01:27:27amiller_:amiller_ is now known as amiller
01:54:31tdryja:tdryja has left #bitcoin-wizards
02:03:55lnr:lnr has left #bitcoin-wizards
02:35:56andytoshi:this made me smile https://bitcointalk.org/index.php?topic=29675.msg11035598#msg11035598 (list of lowest blockhashes, in a thread about it spanning almost 4 years)
04:17:28bsm117532:Hahaaa given a sufficiently large random string of digits, It's no surprise you found pi in it. Also decaf is one digit shorter than c0ffee so occurs more often. And deadbeef is two digits longer so is much more rare. But I'm sure you know this. ;-)
06:22:17gmaxwell:sipa: yea they made their WNAF constant time in a subsiquent update from the code I'd been reading by adding a constant time conditional point sign flip.
06:22:36gmaxwell:sipa: it's still not completely constant time due to other (smaller issues).
06:24:10sipa:i read it has a constant number of iterations
06:29:30gmaxwell:Their code has a unified jacobian add which is different from the one in libsecp256k1.
06:36:08gmaxwell:It looks like 8mul 3sqr, so probably actually the same with the 1/2+2*sqr for a mul substitution.
06:37:34sipa:but you're talking about group operations; not exponentiation algorithms
06:39:10gmaxwell:... Yes, I switch subjects.
06:39:23sipa:oh ok
06:41:05gmaxwell:Why would I have stayed on the old one! I answered you. :) Their WNAF is a 'fixed window', it always double four times and either adds or subtracts. The original version of it branched to flip the sign on the stored point for subtraction but the latest has a conditional sign flip (that I didn't look inside, but I assume is constant time).
07:30:30luigi1111w:luigi1111w is now known as Guest48690
08:05:13verne.freenode.net:topic is: This channel is not about short-term Bitcoin development | http://bitcoin.ninja/ | This channel is logged. | For logs and more information, visit http://bitcoin.ninja
08:05:13verne.freenode.net:Users on #bitcoin-wizards: andy-logbot jtimon phiche dEBRUYNE lclc Guest48690 Tiraspol SDCDev hktud0 luktgf nivah priidu b_lumenkraft Mably koeppelmann unlord binaryatrocity_ HostFat_ c0rw1n TheSeven p15 cluckj bitbumper justanotheruser sadoshi Dr-G2 d1ggy_ CodeShark bliljerk101 grandmaster RoboTeddy omni_ vmatekol_ adam3us jonasschnelli hashtag_ PaulCapestany mengine Starduster pollux-bts Emcy yorick Iriez antgreen merlincorey nuke1989 spinza gielbier arubi_ AllieSenbub
08:05:13verne.freenode.net:Users on #bitcoin-wizards: koshii mkarrer Guest35538 fanquake [ace] eric helo a5m0 Cornholi0 nephyrin null_radix tjader crescendo Sqt wumpus Madars richardus cornus_ammonis mikolalysenko LeMiner GAit cdecker waxwing sturles prodatalab adams_ GreenIsMyPepper harrow melvster vonzipper berndj Zouppen andytoshi Xzibit17 cfields manan19 comboy sneak realcr gmaxwell rustyn jaromil catlasshrugged_ Apocalyptic harrigan bosma Pan0ram1x PRab dasource nullbyte Cory forrestv
08:05:13verne.freenode.net:Users on #bitcoin-wizards: cryptowest_ amiller runeks__ kanzure kefkius throughnothing Graet Eliel veox indolering K1773R Keefe petertodd jcorgan larraboj ryan-c jessepollak MoALTz_ gribble tromp mr_burdell gnusha tromp_ d9b4bef9 wiz aakselrod starsoccer dgenr8 copumpkin btcdrak go1111111 weex SwedFTP Hunger- lmacken dc17523be3 face HM Luke-Jr luny yrashk artifexd kumavis platinuum otoburb huseby midnightmagic BlueMatt warren TD-Linux mariorz hguux___ fenn wizkid057
08:05:13verne.freenode.net:Users on #bitcoin-wizards: ajweiss SubCreative Adlai Anduck iddo poggy nsh kyuupichan Logicwax phantomcircuit EasyAt lechuga_ luigi1111 isis nanotube yoleaux gavinandresen dignork AdrianG s1w livegnik optimator fluffypony Meeh cursive dansmith_btc morcos guruvan BananaLotus bedeho heath roasbeef_ Fistful_of_Coins stonecoldpat afdudley espes__ pigeons sipa warptangent phedny so BrainOverfl0w MRL-Relay azariah btc___ @ChanServ brand0 davout NeatBasis CryptOprah leakypat
08:05:13verne.freenode.net:Users on #bitcoin-wizards: jbenet mappum Oizopower Krellan Taek epscy sdaftuar Alanius nickler gwillen kinlo coryfields_ Muis catcow [d__d] lnovy sl01 null michagogo STRML
09:17:56agorist000:agorist000 is now known as metamarc
09:25:05metamarc:metamarc is now known as agorist000
09:25:41agorist000:agorist000 is now known as metamarket
09:47:46nsh:* nsh muses
09:52:11fluffypony:like the band? or do you literally have a Greek goddess next to you inspiring you?
09:53:02nsh:a little from column A, a little from column Beta
09:53:09sipa:or ARE you a greek goddess?
09:53:16nsh:rats, secret's out
09:54:19nsh:nah, i was thinking of some kind of system to cross-reference -wizards and mailing-list discussions and (a very small subset of) bitcointalk in various ways to try and better systematize the body of emerging thought and knowledge
09:55:01nsh:to create a more manageable ascent of mt. blockchain
09:58:00nsh:* nsh gets distracted epidemiological mountain-climbing analogues and parallel transport
10:01:09nsh:kanzure will hopefully have some ideas about how to go about creating a practical dependency structure for hierarchically-comprehensible complex understanding
10:59:23nsh:Information Geometry and Applications. / Background material -- https://nlsconf.physics.uoc.gr/sites/nlsconf.physics.uoc.gr/files/files/background_4_talk_ghikas_2014_01.pdf
11:06:05nsh:.wik Contrast function
11:06:06yoleaux:"In statistics and information geometry, divergence or a contrast function is a function which establishes the "distance" of one probability distribution to the other on a statistical manifold." — http://en.wikipedia.org/wiki/Contrast_function
12:41:49fanquake:fanquake has left #bitcoin-wizards
14:05:35kanzure:nsh: copy-paste
14:06:29nsh:for what?
14:20:10kanzure:that is the answer
14:29:46nsh:* nsh is doubtful
14:30:59zooko:* zooko is too.
14:31:07zooko:* zooko doesn't know what we're talking about, but is just doubtful in principle.
16:05:20psztorc:psztorc has left #bitcoin-wizards
16:58:44Guest48690:Guest48690 is now known as luigi1111w
18:51:21pampuchy:hi y'all
18:51:39pampuchy:i'm back with my sketchy kripto
19:43:17JoolsX:JoolsX has left #bitcoin-wizards
20:32:40lmatteis:is anybody here an academic?
20:33:24lmatteis:i'm an undergrad and looking for a research career in p2p systems. wondering which path to follow exactly (which classes to take, which labs to look into, etc)
20:37:33kanzure:uh well, here's some stuff you can read http://diyhpl.us/~bryan/papers2/incentives/
20:49:54andytoshi:lmatteis: i'm an acadamic, you can PM me for school advice, but note that i have a math degree and have nothing positive to say about CS departments
20:51:26lmatteis:andytoshi: and you do research in bitcoin?
20:52:36andytoshi:lmatteis: nope :)
20:52:41andytoshi:not in school anyway
20:54:23lmatteis:but you do distribtued stuff?
20:58:17andytoshi:still nope
21:04:45lmatteis:andytoshi: so what do you do :)
21:06:16andytoshi:lmatteis: in the bitcoin space i care about privacy tech; in acadima i do real analysis/information theory
21:07:01lmatteis:i have a hard understanding how to do research in information theory
21:07:07lmatteis:like, most research i read is experimental
21:07:22lmatteis:also, would you say the bitcoin paper is theory?
21:07:29pigeons:andytoshi wrote a coinjoin implementation, a treatice on altcoins, a bitcoin library in rust, an experimental client using the rust library, and a reference on asics for pow
21:08:11kanzure:lmatteis: also see http://diyhpl.us/~bryan/papers2/bitcoin/
21:08:11andytoshi:lol pigeons, thx, im in rust mode right now and honestly couldn't think of what it is i do around here..
21:08:26andytoshi:lmatteis: i've never heard of experimental information theory
21:08:35kanzure:"mostly i just read rust docs"
21:08:38lmatteis:no just experimental research
21:09:41andytoshi:oh, i see. lmatteis click around http://arxiv.org/list/math.IT/recent (don't actulaly try to read it, you'll never get out) to get a feel for it
21:10:13lmatteis:what i meant to say is that i have problems understanding how to perform theoretical research
21:10:29lmatteis:with experimental, you run experiments, observe what happens and report it
21:10:45lmatteis:with theory?
21:10:56andytoshi:oh, all that's cs.IT rather than ma.IT, http://arxiv.org/list/math.FA/recent is a better example of things that are actually theoretical
21:12:07andytoshi:lmatteis: you discover things that are true, then write down what they are and why they're true
21:12:33lmatteis:andytoshi: how do you discover that it's true without experiments?
21:12:47andytoshi:lmatteis: with deductive reasoning
21:12:51kanzure:greg egan claims that even someone just thinking about a problem is an experiment, so you're safe.
21:12:52Apocalyptic:lmatteis, because you get a mathematical proof
21:13:01andytoshi:how could you discover something is true using experiments?
21:13:20kanzure:yeah, aren't experiments only for falsifying possibilities?
21:13:22lmatteis:nothing is true for certain. with experiments you quantify how good it can be
21:13:28kanzure:or s/falsifying/ruling out
21:13:46sipa:experiments are for giving your theory a chance to fail
21:14:06sipa:if they fail to fail, you gain confidence in them
21:14:11kanzure:yes falsifying is definitely the wrong word
21:14:23nsh:* nsh demurs
21:14:29sipa:but that doesn't make them more or less true, just more plausible
21:14:57lmatteis:andytoshi: what do you mean by discovering that something is true? can you give me an example?
21:15:05nsh:experiments are for collecting data. models afford for collected data to differentiate between hypotheses when there is a good coupling between the modelling and the experimental apparatus and reality
21:15:25nsh:but data is always collected
21:17:18andytoshi:lmatteis: sure, "if a schnorr signature is defined as (s, e) for s = k - xe and e = H(m, kG) for x,k,m,G defined in some way for `=` and `-` to be true in a field and H defined as some function, then given (s', e') calculated with the same x and k, x is uniquely determined by (s, s', e, e') and is efficiently computable"
21:17:36andytoshi:sorry, that was simpler in my head..
21:18:09andytoshi:lmatteis: "if a prime number is defined as one for which no positive integers divide it except for 1 and itself, then the natural numbers contain infinitely many primes"
21:18:32andytoshi:also put a bunch of stuff about the natural numbers being a unique factorization demain into the "if"
21:19:02lmatteis:hrm ok
21:19:24nsh:how much stuff about unique factorization can there be to put in?
21:19:34andytoshi:nsh: actually the peano axioms will suffice
21:19:50andytoshi:to imply everything you need, as well as that the natural numbers are themselves infinite
21:20:02andytoshi:and well-ordered
21:21:06nsh:well-ordering only holds as long as you consider only consider finite sets of the natural numbers
21:21:14lmatteis:would you say the bitcoin paper is mostly theoretical?
21:21:24nsh:(or does it)
21:21:46Apocalyptic:nsh, no it doesn't
21:21:52andytoshi:nsh: nope, well-ordering only requires you have a least number; it applies to all subsets of the naturals
21:21:59Apocalyptic:any non empty subset of natural numbers will have a least element
21:22:05nsh:i mean between set
21:22:08nsh:not within
21:22:16andytoshi:i think that is equivalent to the axiom of countable choice tho, it's not obvious
21:22:35andytoshi:lmatteis: well, the bitcoin whitepaper is not super precise ... but yes, it is "theoretical"
21:23:28andytoshi:lmatteis: http://arxiv.org/pdf/1006.1095v5.pdf is something i care about, it's long but the first few pages are easy to read and well-motivated (for a pure math paper)
21:23:58lmatteis:only 12 citations :)
21:26:10nsh:andytoshi, on another note, i was thinking earlier about parallel transport on manifolds and whether you could use it as a kind of scrambling system -- by encoding a message as a set of vectors and the key as a route along some reference manifold -- but i think you'd need the manifold to be nondifferentiable or otherwise freaky to have interesting properties
21:26:27nsh:any thoughts?
21:27:38andytoshi:nsh: "a message" is gonna live in a discrete space, you need a way to embed it and losslessly extract it, or to think about manifolds that are discretized in some way, in either case thinking about the "security" of it seems really hard
21:27:52nsh:* nsh nods
21:31:24lmatteis:wow this is kinda cool http://techcrunch.com/2015/04/10/bittorrent-maelstrom-beta/
21:34:16nsh:Security Applications of Formal Language Theory -- http://www.cs.dartmouth.edu/~sergey/langsec/papers/sassaman-jsys7-3.pdf
21:40:16nsh:(what i'm wondering is if something akin to a generalized pumping lemma can be used to put bounds on the potential for weird machines in programs and protocols accepting complex input)
21:42:42andytoshi:this looks really cool nsh, will read
21:42:54nsh:* nsh smiles
23:41:32JoolsX:JoolsX has left #bitcoin-wizards
23:53:57Guest35538:Guest35538 is now known as maaku