 88 2012-01-17 02:13:23 <etotheipi_> 1/3 are C++ in cppForSwig/BlockUtilsTest.cpp
 89 2012-01-17 02:13:58 <lianj> nice
 90 2012-01-17 02:14:05 <etotheipi_> I don't know how it would be possible to build something like this from the ground up, without having exhaustive tests
 91 2012-01-17 02:14:20 <gruez> lol
 92 2012-01-17 02:14:23 <gruez> bitcoin has test units?
 93 2012-01-17 02:14:23 <lianj> likewise
 94 2012-01-17 02:14:41 <theymos> Bitcoin had no (public) tests...
 95 2012-01-17 02:14:42 <etotheipi_> throughout development, things have constantly been breaking, and constantly needing to be retested and debugged
 96 2012-01-17 02:14:46 <lianj> but most python projects dont test
 97 2012-01-17 02:15:07 <gruez> what's with all the c based naming conventions in bitcoin?
 98 2012-01-17 02:15:08 <etotheipi_> lianj, I'm not sure if that's a thing to do with "python" or the specific programmers
 99 2012-01-17 02:15:19 <gruez> every class is prefixed with C for some reason
100 2012-01-17 02:15:21 <gruez> CWallet
101 2012-01-17 02:15:35 <gruez> is there any other type of wallet?
102 2012-01-17 02:15:56 <lianj> etotheipi_: true, i will stop ranting about python's test culture now
103 2012-01-17 02:16:19 <BlueMatt> gruez: it stands for Class
104 2012-01-17 02:16:41 <gruez> BlueMatt: but still
105 2012-01-17 02:16:46 <gruez> it's sorta redundent
106 2012-01-17 02:16:57 <gruez> same with variable names
107 2012-01-17 02:17:05 <gruez> never do that with my code
108 2012-01-17 02:17:20 <gruez> hover my mouse over the variable, intelisense tells me what variable it is
109 2012-01-17 02:17:32 <gmaxwell> gruez: yea, but you make blue sheds so you're fking evil Q.E.D.
110 2012-01-17 02:17:35 <lianj> etotheipi_: funny though, is there no test lib in stdlib? you sorta did your own bool assert testing
111 2012-01-17 02:18:08 <gruez> :-\n2654924
112 2012-01-17 02:19:18 <etotheipi_> lianj, I have never really used any "official" unit-testing libraries... I've always just done my own thing... but I agree it probably would've been a good idea for such a large project to do it in a more-standardized fashion
113 2012-01-17 02:19:25 <etotheipi_> probably would've been easier on me, too
114 2012-01-17 02:21:29 <etotheipi_> does the Satoshi client ever re-broadcast a transaction if it isn't mined?
115 2012-01-17 02:22:00 <nanotube> yes, at a semi-random interval of about 30min
116 2012-01-17 02:22:25 <etotheipi_> does it ever stop?
117 2012-01-17 02:23:42 <theymos> No.
118 2012-01-17 02:25:04 <gruez> seriously?
119 2012-01-17 02:25:13 <gruez> what's the function that reboardcasts?
120 2012-01-17 02:25:47 <theymos> ResendWalletTransactions
121 2012-01-17 02:27:00 <gruez> lol
122 2012-01-17 02:27:04 <gruez> that's convinient
123 2012-01-17 02:27:50 <theymos> IIRC the function contains the "every ~30 minutes" logic, so you'd have to change that.
124 2012-01-17 02:30:28 <gruez> it does?
125 2012-01-17 02:30:30 <gruez> BOOST_FOREACH(CWallet* pwallet, setpwalletRegistered)         pwallet->ResendWalletTransactions();
126 2012-01-17 02:31:30 <theymos> It definitely does in 0.3.x. I don't know what current versions do.
127 2012-01-17 02:31:43 <gruez> nevermind
128 2012-01-17 02:31:54 <gruez> the logic is in wallet::ResendWalletTransactions
129 2012-01-17 02:32:33 <etotheipi_> so here's a philosophical question:  if transactions are "unconfirmed" until 6 confirmations, why does the Satoshi client allow you spend them and pass on the risk to others?  It seems internally inconsistent
130 2012-01-17 02:32:55 <gruez> you can spend unconfirmed coins even
131 2012-01-17 02:32:56 <gruez> :p
132 2012-01-17 02:33:02 <gruez> check bitcoincharts
133 2012-01-17 02:33:14 <etotheipi_> I'm trying to justify how you can claim the coins "shouldn't be trusted" until 6 conf, but if you can spend them, it seems they are most definitely "yours"
134 2012-01-17 02:33:20 <etotheipi_> I know they can be spent gruez
135 2012-01-17 02:33:37 <etotheipi_> and the Satoshi client only lets you spend zero-conf tx that were sent to yourself
136 2012-01-17 02:33:41 <luke-jr> gruez: rather than a commandline switch, make a JSON-RPC call
137 2012-01-17 02:33:54 <luke-jr> gruez: also note you're not the first to do this. the last attempt was broken though
138 2012-01-17 02:34:00 <gruez> aww
139 2012-01-17 02:34:04 <theymos> The sent transaction will have confirmations >= than the input confirmations, so the value of the recipients confirmations isn't reduced by receiving unconfirmed coins.
140 2012-01-17 02:34:19 <luke-jr> see pull 421
141 2012-01-17 02:34:20 <gruez> i'll see if i can get bitcoin working on windows
142 2012-01-17 02:34:31 <luke-jr> https://github.com/bitcoin/bitcoin/pull/421
143 2012-01-17 02:34:36 <luke-jr> nobody cares about Windows <.<
144 2012-01-17 02:34:49 <theymos> (<=, I mean)
145 2012-01-17 02:34:51 <gruez> but visual studio is my favorite IDE
146 2012-01-17 02:34:59 <Diablo-D3> gruez: diaf.
147 2012-01-17 02:35:13 <etotheipi_> okay, that makes more sense
148 2012-01-17 02:35:19 <luke-jr> gruez: learn to love Qt Creator? :P
149 2012-01-17 02:35:26 <CIA-76> bitcoin: Luke Dashjr 0.5.x * ra5b875f47b3a bitcoind-stable/src/ (8 files in 2 dirs): Merge branch '0.5.0.x' into 0.5.x http://tinyurl.com/7r35pvz
150 2012-01-17 02:35:27 <CIA-76> bitcoin: Luke Dashjr 0.5.x * r382e613ef59f bitcoind-stable/src/bitcoinrpc.cpp: Merge branch '0.5.0.x' into 0.5.x http://tinyurl.com/7h48gn9
151 2012-01-17 02:35:28 <Diablo-D3> use vim goddamnit.
152 2012-01-17 02:35:29 <gruez> qt creator sucks
153 2012-01-17 02:35:37 <luke-jr> gruez: so does Visual Studio though
154 2012-01-17 02:35:40 <gruez> it's missing all my intellisense
155 2012-01-17 02:35:49 <gruez> plus it doesn't have visual assist support
156 2012-01-17 02:36:05 <etotheipi_> agreed, Diablo
157 2012-01-17 02:36:19 <luke-jr> use ed
158 2012-01-17 02:36:26 <gruez> with visual studio, if i dont know a symbol, i can just right click->go to definition
159 2012-01-17 02:36:37 <gruez> plus autocompletion. gotta love that
160 2012-01-17 02:36:43 <etotheipi_> I still haven't opened up a Qt Designer of any sort in Armory... it takes a tad longer, but dynamic content is so much easier
161 2012-01-17 02:36:50 <Diablo-D3> I even have my .vim in a github repo
162 2012-01-17 02:37:00 <roconnor> etotheipi_: amoury is a full client?
163 2012-01-17 02:37:10 <etotheipi_> roconnor, it's pretty damned close
164 2012-01-17 02:37:22 <roconnor> that's impressive; I haven't even got my core client working yet.
165 2012-01-17 02:37:46 <gruez> http://i.imgur.com/SjpIR.png
166 2012-01-17 02:37:58 <gruez> yay, at least the autocompletion is working
167 2012-01-17 02:41:59 <gruez> 
168 2012-01-17 02:42:09 <gruez> cxx = ?
169 2012-01-17 02:42:24 <gruez> i know it's berkely db
170 2012-01-17 02:44:36 <luke-jr> bdb 4.8 specifically
171 2012-01-17 02:44:45 <luke-jr> some doc/* file has dependency info
172 2012-01-17 02:45:08 <luke-jr> gruez: you're aware nobody has built Bitcoin on Windows in a long time, right?
173 2012-01-17 02:45:28 <gruez> is there anything compiler dependent?
174 2012-01-17 02:45:55 <gruez> IntelliSense: cannot open source file "inttypes.h"
175 2012-01-17 02:46:03 <gruez> IntelliSense: cannot open source file "unistd.h"
176 2012-01-17 02:46:09 <gruez> what's with all these weird dependencies
177 2012-01-17 02:47:17 <theymos> GCC is recommended.
178 2012-01-17 02:48:28 <gmaxwell> roconnor: etotheipi_ has the benefit of working in the same language, and can thus follow the bitcoin code very closely.
179 2012-01-17 02:49:12 <Diablo-D3> gmaxwell: hey you, I want your opinion
180 2012-01-17 02:49:13 <etotheipi_> gmaxwell, I didn't follow the code that closely
181 2012-01-17 02:49:20 <etotheipi_> gmaxwell, in fact I barely looked at it
182 2012-01-17 02:49:24 <gmaxwell> etotheipi_: there is some exact copying, at least of trivial stuff.
183 2012-01-17 02:49:36 <gmaxwell> etotheipi_: I hope you're not planning on being dishonest about this.
184 2012-01-17 02:49:40 <Diablo-D3> gmaxwell: think by next monday the price of btc will finally spike?
185 2012-01-17 02:49:41 <etotheipi_> gmaxwell, I copied the page-locked memory code
186 2012-01-17 02:49:48 <luke-jr> gruez: none of the dependencies are very weird.
187 2012-01-17 02:49:50 <etotheipi_> I even got on here to ask about that code
188 2012-01-17 02:50:14 <etotheipi_> to ask if someone knew where i could find something about page-locked memory,  I was told that that code was very generic and not to worry about it
189 2012-01-17 02:50:31 <etotheipi_> maybe I shouldn't have trusted them... but I believe that's the only thing I copied
190 2012-01-17 02:50:51 <gmaxwell> etotheipi_: e.g. the enums for all the script opcodes for example,  I'm not knocking your work of course.
191 2012-01-17 02:51:03 <etotheipi_> let me know if there's something else:  but I wrote the entire thing from scratch
192 2012-01-17 02:51:48 <etotheipi_> the OP_CODE enum was probably copied, as well, but that's... silly
193 2012-01-17 02:52:37 <gmaxwell> The code for getOpCodeName is copied.
194 2012-01-17 02:52:49 <gmaxwell> Bunch of trivial stuff around script it appears.
195 2012-01-17 02:52:52 <etotheipi_> gmaxwell, the point was that I wrote the all the classes and infrastructure from scratch, and following the code wouldn't have help much because I wanted to do differently
196 2012-01-17 02:54:08 <etotheipi_> gmaxwell, I'm confused what you're tryign to prove
197 2012-01-17 02:54:28 <etotheipi_> it's virtually impossible to implement the list of opcodes without making it look like it was copied...
198 2012-01-17 02:54:35 <etotheipi_> it's a list
199 2012-01-17 02:54:43 <theymos> Does Armory verify scripts? I'd be pretty nervous using a client that did that -- seems pretty hard to get absolutely correct.
200 2012-01-17 02:54:59 <etotheipi_> Armory does not do full validation
201 2012-01-17 02:55:03 <etotheipi_> but it could
202 2012-01-17 02:55:19 <etotheipi_> and it does full script eval except for OP_IF/ELSE
203 2012-01-17 02:56:22 <gmaxwell> etotheipi_: I'm not trying to prove anything.
204 2012-01-17 02:56:43 <gmaxwell> etotheipi_: I made a simple statement: 19:48 < gmaxwell> roconnor: etotheipi_ has the benefit of working in the same language, and can thus follow the bitcoin code very closely.
205 2012-01-17 02:56:49 <etotheipi_> once again, I have a ton of unit tests on that... I grabbed some crazy complicated scripts from the testnet and have debugged them through the the script eval
206 2012-01-17 02:56:49 <gmaxwell> and you've had a very defensive response.
207 2012-01-17 02:56:56 <roconnor> etotheipi_: oh, I thought your code was in python or something.
208 2012-01-17 02:56:59 <gmaxwell> etotheipi_: where is your script eval code?
209 2012-01-17 02:57:01 <roconnor> I guess that was someone else then
210 2012-01-17 02:57:27 <etotheipi_> roconnor, my underlying engine is in C++
211 2012-01-17 02:57:27 <gmaxwell> (I'm grepping for magic numbers it should have and not finding it)
212 2012-01-17 02:57:43 <etotheipi_> all blockchain scanning is in C++
213 2012-01-17 02:57:53 <etotheipi_> everythign else is in Python
214 2012-01-17 02:58:09 <roconnor> ah
215 2012-01-17 02:58:14 <etotheipi_> and the reason I did that, was because my goal was to make the absolute fastest blockchain scanning code, ever
216 2012-01-17 02:58:14 <roconnor> so I'm not that confused.
217 2012-01-17 02:58:43 <gmaxwell> etotheipi_: getOpCodeName isn't just a list, it's a big select loop that returns text strings. You reindented it, but it's copied exactly. I think there is absolutely nothing wrong with that.
218 2012-01-17 03:00:19 <gmaxwell> etotheipi_: so, yea, where are is the evaluator? I'm not finding any of its magic values like 10000, 520, or 201.
219 2012-01-17 03:00:35 <etotheipi_> gmaxwell, you may be correct about that... it was a long time ago... my point was you seemed to have narrowed in on a very small piece of 8000 lines of code
220 2012-01-17 03:01:04 <etotheipi_> in reality, the entire C++ implementation is from scratch... except for op-code lists
221 2012-01-17 03:01:08 <etotheipi_> and page-locked memory
222 2012-01-17 03:01:29 <gmaxwell> etotheipi_: It was literally the first place I looked. (admittedly due to intution about what would most likely be identical)
223 2012-01-17 03:01:33 <etotheipi_> I do not have those magic numbers
224 2012-01-17 03:01:52 <gmaxwell> Then you're not implementing a full node then, enh?
225 2012-01-17 03:01:55 <etotheipi_> I don't evaluate with all those checks
226 2012-01-17 03:02:03 <etotheipi_> correct, I was never planning to make a full node
227 2012-01-17 03:02:10 <etotheipi_> but I wanted to be able to verify scripts
228 2012-01-17 03:02:13 <roconnor> I'm still very impressed
229 2012-01-17 03:02:21 <roconnor> I know how hard it is
230 2012-01-17 03:02:29 <roconnor> it's almost impossible :)
231 2012-01-17 03:02:31 <etotheipi_> there's been a lot of girlfriend neglect :|
232 2012-01-17 03:02:37 <etotheipi_> thanks roconnor
233 2012-01-17 03:02:57 <etotheipi_> there were times I wanted to give up... most notably around the time I started the blockchain reorganization unit-tests
234 2012-01-17 03:03:06 <gmaxwell> etotheipi_: You totally should make a full node. We need more of them. Your girlfriend can wait. :)
235 2012-01-17 03:03:08 <etotheipi_> hours and hours of stepping through a debugger
236 2012-01-17 03:03:36 <roconnor> etotheipi_: I don't suppose you have a unit testing node?
237 2012-01-17 03:03:39 <etotheipi_> and many more hours creating a full mini-blockchain
238 2012-01-17 03:03:42 <etotheipi_> with a double-spend
239 2012-01-17 03:04:25 <etotheipi_> roconnor, what do you mean?
240 2012-01-17 03:04:31 <gmaxwell> etotheipi_: I offer to schedule regular flower delivery to placate your girlfriend while you work on adding the rest of the protocol rules. :)
241 2012-01-17 03:04:40 <roconnor> a node that I can connect to who will pass me bad blockchains
242 2012-01-17 03:04:55 <roconnor> in sneaky ways
243 2012-01-17 03:05:20 <etotheipi_> gmaxwell, haha... it wouldn't be bad if she got felt some warmth from the Bitcoin community, that to her feels like a blackhole
244 2012-01-17 03:06:07 <etotheipi_> roconnor, I don't have anything like that.... the unit-tests are all really low-level, just passing in blocks stored in files, and making sure the data structures are updated appropriately
245 2012-01-17 03:06:59 <etotheipi_> gmaxwell, full-verification may be in my future... but my goal was the features.  And I know the future of bitcoin lies in lite-nodes, so I had no reservations
246 2012-01-17 03:07:31 <etotheipi_> really, I just wanted a simple interface for multiple wallets, importing private keys and managing offline transactions... it just so happened that a bunch of other people did too :)
247 2012-01-17 03:08:25 <roconnor> Ah
248 2012-01-17 03:08:31 <roconnor> I haven't done any negative testing :(
249 2012-01-17 03:09:14 <gmaxwell> roconnor: hey, you made it fail on testnet.
250 2012-01-17 03:10:02 <roconnor> okay, then I haven't done any positive testing
251 2012-01-17 03:10:04 <roconnor> one of the two
252 2012-01-17 03:10:25 <etotheipi_> and I hope someone will help me test offline transactions... I believe it is critical for businesses... and I think I made it as simple as it can possibly get
253 2012-01-17 03:10:33 <etotheipi_> that was more absolute #1 priority in all this
254 2012-01-17 03:10:43 <roconnor> what are offline transactions?
255 2012-01-17 03:10:44 <etotheipi_> http://bitcoinarmory.com/index.php/using-offline-wallets-in-armory
256 2012-01-17 03:11:01 <etotheipi_> full wallet on offline computer, watching-only wallet on online computer
257 2012-01-17 03:11:20 <etotheipi_> create transaction online, take it to offline computer for signing, bring back and broadcast
258 2012-01-17 03:11:28 <etotheipi_> AND offline computer does not need blockchain1
259 2012-01-17 03:11:29 <etotheipi_> !
260 2012-01-17 03:11:32 <roconnor> oh
261 2012-01-17 03:12:04 <roconnor> what about recieving third party transactions?
262 2012-01-17 03:12:20 <roconnor> (or rather second party transactions)
263 2012-01-17 03:12:36 <etotheipi_> this is why I created BIP 0010:  https://en.bitcoin.it/wiki/BIP_0010
264 2012-01-17 03:12:57 <etotheipi_> it's not only perfect for multi-sig transacitons... but it's perfect for offline transactions (it's just a 1-of-1 transaction that you don't have the private keys for)
265 2012-01-17 03:13:42 <roconnor> I can take anyone's transaction and put it into Armory?
266 2012-01-17 03:14:12 <etotheipi_> I'm not sure what that means... any transaction can be saved to BIP0010 format
267 2012-01-17 03:14:27 <etotheipi_> if it's unsigned, you can give it to the person who has the private keys and they can add their signature to it (without needing the blockchain)
268 2012-01-17 03:14:35 <gmaxwell> etotheipi_: "the subsequent signature will be valid no matter what inputs values were provided"!!! good spotting but I don't think thats minor.
269 2012-01-17 03:14:38 <etotheipi_> they can then send it back or broadcast it if it has enough sigs
270 2012-01-17 03:14:51 <roconnor> well I was talking before with someone; I said that if someone sends me bitcoins, I'd rather them give me the transaction myself, since it is in my interest for it to be mined properly.
271 2012-01-17 03:15:09 <etotheipi_> roconnor, that is possible with Armory
272 2012-01-17 03:15:14 <roconnor> good
273 2012-01-17 03:15:15 <roconnor> A+
274 2012-01-17 03:15:25 <gmaxwell> etotheipi_: when I look at a offline wallet my thought is that the offline system is the _only_ trusted system, it's trusted because it's offline and can't get hacked.
275 2012-01-17 03:15:27 <etotheipi_> they just click the "Create Unsigned Transaction" button instead of the "Send" button
276 2012-01-17 03:16:02 <etotheipi_> gmaxwell, that's why I made it... not only do I want it, I believe that businesses and anyone with serious interest in Bitcoin needs the 99% security that offline transactions offers
277 2012-01-17 03:16:18 <etotheipi_> there's just no better way to keep your money
278 2012-01-17 03:16:27 <etotheipi_> but there's just no good way to do it ... until now :)
279 2012-01-17 03:17:12 <gruez_> Just wondering, what sort of IDE do the core devs work with?
280 2012-01-17 03:17:15 <etotheipi_> gmaxwell, with regards to the input values I believe it's minor because it's not a dealbreaker:  most people didn't expect to be able to do these kinds of transactions without the blockchain
281 2012-01-17 03:17:34 <etotheipi_> and most of these transactions will be filtered through a node with the blockchain, before it ever gets to the signing node
282 2012-01-17 03:17:38 <gruez_> Just plain text editor with syntax hilighting?
283 2012-01-17 03:17:45 <doublec> etotheipi_: doesn't that kinda mean the "BIP 16 and variants are needed for secure wallets" is not true?
284 2012-01-17 03:17:58 <doublec> etotheipi_: since you seem to be able to do the same thing with your offline transactions
285 2012-01-17 03:18:06 <roconnor> doublec: probably
286 2012-01-17 03:18:09 <etotheipi_> doublec, it's a different thing
287 2012-01-17 03:18:15 <etotheipi_> but there is certainly overlap
288 2012-01-17 03:18:24 <roconnor> doublec: etotheipi_'s solution is far superiour
289 2012-01-17 03:18:33 <etotheipi_> but a little less convenient
290 2012-01-17 03:18:34 <roconnor> though I guess it does require an offline computer
291 2012-01-17 03:18:36 <doublec> yeah I see it's a different approach to the same problem, but I like the approach
292 2012-01-17 03:18:47 <etotheipi_> (well, a lot less convenient)
293 2012-01-17 03:18:55 <roconnor> BIP 0016 also solves some other minorer problems
294 2012-01-17 03:18:59 <doublec> right
295 2012-01-17 03:19:08 <etotheipi_> but I have plans for supporting two-factor authentication without a third-party :)
296 2012-01-17 03:19:18 <etotheipi_> if I ever get an android app developed
297 2012-01-17 03:19:24 <k9quaint> gmaxwell: offline systems can be hacked
298 2012-01-17 03:19:25 <josephcp> I wish OP_CAT could be enabled and we can solve this problem EZPZ :-/
299 2012-01-17 03:19:29 <gmaxwell> etotheipi_: but the node with the blockchain is compromised.
300 2012-01-17 03:19:54 <gmaxwell> k9quaint: yes, but it's a much taller barrier, especially when the only channel is simple ascii.
301 2012-01-17 03:20:21 <etotheipi_> gmaxwell, but if your host system is hacked, then you have a problem regardless of the offline system
302 2012-01-17 03:20:24 <gmaxwell> etotheipi_: you could pretty much solve that by including the whole inputs and then the offline node could choose to check them and hash them itself.
303 2012-01-17 03:20:37 <gmaxwell> etotheipi_: the attaker could only DOS you. Not spend or burn your money though.
304 2012-01-17 03:20:46 <k9quaint> gmaxwell: no, the barrier is unchanged, the vector of attack changes
305 2012-01-17 03:21:00 <roconnor> I guess the first person to make a cheap $5 handheld offline wallet device will be rich.
306 2012-01-17 03:21:02 <k9quaint> picture an old lady with a USB stick with her bitcoin wallet on it
307 2012-01-17 03:21:07 <roconnor> s/$5/1BTC
308 2012-01-17 03:21:13 <gmaxwell> k9quaint: it's a harder attack vector. Your communication channel is a limited ascii channel with very carefully audited validation code.
309 2012-01-17 03:21:17 <etotheipi_> the biggest threat to my system is USB key viruses... they'd have to be targeted, but it could happen
310 2012-01-17 03:21:24 <etotheipi_> though I could solve this with QR codes
311 2012-01-17 03:21:36 <gmaxwell> I was about to say "2d barcodes"
312 2012-01-17 03:21:39 <etotheipi_> both online and offline computer only communicate via webcams and QR codes :)
313 2012-01-17 03:21:40 <k9quaint> gmaxwell: unless I trick the old lady into giving up the cherry ;)
314 2012-01-17 03:22:06 <roconnor> etotheipi_: sounds idea
315 2012-01-17 03:22:09 <roconnor> *ideal
316 2012-01-17 03:22:13 <cjdelisle> k9quaint: congradulations, you have used "old lady" and "cherry" in a sentence.
317 2012-01-17 03:22:18 <etotheipi_> lol
318 2012-01-17 03:22:26 <luke-jr> etotheipi_: you could have a monitor and camera pair for exchanging data between online and offline <.<
319 2012-01-17 03:22:35 <k9quaint> cjdelisle: do I win a prize?
320 2012-01-17 03:23:03 <cjdelisle> yeap
321 2012-01-17 03:23:04 <etotheipi_> luke-jr, exactly ... and I've already got QR code generation working, and QR reading isn't terribly hard
322 2012-01-17 03:23:05 <josephcp> yeah QR+Camera is the safest, but it's like a 3-step process
323 2012-01-17 03:23:07 <cjdelisle> a wet kiss
324 2012-01-17 03:23:12 <k9quaint> old people are vulnerable to social engineering, and in fact a lot of attacks on their identity shun automated systems and go for the paper and pencil variety
325 2012-01-17 03:23:17 <luke-jr> etotheipi_: Bitcoin-Qt 0.6 has a QR code generator builtin
326 2012-01-17 03:23:26 <cjdelisle> redeamable anywhere old ladies are sold
327 2012-01-17 03:23:30 <etotheipi_> but I hate the thought of having to try to support webcam interfaces on arbitrary systems
328 2012-01-17 03:23:32 <luke-jr> btw, I was trolling about the camera/monitor thing
329 2012-01-17 03:23:46 <josephcp> you need to get a list of available output hashes, then sign on PDA, then read from pda, too hard
330 2012-01-17 03:23:46 <luke-jr> that's just as networked as a serial port
331 2012-01-17 03:23:48 <k9quaint> cjdelisle: I havent found any vendors that sell old women for BTC :(
332 2012-01-17 03:23:54 <etotheipi_> luke-jr, I don't actually plan on doing it, but the thoguht crossed my mind as feasible
333 2012-01-17 03:24:07 <cjdelisle> k9quaint: sounds like there's a gap in the market
334 2012-01-17 03:24:07 <luke-jr> etotheipi_: serial port is just as secure, and much easier
335 2012-01-17 03:24:13 <cjdelisle> which you can fill
336 2012-01-17 03:24:48 <k9quaint> cjdelisle: I was trying to fill a gap, thats for sure
337 2012-01-17 03:24:50 <etotheipi_> I'm not much of a hardware guy... but my job is image/video processing so the whole QR+camera thing was intriguing to me
338 2012-01-17 03:25:36 <k9quaint> QR codes give me the willies
339 2012-01-17 03:25:53 <etotheipi_> for now, the USB key technique I've implemented is miles ahead of the security of any online system
340 2012-01-17 03:25:57 <k9quaint> the idea that I am running random code on a consumer device....yuck :(
341 2012-01-17 03:26:37 <etotheipi_> I agree it's not 100%, but it's still a huge step up... and I'm very excited I finally got the pipeline set up and debugged
342 2012-01-17 03:28:26 <etotheipi_> a business owner can create all his employees wallets at home on an offline computer, create the watching-only copies and import them to the employees computers
343 2012-01-17 03:28:48 <etotheipi_> they can all work with those wallets EXACTLY as they would any other wallet, but they can't spend the money
344 2012-01-17 03:29:02 <doublec> I wonder if there will be a market for signed transactions
345 2012-01-17 03:29:03 <etotheipi_> and the business owner doesn't have to worry about anyone stealing any of it
346 2012-01-17 03:29:09 <doublec> trading outside the blockchain
347 2012-01-17 03:29:16 <k9quaint> etotheipi_: you are trusting the average every day person to oversee a chain of events that is intregral to security
348 2012-01-17 03:29:24 <doublec> especially ones originating from fresh generated coins
349 2012-01-17 03:30:06 <etotheipi_> k9quaint, I'm simply, dramatically, lowering the learning curve for someone who wants to use "offline transactions"
350 2012-01-17 03:30:19 <CIA-76> bitcoin: Anthony G. Basile maintree * r8f5c67ff7460 gentoo/net-p2p/ (9 files in 2 dirs): Bumps from release candidates http://luke.dashjr.org/programs/bitcoin/w/gentoo_ebuild.git/commitdiff/8f5c67ff74608348dbcd1b405e22079b3dad013f
351 2012-01-17 03:30:39 <k9quaint> and offline computers are also vulnerable (voting machines for instance), and they become less secure over time since they are not patched
352 2012-01-17 03:31:02 <k9quaint> I am just playing devils advocate
353 2012-01-17 03:31:13 <etotheipi_> human error will always be a problem, but at least you don't need to be an uber-linux nerd with crypto backgroudn and scripting/CLI capabilities to use my system
354 2012-01-17 03:31:28 <k9quaint> human error is always the problem with security ;)
355 2012-01-17 03:31:36 <etotheipi_> it's actually comprehendable and usable by someone without much understanding of the inner workings of Bitcoin at all
356 2012-01-17 03:31:40 <k9quaint> since the idea of security is an emotion inside a persons head
357 2012-01-17 03:35:38 <etotheipi_> well I got a lot more work to do to bring armory to the "average" user, but right now it's very usable (without zero-conf tx), so it'd be great if anyone could help me test Armory
358 2012-01-17 03:35:56 <etotheipi_> if you got the RAM and (Ubuntu || patience-for-the-windows-build-instructions)
359 2012-01-17 03:36:42 <etotheipi_> just print a paper wallet if you use it with real money... you'll always be able to recover all your keys with it, even if something went terribly wrong with the software
360 2012-01-17 03:37:00 <etotheipi_> (but so far it's been pretty robust)
361 2012-01-17 03:38:15 <k9quaint> can it survive a 7 year old with crayons? :)
362 2012-01-17 03:38:44 <k9quaint> that is a persistent attack vector that I am subject to for the next 331 days
363 2012-01-17 03:39:15 <k9quaint> (then it metamorphs into an 8 year old with sharpies)
364 2012-01-17 03:39:53 <etotheipi_> haha
365 2012-01-17 03:40:50 <etotheipi_> I keep my paper backups tucked away in a book on my bookshelf
366 2012-01-17 03:41:13 <etotheipi_> I have a feeling a 7-yo kid won't be opening my general relativity textbook
367 2012-01-17 03:42:11 <k9quaint> you are attributing rationality to an elemental force of chaos
368 2012-01-17 03:43:29 <etotheipi_> but seriously, I desperately need people to help me with testing... though I will have a more robust build system setup soon
369 2012-01-17 03:44:18 <etotheipi_> and I know you all want to import your vanitygen addresses, sweep keys, and view all the goodies available in Developer mode
370 2012-01-17 06:00:51 <CIA-76> bitcoin: Con Kolivas * r4cf433c91706 cgminer/util.c: Remove TCP_NODELAY from curl options as many small packets may be contributing to network overload. http://tinyurl.com/78lppak
371 2012-01-17 06:00:52 <CIA-76> bitcoin: Con Kolivas * r12afb479d375 cgminer/main.c: Invalidating work after longpoll made hash_pop return no work giving a false positive for dead pool. Rework hash_pop to retry while finds no staged work until the abstime timeout really expires. http://tinyurl.com/76wvnpo
372 2012-01-17 06:17:14 <is4tomj> Is there a way to get bitcoind to issue a callback after a transaction has been detected whether or not it is confirmed?
373 2012-01-17 06:18:54 <is4tomj> really what I'm looking for is a mechanism to perform an HTTP Post.
374 2012-01-17 06:40:28 <CIA-76> bitcoin: Con Kolivas * r5d6159f55ad4 cgminer/main.c: Don't consider a pool lagging if a request has only just been filed. This should decrease the false positives for "pool not providing work fast enough". http://tinyurl.com/7rzjc3u
375 2012-01-17 07:32:23 <gribble> New news from bitcoinrss: laanwj opened pull request 762 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/762>
376 2012-01-17 07:35:19 <CIA-76> bitcoin: Wladimir J. van der Laan master * rb2a967c / src/qt/bitcoingui.cpp : Revert to global progress indication (see #753) - http://git.io/vxGB_A https://github.com/bitcoin/bitcoin/commit/b2a967cd0b158f40787cc1acb4f1e245b5066cdf
377 2012-01-17 07:35:20 <CIA-76> bitcoin: Wladimir J. van der Laan master * r06706ab / src/qt/optionsmodel.cpp : Remove erroneous ":" in front of port in options dialog (introduced with network refactor) - http://git.io/C2VU5g https://github.com/bitcoin/bitcoin/commit/06706ab8ef3a5fc7448ddfc1f64f92505cd710ee
378 2012-01-17 07:35:21 <CIA-76> bitcoin: Wladimir J. van der Laan master * r43cda5f / (src/qt/bitcoingui.cpp src/qt/optionsmodel.cpp): Merge pull request #762 from laanwj/qtprogressfix ... https://github.com/bitcoin/bitcoin/commit/43cda5f325f53e8941efa7f712aed66e3fde172a
379 2012-01-17 07:37:36 <da2ce7> if bitcoin is unable to connect to a proxy as specified in the coinfiguration, maybe we should include a popup saying something like 'unable to connect to proxy, check configuration.'
380 2012-01-17 07:37:51 <da2ce7> or a banner at the bottom of the window.
381 2012-01-17 07:37:56 <wumpus> da2ce7: agreed, file an issue :)
382 2012-01-17 07:42:29 <gribble> New news from bitcoinrss: da2ce7 opened issue 763 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/issues/763>
383 2012-01-17 07:42:43 <da2ce7> there we are :)
384 2012-01-17 08:12:53 <BlueMattBot> Yippie, build fixed!
385 2012-01-17 08:12:54 <BlueMattBot> * laanwj: Revert to global progress indication (see #753)
386 2012-01-17 08:12:55 <BlueMattBot> * laanwj: Remove erroneous ":" in front of port in options dialog (introduced with network refactor)
387 2012-01-17 10:17:55 <da2ce7> Total BTC in?: 50
388 2012-01-17 10:18:07 <da2ce7> tx 67875b261073d5718e3621c17f71b588d8cb1654ad97593ce5c40f20d645b674
389 2012-01-17 10:18:27 <da2ce7> any ideas?
390 2012-01-17 10:19:45 <da2ce7> OneFixt, gmaxwell, luke-jr, MrTiggr?
391 2012-01-17 10:22:10 <OneFixt> hm
392 2012-01-17 10:22:27 <OneFixt> fees?
393 2012-01-17 10:22:32 <OneFixt> or, let me see
394 2012-01-17 10:23:43 <OneFixt> da2ce7: what's wrong with it?
395 2012-01-17 10:24:06 <da2ce7> hopefully nothing... just bitcoin explorers addingup.
396 2012-01-17 10:24:19 <da2ce7> it has one input 50... generation...
397 2012-01-17 10:24:25 <da2ce7> however it's output is more than 50.
398 2012-01-17 10:24:25 <OneFixt> + fees
399 2012-01-17 10:24:52 <OneFixt> or is it not showing the fees...
400 2012-01-17 10:24:58 <da2ce7> no fees.... negitive fees.
401 2012-01-17 10:25:11 <OneFixt> compare to http://blockexplorer.com/tx/a568357e8d43c06b1a4b966f01ab7f1e36c8f8744cc5a4d390c7902e25a454bd
402 2012-01-17 10:25:30 <da2ce7> hmmm
403 2012-01-17 10:25:46 <OneFixt> seems like fee shouldn't be shown as negative though
404 2012-01-17 10:26:18 <OneFixt> i have a feeling it's just fees, i got confused by that once
405 2012-01-17 10:28:28 <da2ce7> if you have no inputs where do the fees come from?
406 2012-01-17 10:28:46 <OneFixt> that's the confusing part
407 2012-01-17 10:29:16 <OneFixt> i think they come from the block inputs, but those aren't shown in the particular tx
408 2012-01-17 10:31:37 <da2ce7> ah... now I get it...
409 2012-01-17 10:31:57 <da2ce7> the fees are from annother tx in the same block.
410 2012-01-17 10:32:02 <OneFixt> yep
411 2012-01-17 10:32:31 <da2ce7> however block-explorer don't show em propper.
412 2012-01-17 12:15:24 <roconnor__> mental note: find etotheipi_'s house and steal his general relativity textbook.
413 2012-01-17 12:30:44 <CIA-76> bitcoin: p2k * r0525a4da1e6f ecoinpool/apps/ecoinpool/src/mysql_sharelogger.erl: PSJ Compatibility Fix http://luke.dashjr.org/programs/bitcoin/w/ecoinpool.git/commitdiff/0525a4da1e6f4bed64d7b49e63943698d96eb70b
414 2012-01-17 12:42:04 <sipa> t
415 2012-01-17 13:04:02 <Joric> are you going to black out bitcoin.org in SOPA protests?
416 2012-01-17 13:04:25 <edcba> let's blackout the bitcoin network !
417 2012-01-17 13:06:30 <UukGoblin> NOO
418 2012-01-17 13:10:26 <ThomasV> luke-jr: you here?
419 2012-01-17 13:43:30 <diki> Now I see what Tycho meant by cgminer discarding work...
420 2012-01-17 13:44:27 <diki> 1550 discarded works...is..
421 2012-01-17 13:44:46 <diki> too much. I will have to remove the code for work discarding..
422 2012-01-17 13:45:09 <diki> a few rejects is nothing compared ot my 150% efficiency
423 2012-01-17 13:45:38 <helo> wow that's really efficient
424 2012-01-17 13:48:03 <diki> The reason is, work rolling, nothing important, but it also helps the pool reduce load
425 2012-01-17 14:55:40 <luke-jr> Joric: No.
426 2012-01-17 14:55:43 <luke-jr> ThomasV: ?
427 2012-01-17 14:55:52 <ThomasV> hi luke
428 2012-01-17 14:56:09 <luke-jr> diki: 150% efficiency is crap; I get over 600% with my poclbm
429 2012-01-17 14:56:12 <luke-jr> ThomasV: hi
430 2012-01-17 14:56:23 <ThomasV> luke-jr: I was wondering what happened with your forum patch for URI
431 2012-01-17 14:56:40 <luke-jr> ThomasV: which one?
432 2012-01-17 14:56:55 <luke-jr> ThomasV: the branch is still there, unmaintained; upstream refused to merge it
433 2012-01-17 14:57:10 <ThomasV> luke-jr: this one: https://bitcointalk.org/index.php?topic=14947.msg696654#msg696654
434 2012-01-17 14:57:27 <ThomasV> why did they refuse it?
435 2012-01-17 14:57:33 <luke-jr> oh, for forum
436 2012-01-17 14:57:46 <ThomasV> yes, for forum
437 2012-01-17 14:57:48 <luke-jr> I think theymos was just waiting for someone to test it locally before applying
438 2012-01-17 14:57:51 <luke-jr> <.<
439 2012-01-17 14:58:08 <ThomasV> waiting 7 months...
440 2012-01-17 14:58:24 <luke-jr> you can test it ;)
441 2012-01-17 14:58:28 <ThomasV> why can't he test it?
442 2012-01-17 14:58:34 <luke-jr> no idea
443 2012-01-17 14:58:37 <ThomasV> I don't have a smf install
444 2012-01-17 14:58:50 <ThomasV> and the patch seems easy to review
445 2012-01-17 14:59:08 <ThomasV> I mean, it's clear it will not display a cosby face
446 2012-01-17 15:00:30 <ThomasV> luke-jr: having bitcoin URIs is a minimum requirement for a Bitcoin forum, IMO
447 2012-01-17 15:00:40 <luke-jr> ThomasV: so bug theymos.
448 2012-01-17 15:00:58 <ThomasV> he's not here atm
449 2012-01-17 15:01:01 <ThomasV> I will
450 2012-01-17 15:07:32 <MC1984> i was thinking bitcoin should register a uri handler with browsers on install
451 2012-01-17 15:07:48 <ThomasV> luke-jr: did you see https://bitcointalk.org/index.php?topic=58534.msg689190#msg689190 ?
452 2012-01-17 15:08:14 <ThomasV> would be nice to have your opinion
453 2012-01-17 15:11:10 <gavinandresen> gmaxwell: thanks for the fuzzer feedback, I started working on a quick&dirty fuzzer this morning
454 2012-01-17 15:11:16 <luke-jr> MC1984: my patch did that
455 2012-01-17 15:11:25 <luke-jr> gavinandresen: good to go with 0.5.2?
456 2012-01-17 15:11:46 <gavinandresen> luke-jr: Did anybody reproduce the build?
457 2012-01-17 15:12:05 <gavinandresen> (my build machine's network is still broken...)
458 2012-01-17 15:12:49 <gavinandresen> luke-jr: ... if somebody else trustworthy will vouch for the SHA's on the binaries, then yes, good to go with 0.5.2
459 2012-01-17 15:13:05 <Latoshi> sweet! new bitcoin version!
460 2012-01-17 15:13:07 <luke-jr> devrandom did, I think
461 2012-01-17 15:13:13 <Joric> what is this fuzzer for? getting rid of ascii art?
462 2012-01-17 15:13:14 <luke-jr> he uploaded sigs to gitian repo or something
463 2012-01-17 15:13:30 <gavinandresen> Joric: https://gist.github.com/1525448
464 2012-01-17 15:13:38 <luke-jr> I presume he would have said something if they didn't match& maybe he can confirm
465 2012-01-17 15:16:00 <MC1984> nice bip but why do you persist with that tonal shit that no one understands
466 2012-01-17 15:16:03 <MC1984> i have ten fingers ffs
467 2012-01-17 15:16:17 <luke-jr> MC1984: hm?
468 2012-01-17 15:16:21 <copumpkin> not sure why ten fingers is relevant unless you count on your fingers a lot
469 2012-01-17 15:16:28 <luke-jr> copumpkin: even then, not relevant
470 2012-01-17 15:16:32 <copumpkin> yeah
471 2012-01-17 15:16:43 <luke-jr> copumpkin: much easier to count in base 6 or 2
472 2012-01-17 15:16:44 <copumpkin> it's an explanation for why we have decimal, but it isn't an explanation of why we should keep it
473 2012-01-17 15:16:44 <luke-jr> on fingers
474 2012-01-17 15:17:04 <MC1984> its relevant because i evolved for base 10 primarily
475 2012-01-17 15:17:08 <MC1984> because of ten fingers
476 2012-01-17 15:17:09 <copumpkin> wat
477 2012-01-17 15:17:11 <mtrlt> lol
478 2012-01-17 15:17:23 <mtrlt> no, you evolved for base 20
479 2012-01-17 15:17:27 <mtrlt> because of ten fingers + ten toes
480 2012-01-17 15:17:33 <mtrlt> or binary, because two arms!
481 2012-01-17 15:17:44 <Latoshi> more like base 4 because of legs
482 2012-01-17 15:17:45 <MC1984> who uses thier toes to count are you retarded
483 2012-01-17 15:17:46 <copumpkin> MC1984: so earlier americans that use all those imperial measurements with /8 fractions underwent a different evolution process?
484 2012-01-17 15:17:52 <mtrlt> MC1984: you don't?
485 2012-01-17 15:17:54 <copumpkin> like the stock ticks of 1/8 dollar
486 2012-01-17 15:18:16 <luke-jr> MC1984: naturally, humans avoid base 10 units
487 2012-01-17 15:18:17 <mtrlt> MC1984: also i think it's the retarded people who use fingers to count
488 2012-01-17 15:18:20 <mtrlt> ;)
489 2012-01-17 15:18:27 <mtrlt> i count in my head
490 2012-01-17 15:18:29 <MC1984> merika is still using imperial when literally the rest of the planet moved to base 10 units, i think that speaks for itself
491 2012-01-17 15:18:37 <copumpkin> MC1984: don't get me wrong, I have objections to tonal, but the number of fingers isn't one of them
492 2012-01-17 15:18:42 <luke-jr> MC1984: until SI happened, almost every culture was using tonal and dozenal units
493 2012-01-17 15:19:00 <copumpkin> MC1984: every base is base 10 :)
494 2012-01-17 15:19:01 <luke-jr> MC1984: also, no country has moved to SI except by force
495 2012-01-17 15:19:06 <copumpkin> or base "10" I should say
496 2012-01-17 15:19:06 <MC1984> yes and we also used to trepann each other to let bad spirits out
497 2012-01-17 15:19:23 <mtrlt> luke-jr: still SI is a lot better than the fuckwad of imperial units :)
498 2012-01-17 15:19:30 <luke-jr> MC1984: even today, people in Europe are willing to go to jail for opposing SI
499 2012-01-17 15:19:34 <luke-jr> mtrlt: nope
500 2012-01-17 15:19:36 <mtrlt> luke-jr: if imperial units were logical as well, like always base 2, i'd probably not oppose them
501 2012-01-17 15:19:51 <mtrlt> but they are like 3 feet by inch, 5139 inches in a yard, 5781 yards in a mile
502 2012-01-17 15:19:54 <Latoshi> I like how 1 satoshi = 1 bitcoin bong
503 2012-01-17 15:19:55 <luke-jr> mtrlt: they're almost all binary, dozenal, or tonal ;)
504 2012-01-17 15:20:05 <mtrlt> they should be all in one base
505 2012-01-17 15:20:15 <mtrlt> otherwise they're too complicated
506 2012-01-17 15:20:17 <luke-jr> mtrlt: I agree, that's why we have the tonal system.
507 2012-01-17 15:20:24 <luke-jr> mtrlt: that is the natural progression of units
508 2012-01-17 15:20:49 <mtrlt> everything in base10 is better than something in base12, something in base16 and something in base2
509 2012-01-17 15:20:55 <Latoshi> but what I don't get, luke-jr, is that not everybody smokes weed, so why should we be using notations like "bitcoin bong" and "bong bitcoin"?
510 2012-01-17 15:20:56 <luke-jr> not "let's throw it all out, and start with decimal despite its unfitness for human use":
511 2012-01-17 15:20:59 <mtrlt> but everything in base16 might be better than base10
512 2012-01-17 15:21:02 <mtrlt> or worse
513 2012-01-17 15:21:03 <mtrlt> dunno
514 2012-01-17 15:21:11 <luke-jr> Latoshi: what does weed have to do with it?
515 2012-01-17 15:21:12 <mtrlt> i dont think it'd make a difference :P
516 2012-01-17 15:21:19 <ThomasV> I hope the 'tonal' question did not pollute the debate about your forum patch; there's no 'tonal' code in it afaict
517 2012-01-17 15:21:33 <luke-jr> mtrlt: even with the forced decimal, people are STILL finding ways to use non-decimal
518 2012-01-17 15:21:35 <MC1984> fuck it
519 2012-01-17 15:21:35 <mtrlt> it always pollutes everything!
520 2012-01-17 15:21:37 <luke-jr> mtrlt: for example, quarters
521 2012-01-17 15:21:40 <MC1984> lets count bitcoin in hex
522 2012-01-17 15:21:50 <luke-jr> USD is a decimal currency, yet they still can't avoid binary divisions like quarters
523 2012-01-17 15:21:53 <Latoshi> luke-jr, you use a bong to smoke weed... sorry I didn't know if this was obvious or not
524 2012-01-17 15:22:05 <mtrlt> luke-jr: not in here. coins of 1 EUR, 50c, 20c, 10c, 5c ;)
525 2012-01-17 15:22:11 <luke-jr> ThomasV: none of the URI stuff is tonal-related.
526 2012-01-17 15:22:15 <mtrlt> even though there _is_ a binary progression with 20 -> 10 -> 5
527 2012-01-17 15:22:24 <luke-jr> MC1984: that's what TBC is
528 2012-01-17 15:22:28 <ThomasV> luke-jr: yes, the wiki page
529 2012-01-17 15:22:51 <luke-jr> ThomasV: at most, some examples are for TBC
530 2012-01-17 15:23:02 <MC1984> as if bitcoin wasnt inaccessable enough as it is
531 2012-01-17 15:23:20 <ThomasV> MC1984: :-)
532 2012-01-17 15:23:21 <MC1984> lets count in some crazy mesopotemian system too"
533 2012-01-17 15:23:43 <luke-jr> MC1984: Tonal is more important than Bitcoin
534 2012-01-17 15:23:51 <mtrlt> hah :P
535 2012-01-17 15:24:22 <ThomasV> yes, Bitcoin is just a way to propagate the Tonal doctrine
536 2012-01-17 15:24:23 <mtrlt> it's not better enough for people to give a shit
537 2012-01-17 15:24:40 <luke-jr> ThomasV: to me, it is
538 2012-01-17 15:24:50 <ThomasV> luke-jr: that's what I meant
539 2012-01-17 15:24:54 <MC1984> bottom line, tonal = ultra hipster
540 2012-01-17 15:25:09 <MC1984> denary is just too mainstream man!
541 2012-01-17 15:26:38 <luke-jr> wtf is denary
542 2012-01-17 15:27:29 <ThomasV> the counting base for denizens
543 2012-01-17 15:28:34 <sipa> nicest counting base: http://en.wikipedia.org/wiki/Golden_ratio_base
544 2012-01-17 15:28:59 <MC1984> shit wiki really is going down
545 2012-01-17 15:29:35 <ThomasV> GPUs are going down tomorrow too
546 2012-01-17 15:29:47 <copumpkin> I love that http://en.wikipedia.org/wiki/Proof_that_0.999..._equals_1 has an actual wikipedia page
547 2012-01-17 15:29:51 <MC1984> ?
548 2012-01-17 15:30:19 <edcba> why not ?
549 2012-01-17 15:30:21 <MC1984> .99999999..... does not equal 1
550 2012-01-17 15:30:25 <copumpkin> oh god
551 2012-01-17 15:30:27 <Latoshi> all pool servers must shutdown in protest of sopa
552 2012-01-17 15:30:53 <copumpkin> MC1984: so I take it you've read that whole page and have found flaws in all their proofs
553 2012-01-17 15:31:03 <roconnor> copumpkin: phi^2 is an even better base.  Adding numbers becomes a game of atomic.
554 2012-01-17 15:31:05 <MC1984> how can it equal 1
555 2012-01-17 15:31:06 <ThomasV> ATI have a remote control; they will use it on SOPA day
556 2012-01-17 15:31:14 <copumpkin> MC1984: how can 1/3 equal 0.3333333?
557 2012-01-17 15:31:19 <copumpkin> (repeating)
558 2012-01-17 15:31:37 <MC1984> because fractions are retarded weve already established that
559 2012-01-17 15:31:39 <copumpkin> anyway, read the article
560 2012-01-17 15:31:44 <copumpkin> lolwut
561 2012-01-17 15:31:44 <mtrlt> okay that does it
562 2012-01-17 15:32:35 <MC1984> he mad
563 2012-01-17 15:33:19 <copumpkin> yep, you clearly won by intellectual superiority and he parted in acknowledgment :)
564 2012-01-17 15:33:39 <MC1984> thats what i was thinking
565 2012-01-17 15:41:54 <devrandom> gavinandresen, luke-jr: I'm the only one that did a gitian build of 0.5.2, so I didn't compare my build against anybody else's
566 2012-01-17 15:42:28 <luke-jr> devrandom: BlueMatt's build was first O.o
567 2012-01-17 15:42:50 <luke-jr> devrandom: http://luke.dashjr.org/programs/bitcoin/files/bitcoin-0.5.2/
568 2012-01-17 15:42:55 <luke-jr> devrandom: can you check those against yours?
569 2012-01-17 15:46:13 <devrandom> luke-jr: I can confirm that the linux binaries match
570 2012-01-17 15:46:24 <luke-jr> devrandom: but not the others? or still checking those?
571 2012-01-17 15:46:27 <luke-jr> gavinandresen: %
572 2012-01-17 15:46:29 <luke-jr> ^
573 2012-01-17 15:46:37 <devrandom> I don't think we have a process for building macosx
574 2012-01-17 15:46:53 <luke-jr> gavinandresen: how do you usually check OS X binaries?
575 2012-01-17 15:46:56 <luke-jr> devrandom: Win32?
576 2012-01-17 15:47:06 <gavinandresen> I trust myself that the osx binaries are OK.
577 2012-01-17 15:47:32 <gavinandresen> (if anybody has brilliant ideas for reproducible osx binaries.....   go for it....)
578 2012-01-17 15:48:22 <devrandom> I could try to build win32, but I thought we were still chasing down non-determinism in that process
579 2012-01-17 15:48:23 <sipa> i don't think there are OSX-crosscompilers for other unix systems, or are there?
580 2012-01-17 15:48:34 <sipa> and it is not allowed to run OSX in a VM
581 2012-01-17 15:49:08 <devrandom> I wonder if the ports system could be used...  I don't know anything about it
582 2012-01-17 15:49:24 <luke-jr> gavinandresen: I trust you too, but you said you wanted double-checked binaries always, so& :P
583 2012-01-17 15:49:37 <luke-jr> I don't care for delays for OS X tho
584 2012-01-17 15:49:39 <devrandom> http://www.macports.org/
585 2012-01-17 15:49:46 <ageis> there's also fisk
586 2012-01-17 15:49:48 <ageis> err]
587 2012-01-17 15:49:49 <ageis> fink
588 2012-01-17 15:50:08 <luke-jr> sipa: OS X is too niche for anyone to bother with crossdev I think :/
589 2012-01-17 15:50:17 <sipa> maybe
590 2012-01-17 15:50:19 <ageis> I have an OSX build environment with fink and XCode if anyone needs somethin
591 2012-01-17 15:51:55 <luke-jr> http://code.google.com/p/toolwhip/
592 2012-01-17 15:52:05 <luke-jr> ageis: setup to build Bitcoin?
593 2012-01-17 15:52:21 <sipa> interesting!
594 2012-01-17 15:53:24 <ageis> luke-jr: i can try
595 2012-01-17 15:53:31 <ageis> haven't done it yet
596 2012-01-17 15:53:41 <sipa> the problem is not a setup to build bitcoin
597 2012-01-17 15:53:44 <ageis> fink should supply a lot of the dependencies
598 2012-01-17 15:54:18 <gavinandresen> I use macports to get most of the osx dependencies.  But as sipa said, that's not the problem-- the problem is getting reproducible builds.
599 2012-01-17 15:55:19 <luke-jr> devrandom: anyhow, better to try to build win32 than not try at all
600 2012-01-17 15:55:57 <devrandom> luke-jr: ok
601 2012-01-17 15:58:18 <devrandom> building
602 2012-01-17 16:01:13 <ageis> which source should i try to build?
603 2012-01-17 16:03:42 <devrandom> ageis: there's no way to get a reproduceable build on OSX, so there's no point (see https://gitian.org/ for why we are trying to do that)
604 2012-01-17 16:04:32 <ageis> gotcha
605 2012-01-17 16:04:56 <luke-jr> devrandom: how's non-Ruby gitian coming? :P
606 2012-01-17 16:08:01 <devrandom> luke-jr: slowly
607 2012-01-17 16:08:36 <devrandom> luke-jr: i.e. not happening yet :-P
608 2012-01-17 17:12:14 <Diablo-D3> http://www.newegg.com/Product/Product.aspx?Item=N82E16812123319
609 2012-01-17 17:12:16 <Diablo-D3> lol 28"
610 2012-01-17 17:27:13 <DrHaribo> Anyone know which miner support the noncerange extension?
611 2012-01-17 17:27:18 <DrHaribo> *miners
612 2012-01-17 17:28:18 <luke-jr> DrHaribo: I think gMinor claims to
613 2012-01-17 17:33:10 <DrHaribo> luke-jr: Did you implement noncerange on your server? Sounds like there is very little miner support.
614 2012-01-17 17:33:35 <luke-jr> DrHaribo: I wrote a reference implementation for pushpool a while ago, but never got it merged with my live server
615 2012-01-17 17:34:33 <Diablo-D3> what the fuck
616 2012-01-17 17:34:39 <Diablo-D3> doesnt newegg sell a pci-e 6 to 8?
617 2012-01-17 17:35:11 <DrHaribo> Too bad, seems like it could be a useful extension
618 2012-01-17 17:35:22 <luke-jr> DrHaribo: indeed
619 2012-01-17 17:35:36 <luke-jr> Diablo-D3: how about being the first mainstream miner to support it? :P
620 2012-01-17 17:38:14 <Diablo-D3> luke-jr: supporting what?
621 2012-01-17 17:42:54 <luke-jr> Diablo-D3: noncerange
622 2012-01-17 17:54:49 <devrandom> luke-jr, gavinandresen: for win32, no match on bitcoin-qt or setup, but match on bitcoind
623 2012-01-17 17:55:14 <luke-jr> O.o
624 2012-01-17 17:55:26 <sipa> is the reason for that indeterminism found already, actually?
625 2012-01-17 17:55:33 <luke-jr> I *guess* that means there's still deterministic bugs?
626 2012-01-17 17:55:48 <luke-jr> devrandom: are the bitcoin-qts *close*?
627 2012-01-17 17:56:07 <luke-jr> devrandom: want to post yours so I can analyze it?
628 2012-01-17 17:58:02 <gmaxwell> it can be useful to objdump the files and diff them.
629 2012-01-17 17:58:17 <luke-jr> gmaxwell: my plan exactly
630 2012-01-17 18:00:52 <sipa> MC1984: convinced already that 0.9999999... == 1 ?
631 2012-01-17 18:01:15 <MC1984> no
632 2012-01-17 18:01:36 <sipa> is 0.3333333... == 1/3 ?
633 2012-01-17 18:02:25 <Diablo-D3> yes
634 2012-01-17 18:02:30 <luke-jr> sipa: no
635 2012-01-17 18:02:32 <MC1984> fuck fractions
636 2012-01-17 18:02:45 <luke-jr> 1/3 is 0.555&
637 2012-01-17 18:02:59 <luke-jr> 2/3 is 0.999&
638 2012-01-17 18:03:16 <sipa> is it unreasonable to assume i'm using decimal here?
639 2012-01-17 18:03:21 <luke-jr> yes. :P
640 2012-01-17 18:03:25 <sipa> please
641 2012-01-17 18:03:38 <MC1984> he is the brige troll of bitcoin i swear
642 2012-01-17 18:04:09 <luke-jr> it's unreasonable to be arguing over whether 0.999& == 1 or not
643 2012-01-17 18:04:10 <luke-jr> ;)
644 2012-01-17 18:05:39 <MC1984> ok 1/3 appears to be .33333
645 2012-01-17 18:06:06 <sipa> now, multiply the fraction by 3, and al digits in the expansion by 3
646 2012-01-17 18:06:13 <sipa> you get 3/3 == 0.999999...
647 2012-01-17 18:06:20 <MC1984> somethings tells me thats a failure of representing a third though
648 2012-01-17 18:06:30 <sipa> it is not
649 2012-01-17 18:06:31 <Diablo-D3> its, btw, 0.3333~
650 2012-01-17 18:06:36 <Diablo-D3> the 3s go on forever
651 2012-01-17 18:06:41 <sipa> yes, of course
652 2012-01-17 18:06:47 <sipa> hence the ...
653 2012-01-17 18:07:18 <MC1984> yes so 1/3 is .333333...
654 2012-01-17 18:07:28 <MC1984> and yet i cant accept that .9999999 is 1
655 2012-01-17 18:07:32 <MC1984> this is fuckery
656 2012-01-17 18:07:36 <Diablo-D3> sipa: two different ways of writing it I think
657 2012-01-17 18:07:40 <Diablo-D3> when I was in school, it was ~
658 2012-01-17 18:07:45 <sipa> Diablo-D3: sure, can be
659 2012-01-17 18:08:10 <Diablo-D3> mc1984: .99999 is 1 depending on your rounding rules. ;)
660 2012-01-17 18:08:16 <luke-jr> .333& * 3 = 1
661 2012-01-17 18:08:20 <sipa> MC1984: the series 0.9, 0.99, 0.999, 0.9999, 0.99999..., is a series that converges to 1; it comes closer and closer (and arbitrarily close), but does not reach it
662 2012-01-17 18:08:30 <luke-jr> .999& != 1
663 2012-01-17 18:08:40 <MC1984> thats what i mean
664 2012-01-17 18:08:42 <sipa> MC1984: now, 0.9999... (with the ...) represents the limit of that series
665 2012-01-17 18:08:44 <luke-jr> sipa: which means it isn't 1
666 2012-01-17 18:08:49 <MC1984> it approaches 1
667 2012-01-17 18:08:58 <sipa> MC1984: the series approaches it, yes
668 2012-01-17 18:09:01 <MC1984> but it is not exactly equal to 1 and never will be
669 2012-01-17 18:09:18 <sipa> but with the ..., we are talking about the limit of the series, and not the series itself
670 2012-01-17 18:09:27 <luke-jr> sipa: not most people
671 2012-01-17 18:09:27 <MC1984> its an infinite number, the moment you stop counting it out to define it, it does not equal one
672 2012-01-17 18:09:37 <MC1984> disrupt that process and the whole premise is fucked
673 2012-01-17 18:09:38 <sipa> MC1984: but it does not stop
674 2012-01-17 18:09:58 <luke-jr> 0.999& to me means the series itself
675 2012-01-17 18:10:07 <luke-jr> a single number
676 2012-01-17 18:10:11 <luke-jr> infinitely close to 1, but not 1 itself
677 2012-01-17 18:10:26 <MC1984> fuck maths i swear
678 2012-01-17 18:10:32 <sipa> in the set of real numbers there does not exist something infinitely close to 1 but not 1
679 2012-01-17 18:10:44 <luke-jr> fine, so maybe it isn't a real number.
680 2012-01-17 18:10:51 <luke-jr> but it's still a concept.
681 2012-01-17 18:10:55 <sipa> in that case, you are right
682 2012-01-17 18:11:07 <sipa> i was assuming real numbers here
683 2012-01-17 18:15:01 <MC1984> so if we jump through a black hole into some strange universe, .99999..... really does =1?
684 2012-01-17 18:15:25 <copumpkin> I take it you still haven't read that wikipedia article
685 2012-01-17 18:15:40 <MC1984> its over my head
686 2012-01-17 18:15:56 <copumpkin> so you don't understand math but you insist on having opinions about it?
687 2012-01-17 18:16:45 <copumpkin> the basic point is that decimal notation isn't unique. More than one decimal sequence can represent the same underlying number
688 2012-01-17 18:17:07 <copumpkin> the same idea applies in any base, for what it's worth
689 2012-01-17 18:17:29 <sipa> copumpkin: not every real number has multiple representations
690 2012-01-17 18:17:41 <copumpkin> sipa: I didn't say that :P
691 2012-01-17 18:17:42 <sipa> (but some do, like 1 and 0.99999...)
692 2012-01-17 18:17:45 <sipa> ok than
693 2012-01-17 18:17:48 <sipa> then
694 2012-01-17 18:17:51 <MC1984> i have no idea what you just said
695 2012-01-17 18:18:16 <MC1984> and having opinions on things we dont know about is what makes us all citizens of the glorious west
696 2012-01-17 18:18:17 <copumpkin> 1 and 0.9 repeating are the same number, just like 0.1 and 0.09 repeating are the same number
697 2012-01-17 18:18:33 <sipa> haha
698 2012-01-17 18:18:46 <copumpkin> 
699 2012-01-17 18:18:58 <copumpkin> (Asimov)
700 2012-01-17 18:19:13 <copumpkin> </elitist_bastard>
701 2012-01-17 18:19:36 <gmaxwell> .999&==1 is one of the uncountably infinite repeated arguements on the internet.
702 2012-01-17 18:20:13 <copumpkin> gmaxwell: which is more controversial, the monty hall problem or that one?
703 2012-01-17 18:20:26 <gmaxwell> There is an elegant semi-proof of the equality, 1/3 == .3333   multiply both by 3   1/3*3 = 1  ; 3*.333 = .999
704 2012-01-17 18:20:35 <copumpkin> gmaxwell: he seems to reject fractions
705 2012-01-17 18:20:59 <copumpkin> I tried that one earlier, and got [11:31:47] <MC1984> because fractions are retarded weve already established that
706 2012-01-17 18:21:08 <gmaxwell> copumpkin: oh sorry, I'm LIFO.
707 2012-01-17 18:21:20 <gmaxwell> Monty hall is easily settled with simulation.
708 2012-01-17 18:21:24 <sipa> luke-jr: if you claim that 0.9999... is "slightly less than 1", 0.33333... is also "slightly less" than 1/3, by the way.
709 2012-01-17 18:21:42 <luke-jr> gmaxwell: no, 3 * .333& = 1
710 2012-01-17 18:21:48 <nanotube> hehe at least the OT discussion on -dev is more fun than OT stuff elsewhere.
711 2012-01-17 18:21:55 <MC1984> fractions dont seem to wholly reconcile with denary system
712 2012-01-17 18:22:03 <copumpkin> there are alternate systems in which 0.9 repeating actually is "slightly less" than 1, but they aren't mainstream and most of math isn't built on top of them
713 2012-01-17 18:22:12 <luke-jr> sipa: how about "neither less than, nor equal to"? ;p
714 2012-01-17 18:22:14 <MC1984> theyre both just human constructs for the representation of numbers
715 2012-01-17 18:22:18 <luke-jr> (nor greater than obviously)
716 2012-01-17 18:22:21 <Diablo-D3> copumpkin: well
717 2012-01-17 18:22:23 <copumpkin> luke-jr: ooh, the IEEE 754 approach
718 2012-01-17 18:22:29 <sipa> copumpkin: haha
719 2012-01-17 18:22:31 <Diablo-D3> yeahm ieee 754 gets it "right"
720 2012-01-17 18:22:40 <MC1984> i dont think theyre comparable and .99999...=1 fuckery shows that it hink
721 2012-01-17 18:23:02 <copumpkin> there's a simple formula for summing an infinite geometric series that most high-schoolers learn
722 2012-01-17 18:23:09 <Diablo-D3> the problem is, decimal representation of fractions is inherently wrong
723 2012-01-17 18:23:14 <copumpkin> that also gives you an obvious "proof" that 0.999& = 1
724 2012-01-17 18:23:25 <sipa> Diablo-D3: it is not
725 2012-01-17 18:23:27 <MC1984> Diablo-D3 thats what i said
726 2012-01-17 18:23:38 <gmaxwell> I heard that if you put an airplane on a treadmill that it can't take off.
727 2012-01-17 18:23:45 <copumpkin> gmaxwell: lol
728 2012-01-17 18:23:49 <Diablo-D3> gmaxwell: oh christ not that one
729 2012-01-17 18:23:49 <luke-jr> gmaxwell: lol
730 2012-01-17 18:23:54 <amiller> gmaxwell, i saw a youtube video with a counter example
731 2012-01-17 18:23:54 <gmaxwell> (assuming the treadmill matches the planes speed)
732 2012-01-17 18:23:54 <sipa> 0.333... describes exactly one real number, one that equals 1/3
733 2012-01-17 18:23:59 <copumpkin> gmaxwell: did you see the whole mythbusters episode on that? I can't believe they made a whole episode on such an obvious question
734 2012-01-17 18:24:08 <gmaxwell> amiller: (Yes, I'm trolling)
735 2012-01-17 18:24:13 <Joric> they actually took off in the end
736 2012-01-17 18:24:14 <amiller> bah
737 2012-01-17 18:24:15 <Diablo-D3> before mythbusters we had... cecil adams.
738 2012-01-17 18:24:17 <sipa> copumpkin: what was the outcome?
739 2012-01-17 18:24:21 <copumpkin> sipa: guess :P
740 2012-01-17 18:24:35 <copumpkin> sipa: OBVIOUSLY it didn't take off, duh
741 2012-01-17 18:24:44 <copumpkin> the treadmill was slowing it down
742 2012-01-17 18:24:52 <gmaxwell> It took off, of course. It helps if you imagine the plane to be a hovercraft (the wheels just roll with a little friction)
743 2012-01-17 18:24:59 <copumpkin> :P
744 2012-01-17 18:25:08 <gmaxwell> frighteningly, the pilot thought it wouldn't take off.
745 2012-01-17 18:25:29 <copumpkin> many mathematicians have argued against the switching doors answer for monty hall
746 2012-01-17 18:25:38 <copumpkin> being an "expert" doesn't always help :P
747 2012-01-17 18:25:49 <gmaxwell> copumpkin: yea, comes from a language disgreement about what monty hall knows/does.
748 2012-01-17 18:25:52 <copumpkin> especially if your subfield doesn't involve careful thinking about that kind of problem
749 2012-01-17 18:25:54 <sipa> history is full of very smart people who said very stupid things
750 2012-01-17 18:26:06 <gmaxwell> copumpkin: if you make monty random (he can show the car) then it doesn't work.
751 2012-01-17 18:26:16 <copumpkin> oh, sure
752 2012-01-17 18:26:25 <copumpkin> but I think that's a fairly silly assumption :P
753 2012-01-17 18:26:30 <copumpkin> since everyone says it's a game show
754 2012-01-17 18:26:35 <MC1984> i think the plane will take off
755 2012-01-17 18:26:48 <copumpkin> MC1984: depends where its propulsion is coming from!
756 2012-01-17 18:26:50 <gmaxwell> yea, but it's hard to escape once you believe that perspective.
757 2012-01-17 18:26:54 <copumpkin> but yeah, for most real planes it will :P
758 2012-01-17 18:27:10 <MC1984> its pushing against the air not the ground
759 2012-01-17 18:27:11 <copumpkin> say you designed an um
760 2012-01-17 18:27:17 <copumpkin> a glider that has a race car tow it
761 2012-01-17 18:27:29 <copumpkin> and they're both on the treadmill
762 2012-01-17 18:27:34 <gmaxwell> yea, if the plane had engines attacked to the wheels, then it wouldn't.
763 2012-01-17 18:27:42 <copumpkin> or someone put a massive engine into the glider's landing gear to help it take off
764 2012-01-17 18:27:44 <sipa> yeah attacking engines!
765 2012-01-17 18:27:57 <MC1984> put the plane in a win tunnel, and it will still take off but wont go anywhere
766 2012-01-17 18:28:12 <helo> that would be awesome
767 2012-01-17 18:28:15 <gmaxwell> hahah
768 2012-01-17 18:28:18 <gmaxwell> attached.
769 2012-01-17 18:28:27 <sipa> MC1984: at exactly the correct wind speed, yes :)
770 2012-01-17 18:28:35 <MC1984> thats what i mean
771 2012-01-17 18:28:48 <MC1984> you get some great troll physics threads on /sci/
772 2012-01-17 18:28:56 <gmaxwell> the control loop for that would be fun.
773 2012-01-17 18:29:07 <MC1984> there was one to do with ramming a cube through a portal, was genuinely perplexing
774 2012-01-17 18:29:22 <amiller> copumpkin, look i made a theorem that [OP_PUSHDATA OP_DROP] is equivalent to [OP_NOP]   (according to my pure functional definition of those ops) https://github.com/amiller/CoinCoq/blob/master/Examples.v
775 2012-01-17 18:29:41 <sipa> amiller: nice!
776 2012-01-17 18:29:57 <copumpkin> cool :)
777 2012-01-17 18:29:57 <devrandom> sipa: luke-jr: I think the problem is in the qt build...
778 2012-01-17 18:30:12 <devrandom> will post a qt build later
779 2012-01-17 18:30:33 <luke-jr> devrandom: & please post your results so I can look too :P
780 2012-01-17 18:30:37 <copumpkin> (* Helper functions, would probably be easier just to use monads *)
781 2012-01-17 18:30:38 <copumpkin> zomg
782 2012-01-17 18:30:41 <sipa> devrandom: but what is different? a randomly generated number? the time? the phase of the moon?
783 2012-01-17 18:30:42 <copumpkin> use the monad, luke
784 2012-01-17 18:30:45 <copumpkin> oh wait, not luke
785 2012-01-17 18:30:48 <amiller> lol
786 2012-01-17 18:30:50 <copumpkin> use the monad, amiller
787 2012-01-17 18:31:05 <copumpkin> amiller: was it you who linked to the hoare state monad?
788 2012-01-17 18:31:13 <amiller> yeah
789 2012-01-17 18:31:16 <copumpkin> cool
790 2012-01-17 18:31:20 <amiller> i went on a whole tour of monads and how they would work in the theorem prover
791 2012-01-17 18:31:22 <copumpkin> that'll work nicely
792 2012-01-17 18:31:29 <amiller> to the best i understand it, i would still have to write a pure functional specification
793 2012-01-17 18:31:45 <copumpkin> yeah, pretty much anything in a theorem prover is purely functional
794 2012-01-17 18:31:47 <devrandom> sipa: I don't know, I just remember that BlueMatt and I had diffs there
795 2012-01-17 18:31:53 <amiller> and what i could use the hoarestatemonad to do is to show that a monadic form (x <- pop; push x; push x) is euqivalent to the pure functional one
796 2012-01-17 18:31:55 <copumpkin> the best you can get to move away from that is to model effects explicitly
797 2012-01-17 18:32:06 <amiller> but that i can't use the monads to circumvent the need to write an entire pure functional spec
798 2012-01-17 18:32:12 <copumpkin> yeah, tru
799 2012-01-17 18:32:18 <amiller> yeah i think we're in synch there
800 2012-01-17 18:32:21 <amiller> cool :p
801 2012-01-17 18:32:28 <devrandom> luke-jr: I'm uploading the bitcoin build to https://gitian.org/0.5.2-win32.zip (will take another 5 min to complete)
802 2012-01-17 18:33:13 <amiller> i don't understand #haskell, man
803 2012-01-17 18:33:18 <devrandom> will upload the qt build tonight or tomorrow
804 2012-01-17 18:33:19 <amiller> they have more people in there than in #python