--- Log opened Sun Nov 16 00:00:52 2014 00:18 < archels_> not sure why this is anything but sensationalism http://www.nature.com/ncomms/2014/141111/ncomms6392/full/ncomms6392.html 00:18 -!- archels_ is now known as archels 00:19 -!- archels [charl@toad.stack.nl] has quit [Changing host] 00:19 -!- archels [charl@unaffiliated/archels] has joined ##hplusroadmap 00:30 -!- altersid_ is now known as altersid 01:07 -!- Dumpster_D1ver [~loki@50.242.254.37] has quit [Quit: Leaving] 01:07 -!- Dumpster_D1ver [~loki@50.242.254.37] has joined ##hplusroadmap 01:12 -!- justanotheruser [~Justan@unaffiliated/justanotheruser] has quit [Ping timeout: 240 seconds] 01:46 -!- lichen [~lichen@c-50-139-11-6.hsd1.or.comcast.net] has joined ##hplusroadmap 01:54 -!- justanotheruser [~Justan@unaffiliated/justanotheruser] has joined ##hplusroadmap 02:49 -!- JayDugger [~jwdugger@pool-173-57-55-138.dllstx.fios.verizon.net] has quit [Quit: Leaving.] 02:54 -!- ebowden [~ebowden@147.69.147.27] has quit [Remote host closed the connection] 03:36 -!- justanotheruser is now known as yolandi 03:36 -!- yolandi is now known as yolandiVisser 04:16 < fenn> archels: the paranoid schizos are going to have a fun time with that 04:21 < fenn> "Synthetic devices for traceless remote control" "mind-controlled gene switch that enables human brain activities and mental states" 04:24 < fenn> i'm kinda suprised it got into Nature 05:12 -!- qu-bit [~shroedngr@unaffiliated/barriers] has quit [Read error: Connection reset by peer] 05:12 -!- FourFire [~FourFire@77.88.71.253] has quit [Quit: Leaving] 05:33 < kanzure> hrm. 05:34 < kanzure> hmm james/cubespawn is using counterparty 05:35 < kanzure> http://www.cubespawn.com/ 05:36 < kanzure> "cubecoin" ugh 05:38 < kanzure> http://thecoinfront.com/crypto-swartz-will-get-you-paid-for-your-great-content/ 05:38 < kanzure> "It works by allowing anyone to upload content to a decentralized network, tag the content with a short but meaningful sentence and then allow anyone to ‘vote’ on that tag." 05:38 < kanzure> that's an awful idea :( 05:40 < fenn> it's reddit 05:40 < fenn> can you add new tags? 05:42 < fenn> oh it gets better, "votes have a minimum cost in order to prevent vote spamming" 05:43 < fenn> so now you get to make spammy content, pay yourself to publish it, and everyone sees it 05:43 < fenn> why do we still not have a trust network 05:43 < kanzure> so they have created a bad idea for a bad advertizing channel? 05:44 < kanzure> i don't know what a "trust network" is supposed to be but everything i've seen claiming to be one has just been riddled with terrible ideas everywhere 05:45 < fenn> you do graph traversal until you find the source of the thing you're judging, and you add up all the links voting it up/down 05:45 < kanzure> yes i know that's the idea, but in practice you get sybil attacks 05:45 < fenn> if there's a blob of spam-bots hanging off the side of the graph, it doesn't get counted in the traversal 05:45 < fenn> why "in practice you get sybil attacks"? 05:46 < kanzure> if you don't care about the other parts of the network then why do you need a network 05:46 < fenn> because otherwise people will do sybil attacks? 05:46 < kanzure> no, i mean if you don't need a network, don't use it 05:46 < fenn> you don't care about the spambots because they're spam 05:47 < fenn> i'm confused 05:47 < fenn> we need a network to prevent sybil attacks 05:48 < kanzure> haha what 05:48 < kanzure> all the cool kids these days just assume that everyone is a sybil attack and then proceed from there, instead of using some central authority to grant immunity from sybil 05:48 < fenn> "The Sybil attack in computer security is an attack wherein a reputation system is subverted by forging identities in peer-to-peer networks." 05:48 < kanzure> "forging" is wrong 05:48 < fenn> find, "creating" 05:48 < fenn> fine* 05:49 < kanzure> well you weren't the quote author 05:49 < fenn> i can fix it if you want 05:49 < kanzure> nope we're good now 05:49 -!- CheckDavid_ [uid14990@gateway/web/irccloud.com/x-hmxqbprekuycilkv] has joined ##hplusroadmap 05:49 < fenn> "In a Sybil attack the attacker subverts the reputation system of a peer-to-peer network by creating a large number of pseudonymous identities, using them to gain a disproportionately large influence." 05:50 < fenn> so what are the cool kids doing? 05:50 < fenn> ignoring everyone? 05:50 < kanzure> bitcoin assumes everyone is an adversary or something 05:50 < kanzure> or potential adversary 05:50 < fenn> bitcoin isn't a reputation system 05:50 < kanzure> and instead each node has to run "eternal vigilance" or w/e 05:51 < kanzure> yes but sybil attacks are important, however 05:51 < kanzure> *are important in bitcoin 05:51 < fenn> maybe you're thinking denial of service 05:51 < kanzure> indeed no 05:51 < kanzure> although 05:51 * fenn reads https://en.bitcoin.it/wiki/Weaknesses#Sybil_attack 05:51 < kanzure> actually yes, arguably a sybil attack is a denial of service attack 05:52 < kanzure> if you populate a system's database with lots of junk, that sounds like denial of service to me 05:52 < kanzure> although not in the conventional sense 05:52 < fenn> this looks like a denial of service to me 05:53 < kanzure> this conversation should probably be moved to -wizards since they think about reputation networks and sybil attacks more regularly than i do 05:53 < fenn> bleh wizards is too smartypants 05:53 < kanzure> e.g. "yo dawgs can you link me to good ideas or implementations for reputation systems? why doesn't exist?" 05:54 < kanzure> wait what do you mean 05:54 < fenn> they expect me to have read 30 years of academic crypto literature 05:54 < kanzure> well i haven't 05:55 < fenn> also i think they'd misunderstand and think i was talking about key verification or something 05:55 < fenn> when i really mean whuffie 05:56 < kanzure> then just say whuffie, jeesh 05:56 < fenn> jeesh 06:03 < kanzure> http://www.waag.org/biohackacademy "The registrations have opened for an intense biohack bootcamp that will start on February 17, 2015: 06:03 < kanzure> "Build your own biofactory in 10 weeks in Waag Society’s FabLab and Open Wetlab. Make all the equipment you need and use an organism from our collection to produce biomaterials such as pigments, cellulose, algae, fuel and more. During weekly lectures you will also learn the basics of biotechnology, 3D design, digital manufacturing, programming and electronics." 06:04 -!- bbrittain [~tiktaalik@2601:6:1781:1800:c0bf:5e26:5bb1:a21b] has quit [Ping timeout: 265 seconds] 06:07 < fenn> that's a lot to learn 06:08 < kanzure> might be possible if it's not learning (just follow instructions) 06:09 < fenn> experience from following instructions is surprisingly valuable 06:09 < kanzure> postdocs rejoice 06:09 < fenn> it gives you a basis to anchor book learning 06:09 -!- Dumpster_D1ver [~loki@50.242.254.37] has quit [Quit: Leaving] 06:09 -!- Dumpster_D1ver [~loki@50.242.254.37] has joined ##hplusroadmap 06:10 < fenn> "here is your clue certification, we have 90% confidence you have a clue" 06:11 < fenn> certifications should come with error bars 06:11 < kanzure> hey do you know any wsgi integration testing frameworks/libraries/things for python? 06:12 < kanzure> i need some gevent-based integration testing framework thing.. except each test must be executed in order (or rather, synchronously, and never in parallel). 06:12 < fenn> no, gevent? 06:13 < kanzure> gevent is just a coroutine library from before asyncio existed 06:13 < kanzure> https://docs.python.org/3/library/asyncio.html 06:13 < kanzure> http://gevent.org/ 06:14 < fenn> "Subroutines are special cases of ... coroutines." great 06:14 < fenn> still doesn't tell me what it is 06:15 < kanzure> .wik coroutine 06:15 < yoleaux> "Coroutines are computer program components that generalize subroutines for nonpreemptive multitasking, by allowing multiple entry points for suspending and resuming execution at certain locations." — http://en.wikipedia.org/wiki/Coroutine 06:15 < fenn> this sounds like GOTO spaghetti 06:16 < kanzure> in practice it ends up looking something like: gevent.sleep() instead of sleep() 06:17 < fenn> so are you doing something like task queues? 06:17 < kanzure> not at the moment 06:18 * kanzure pushes some code 06:18 < kanzure> https://github.com/kanzure/bitcoin-reorg-compatibility-toy 06:19 < kanzure> so, one of the reasons why none of the bitcoin companies from 2010-2011 are still around is because they all failed to implement reorg-handling code 06:19 < kanzure> so i figured i would write an example of an application that can correctly handle rollbacks 06:19 < kanzure> and the demonstration of this correct handling would be in the integration tests 06:20 < fenn> do you know about operational transforms, like in etherpad/google docs 06:20 < kanzure> if i do then not by that name.. 06:20 < fenn> basically small-scale diffs that can be applied atomically 06:21 < kanzure> to solve my current problem i could just write plain python scripts without a testing framework. and then execute those files individually. 06:21 < fenn> anyway it strikes me as similar to reverting a series of commits in a version control system 06:22 < kanzure> but i strongly prefer a testing framework (like unittest?) that can run wsgi applications and have normal testing framework things that i don't want to manually do. 06:22 < fenn> what does wsgi have to do with anything 06:22 < kanzure> well, bitcoind has some event notify hook stuff 06:23 < kanzure> where it calls a script, which i figured would be a curl/wget call to an http web server, which would then perform the right rollbacks in the application itself 06:23 < fenn> does bitcoind exchange data over http? 06:23 < kanzure> you can pass it a path to a script for alertnotify, blocknotify, etc. 06:23 < kanzure> that script can load business code i guess, but ideally i can be running a wsgi application and bitcoind will trigger alertnotify which will trigger an http request to that wsgi application being executed by the test framework. 06:24 < fenn> are reorgnizations that rare that it's an "alert"? 06:24 < kanzure> no, alert is something else 06:25 < kanzure> iirc it's satoshi nakamoto's bat signal 06:25 < kanzure> https://en.bitcoin.it/wiki/Alerts 06:27 < fenn> sounds like a single point of failure 06:27 < kanzure> "The known private key holders are Satoshi Nakamoto, Gavin Andresen and theymos. There are other people able to issue alerts in the event of the incapacitation of the aforementioned.[1]" 06:27 < kanzure> not quite single? 06:27 < fenn> you should be able to configure your client to receive alerts from whatever key you want 06:28 < kanzure> "There are people other than Gavin and myself who have the alert key, but we're the public contact points. If both of us are incapacitated, there are people who will still be able to issue alerts. If one of the alert key holders goes evil, there's a failsafe in the code to handle this." 06:28 < fenn> "it's open source" he says 06:28 < kanzure> sure, but this is just a default 06:29 < fenn> BNN the bitcoin news network 06:29 < fenn> i'm reporting from rural cyberia where an outbreak of mad miner's disease is crippling the local yield 06:29 < kanzure> -blocknotify= Execute command when the best block changes (%s in cmd is replaced by block hash) 06:29 < kanzure> -walletnotify= Execute command when a wallet transaction changes (%s in cmd is replaced by TxID) 06:31 < fenn> can you explain "correctly handle rollbacks" 06:31 < kanzure> i should really say "reorgs" instead of "rollbacks" because "rollbacks" is not in the jargon apparently 06:32 < kanzure> "The term "blockchain reorganization" is used to refer to the situation where a client discovers a new difficultywise-longest well-formed blockchain which excludes one or more blocks that the client previously thought were part of the difficultywise-longest well-formed blockchain. These excluded blocks become orphans. Chain reorganization is a client-local phenomenon; the entire bitcoin network doesn't "reorganize" simultaneously." 06:33 < fenn> is there a point at which you're certain you're on the longest chain? 06:33 < fenn> is there ever* 06:33 < kanzure> probability dictates that for most transactions waiting 6 blocks (6 confirmations) is enough 06:34 < kanzure> https://people.xiph.org/~greg/attack_success.html 06:34 < kanzure> as the blockheight increases the probability of a block being reverted/changed/modified drops dramatically 06:35 < fenn> so, "no" 06:35 < kanzure> when a block is replaced in a reorg there are a number of possibilities.. the original transactions may be included, some may be excluded, other transactions may be included (like those that existed in a prior higher block... which as you can imagine looks funky for deposits to a BTC-holding service) 06:35 < kanzure> yes the answer is "no" but most people running a bitcoin-accepting service do not care about a 1 / 10^1000000 chance or w/e. 06:36 < fenn> just trying to get my mental models straightened out 06:36 < kanzure> attack_success.html says that if an attacker has 1% of the total network hashrate and is trying to rewrte a block from 200 blocks ago, attack success probability is "1.23553e-16" 06:37 < fenn> i dont think they need to be an attacker to cause a reorganization 06:37 < kanzure> the cost of that sort of attack is pretty high, so if you are only dealing with low-value transactions you can usually just assume that nobody is going to bother spending $50M/day or whatever attacking you to revert a $5 transaction 06:38 < fenn> i mean it's just something that happens due to the algorithm 06:38 < kanzure> yep that's right 06:38 < kanzure> but by default you are not attempting to rewrite very deep blocks 06:38 < kanzure> s/rewrite/redo 06:38 < fenn> so you're saying companies failed because of this? because customers lost faith in companies due to transactions that were reverted? 06:39 < kanzure> although justanotheruser pointed out one scenario where a miner sees that another miner got a block that had a huge transaction fee (like say 10,000 BTC maybe)-- in which case, it would be in their best interest to try to mine that block again and so on... if they have enough hasrate, at least. 06:39 < kanzure> well... there are a number of reasons that companies failed, some related to reorgs directly and some that are only slightly related. 06:40 < kanzure> wasn't about losing faith in companies though (afaik) 06:41 < kanzure> more like, "oops a reorg happened, and the original deposit into the site is no longer in the blockchain (and instead those outputs are spent some other way by the original depositor), and you just let the depositor withdraw BTC from your site..." 06:42 < kanzure> (which as you can imagine is not a sustainable or good business practice) 06:44 < fenn> but how would anyone be able to predict and thus exploit that? 06:44 < fenn> if they can't exploit it, it's just noise 06:46 < fenn> the exchange would want to be able to detect that their blockchain model is more up-to-date with global consensus than the user, so presumably they'd spend more on bandwidth and cpu than the user to counter that scenario 06:46 < fenn> global longest-difficulty 06:48 < fenn> also you'd want to ban users that repeatedly deposit/withdraw huge amounts in a spammy way, since they're probably trying to exploit this 06:48 < kanzure> that sounds like a sane thing to do, sure 06:48 < kanzure> another option is to increase the number of confirms that you wait before you let the user do anything with their deposits or payment or w/e 06:48 < kanzure> as for your prediction question.... https://people.xiph.org/~greg/attack_success.html 06:49 < fenn> or implement a fixed percentage deposit/withdrawal fee proportional to the reorg rate 06:49 < kanzure> hmm i don't think proportional to the reorg rate is the right thing there, but i see what you mean. 06:49 < fenn> the rate is variable depending on the age of the transaction 06:50 < kanzure> and the BTC size 06:50 < kanzure> s/size/amount 06:50 < fenn> theoretically, but iirc you haven't seen any evidence of the stealthy mining pool attack idea 06:50 < kanzure> anywho, in the event that a reorg happens and there's a missing transaction or something, a service should still do the right thing 06:51 < kanzure> instead of just not recognizing anything happened, and then carrying on... 06:51 < fenn> what is the right thing? 06:52 < kanzure> depends on the exact service you are running, so my example is just going to be something like "make sure the internal balance for that user in the application is updated to reflect that they never had a deposit" 06:52 < fenn> there will also be scenarios where the user loses their coins to the exchange 06:52 < kanzure> s/never/that it is now invalidated 06:54 < fenn> this is some terrible time-travel paradox voodoo 06:54 < kanzure> yeah.... 06:54 < kanzure> feels like alien technology stuff 06:55 < fenn> more like quantum physics 06:55 < kanzure> "oh but better be careful because reality never happened, apparently" 06:55 < fenn> virtual photons n stuff 06:55 -!- Zinglon [~Zinglon@5351F2A8.cm-6-2d.dynamic.ziggo.nl] has joined ##hplusroadmap 06:57 < fenn> quick hire a physicist 06:58 < kanzure> "CVEs for privacy" https://github.com/openrightsgroup/common-privacy-violations 06:59 < fenn> i wish people would stop talking about "privacy" and instead talk about "deanonymization" or "libel" 07:00 < kanzure> anyway as you can imagine this time paradox stuff does not make bitcoin any easier to understand 07:01 < fenn> wtf 1 commit with 1 file with 1 line 07:01 < fenn> how do i down-vote 07:01 < kanzure> whoops 07:01 < kanzure> blame grr in swhack 07:01 < fenn> i left swhack because it was too noisy 07:02 < kanzure> sounds right 07:03 < fenn> actually the time paradox stuff is good because it keeps you thinking about how the consensus algorithm is distributed 07:03 < fenn> it's fundamentally different from an atomic centralized database 07:04 < fenn> er, atomic is the wrong word 07:04 < fenn> canonical? 07:04 < kanzure> it seems to me that you can /probably/ safely assume it's a regular centralized ledger if you are only dealing with very tiny transactions, i think 07:04 < kanzure> and where transaction loss is not detrimental to your operations 07:04 < fenn> no, transaction size doesn't matter, it's transaction age that matters 07:05 < kanzure> (not all possible services need to hold a bitcoin balance. some of them just broadcast transactions or something.) 07:06 < fenn> you'd need a significant quantity of "smart" miners to go after large transactions (technically blocks with large transaction fees) but these miners don't exist yet or can't compete with dumb miners 07:06 < kanzure> eligius is a well-maintained large mining pool that does some clever incentive-compatible things because luke-jr is regularly looking out for optimizations to make 07:07 < fenn> well, is he doing this? 07:07 < kanzure> i haven't asked. 07:08 < kanzure> you can make certain assumptions that make my toy/demo/example unnecessary i'm sure, but i'm interested in the toy/demo for mostly "pie in the sky" thinking reasons 07:08 < kanzure> and reasons like "why hasn't anyone implemented this sort of example yet?" 07:09 * fenn wonders if he just tickled the dragon 07:10 < kanzure> https://en.bitcoin.it/wiki/Eligius 07:11 -!- CheckDavid_ [uid14990@gateway/web/irccloud.com/x-hmxqbprekuycilkv] has quit [] 07:12 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-crdvwsjxpjwzlbui] has joined ##hplusroadmap 07:14 < fenn> http://eligius.st/ don't work for me 07:14 < fenn> it's returning a badly mime-typed 503 07:14 < kanzure> hmm back to trying to find an acceptable python integration testing framework. 07:15 -!- comma8 [comma8@gateway/shell/yourbnc/x-rgbzddwlzjvweayr] has quit [Ping timeout: 272 seconds] 07:18 -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 264 seconds] 07:21 < fenn> https://people.xiph.org/~greg/attack_success.html has some atrocious code in it 07:23 < kanzure> emscripten generated that code 07:23 < kanzure> you can tell because it says "emscripten" 07:23 < fenn> i see that, but still, fuck that 07:23 < fenn> this is a clear case for a simple equation 07:24 -!- comma8 [comma8@gateway/shell/yourbnc/x-vsuutkxukgollmot] has joined ##hplusroadmap 07:24 < kanzure> andytoshi: do you have a link to the equation used in https://people.xiph.org/~greg/attack_success.html ? 07:26 -!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap 07:27 < fenn> do symbolic regression on the output ;P 07:32 < fenn> this reminds me of some greg egan novel where they created a new universe out of a different set of computational primitives and went exploring. basically every egan novel ever 07:33 < kanzure> 16:32 < justanotheruser> probability of winning is in the satoshi whitepaper and here https://people.xiph.org/~greg/attack_success.html 07:33 < kanzure> page 6 https://bitcoin.org/bitcoin.pdf 07:34 < kanzure> something that has been bugging me is that an attacker can exclude his deposit transaction but a withdraw transaction published by the service may still be usable... 07:35 < fenn> so (hashrate_attacker/hashrate_network)^num_blocks_behind 07:35 < kanzure> one option might be to include all previous outputs by the exchange in every transaction (and of course, only spend the right amount to whatever external address, while sending the rest back to the exchange). however, this will destroy any privacy benefits of using separate transactions... 07:36 < fenn> are you telling me we need a 3000 line script to calculate that? 07:36 < fenn> jesus 07:36 < kanzure> what you should be asking is why emscripten was used 07:36 < kanzure> he was probably using some source code from bitcoind for some reason 07:36 < fenn> thats what i am asking 07:36 < andytoshi> fenn: it's a sum, your formula is too specific 07:37 < andytoshi> but no, it's not 3000 lines, there is C code in the bitcoin.pdf it's like 20 lines 07:37 < kanzure> then why is gmaxwell's version 3000 lines? 07:37 < andytoshi> paging through, i see a loot of boilerscript 07:37 < fenn> it has a bunch of unrelated crap 07:38 < andytoshi> eg `function dedup` is only used in `calculateStructAlignment` which is only used recursively by itself 07:38 < fenn> anyway "see page 7 of bitcon paper" would have sufficed 07:40 < kanzure> andytoshi: could you also look over my message spam in #bitcoin from a few seconds ago? 07:41 < andytoshi> sure 07:43 < fenn> i think satoshi's equation on page 6 is wrong anyway, it shouldn't be "1" for p <= q, it should still be (q/p)^z 07:43 < fenn> but i am just a n00b what do i know 07:46 < Qfwfq> Fingerprints of Tor exit nodes providing JSTOR access: http://sprunge.us/WWDB 07:48 < fenn> for completeness, Luke-jr says "that wouldn't work" wrt mining stale blocks with high transaction fees 07:51 < yorick> Qfwfq: <3 07:53 < fenn> "the internet treats censorship as damage and routes around it" 07:54 < andytoshi> fenn: it's 1 07:57 < fenn> i guess the formulation of "will catch up" doesn't make sense, because the honest nodes can always "catch up" again, with some small probability 07:58 < fenn> there's no real difference between the attacker and the rest of the network 07:58 < fenn> consider a network with two nodes; which one is the attacker and which one is the honest node? 08:00 < andytoshi> "will catch up" makes sense, it means there exists a time T at which the attacker's chain has same total work as the honest one 08:00 < fenn> no because only blocks with more work are accepted as valid 08:01 < fenn> uh, pedantry 08:01 < andytoshi> ?? 08:02 < andytoshi> blocks are valid if their included transactions are valid, if the work on them exceeds the required work, and the required work is correctly calculated from the chain they are included in 08:03 < fenn> oh, my misunderstanding/unfamiliarity with terminology 08:03 < fenn> what's it called when a node looks at blocks and picks the one with the most work? 08:05 < andytoshi> i don't think there's a term for that, we say things like "determining the true history" which is obvs not good terminology :) 08:05 < fenn> maybe "finding consensus" or "consensus chain"? 08:09 < andytoshi> yeah, that's better, "determining the consensus history" 08:10 < fenn> true/false there can exist two chains with equal proof of work 08:12 < fenn> o great kanzoracle hear my prayers 08:13 * fenn goes back to squinting at hieroglyphics 08:13 < andytoshi> there can, the way bitcoin measures proof-of-work 08:13 < andytoshi> in fact almost all single-block forks are of this nature 08:15 < fenn> hm ok i was thinking proof of work was some large integer value 08:15 -!- bbrittain [~bbrittain@172.245.212.12] has joined ##hplusroadmap 08:16 < fenn> is length of difficulty just counting the number of F's in all the hash? 08:16 < fenn> hashes* 08:18 < andytoshi> no, it computes a target [0, T] that the hash has to fall into, when intepreted as a 256-bit number 08:19 < andytoshi> and T is recalculated every 2016 blocks so that the last 2016 blocks, had they used target T, would've taken 2 weeks 08:19 < andytoshi> (by multiplying T by <2 weeks>/) 08:20 < andytoshi> and this target is included in the blockheader for each block, and has to be exactly correct for the block to be a valid part of its chain 08:20 < fenn> hm maybe i meant number of 0's 08:21 -!- Merovoth [~Merovoth@gateway/tor-sasl/merovoth] has joined ##hplusroadmap 08:22 < fenn> i didnt mean "how do you find the network difficulty rate" i meant "how do you find the length of the chain" 08:23 < andytoshi> there is a value T in each block ... 2^256/T is the expected time (in hash attempts) to find each block, so you sum up all the 2^256/T's 08:23 < fenn> if anything within [0, T] qualifies as valid, then any chain with an equal number of valid block has equal length? 08:23 < andytoshi> the actual value of the hash has to be ≤ T, but even if it's super tiny this is irrelevant, all that's checked is whether it's less than T 08:23 < andytoshi> yup 08:23 < fenn> so why do they say "length of difficulty" instead of just "length of chain" 08:23 < andytoshi> exactly, except when you cross difficulty-periods 08:24 < andytoshi> because when the difficulty changes, the T values change 08:24 < andytoshi> so long chains can have different total work even with same number of blocks 08:25 < fenn> wouldnt their blocks be considered invalid because they had the wrong target? 08:25 < andytoshi> no, if they are different chains their blocks might have different timestamps, so they'd have different "correct targets" 08:26 < kanzure> haha time travel 08:26 < fenn> literally 08:26 < andytoshi> each block has a (miner-determined) timestamp in it which is used to compute "<2 weeks>/" in the diffchange equation 08:26 < andytoshi> so "actual time" is a bad term, should be "miner-claimed time" :) 08:27 < maaku> with a lot of leeway... 08:27 < fenn> do people forge timestamps in practice? 08:28 < maaku> it would be stupid for them not to 08:28 < fenn> why even call it timestamps if it's just an arbitrary number 08:28 < maaku> any miner working on the first or last block of an adjustment period can guarantee 0.5% more income for the next 2k blocks by lying on their timestamps 08:29 < fenn> back in the real world we have an incentive to make timestamps as accurate as possible 08:29 < maaku> fenn: it isn't arbitrary. it's roughly +/- 2hrs of actual time 08:29 < maaku> fenn: there are incentives here too 08:29 < maaku> it just doesn't result in a clock that is more than a few hours in sync 08:30 < fenn> timelikestamps 08:37 < fenn> this is an awful diagram http://en.bitcoinwiki.org/images/thumb/4/41/Mining.png/500px-Mining.png 08:42 < fenn> this whole page is terrible; i'd rewrite it if i were more sure of how stuff actually works 08:42 < fenn> http://en.bitcoinwiki.org/Mining 08:43 < fenn> it reads like someone ran it through google translate a couple times 08:44 < maaku> *the whole wiki is terrible 08:44 < maaku> fixed that for you 08:44 < fenn> thanks 08:44 < maaku> did you know you have to pay the admin to get edit rights? 08:45 < fenn> welcome to our brave new cryptocurrency world 08:45 < maaku> f$*k that 08:45 < fenn> why do people still use it then 08:47 < fenn> oh bitcoin.it is different 08:51 < maaku> oh ok 08:51 < maaku> well that explains why this one exists :) 08:54 < fenn> i'm dreading the proliferation of unicode TLD's and domains 08:55 < fenn> see here's a perfect situation for a distributed trust metric 08:56 < fenn> there's no reason my browser should have to rely on the meatware to detect spammy domains 09:01 < kanzure> i have edit rights and i didn't pay nothing 09:03 < kanzure> maaku: were you following this conversation in #bitcoin a few minutes ago? i am interested in feedback http://diyhpl.us/~bryan/irc/bitcoin/reorgs.log.txt 09:04 < maaku> I'm not on #bitcoin; is there a one-sentence summary q? 09:05 < kanzure> in these logs i am proposing that every bitcoin-using service/company should correctly handle reorgs by using a transaction that references all deposited unspents, and that without doing this you have a good recipe for losing reserves. 09:05 < kanzure> and that every withdrawal should reference this single consolidated tx out that references the deposited unspents 09:05 < kanzure> because if a deposit goes missing you want withdrawals to go missing. anything else would be inappropriate. so it's just my thoughts on implementation details etc. 09:05 < maaku> fees would be insane 09:06 < kanzure> i think there's a way to do this sort of transaction without directly referencing all previous deposited unspents, you just need it referenced in the history tree somewhere 09:06 < maaku> well you can just have one internal token input/output 09:06 < maaku> you don't have to reference them all 09:06 < maaku> impose transaction serialization yourself by having a single input which is used as a semaphore 09:06 < kanzure> "the rule is basically: you should never allow a withdrawal from an unspent output that is unrelated to the latest set of deposits" 09:07 < kanzure> "unrelated" is not quite the right word 09:07 < kanzure> i think semaphore is probably more accurate...? 09:07 < maaku> sounds too complicated 09:07 < kanzure> but you see the problem though, right? 09:07 < maaku> no, i just gave a simple solution 09:07 < maaku> stick 1 btc in an output 09:08 < maaku> every exchange transaction involves (successive verisons of) that output 09:08 < maaku> and each transaction spend the deposits which have accumulated since the last 09:09 < kanzure> so your scheme puts all deposited BTC in a single address? 09:10 < kanzure> i mean, a single key in control of all the unspent outputs 09:14 < maaku> what? no 09:14 < maaku> there's just one output that gets recycled on every transaction 09:14 < maaku> which forceably serializes the txns 09:18 < kanzure> yep, okay, that's what i proposed in that log excerpt 09:18 < kanzure> cool 09:21 < kanzure> fenn: "Mystery trader buys all Europe's cocoa 17 Jul 2010 The purchase was enough to move the entire global cocoa market, sending the price to the highest level since 1977, and triggering rumours and intrigue in the City. It is unclear which person, or group of traders, was behind the deal, but it was the largest single cocoa trade for 14 years. The cocoa beans, which are sitting in warehouses either in The Netherlands, Hamburg, or closer to ... 09:21 < kanzure> ... home in London, Liverpool or Humberside is equivalent to the entire supply of the commodity in Europe, and would fill more than five Titanics. They are worth £658 million." http://www.telegraph.co.uk/finance/markets/7895242/Mystery-trader-buys-all-Europes-cocoa.html 09:21 < kanzure> someone's got a chocolate problem 09:22 < fenn> i've got a chocolate probem 09:22 < maaku> someon cornered the market 09:22 < maaku> i tihnk we all have chocolate problems now 09:23 < fenn> what's the expiration date on a cocoa bean 09:24 < fenn> maybe its reptilians from sirius B 09:26 < fenn> In one of the Race's bases in Siberia, morale is at an all time low. The weather is a truly miserable condition from the hot one the Lizards are used to, and the Race's soldiers feel they're constantly being sent to their deaths by incompetent commanders. Many have fallen into abusing ginger, which works as a narcotic for them, even though it has been outlawed by Atvar's orders (such disobedience 09:26 < fenn> would have been considered unthinkable before they came to Earth). 09:27 < fenn> s/ginger/chocolate/ 09:27 < kanzure> hmm what were those changes to paperbot from yesterday 09:27 < kanzure> i didn't look. /me goes to look 09:32 -!- Jaakko9113 [~Jaakko@host217-42-47-122.range217-42.btcentralplus.com] has quit [Ping timeout: 240 seconds] 09:33 -!- Jaakko9113 [~Jaakko@host86-190-181-79.range86-190.btcentralplus.com] has joined ##hplusroadmap 09:44 -!- maaku [~quassel@50-0-37-37.dsl.static.fusionbroadband.com] has quit [Remote host closed the connection] 09:45 < kanzure> all of this stuff reminds me of doc brown's chain reactions: https://www.youtube.com/watch?v=KJRh-37H4fA 09:45 < fenn> kanzure i think you should write up your proposal for how to deal with reorganizations 09:46 < fenn> the log is kinda rambly and i don't know enough about transactions yet to really understand what you're saying 09:47 < kanzure> i think it would be more obvious to if you had a good intuition about how current wallets are implemented and how people often create transactions at the moment 09:47 < fenn> 1) description of the attack 2) description of the loss-avoidance scheme 3) market effects and incentives resulting from implementing 2 09:48 < kanzure> "oh you want a withdrawal? sure have some free money (also we are now debiting your internal account on our site)" 09:48 < kanzure> but the problem with this is that in the event of a reorg, the deposit may no longer exist but your withdrawal sure does 09:48 < fenn> i meant a whitepaper, not telling me over irc 09:48 < fenn> doesn't have to be latex or whatever 09:48 < kanzure> oh man this is sounding like work 09:49 < fenn> sorry 09:49 < kanzure> hmm 09:49 < kanzure> maybe an email to bitcoin-development 09:49 < fenn> the attack would probably involve 2 "users" 09:50 < fenn> accounts really 09:50 < kanzure> oh the 2 users thing in the logs was just a reference to a way to bypass one of the proposed schemes 09:50 < kanzure> my original proposal was to link all deposits with all withdrawals 09:50 < fenn> right, but it's important to keep in mind 09:50 < kanzure> and i was trying to weaken it to be just "all deposits from that one user" but this doesn't work because in some cases a user can withdraw more than they deposited anyway 09:51 < kanzure> (actually that reason might be wrong. there's possibly a better reason that i am forgetting at the moment.) 09:51 < fenn> money laundring/anonymization schemes want to do the opposite of this, so you'd have to wait some period to be certain there are no reorgs 09:51 < kanzure> "just wait 6 blocks" is what the consensus advice has been so far to other developers 09:52 < fenn> yeah usually the point of an exchange is to send or receive money 09:52 < kanzure> it's not just for exchanges though, this is broadly applicable to many bitcoin services 09:52 < fenn> i dunno what a service is 09:52 < kanzure> an exchange is a service 09:52 < kanzure> bitcoin-using companies/services/things 09:52 < kanzure> applications? 09:53 < fenn> is there a valid use case for withdrawing deposits less than 6 blocks old? 09:54 < kanzure> my threat model assumes adversaries that can generate 6 blocks pretty easily 09:55 < kanzure> "in the event of a disappearance of a deposit, like during a reorg with some unprobably-high-but-otherwise-happening depth, it is important that withdrawals become invalidated because otherwise you are giving away money" 09:55 < fenn> oh right, some kind of denial of service to gain hashrate share 09:55 < kanzure> nah just "we spent 20 billion dollars building mining equipment" and stuff 09:55 < fenn> easier to fuck up the exchange's connection to the bitcoin network 09:56 < kanzure> argh stop talking about exchanges 09:56 < fenn> exchanges is all i know :( 09:56 < kanzure> coinbase 09:56 < kanzure> (the site) 09:56 < fenn> easier to sabotage N miners' connections to $bitcoin_company than to buy N miners 09:56 < kanzure> so one of the properties of the bitcoin network is that you don't really know easily which bitcoind client is related to a company 09:57 < kanzure> unless they are hosting it on the same ip address 09:57 < fenn> that doesnt seem hard to discover 09:57 < kanzure> or unless you are monitoring an overwhelmingly large chunk of the bitcoin network directly and can identify which p2p nodes in the network are broadcasting company-related transactions earlier than the rest of the network 09:57 < fenn> .wik coinbase 09:57 < yoleaux> "Coinbase is a bitcoin wallet and exchange service headquartered in San Francisco California, founded by Brian Armstrong and Fred Ehrsam. Users of Coinbase can buy and sell bitcoin at the current market rate with a bank transfer in both the United States and eighteen European countries for a fee of about 1% of the value of the transaction." — http://en.wikipedia.org/wiki/Coinbase 09:58 < kanzure> i would hesitate to call coinbase.com an exchange 09:58 -!- snuffeluffegus [~snuff@5.150.254.180] has joined ##hplusroadmap 09:58 < kanzure> https://en.bitcoin.it/wiki/Coinbase 09:58 < fenn> ok there are two kinds of exchanges, BTC-BTC exchanges and BTC-fiat exchanges 09:58 < kanzure> "The coinbase is the content of the 'input' of a generation transaction. While regular transactions use the 'inputs' section to refer to their parent transaction outputs, a generation transaction has no parent, and creates new coins from nothing." 09:58 < kanzure> no that's not what i meant 09:58 < kanzure> i mean, yes you can buy and sell bitcoin on coinbase.com, that's true... 09:58 < kanzure> but there's no order book that you are looking at 09:59 < kanzure> and you can only buy/sell you can't make an order above whatever price coinbase.com offers you 09:59 < kanzure> s/above/above or below 09:59 < fenn> i hate when people name things after domain-specific terminology 09:59 < kanzure> blockchain.info is also guilty of this 09:59 < kanzure> and chain.com 10:00 < fenn> unspent-txo.com 10:00 < fenn> bitcoinwallet.net 10:00 < fenn> or whatever 10:00 < kanzure> yeah... 10:00 < kanzure> as you can imagine the name blockchain.info has caused a lot of problems with introducing people to the concepts 10:01 < kanzure> "so you mean there is a company that maintains the blockchain?" 10:01 < kanzure> "sigh.... no. but yes that company exists.... i am not denying their existence." 10:02 < fenn> there really needs to be some kind of interactive game where you buy and sell fake stuff with fake bitcoins and see what's going on in the servers and what information is transmitted around] 10:02 < kanzure> there's testnet and regtest for testing bitcoin things 10:02 < fenn> because right now 99% of what's happening is invisible 10:03 < kanzure> and there's some games that claim to be related to bitcoin but they are just games (they don't actually implement bitcoin stuff) 10:03 < fenn> add invisibility to faulty mental models and incentives for confusion and you get confusion 10:03 < kanzure> http://www.beepboopbitcoin.com/ 10:04 < kanzure> there's lots of content published everywhere but the problem is that it's faster and cheaper to create bullshit 10:04 < fenn> it doesnt have to actually implement bitcoin; it would be hard to follow ECDSA calculations anyway 10:04 -!- qu-bit [~shroedngr@unaffiliated/barriers] has joined ##hplusroadmap 10:05 < fenn> reading "why johnny can't encrypt" illustrates how hard it is to convey accurate metaphors for cryptographic operations 10:08 < fenn> why is coinbase not an exchange? they just have a manually-operated buy/sell setpoint 10:08 < fenn> could write one in 5 minutes 10:08 < kanzure> i suppose wikipedia calls it one so i should too 10:09 < kanzure> coinbase.com has a few other features that are not exchange-related, at least, like wallets and merchant payment processing 10:09 < kanzure> (typical paypal.com feature-set for integration with iframes or whatever else) 10:09 < fenn> so they do BTC-BTC and BTC-fiat 10:09 < kanzure> i don't know what BTC-BTC is 10:10 < kanzure> they let you send BTC to other users using their internal ledger 10:10 < kanzure> does that count? 10:10 < fenn> trading bitcoins within a private site without doing transactions on the bitcoin network, to get around the 10 minute block latency 10:10 < kanzure> ah okay. yes. 10:10 < kanzure> but they have like a 4 day waiting period after every operation i think 10:10 < fenn> also so people dont have to set up their own client 10:10 < fenn> wtf 4 days? 10:11 < kanzure> *after every buy/sell 10:11 < kanzure> yeah, ACH 10:11 < fenn> ok that's because of the retarded banking system that's been around since 1890 10:11 < kanzure> or er it may be ACH... i dunno. 10:11 < fenn> .wik ACH 10:11 < yoleaux> "Disambiguation: ACH" — http://en.wikipedia.org/wiki/ACH 10:11 < fenn> rawr 10:11 < fenn> .wik automated clearing house 10:11 < yoleaux> "Automated Clearing House (ACH) is an electronic network for financial transactions in the United States. ACH processes large volumes of credit and debit transactions in batches. ACH credit transfers include direct deposit payroll and vendor payments." — http://en.wikipedia.org/wiki/Automated_clearing_house 10:12 < kanzure> yoleaux should make a guess when there's a disambiguation required 10:12 < fenn> nevermind 10:13 < fenn> i have a feeling we're putting more thought into the bitcoin network and exchange operator policies than was ever put into "real" banking 10:14 < kanzure> "sure just have my buddy audit their books i'm sure everything is fine" 10:14 < fenn> Some common Standard Entry Class (SEC) codes: 10:14 < fenn> SEC code, that's not confusing at all 10:16 < kanzure> a lot of banking and accouning quickly intersects professions like law 10:16 < kanzure> because they are all entangled in this big messy confusion 10:16 < kanzure> with lots of principle agent problems and stuff 10:18 -!- rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has joined ##hplusroadmap 10:18 < fenn> "Coinbase has received US$31 million in capital investments, nearly one-third of the total publicly disclosed investment in Bitcoin companies made as of February 2014" 10:18 < fenn> thats smaller than i expected 10:19 < kanzure> hmm well they just raised $60M like yesterday 10:24 -!- rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has quit [Ping timeout: 245 seconds] 10:54 < bbrittain> I've been kinda bored for the last hour, and I started looking for bio-inspired fiction. holy shit, nobody thinks it's gonna be a good thing 10:54 < bbrittain> It's all so... depressing 10:55 < kanzure> i have really not seen any great biology-inspired fiction 10:55 < kanzure> maaaybe some stuff on orion's arm 10:55 < kanzure> http://orionsarm.com/ 10:57 < bbrittain> zomg, it's like stepping back into the late 90s internet 10:57 -!- TheShadowFog [~TheShadow@2601:8:3e80:c:645e:de36:d7cf:9c2f] has joined ##hplusroadmap 10:58 < fenn> bbrittain: greg egan has some interesting books 10:58 < bbrittain> fenn: I've heard mixed reviews of egan's stuff, so I've always refrained 10:58 < bbrittain> worth it? 10:58 < fenn> i liked it 10:58 < bbrittain> a series/book in particular? 10:58 < fenn> teranesia, chaff 10:59 < kanzure> greg egan's biology stuff never seemed particularly impressive 10:59 < kanzure> but i haven't looked at those.. what's in them? 11:00 < fenn> i forget which is what, it all gets mushed up in my memory 11:00 < bbrittain> "Chaff — An agent is sent to kill a geneticist who is working in a drug lord-controlled stronghold in the jungles of Colombia, and working on important brain-altering research." 11:01 < kanzure> was that his short story one? 11:01 < kanzure> "white knights" etc 11:01 < fenn> yes 11:01 < bbrittain> I saw some mention of this biofiction film festival on twiterr a few weeks ago, I watched a couple of the winning ones. Essentially boils down to "don't fuck with nature. nature is special" 11:02 < kanzure> yeah, it's totally boring stuff that has been done hundreds of itmes 11:02 < kanzure> *times 11:02 < kanzure> i think the problem is that it's just genuinely hard to write good biology-inspired fiction 11:02 < kanzure> biology is hard enough as it is 11:02 < bbrittain> I read ribofunk, which I thought had mixed stories, but good scifi qualities 11:03 < kanzure> who was the author that used a "one short story per new technology introduced" strategy? 11:03 < kanzure> hm. 11:04 < bbrittain> I liked windup girl 11:04 < bbrittain> roo'd was a meh for writting, story was good 11:06 < bbrittain> hmmm... the fact that I can't think of anything else is telling 11:06 < bbrittain> makes me want to give it a go, but it is such a time commitment 11:06 < bbrittain> and I have better things to be doing 11:07 < kanzure> make them very very short stories 11:07 < fenn> various levels of neurosciencey biotech stuff: ted chiang's "understand", john mcloughlin "the helix and the sword, joan slonczewski "the brain plague", cory doctorow "0wnz0red" 11:08 < fenn> "the helix and the sword" is more about post-earth space colonization 11:09 < fenn> hard to reconcile the huge number of books i havent read with the books i have read, so i have a bias to assume that i've read the good ones :\ 11:10 < kanzure> you can also factor in recommendations from your peer group though 11:10 < fenn> hugo and nebula awards are pretty reliable 11:10 < kanzure> it is unlikely for terribly relevant books to have escaped our collective attention for a while now 11:10 < fenn> anything not SF is not worth reading :P 11:10 < bbrittain> hugo nebula rarely give false positives, but I bet they give a whole bunch of false negatives :P 11:11 < fenn> oh "rainbows end" by vernor vinge has some terribly bad biotech references 11:11 < kanzure> "accelerando" if you can tolerate it (apparently some people in here can) 11:11 < bbrittain> ugh, and some terribly bad codec/network latency tech 11:11 < kanzure> "neverness" had biology stuff heh 11:11 < fenn> no it doesn't 11:11 < fenn> it's all about monks and space piots 11:11 < fenn> pilots* 11:11 < fenn> and "entities" 11:12 < kanzure> what about the agathanians? 11:14 < bbrittain> movie recs? 11:14 < kanzure> a whole bunch of junk 11:15 < kanzure> i think the most biologically accurate you could hope for is like x-men or something 11:16 * bbrittain likes x-men 11:19 -!- SolGr [~SolGr@c-69-141-24-242.hsd1.nj.comcast.net] has joined ##hplusroadmap 11:20 -!- _Sol_ [~SolGr@c-69-141-24-242.hsd1.nj.comcast.net] has quit [Ping timeout: 256 seconds] 11:27 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap 11:30 < archels> which of Stross' works should I read after Accelerando? 11:32 < fenn> his obituary 11:33 < fenn> oops did i say that 11:40 < fenn> "singularity sky" won a Hugo award so there's that 11:48 < kanzure> i need more of you guys to read more vernor vinge please 11:48 < bbrittain> I love about half of vinges stuff 11:48 < bbrittain> the other half... :/ 11:49 < kanzure> i started on that other half 11:49 < fenn> which half 11:49 < kanzure> a deepness in the sky 11:50 < yorick> yeah, Rainbow's End was terrible 11:50 < fenn> i read true names; it was more about transhumanism and becoming one with the computer network than it was about anonymity and hacker culture 11:51 < yorick> fenn: was it the good half? 11:52 < fenn> um, it might have been original in 1980 but now it just sounds like every hollywood "cyberspace teenage hacker" plot 11:52 < fenn> also the illustrations were really bad 11:53 < kanzure> nihei is the only appropriate illustrator choice 11:53 < fenn> this wasn't that sort of book 11:54 < kanzure> don't impose your fascist manga ideas on me 11:54 < kanzure> er... or something. 11:54 < fenn> i mean, nihei's works set on 21st century earth aren't particularly amazing 11:55 < kanzure> "books can be like, whatever we want. just look at borges' junk." 11:56 < fenn> nihei would be a good illustrator for borges 11:57 < fenn> in general, if there's new architecture or bodily forms 11:58 < fenn> the premise of True Names was that cyberspace looked like your average dungeon crawler 11:58 < kanzure> "from now on every illustration is tensegrity" 11:58 < kanzure> "and hexagons" 11:59 < fenn> tread lightly, lest ye offend the trans-aetheric modalities 12:01 < kanzure> why aren't there any scifi authors in here trying to exploit this channel 12:01 < yorick> at least three of us are stross, probably 12:02 < yorick> but it's not *that* popular 12:02 < kanzure> i strongly doubt stross is in here 12:02 < yorick> http://www.antipope.org/charlie/blog-static/2013/02/rokos-basilisk-wants-you.html 12:03 < fenn> because you crosspost everything to tt 12:03 < yorick> yeah, he's probably not here 12:03 < kanzure> for other reasons too though 12:04 < kanzure> like he's gone off on the crazy end anyway 12:04 < fenn> is roko's basilisk just pascal's mugging 12:04 < kanzure> (social justice outrage machining) 12:06 < fenn> heh "if people get the afterlife they believe they will get, missionaries should be shot on sight." 12:13 -!- Zinglon [~Zinglon@5351F2A8.cm-6-2d.dynamic.ziggo.nl] has quit [Remote host closed the connection] 12:15 < fenn> huh "In 1995, Nick Szabo proposed a challenge to build a macroscale replicator from Lego robot kits and similar basic parts." 12:16 < fenn> i probably already pasted that in this channel 12:18 < fenn> corollary to Saunt Lora's Proposition: any unique insight you thought you should have known by know has actually been pasted into ##hplusroadmap 12:19 < fenn> by now* 12:19 < kanzure> there was a handful of usenet posts about self-relication 12:19 < kanzure> *self-replication 12:20 < fenn> "i am a relic" <- self-relication 12:20 < kanzure> i'm sure that quote is talking about a szabo usenet email 12:20 < fenn> it was to the extropians mailing list 12:20 < kanzure> .title http://szabo.best.vwh.net/nano.musings.html 12:20 < yoleaux> Nick Szabo -- Nanotechnology, Self-Reproduction & Agile Manufacturing 12:20 < kanzure> damn 12:20 < fenn> http://web.archive.org/web/20060307220916/http://www.lucifer.com/~sean/N-FX/macro.html 12:21 < kanzure> hmm i don't remember "extropians@waterville.warwick.com" 12:22 < kanzure> ah the next entry on that page is the usenet one at least 12:22 < kanzure> "In 1998, Chris Phoenix suggested a general idea for a macroscale replicator on the sci.nanotech newsgroup, operating in a pool of ultraviolet-cured liquid plastic, selectively solidifying the plastic to form solid parts. Computation could be done by fluidic logic. Power for the process could be supplied by a pressurized source of the liquid." 12:23 < fenn> how does it generate UV light 12:23 < kanzure> well according to kragenjaviersita this is not one of our interests 12:23 < kanzure> (or his for that matter) 12:24 < kanzure> "Alfredo DeSantis … spoke on “Graph decompositions and secret-sharing schemes,” a silly topic which brings joy to combinatorists and yawns to everyone else." 12:24 < kanzure> .title http://www.scottaaronson.com/blog/?p=2059 12:24 < yoleaux> Shtetl-Optimized » Blog Archive » What does the NSA think of academic cryptographers? Recently-declassified document provides clues 12:24 < fenn> derr, somehow i got stuck in a dependency tree: http://web.archive.org/web/20060313114452/http://www.lucifer.com/~sean/N-FX/depend.html 12:25 < fenn> wtf they were talking about stewart platforms for self-replication 12:25 < kanzure> "I think I have hammered home my point often enough that I shall regard it as proved (by emphatic enunciation): the tendency at IACR meetings is for academic scientists (mathematicians, computer scientists, engineers, and philosophers masquerading as theoretical computer scientists) to present commendable research papers (in their own areas) which might affect cryptology at some future time or (more likely) in some other world. Naturally this ... 12:25 < fenn> i guess i independently came up with the same idea, but for non-nanotech replication 12:25 < kanzure> ... is not anathema to us." 12:30 < kanzure> linked from that page... http://anotherlook.ca/ "n our time one of the dominant paradigms in cryptographic research goes by the name "provable security." This is the notion that the best (or, some would say, the only) way to have confidence in the security of a cryptographic protocol is to have a mathematically rigorous theorem that establishes some sort of guarantee of security (defined in a suitable way) under certain conditions and given ... 12:30 < kanzure> ... certain assumptions. The purpose of this website is to encourage the emergence of a more skeptical and less credulous attitude toward this notion and to contribute to a process of critical analysis of the positive and negative features of the "provable security" paradigm." 12:31 < fenn> merciless crossposting 12:31 < kanzure> http://eprint.iacr.org/2004/152.pdf "We discuss the reasons why the search for mathematically convincing theoretical evidence to support the security of public-key systems has been an important theme of researchers. But we argue that the theorem-proof paradigm of theoretical mathematics is of limited relevance here and often leads to papers that are confusing and misleading." 12:31 < kanzure> yep 12:32 < Qfwfq> plz move discussion 2 ####transhuman-bitcoin-wizards 12:33 < heath> music... 12:33 < heath> promodj.com/tracks 12:57 < juri_> kanzure: what is that pdf editing library you're working with? 12:57 < kanzure> https://github.com/kanzure/pdfparanoia 12:57 < kanzure> yeah i'm not parsing the pdf format 12:57 < kanzure> just manual editing in code 12:57 < kanzure> in other words, don't do as i do 13:01 < fenn> def parse_pdf: """Parses a PDF via pdfminer.""" 13:01 < kanzure> oh 13:01 < kanzure> well, the editing part is manual 13:02 < kanzure> at the time when i looked pdfminer was not capable of writing pdf 13:02 < fenn> probably not 13:03 < fenn> so do you just delete the compressed watermark? 13:03 < kanzure> yep 13:04 < fenn> you know it seems that a lot of stupid file format incompatibility stuff could be solved by better data structure visualization tools 13:04 < fenn> right now we're just monkeys looking at ascii text scratching our heads 13:05 < kanzure> there's theoretically a pdf specification and parsers/generators that might implement it 13:05 < kanzure> but i wasn't able to find a python-compatible one too easily 13:05 < kanzure> and i was not interested in writing my own at the time 13:06 < fenn> hmm i could have used pdfminer in the past 13:06 < fenn> if it actually does what it claims 13:07 < fenn> i ended up rendering to a bitmap and doing OCR :P 13:07 < kanzure> you may also be interested in pdfquery 13:07 < Qfwfq> With PDFs, you have to run through decompression before getting ASCII text. Structure visualisation isn't much more useful than any other structured parse, I think. 13:07 < Qfwfq> Such as that you'd get with PDFMiner? I can't remember it's API that well. 13:07 < Qfwfq> Funnily enough the PDF specification was closed until like five years ago I think 13:08 < kanzure> so far most academic pdfs that i have found have just been using "deflate" for their compression 13:09 < kanzure> but i'm sure some nasty publisher somewhere decided to use some other method 13:10 < yorick> looking at it, I think pdftotext does slightly better at extracting text than pdfminer (for example pdftotext handles broken words well) 13:10 < kanzure> entirely possible 13:10 < fenn> not in my experience 13:11 < kanzure> i don't think pdfminer was made to do text extraction things like that 13:11 < kanzure> i mean, word recombination and stuff 13:16 < Qfwfq> Generation is more of a concern here 13:16 < Qfwfq> Given we want valid PDFs after modification, and it has lots of numeric pointers to objects at positions in the file 13:17 < kanzure> yeah i think technically the output from pdfparanoia is sometimes "invalid" but for some reason various pdf readers still support rendering anyway 13:19 < Qfwfq> s/numeric pointers to/byte offsets of/ 13:19 -!- Coffeenut [~Coffeenut@162.211.217.101] has quit [] 13:38 -!- pete4242 [~smuxi@boole.london.hackspace.org.uk] has joined ##hplusroadmap 13:42 -!- drewbot [~cinch@ec2-54-83-69-194.compute-1.amazonaws.com] has quit [Remote host closed the connection] 13:42 -!- drewbot [~cinch@ec2-54-91-208-9.compute-1.amazonaws.com] has joined ##hplusroadmap 14:21 -!- snuffeluffegus [~snuff@5.150.254.180] has quit [Quit: Leaving] 14:21 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Quit: Ex-Chat] 14:50 < kragenjaviersita> Qfwfq: the PDF specification was never very closed 14:50 < kragenjaviersita> I mean there were undocumented features of Adobe Reader that Acrobat Whatever would exploit 14:51 < kragenjaviersita> but those were mostly details 14:51 < kanzure> ParahSailin_: can you find me the "bitvc qq chatroom" 15:04 -!- loki_ [~loki@50.242.254.37] has joined ##hplusroadmap 15:06 < yorick> bbrittain: I liked Diaspora and Permutation City from Greg Egan the best, but the other two recommendations are about the only ones I haven't read 15:07 -!- yolandiVisser [~Justan@unaffiliated/justanotheruser] has quit [Ping timeout: 256 seconds] 15:15 < fenn> but those aren't about biology at all 15:18 < yorick> oh right, that was a requirement 15:19 < yorick> hm, one of the Orthogonal books has some weird alternative-universe biology, I guess 15:19 -!- loki_ [~loki@50.242.254.37] has quit [Quit: Leaving] 15:20 -!- loki_ [~loki@50.242.254.37] has joined ##hplusroadmap 15:20 < yorick> how about Peter Watts (Blindsight?). he's a biologist, after all 15:21 -!- loki_ [~loki@50.242.254.37] has quit [Client Quit] 15:24 -!- yolandiVisser [~Justan@unaffiliated/justanotheruser] has joined ##hplusroadmap 15:27 -!- Viper168_ [~Viper@unaffiliated/viper168] has joined ##hplusroadmap 15:27 < bbrittain> I'm actually reading Blindsight right now! 15:27 < bbrittain> It's got some cool neuro stuff, nothing more I think 15:28 -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 240 seconds] 15:29 < bbrittain> one of my roommates is eating dunkin donuts and drinking a martini, the other is drinking soylent with a glass of wine 15:31 < bbrittain> this can only be described as something approaching post-modern-absurdism 15:33 -!- Viper168_ is now known as Viper168 15:33 < fenn> which one do you think will live longer 15:34 * kanzure wonders why code has not materialized on his screen 15:35 -!- snuffeluffegus [~snuff@5.150.254.180] has joined ##hplusroadmap 15:36 < fenn> you forgot to connect the brain-computer-interface 15:37 < kanzure> nah i did all the other stuff first and left the "???" for future-me 15:37 < kanzure> which was not a good plan 15:38 < bbrittain> fenn: soylent/glass-o-wine 15:38 < bbrittain> soylent is actually not that bad 15:38 < bbrittain> bland/boring/marketing-hype 15:39 < fenn> i know, i invented it 15:41 < fenn> the waffle form is more palatable tho 15:41 -!- ebowden [~ebowden@147.69.147.27] has joined ##hplusroadmap 15:44 -!- snuffeluffegus [~snuff@5.150.254.180] has quit [Quit: Leaving] 15:46 < kanzure> still having trouble finding a gevent-compatible unittest suite thingy 15:53 < fenn> how about just http://stackoverflow.com/a/23036785 15:53 < fenn> .title 15:53 < yoleaux> unit testing - How to test Python 3.4 asyncio code? - Stack Overflow 15:54 < fenn> decorator to run stuff 15:58 < kanzure> "For you acceptance tests you could start the server in a separate process and send SIGINT when you're done" 15:58 < kanzure> but then how do you insert your mocks >:( 15:59 < kanzure> http://stackoverflow.com/a/11945119/687783 16:02 < bbrittain> "fenn: i know, i invented it" elaborate? 16:03 -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 240 seconds] 16:07 < fenn> typical quantified self hyper-standardization of intakes scenario 16:08 < fenn> rhinehart independently reinvented it, whatever 16:08 < fenn> i'm sure hundreds of other have come up with the same idea 16:09 < fenn> basically, "why do only lab rats get repeatable nutrition?" 16:09 < fenn> or, "why do i have to put effort into thinking about food" 16:10 < fenn> it's really complicated and complex 16:10 < kanzure> and even if you do think about it lots, 16:10 < kanzure> it will still turn out that something is totally not bio-available the way you have been eating it or something 16:11 < kanzure> s/something/something important 16:11 < fenn> i dont trust the RDI values anyway 16:13 < bbrittain> hence why I tried it for just an experiment, it's the aestetic which is apealing, not the reality. 16:16 < fenn> ah but you havent tried it as a waffle 16:17 < fenn> also the soylent formulation is too carb-heavy and has crap oils 16:18 < fenn> a simple reformulation is to blender up oats, concentrated milk, and an egg; cook in waffle iron with a slab of butter; serve with green tea 16:19 < fenn> the butter goes in the waffle iron... 16:19 < fenn> needs potassium and magnesium that still 16:21 < fenn> and C 16:21 < fenn> bleh 16:23 < kanzure> fenn did you see their subdomain thingy here http://diy.soylent.me/recipes/people-chow-301-tortilla-perfection 16:23 < fenn> yes i fought with the stupid javascript webapp thing for long enough already thank you 16:24 < kanzure> i didn't look very closely but it looks like they have other recipes from users for some reason 16:24 < fenn> vanilla corn tortillas yuck 16:25 < fenn> it's missing... all vitamins? 16:26 < fenn> there's a classic study on pellagra from people eating corn tortillas without B vitamins 16:27 < fenn> oh the vanilla stuff has vitamins 16:29 < fenn> kanzure: the idea is you can make your own recipe and it calculates all the nutritional components (assuming you/someone else entered them already) 16:30 < fenn> a glorified spreadsheet 16:31 < kanzure> wow such glory 16:33 < kanzure> grr sqlalchemy will force sqlite to share an in-memory database across threads 16:36 -!- Coldblackice [darkblue@gateway/shell/bouncerstation/x-werdabaeoclhzgnn] has quit [Ping timeout: 265 seconds] 16:48 -!- yolandiVisser is now known as justanotheruser 16:50 -!- snuffeluffegus [~snuff@5.150.254.180] has joined ##hplusroadmap 16:59 -!- Dumpster_D1ver [~loki@50.242.254.37] has quit [Quit: Leaving] 17:10 -!- Dumpster_D1ver [~loki@50.242.254.37] has joined ##hplusroadmap 17:12 -!- Dumpster_D1ver [~loki@50.242.254.37] has quit [Quit: Leaving] 17:12 -!- Dumpster_D1ver [~loki@50.242.254.37] has joined ##hplusroadmap 17:12 < kanzure> https://soundcloud.com/morttagua/morttagua-deeper-sunset 17:28 < ebowden> kanzure: http://www.smbc-comics.com/comics/20141114.png 17:30 < fenn> i tricked someone into doing uncertain future and their result was 40% likely that a super-AI has already taken over all reality 17:31 < ebowden> heh 17:32 < catern> uncertain future is a JAVA APPLET? 17:32 < catern> what. 17:40 < kanzure> yeah sorry about that 17:40 < kanzure> i took too long writing a javascript implementation for them 17:40 < kanzure> so they brought in this java person 17:41 < kanzure> he did okay i guess, but it's just unfortunate that it had to end up a java applet 17:41 < kanzure> anyway here are some screenshots: http://diyhpl.us/~bryan/irc/theuncertainfuture/?C=M 17:44 -!- qu-bit [~shroedngr@unaffiliated/barriers] has quit [Read error: Connection reset by peer] 17:49 -!- Vutral [ss@mirbsd/special/Vutral] has quit [Ping timeout: 245 seconds] 18:00 -!- Dumpster_D1ver [~loki@50.242.254.37] has quit [Quit: Leaving] 18:00 -!- Vutral [ss@mirbsd/special/Vutral] has joined ##hplusroadmap 18:01 -!- sheena [~home@d162-156-158-13.bchsia.telus.net] has quit [Read error: No route to host] 18:06 -!- Dumpster_D1ver [~loki@50.242.254.37] has joined ##hplusroadmap 18:09 < kanzure> "The Moon is given to be moving slower than light because it remains within the "future light cone" propagating from its position at any instant." 18:10 < kanzure> via nsh http://math.ucr.edu/home/baez/physics/Relativity/SpeedOfLight/FTL.html 18:15 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has quit [Remote host closed the connection] 18:17 -!- sheena [~home@d162-156-158-13.bchsia.telus.net] has joined ##hplusroadmap 18:18 -!- Dumpster_D1ver [~loki@50.242.254.37] has quit [Quit: Leaving] 18:18 -!- Dumpster_D1ver [~loki@50.242.254.37] has joined ##hplusroadmap 18:27 < kanzure> http://en.wikipedia.org/wiki/Flyby_anomaly 18:27 < kanzure> "Possible explanations of the flyby anomaly include.... A dark matter halo around Earth;[8]" 18:28 < kanzure> http://arxiv.org/abs/0805.2895 18:28 < kanzure> http://arxiv.org/pdf/0805.2895v4.pdf 18:28 < paperbot> http://libgen.org/scimag/get.php?doi=10.1103%2FPhysRevD.79.023505 18:28 < kanzure> oh come on now you work? 18:29 -!- ebowden [~ebowden@147.69.147.27] has quit [Remote host closed the connection] 18:30 -!- snuffeluffegus [~snuff@5.150.254.180] has quit [Quit: Leaving] 18:32 < kanzure> https://soundcloud.com/morttagua/morttagua-house-mag-series-020 18:49 -!- gene_hacker [~chatzilla@c-50-137-46-240.hsd1.or.comcast.net] has joined ##hplusroadmap 19:14 -!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap 19:31 < kanzure> gene_hacker: hi 19:32 < gene_hacker> hello 19:32 < yoleaux> 8 Nov 2014 09:16Z gene_hacker: solid-state electrically-actuated membrane coatings? something you could apply to a porous ceramic plate, polarise with an electric field, then gently dry or fuse in a kiln maybe (?), finally install as the last side of your vacuum chamber. When you apply power, pumping of nitrogen/gas would occur. Design goals I guess: pumping action; no/low gas permeability, tight crystal structure around … 19:32 < yoleaux> the pump mechanism; electrically ... 19:32 < yoleaux> 8 Nov 2014 09:17Z gene_hacker: ... actuated; works in dry conditions; can be fused with porous substrate forming overall no/low gas permeability 19:32 < gene_hacker> already told him I'll see what we can do 19:33 < gene_hacker> but yeah we might be able to do it, or something like it 19:35 < kanzure> gene_hacker: suspicions of dark matter halo around earth http://arxiv.org/pdf/0805.2895v4.pdf 19:47 < gene_hacker> neat 19:48 < gene_hacker> now hopefully they'll turn out to be ultradense matter, because that'd be cool 19:49 < kanzure> classic jojack: https://www.youtube.com/watch?v=xCjbZic2Tmg#t=38m10s 19:51 < sheena> fenn: do you know JDM Engine, as a term for a.. type of imported engine? 20:01 -!- Jaakko9113 [~Jaakko@host86-190-181-79.range86-190.btcentralplus.com] has quit [Quit: Nettalk6 - www.ntalk.de] 20:10 -!- gde33 [kvirc@546A0B75.cm-12-3a.dynamic.ziggo.nl] has quit [] 20:13 -!- Maelstro [~beatzebub@S0106b81619e8ecee.gv.shawcable.net] has quit [Read error: Connection reset by peer] 20:17 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-crdvwsjxpjwzlbui] has quit [Quit: Connection closed for inactivity] 20:18 -!- ebowden [~ebowden@147.69.147.27] has joined ##hplusroadmap 20:23 -!- ebowden [~ebowden@147.69.147.27] has quit [Ping timeout: 244 seconds] 20:28 -!- Beatzebub [~beatzebub@S0106b81619e8ecee.gv.shawcable.net] has joined ##hplusroadmap 20:33 -!- justanotheruser [~Justan@unaffiliated/justanotheruser] has quit [Ping timeout: 244 seconds] 20:40 < kanzure> "The attack relies primarily on the fact that a given input/output difference pattern only occurs for certain values of inputs. Usually the attack is applied in essence to the non-linear components as if they were a solid component (usually they are in fact look-up tables or sboxes). Observing the desired output difference (between two chosen or known plaintext inputs) suggests possible key values." 20:40 -!- justanotheruser [~Justan@unaffiliated/justanotheruser] has joined ##hplusroadmap 20:44 < dingo_> hehe i was mentioned but i was at yosemite all weekend and my buffer can't recall it :) 20:47 < gene_hacker> did you manage to pirate some thermus aquaticus? 20:48 < kanzure> dingo_: it may have been about a 91 chevy sprint 20:58 -!- Burnin8 [~Burn@pool-71-191-174-26.washdc.fios.verizon.net] has joined ##hplusroadmap 20:59 -!- Burnin8 [~Burn@pool-71-191-174-26.washdc.fios.verizon.net] has quit [Client Quit] 21:01 -!- Burn_ [~Burn@pool-71-191-174-26.washdc.fios.verizon.net] has quit [Ping timeout: 264 seconds] 21:12 < nmz787> delinquentme: ya I think scripting, kanzure recommended to me most recently cadquery which uses freecad which uses opencascade 21:13 < nmz787> delinquentme: guy at the local FIB manufactory had a demo using freecad to slice 3d models into layers that he fed sequentially to the FIB 21:18 < nmz787> .tell delinquentme check logs ^... also you should start looking up nanofluidics... you can literally exclude chemistry from happening and stuff by just making a hallway too small, do separations by tuning surface charge of the channel and buffer make-up 21:18 < yoleaux> nmz787: I'll pass your message to delinquentme. 21:18 -!- ebowden [~ebowden@147.69.147.27] has joined ##hplusroadmap 21:18 < nmz787> paperbot: http://iopscience.iop.org/0957-4484/20/16/165302/pdf/0957-4484_20_16_165302.pdf 21:19 < nmz787> paperbot: http://iopscience.iop.org/0957-4484/20/16/165302 21:32 -!- ebowden [~ebowden@147.69.147.27] has quit [Remote host closed the connection] 21:46 -!- TheShadowFog [~TheShadow@2601:8:3e80:c:645e:de36:d7cf:9c2f] has quit [Quit: Leaving] 22:08 -!- Burninate [~Burn@pool-71-191-174-26.washdc.fios.verizon.net] has joined ##hplusroadmap 22:14 -!- qu-bit [~shroedngr@unaffiliated/barriers] has joined ##hplusroadmap 22:16 -!- Dumpster_D1ver [~loki@50.242.254.37] has quit [Quit: Leaving] 22:16 -!- Dumpster_D1ver [~loki@50.242.254.37] has joined ##hplusroadmap 22:16 -!- Dumpster_D1ver [~loki@50.242.254.37] has quit [Client Quit] 22:18 -!- ebowden [~ebowden@147.69.147.27] has joined ##hplusroadmap 22:37 -!- ebowden [~ebowden@147.69.147.27] has quit [Remote host closed the connection] 23:16 -!- ebowden [~ebowden@147.69.147.27] has joined ##hplusroadmap 23:28 -!- Burnin8 [~Burn@pool-71-191-174-26.washdc.fios.verizon.net] has joined ##hplusroadmap 23:31 -!- Burninate [~Burn@pool-71-191-174-26.washdc.fios.verizon.net] has quit [Ping timeout: 240 seconds] 23:39 -!- Beatzebub [~beatzebub@S0106b81619e8ecee.gv.shawcable.net] has quit [Read error: Connection reset by peer] 23:40 -!- Qfwfq [~WashIrvin@unaffiliated/washirving] has quit [Ping timeout: 240 seconds] 23:45 -!- Qfwfq [~WashIrvin@cm58.kappa38.maxonline.com.sg] has joined ##hplusroadmap 23:45 -!- Qfwfq [~WashIrvin@cm58.kappa38.maxonline.com.sg] has quit [Changing host] 23:45 -!- Qfwfq [~WashIrvin@unaffiliated/washirving] has joined ##hplusroadmap --- Log closed Mon Nov 17 00:00:53 2014