--- Log opened Mon Nov 17 00:00:53 2014
00:24 -!- gene_hacker [~chatzilla@c-50-137-46-240.hsd1.or.comcast.net] has quit [Quit: ChatZilla 0.9.91 [Firefox 33.1/20141106120505]]
00:25 -!- ThomasEgi [~thomas@2a02:810b:33f:dc18:55d4:46dd:e1a3:3dcf] has joined ##hplusroadmap
00:25 -!- ThomasEgi [~thomas@2a02:810b:33f:dc18:55d4:46dd:e1a3:3dcf] has quit [Changing host]
00:25 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has joined ##hplusroadmap
00:38 -!- delinquentme [~dingo@74.61.157.78] has quit [Ping timeout: 264 seconds]
01:02 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-vzpcvxwvwpvtvmln] has joined ##hplusroadmap
01:55 -!- pete4242 [~smuxi@boole.london.hackspace.org.uk] has quit [Remote host closed the connection]
03:16 -!- Viper168_ [~Viper@unaffiliated/viper168] has joined ##hplusroadmap
03:17 -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 245 seconds]
03:23 -!- Viper168_ is now known as Viper168
03:44 -!- eudoxia [~eudoxia@r190-134-40-20.dialup.adsl.anteldata.net.uy] has joined ##hplusroadmap
04:05 -!- ebowden [~ebowden@147.69.147.27] has quit [Remote host closed the connection]
04:09 -!- ebowden [~ebowden@147.69.147.27] has joined ##hplusroadmap
04:17 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap
04:31 -!- ebowden [~ebowden@147.69.147.27] has quit [Remote host closed the connection]
05:02 -!- qu-bit [~shroedngr@unaffiliated/barriers] has quit [Read error: Connection reset by peer]
05:15 < fenn> flyby anomaly is probably due to an unaccounted for electric charge interacting with the earth's magnetosphere
05:17 < fenn> sheena: JDM means "japanese domestic market" and usually refers to engines which were not sold in the rest of the world for marketing and logistics reasons; usually the engines people import are somehow larger or more powerful than the north american versions
05:23 -!- Zinglon [~Zinglon@D549A77D.cm-10-1a.dynamic.ziggo.nl] has joined ##hplusroadmap
05:39 -!- Merovoth [~Merovoth@gateway/tor-sasl/merovoth] has quit []
05:49 < fenn> Zinglon: welcome, welcome welcome
05:53 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Ping timeout: 240 seconds]
05:55 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap
05:58 -!- snuffeluffegus [~snuff@5.150.254.180] has joined ##hplusroadmap
06:08 < Zinglon> Hello!
06:08 < Zinglon> Sorry was afk a bit
06:10 < fenn> what brings you to this corner of the interwebs
06:17 < Zinglon> Randomly found this channel while browsing
06:18 < Zinglon> A friend of mine was talking about biohacking so i decided to look it up
06:19 < fenn> did you see http://waag.org/en/project/biohack-academy-1-biofactory
06:21 < Zinglon> I'll check it out
06:21 < Zinglon> Thank you!
06:24 < kanzure> die mensen zijn erg vermakelijk
06:26 < fenn> oof it's 850 euro
06:26 < kanzure> there should be an hplusroadmap version of that eventually
06:26 < kanzure> classes should include "defensive irc"
06:26 < fenn> how do build a computer out of a dremel and a potato
06:27 < kanzure> how to inject dna into bacteria by yelling
06:27 < kanzure> marksmanship and dna transformation
06:28 < kanzure> (that one can be jrayhawk)
06:29 < kanzure> so, no to the dark matter halo, eh?
06:29 < archels> where did that Dutch suddenly come from o_O
06:30 < archels> or I should say Belgian because no Dutch person says "vermakelijk"
06:30 < archels> unless they're being sarcastic
06:30 < yorick> I have the feeling half of us are dutch or something
06:31 < kanzure> weet je over "transhumanismus" op EFnet?
06:31 < yorick> nee maar dat klinkt duits
06:31 < kanzure> "transhumani"
06:31 < yorick> what?
06:31 < kanzure> een taal is als ruwe tong gebeten door boze pad
06:32 < kanzure> i don't know man, i didn't take fucking classes
06:32 < yorick> kanzure: we're banning you from speaking dutch ever again
06:32 < archels> chuckle
06:32 < Qfwfq> le mi varkiclaflo'i cu culno lo angila
06:32 < kanzure> yorick: understood
06:32 < yorick> Qfwfq: is this lojban
06:33 < yorick> (no, but it should be!)
06:33 < fenn> "a language is as rough tongue bitten by evil path"??
06:33 < yorick> fenn: I think it's a toad
06:33 < yorick> Qfwfq: oh it is lojban!
06:33 < yorick> Qfwfq: <3
06:34 < fenn> Germans: Oh you’re learning German? Hey, you’re not so bad at it. Don’t fuck it up though. 
06:34 < fenn> French: About time you learned French. 
06:34 < fenn> Dutch: but why would you do this
06:34 < kanzure> yorick: have you met archels
06:34 < yorick> kanzure: no
06:34 < kanzure> well why not
06:34 < yorick> reasons!
06:35 < Qfwfq> Specifically, it's a Lojban translation of "May I have a box of matches?".
06:35 < eudoxia> Spanish: it's my native language, but what's your excuse?
06:35 < archels> yorick: aren't you near Nijmegen, actually?
06:35 < yorick> archels: yeah, I have probably met people who met you
06:35 < ThomasEgi> fenn, .. fins...
06:35 < yorick> Qfwfq: liar
06:35 < archels> haha
06:35 < archels> RU?
06:35 < yorick> archels: ja
06:36 < archels> what programme?
06:36 < yorick> computer science
06:36 < archels> they still have that? I thought it was all Information Science these days
06:37 < yorick> archels: whatever, they call it "informatica". they got rid of informatiekunde recently
06:37 < fenn> i get the impression informatics is more about libraries and data handling, whereas computer science is applied math
06:38 < yorick> fenn: the programme is called "computer science" in english
06:38 < archels> huh. just a few years ago they were talking about cutting CS (because it was so small and not performing well) and putting all their eggs in Information Science
06:39 < yorick> archels: their CS programme was voted the best in NL last year
06:39 < fenn> who needs science when you have engineering! pah
06:39 < yorick> fenn: yeah I studied a 'computer engineering' but it was secretly information sciences
06:39 < archels> yeah by the students partaking in it. I don't put much weight on that
06:39 < yorick> archels: true, they are idiots.
06:40 < archels> :)
06:40 < yorick> archels: but no other place is any better :D
06:41 < archels> not so sure about that one
06:41 < archels> but as my alma mater is still employing me, I should probably shut up at this point
06:41 -!- kragenjaviersita is now known as kragen
06:41 < yorick> archels: I checked utrecht, amsterdam(x2) and delft
06:41 < yorick> it's possible eindhoven is better
06:42 < archels> yeah, I did my first Master's in Eindhoven
06:42 < archels> in the end it comes down to what you want to do with your degree
06:43 < archels> if you want to do Information Sciencey stuff, then the RU is probably a good place to be
06:43 < yorick> I want to get this degree over with asap, I suppose
06:44 < archels> haha
06:44 < archels> where to next?
06:45 < fenn> now we try and take over the WORLD
06:45 < yorick> no idea!
06:45 < yorick> archels: oh hey you're at donders. I was there last week for an MRI thing
06:46 < archels> yes
06:46 < archels> I don't work with wet stuff though (people, animals or otherwise)
06:46 < archels> my office is in the Huygens as well
06:47 < yorick> oh right you're behind the scary door that goes "NO COFFEE FOR STUDENTS"
06:48 < archels> haha yeah
06:48 < archels> not sure whose great idea it was to put that up there
06:48 < archels> (the coffee from that machine takes like gunk, anyway)
06:48 < yorick> (the fun part is if you walk 30 meters to mercator there is great free coffee for students)
06:50 < archels> Mercator has free coffee? interesting
06:51 < archels> were you at the Moenenspace meeting, by the way?
06:51 < yorick> technically only for ICIS people
06:51 < yorick> I was not
06:52  * yorick wasn't studying there back then
06:52 -!- Vutral [ss@mirbsd/special/Vutral] has quit [Ping timeout: 245 seconds]
06:53 < fenn> oh damn i am mixing you up with "yorik"
06:53 < archels> ah right
06:53 < yorick> fenn: there is a yorik?
06:53 < archels> what year are you in, then?
06:54 < yorick> archels: ...first :/
06:54 < fenn> yorik van havre is a freecad dev
06:54 < yorick> fenn: yorick is also a LoL champion and a shakespeare character and a programming language
06:55 < fenn> yes but i dont talk to them on irc
06:56 -!- superobserver [~superobse@gateway/tor-sasl/yorrick] has joined ##hplusroadmap
06:56 -!- superobserver is now known as Guest46018
06:57 -!- Guest46018 [~superobse@gateway/tor-sasl/yorrick] has quit [Remote host closed the connection]
06:57 < yorick> oh yeah we were just in need of a 'yorrick'
06:57 < fenn> heh
06:58 < fenn> yo dawg you just got yorickrolled
07:00 < fenn> what the heck is going on in this picture? http://waag.org/sites/waag/files/public/styles/detailpage/public/Projects/picnic12_eye.jpg
07:00 < yorick> fenn: weird lamps
07:00 < archels> biannual M.C. Escher convention
07:01 < fenn> it looks like they're being bombed by space invaders
07:01 < yorick> also there is a stair with a table and a slope in the middle of the thing
07:18 < kanzure> my talent agent is in the new yorker http://www.newyorker.com/magazine/2014/11/24/programmers-price
07:19 < kanzure> comments https://news.ycombinator.com/item?id=8616952
07:24 < fenn> "lying on the beach gasping because they can’t get enough talented people in for these jobs." uh huh sure they are
07:26 < kragen> saw an item recently by a recruiter about why agents for programmers don't work out
07:26 < kragen> he'd tried to do it
07:26 < fenn> of course a recruiter would say that
07:27 < kragen> well, he'd apparently spent a couple of years failing at it
07:27 < fenn> they're the enemy! competition!
07:27 < kragen> at working for programmers rather than companies, I mean
07:28 < kragen> he said basically the problem is that it's easy for programmers to find jobs, and it's hard for jobs to find programmers
07:28 < kragen> which is the reverse of the situation in fields like actors and novelists where agents are a thing
07:30 < fenn> an agent's job is to get you the best terms possible by doing all the social legwork bullshit
07:30 < kragen> yes
07:30 < fenn> it doesn't matter what the capital/labor ratio is
07:31 < kragen> how much do you think you're losing by doing it yourself?
07:31 < fenn> infinity
07:31 < kragen> your income would be infinite if you only had an agent?
07:31 < fenn> i mean, i can't get a job because i suck at doing the social legwork bullshit
07:31 < kragen> oh
07:31 < fenn> so $salary/0
07:31 < kragen> so, your entire income
07:32 < kragen> in the range of US$40k to US$200k/year
07:32 < kragen> in fields where agents are a thing, the *vast majority* of people in the field are in that situation
07:32 < kragen> aspiring actors who can't get an auditiion
07:33 < kragen> insurance salesmen with a drawer full of rejection slips for their novel
07:34 < fenn> all these people trying to write novels should just be programmers instead
07:34 < kragen> because there are so many of them — and more importantly, so many really good ones — agents are able to make a living in those fields
07:34 < kanzure> fenn, i don't think a talent agent would make you able to tolerate having a ob
07:34 < kanzure> *job
07:35 < fenn> a good agent would listen to me and find a tolerable job
07:35 < kanzure> your tolerable job doesn't exist as far as i know
07:35 < fenn> wah
07:35 < fenn> wah!
07:36 < fenn> this is why i think "fuck the world" is a reasonable answer
07:36 < kragen> it may also be why you don't actually have an agent
07:36 < kanzure> he doesn't have an agent because i haven't introduced him to 10x yet
07:36 < fenn> agents for software people didn't exist when i was looking for them
07:37 < fenn> i was friends with some guy who eventually became an agent, but somehow it never came up in conversation
07:37 < kanzure> when i think back to the types of gigs i've been exposed to (including the ones i've said no to) from 10x, i really just can't see fenn doing any of those
07:37 < kanzure> he is technically skilled enough to do them, but i strongly doubt he would want to for only $30k/mo
07:38 < kanzure> especially given the sorts of demands on attention or context switching that are often made
07:38 < kragen> that's a pretty high billing rate
07:38 < kanzure> well i'm very good at what i do
07:38 < fenn> altay guvench apparently works at 10x
07:38 < kanzure> altay owns 10x
07:39 < fenn> right, see
07:39 < kanzure> you know him too?
07:39 < fenn> yeah he was at langton a lot
07:39 < kanzure> oh, i'll hook you up then
07:39 < kragen> and apparently what you do is valuable
07:40 < Qfwfq> langton?
07:41 < eudoxia> i'm doomed to always confuse linden labs with langton labs
07:41 < fenn> langton is the fuzzy center of the neuroscience startup burningman bay area cocktail party nexus
07:43 < Qfwfq> lol
07:43 < Qfwfq> i mean, that's a mouthful, but it sounds like a party
07:45 < fenn> kragen: i'm not sure if what i do is really valuable because there are so many ridiculously talented and hardworking people out there with pages and pages of stuff they've done
07:45 < fenn> i realize this is a logic error but it still seems that way
07:46 < fenn> when actually looking for opportunities i get a lot of interest but no returned calls
07:47 < fenn> i mean, smarmy business people seem interested in me (the way a dog looks at a hamburger) but they don't actually follow through
07:48 < kanzure> do you really want to be mucking around in css for 14 hours a day
07:48 < kanzure> i'm afraid that would kill you
07:48 < eudoxia> css is not so bad once you lower your expectations
07:48 < fenn> i actually don't mind css
07:49 < fenn> it's much easier now with all the inspector tools like firebug
07:51 < fenn> but no, spending my life writing a web page doesn't seem filling
07:51 < fenn> fulfilling*
07:51 < kanzure> yes but would it kill you
07:51 < kanzure> hmm this article mentions langton
07:51 < kanzure> or, er, phage
07:51 < fenn> same thing
07:53 -!- eudoxia [~eudoxia@r190-134-40-20.dialup.adsl.anteldata.net.uy] has quit [Quit: Lost terminal]
07:56 -!- delinquentme [~dingo@74.61.157.78] has joined ##hplusroadmap
07:56 < kanzure> does he know you as fenn?
07:59 < fenn> i dont know
08:00 < fenn> keep in mind this was three years ago
08:00 < fenn> or four years ago
08:10 -!- Merovoth [~Merovoth@gateway/tor-sasl/merovoth] has joined ##hplusroadmap
08:11 -!- Boscop_ [~me@e102.stw.stud.uni-saarland.de] has joined ##hplusroadmap
08:15 -!- Boscop [~me@unaffiliated/boscop] has quit [Ping timeout: 255 seconds]
08:27 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-vzpcvxwvwpvtvmln] has quit [Quit: Connection closed for inactivity]
08:39 < fenn> this "vetted for interpersonal skills" sounds like anti-autism discrimination
08:39 < fenn> "we don't want a coder who drools"
08:39 < fenn> fuck you, you smell like cologne
08:42 < fenn> "It turns out that negotiating is a lot easier when you're doing it for someone else."
08:43 < fenn> i wonder why that is
08:46 < Qfwfq> Intuitively: you'd worry less about appearing greedy in asking for things? Lots of people are bad at negotiation for that reason, including me.
08:47 < kanzure> no it's not because of feeling greedy
08:47 < kanzure> it's things like "easier to walk away from a bad deal"
08:48 < archels> merit-based negotiation always turns out better
08:48 < archels> it's easier to negotiate purely based on merits when you're not personally/emotionally involved
08:50 < kanzure> hmm https://github.com/kanzure/pdfparanoia/pull/43
08:51 < fenn> yeah i think the stress response shutting down half your brain has something to do with it
08:55 < Qfwfq> pdfminer3k hasn't been updated since my last patch
08:55 < kanzure> what about the other one
08:55 < Qfwfq> march update to pdfminer removed initialize() http://euske.github.io/pdfminer/index.html#changes
08:55 < kanzure> so they are inconsistent?
08:55 < Qfwfq> yeah
08:55 < kanzure> well that sucks
08:56 < kanzure> i don't know what to do about that
08:56 < Qfwfq> i say stop supporting 3, rather than litter the source code with conditionals
08:56 < kanzure> why are there two separate versions of pdfminer
08:56 < kanzure> i think abandoning python3 support is bad
08:56 < kanzure> python2 should be the one to suffer, if anything
08:57 < Qfwfq> given it's being used as a library in paperbot.. yeah
08:58 < Qfwfq> it's just that pdfminer for 3 has been out of maintenance for 2 years
08:58 < Qfwfq> and calling the script with a different python runtime is always an option
08:59 < kanzure> is there an alternative to pdfminer that works
08:59 < Qfwfq> i don't know the python ecosystem
09:00 < fenn> "Paul Cretu, and he and his partner were working on transcription software that records everything you say, leaving you with a searchable record of your thoughts and conversations"  sign me up
09:00 < kanzure> i am a little busy at the moment, it would be helpful if you or someone else could figure that out and just do the right thing
09:00 < kanzure> and then either remove python2 support or python3 support from pdfparanoia
09:00 < kanzure> and then remove pdfminer3k and possibly also pdfminer
09:01 < Qfwfq> kanzure: https://pypi.python.org/pypi/pdfminer.six/20140915
09:02 < Qfwfq> that version works with 2 and 3
09:02 < kanzure> sure, throw that in
09:02 < Qfwfq> k, i'll try and make time for it
09:03 < kanzure> https://github.com/kanzure/pdfparanoia/issues/44
09:05 < Qfwfq> feel free to assign me, though i'm kinda busy for the next fortnight
09:11 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-butqbhoardxhwipg] has joined ##hplusroadmap
09:13 < fenn> i would not have pegged Mark Mian as "a branding and marketing specialist"
09:14 < fenn> i guess he does have sort of a guru vibe
09:16  * fenn highfives the ghost of wittgenstein
09:18 < fenn> korzybski too
09:22 < fenn> http://en.wikipedia.org/wiki/File:Duck-Rabbit_illusion.jpg
09:28 < delinquentme> nmz787, just sent you stuff
09:28 < yoleaux> 05:18Z <nmz787> 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
09:35 < fenn> memoserv people
09:36 -!- Boscop__ [~me@e102.stw.stud.uni-saarland.de] has joined ##hplusroadmap
09:39 -!- Boscop_ [~me@e102.stw.stud.uni-saarland.de] has quit [Ping timeout: 250 seconds]
09:44 -!- nmz787_i [~nmccorkx@134.134.139.74] has joined ##hplusroadmap
09:45 < nmz787_i> delinquentme: how do I open it? (at work now and have class this eve... prob won't have time to open it till around 9pm)
09:45 < nmz787_i> ah, solidworks
09:45 < nmz787_i> ok
09:45 < nmz787_i> I think I can open that
09:46 < nmz787_i> delinquentme: you might also try exporting a STEP or IGES file and send that over
09:48 < nmz787_i> downloading this now http://www.solidworks.com/sw/support/edrawings/e2_download.htm
09:50 -!- nmz787_i [~nmccorkx@134.134.139.74] has quit [Remote host closed the connection]
10:01 -!- nmz787_i [~nmccorkx@134.134.139.76] has joined ##hplusroadmap
10:10 -!- delinquentme [~dingo@74.61.157.78] has quit [Remote host closed the connection]
10:21 -!- crescend1 is now known as crescendo
10:24 < nmz787_i> paperbot: http://www.nature.com/srep/2014/140417/srep04717/pdf/srep04717.pdf
10:27 < nmz787_i> paperbot: http://www.nature.com/srep/2014/140417/srep04717/extref/srep04717-s1.pdf
10:27 < nmz787_i> fenn : FYI for coding (those are both open articles and thus the pdfs should be accessible to anyone)
10:27 -!- snuffeluffegus [~snuff@5.150.254.180] has quit [Quit: Leaving]
10:33 < fenn> paperbot: dude
10:34 < fenn> works w/2.0
10:38 < kanzure> confirmed works?
10:47 < fenn> /tmp/ffe79d50b58cde38644e6b31db6587d1.pdf
10:48 < fenn> no doi extractor so no metadata
10:48 < kanzure> i'm a little uneasy about replacing paperbot with paperbot2 because of the no zotero integration
10:48 < kanzure> means losing all the journal parsers
10:50 < fenn> actually there's no way to extract a doi from the pdf link without downloading the pdf and converting it to text etc
10:51 < fenn> not sure how zotero would handle that any better
10:51 < kanzure> most url structures for pdf links to publisher sites can be converted to non-pdf links automatically by inspecting the url and knowing the publisher and knowing the publisher's url scheme
10:53 < fenn> is that what zotero does?
10:54 < kanzure> nope
10:55 < fenn> does this code exist anywhere?
10:56 < fenn> publisher pdf-url <-> html-url mapping
10:56 < kanzure> nope
10:56 < fenn> ok :)
10:56 < kanzure> my comment about uneasiness was regarding losing the journal parsers
10:57 < fenn> do you think there are enough papers without DOIs to worry about not having a DOI?
10:57 < kanzure> i don't see the relevance really? store doi when it's known, otherwise don't care too hard
10:58 < kanzure> the stuff in model.py is probably a little wrong
10:58 < fenn> because you can use the DOI to get metadata in a uniform format, thus you can get rid of zillions of journal-specific parsers
10:58 < kanzure> a single paper doesn't have just one doi; there's a doi for the paper but also the issue/volume of the jorunal
10:58 < kanzure> and then a doi for each supplementary document or something
10:58 < kanzure> so it should be a list or dict of dois, i guess
10:59 < kanzure> i don't want to force paperbot users to type in dois
10:59 < fenn> that's not what i'm saying
10:59 < fenn> paperbot looks at the link, finds a DOI somewhere in the content of the link, then looks up the doi on doi.org to get the metadata
10:59 < kanzure> i don't want to force paperbot to only extract metadata when the publisher submitted a paper's metadata to doi's system
11:01 < fenn> it's an easy 80% solution
11:02 < kanzure> then do it
11:06 < fenn> maybe i'm wrong about how DOI works; when i search for a DOI it just forwards me to the Nature page
11:06 < fenn> but somehow wikipedia has bots that populate DOI citations with metadata
11:09 < kanzure> doi is not omnipotent
11:10 < fenn> ew they want $40k for metadata lookup
11:10 < kanzure> haha
11:10 < fenn> http://www.crossref.org/04intermediaries/34affiliate_fees.html#CMS_2012_Fees
11:11 < nmz787_i> kanzure: why not have both v1 and v2 active here until we work out the transitional bugs?
11:12 < kanzure> right now neither of them work
11:12 < kanzure> so having two things that don't work wont be helpful
11:12 < nmz787_i> fenn said the last  links worked with v2
11:13 < fenn> it downloaded the pdf at least
11:13 < nmz787_i> their working status changed since his message?
11:13 < kanzure> do whatever you want, within the limits of 1) do the right thing, 2) don't be lazy, 3) pretend i'll kickban you if you make bad implementation choices
11:17 < fenn> oh my god i am so underwhelmed by the crossref.org website
11:17 < fenn> not to mention their business model is just as bad as the closed publishing model
11:17 < fenn> they charge publishers to get their data into the system, then they charge libraries to get the metadata back out
11:18 < fenn> s/data/metadata/
11:18 < fenn> why is it every time i think there's a reasonable metadata system it turns out to be some fucked up closed model
11:19 < kanzure> because librarians are terrorists
11:19 < fenn> librarians are pussies
11:19 -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 264 seconds]
11:20 < fenn> it's time for the terrorists to step up their game
11:20 < kanzure> terrorize american by hijacking the library of congress?
11:20 < fenn> yes..
11:20 < fenn> by making its metadata available to all...
11:21 < fenn> MUWAHAHAHAHA
11:21 < fenn> "You can search for papers in Pubget just like you would in PubMed, yet instead of the search results just linking to papers, with Pubget the search results *are* the papers." hrm
11:22 < kanzure> "sometimes"
11:22 < kanzure> that doesn't always work
11:24 < fenn> it seems to convert DOI to metadata at least
11:25 < fenn> i need more DOI test cases
11:26 < kanzure> grep hplusroadmap logs for jstor.org, their urls are doi numbers
11:26 < fenn> but that's just stuff in jstor
11:26 < fenn> i don't know what range of journals pubget covers
11:26 < fenn> is it just pubmed?
11:27 < Zinglon> http://www.medicalnewstoday.com/articles/257898.php
11:27 < Zinglon> hmm
11:27 -!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap
11:27 -!- Vutral [ss@mirbsd/special/Vutral] has joined ##hplusroadmap
11:27 < fenn> "Pubget covers over 20 million research papers" but that's only 20% of the 100 million DOIs
11:28 < fenn> is wireless blood sensors a new thing?
11:29 < fenn> so much of this medical device startup technology just gets buried under the noise of history
11:30 < fenn> .title
11:30 < yoleaux> "Tiny Lab" Implanted Under Skin Transmits Blood Marker Levels - Medical News Today
11:30 < kanzure> subdermal sensor plus inductive powering is sorta novel, i mean it's not groundbreaking but not trivial to setup either
11:31 < fenn> it seems trivial from my perspective
11:31 < kanzure> depends on the sensor and what sort of math they are doing before transmission ove what protocol
11:31 < kanzure> *over
11:31 < fenn> i'm curious about the "nano-sized sensors" though
11:32 < fenn> integrated plasmonics was working on an array of such sensors
11:33 < fenn> if it's just an antibody bound to a semiconductor, i'd expect it to get covered with all sorts of junk by the immune system
11:34 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Ping timeout: 264 seconds]
11:34 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap
11:34 < fenn> you might be able to get around that by encapsulating the whole thing in dialysis tubing or some kind of barrier that will let macromolecules through but not cells
11:36 -!- justanotheruser [~Justan@unaffiliated/justanotheruser] has quit [Ping timeout: 272 seconds]
11:44 -!- Vutral [ss@mirbsd/special/Vutral] has quit [Ping timeout: 244 seconds]
11:44 < nmz787_i> there was a wireless sensor i read about a few days ago that said it was small enough to get between cells
11:45 < nmz787_i> it seemed like basically a small IC chip die that had sensors and wireless onboard
11:48 < fenn> randall koene's thing? neualinkco?
11:48 < jrayhawk> fenn, sheena: due to smog, the JDM is warped by an extremely onerous inspection schedule that makes it generally untenable to own a car more than ten years, so engine longevity is not a big priority.
11:49 < fenn> why do hondas run for hundreds of thousands of miles then
11:50 < jrayhawk> Because the USDM cares deeply. The JDM is a testbed.
11:51 < nmz787_i> http://www.urbandictionary.com/define.php?term=JDM
11:55 < jrayhawk> It's a very nice product development arrangement. You get to beta test everything in a smaller market that doesn't care, then you get to crush the larger market with a reputation for quality.
11:56 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Quit: Ex-Chat]
11:57 -!- Vutral [ss@mirbsd/special/Vutral] has joined ##hplusroadmap
11:59 -!- justanotheruser [~Justan@unaffiliated/justanotheruser] has joined ##hplusroadmap
12:18 < fenn> you'd think american car companies would have figured this out and started selling JDM models too
12:20 < jrayhawk> american car companies don't really understand the JDM market very well and have a lot of failed products over there
12:21 < jrayhawk> they did at one point do a marketing exchange; an USDM american-badged corolla for a JDM japanese-badged cavelier
12:25 < Zinglon> I have to be going now, bye guys!
12:25 < nmz787_i> I can't see a japanese person driving an old cavalier (80s version)
12:26 < nmz787_i> the carburetors on them were ridiculous hybrid beasts (horrible)
12:26 < nmz787_i> some smashup of carbs and EFI
12:27 < Zinglon> *get going
12:27 < fenn> bye
12:27 -!- Zinglon [~Zinglon@D549A77D.cm-10-1a.dynamic.ziggo.nl] has quit [Quit:  I love my HydraIRC -> http://www.hydrairc.com <-]
12:29 < fenn> what's the word for something like a unit test, but for scraping websites?
12:30 < fenn> to test if the scraper still works for a single website
12:30 < kanzure> against the real website?
12:30 < fenn> yeah
12:30 < kanzure> "integration test"
12:30 < fenn> hrmph
12:34 < kanzure> going out to the live web is usually considered bad though
12:34 < kanzure> so usually you download it once, commit the html file, and just parse that one forever into the future
12:34 < kanzure> and then when they change, you fix your test
12:35 -!- ebowden [~ebowden@147.69.147.27] has joined ##hplusroadmap
12:35 < Qfwfq> That's what I've always done, but now I'm curious how people handle notification of changed structure.
12:36 < nmz787_i> you should probably note the date at which the HTML was retrieved, and display that somewhere when the tests are run
12:36 < kanzure> notifications are usually people enraged at you over irc
12:36 < Qfwfq> The content might change without you needing to revise the selector, so you want a larger diff on the DOM structure
12:36 < kanzure> "fuck you kanzure" etc
12:36 < Qfwfq> lol
12:37 < Qfwfq> "What if nobody uses your stuff?" "Why are you maintaining it, then?"
12:39 < Qfwfq> "Past some popularity threshold, integration tests may be replaced by an IRC handle."
12:39 -!- maaku [~quassel@50-0-37-37.dsl.static.fusionbroadband.com] has joined ##hplusroadmap
12:40 -!- maaku is now known as Guest49700
12:42 -!- Guest49700 is now known as maaku
12:46 -!- ebowden [~ebowden@147.69.147.27] has quit [Remote host closed the connection]
12:56 -!- qu-bit [~shroedngr@unaffiliated/barriers] has joined ##hplusroadmap
13:09 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has quit [Ping timeout: 265 seconds]
13:14 -!- nmz787_i [~nmccorkx@134.134.139.76] has quit [Quit: Leaving.]
13:25 -!- nmz787_i [~nmccorkx@192.55.54.40] has joined ##hplusroadmap
13:27 < fenn> should i install pdfminer from ubuntu (apt-get install python-pdfminer) or from pip-python (pip install <something>)?
13:29 < fenn> apt-get version seems to work
13:33 -!- Vutral [ss@mirbsd/special/Vutral] has quit [Excess Flood]
13:35 -!- TheShadowFog [~TheShadow@2601:8:3e80:c:7d71:2e6c:7049:6035] has joined ##hplusroadmap
13:38 -!- Vutral [ss@mirbsd/special/Vutral] has joined ##hplusroadmap
13:45 < fenn> woah this is a nutty piece of software
13:47 -!- Vutral [ss@mirbsd/special/Vutral] has quit [Excess Flood]
13:51 < Viper168> squirrels been in it?
13:54 < TheShadowFog> fenn, what software?
13:55 < fenn> pdfminer
13:55 < TheShadowFog> ooh
13:55 < TheShadowFog> yeah i've used that before
13:55 < fenn> apparently you have to do all these steps just to get text out http://www.garysieling.com/blog/scraping-pdf-text-with-python
13:55 < TheShadowFog> oh wait that was a different program
13:56 < TheShadowFog> yeah sorry that looks annoying damn son
14:05 < kanzure> fenn: pdfminer from pypi because that's the one other pdfparanoia users are installing by default
14:07 -!- delinquentme [~dingo@74.61.157.78] has joined ##hplusroadmap
14:33 -!- Qfwfq [~WashIrvin@unaffiliated/washirving] has quit [Ping timeout: 255 seconds]
14:34 -!- snuffeluffegus [~snuff@5.150.254.180] has joined ##hplusroadmap
14:35 -!- Vutral [jxy71QysoJ@mirbsd/special/Vutral] has joined ##hplusroadmap
14:38 -!- Qfwfq [~WashIrvin@cm58.kappa38.maxonline.com.sg] has joined ##hplusroadmap
14:38 -!- Qfwfq [~WashIrvin@cm58.kappa38.maxonline.com.sg] has quit [Changing host]
14:38 -!- Qfwfq [~WashIrvin@unaffiliated/washirving] has joined ##hplusroadmap
14:41 < kanzure> ouch https://github.com/bitcoin-abe/bitcoin-abe/blob/master/Abe/DataStore.py#L2530
14:41 < kanzure> err i mean, https://github.com/bitcoin-abe/bitcoin-abe/blob/f3a018120795e2c3524b528116fc7b0f18ebbe79/Abe/DataStore.py#L2530
14:42 < fenn> pdftotext is about 30 times faster and they are nearly identical except for whitespace; however pdfminer managed to translate a figure caption and pdftotext did not
14:43 < kanzure> i would prefer a pure-python implementation (even if it is slower) so that i don't have to use non-python dependencies in that library/tool
14:44 < fenn> i hope there is a better way to do this than just converting the whole document ; i don't really care about formatting since i am just looking for the DOI
14:45 < fenn> anyway it's good to know that pdfminer works as expected (once you figure out the correct sequence of hoops to jump through)
14:45 < kanzure> unfortunately you can't really assume every pdf is using "deflate" for each text section..
14:48 < fenn> you mean i can't just zcat foo.pdf | grep ??? :P
14:48 < kanzure> "it's a 5% solution, take it"
14:48 < kanzure> these partial solutions multiply together and then nothing works
14:50 -!- drewbot [~cinch@ec2-54-91-208-9.compute-1.amazonaws.com] has quit [Remote host closed the connection]
14:51 -!- drewbot [~cinch@ec2-54-90-152-247.compute-1.amazonaws.com] has joined ##hplusroadmap
14:52 < fenn> "nside a PDF document, text is in no particular order (unless it is importing for printing), most of the time the original text structure is lost (letters may not be grouped as words and words may not be grouped in sentences, and the order they are placed in the paper often is random)."
14:54 < fenn> hmm looking at pypdf it says "extracting document information (title, author, ...)"
15:00 < fenn> nevermind it's a lie
15:01 -!- qu-bit [~shroedngr@unaffiliated/barriers] has quit [Read error: Connection reset by peer]
15:01 < Qfwfq> pdftotext is decent if you want ascii with preserved layout
15:02 < Qfwfq> nevermind, you mention it in scrollback
15:18 < nmz787_i> fenn: here is some code I wrote using pdfminer and pypdfocr http://paste.pound-python.org/show/7rOIMaAUtg8uB0UoOIW5/
15:19 < nmz787_i> which ended up looking something like this later http://umap.fluv.io/en/map/untitled-map_2096#11/45.4960/-122.7818
15:20 < fenn> i can't see that but i can imagine
15:20 < fenn> is that real OCR or fake OCR
15:21 < fenn> hm it uses tesseract
15:23 < nmz787_i> in that case the PDFs were just images of scanned papers
15:24 -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 244 seconds]
15:38 -!- Qfwfq [~WashIrvin@unaffiliated/washirving] has quit [Ping timeout: 264 seconds]
15:42 < kragen> kanzure: I just tweeted https://twitter.com/kragen/status/534491597371879424 quoting you anonymously
15:42 < kragen> if you'd prefer that I delete the tweet or credit you, let me know
15:42 -!- Qfwfq [~WashIrvin@unaffiliated/washirving] has joined ##hplusroadmap
15:46 -!- Vutral [jxy71QysoJ@mirbsd/special/Vutral] has quit [Ping timeout: 256 seconds]
15:48 < kanzure> shrug, if anyone asks it's okay to say, i suppose, but otherwise don't bother
15:49 < kanzure> i think more accurately i would be worried about that sort of work draining the life force out of fenn
15:51 < kanzure> there's only so many technical compromises you can ask him to make before he dies or something
15:51 < kragen> heh
15:52 < kragen> I wonder if that's true of djb too
15:52 < kanzure> djwho?
15:54 -!- rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has joined ##hplusroadmap
15:56 < maaku> i'll work for $30k/mo ;)
15:56 < maaku> kanzure: cryptographer
15:56 < maaku> http://cr.yp.to/djb.html
16:01 -!- Vutral [ss@mirbsd/special/Vutral] has joined ##hplusroadmap
16:02 < fenn> i could probably survive a month of 9 to 5
16:06 < kanzure> maaku: you're booked up at blockstream
16:06 < kragen> kanzure: qmail convinced me that we don't know how to program yet
16:06 < kanzure> oh, bernstein
16:07 < maaku> true, but $30k/mo would be awfully temping ;)
16:07 < kanzure> when did your cliff/vesting schedule start?
16:07 < maaku> i know that's standard around wall st, but it's hard to believe anyone gets paid that much for writing code
16:07 < maaku> a few months into an 18 mo cliff
16:07 < kanzure> that's not hard at all, it's barely more than your lawyer geeze
16:08 < kanzure> or your plumber for that matter
16:08 < fenn> yeah plumbers get outrageous cash
16:08 < kanzure> s/hard/hard to believe
16:08 < heath> 360k/yr?
16:08 < heath> for a plumber...
16:08 < heath> i don't believe it
16:08 < maaku> if you assume 40hr work week.. i don't think my plumber gets 8 billable hours a day
16:08 < kanzure> have you guys been living under a rock
16:08 < fenn> for new construction, not the guy with the butt crack
16:08 < kanzure> yeah, maybe not 40 hours/week, i don't know how plumbers schedule themselves
16:09 < kragen> they don't bill for travel time
16:09 < kanzure> new construction types might?
16:09 < kragen> and there are plumbers who do indeed work 8hour days onsite
16:10 < kanzure> well, look, if you're working on applications and source code that make $XX million/year, and you're increasing revenue by entire basis points, there's a strong argument to be made for "pay me"
16:10 < kragen> but we're talking about my friend Dann who fixes Boston's sewage plant when it breaks, not the guy who to your house with a Roto-Rooter
16:10 < kragen> who drives to your house
16:11 < kragen> I'm pretty sure he doesn't bill US$30k/month ever though
16:12 < maaku> kanzure: right, but there's also a hundred other people who will happily step in and underbid you
16:13 < kanzure> maaku: there's a scarcity of good developers, so what do you think happens when you're a bitcoin developer too?
16:13 < maaku> well specializing certainly does help
16:13 < kanzure> that's like double unicorn
16:13 < kanzure> there's probably <100 people total on the market right now as a bitcoin developer available to be hired for anything in the entire space
16:13 < kanzure> and also, we probably know all of them
16:15 < maaku> https://news.ycombinator.com/item?id=8620201
16:16 -!- Vutral [ss@mirbsd/special/Vutral] has quit [Ping timeout: 244 seconds]
16:16 < kanzure> i haven't looked at their reputation stuff. is it sane?
16:16 < maaku> Maybe.
16:16 < kanzure> hah "Why do you include floating point operations in the consensus code?"
16:16 < maaku> There's no adversarial analysis as far as I can see.
16:16 < kanzure> "wtf is code review?"
16:17 < maaku> yeah well that *could* be fixed
16:17 < maaku> i'm more concerned that the basic model might not be safe
16:17 < maaku> but if it is, it's pretty amazing what can be done
16:17 < maaku> not as a prediction market, but as a way of sourcing external data for smart contracts
16:19 < kanzure> are there any working implementations of whuffie? fenn was asking
16:19 < maaku> E.g. the canonical example of a usdcoin built by using the prediction market to issue shares against the future btc/usd price.
16:19 < maaku> not familiar with it
16:20 < kanzure> .wik whuffie
16:20 < yoleaux> "Whuffie[pronunciation?] is the ephemeral, reputation-based currency of Cory Doctorow's science fiction novel Down and Out in the Magic Kingdom and his short story Truncat." — http://en.wikipedia.org/wiki/Whuffie
16:20 < kanzure> "The Whuffie Bank launched as a non-profit company at the TechCrunch50 conference on September 15, 2009. The organization aims to create a reputation currency for social networks.[4] However, their URL and Twitter feed became inactive on April 2012. As of December 2013 the website is inaccessible and relays the message "No more Whuffies :'(."[5]"
16:23 < fenn> ugh not "The Whuffie Bank" no no no
16:24 < maaku> repuation a la ripple?
16:24 < fenn> yeah that's closer
16:25 < fenn> reputation is in the eye of the beholder
16:25 < kanzure> the first ripple or the current ripple?
16:25 < maaku> first ripple. ryan fugger's ripple
16:25 < fenn> who your friends are determines the reputation of the target being judged
16:25 < maaku> the only ripple :P
16:26 < fenn> anyway i'm not as interested in whuffie the social currency as just a simple way to determine if something is bullshit or not
16:26 < fenn> s/as/as much as/
16:27 < kanzure> if reputation can be purchased then wouldn't that just mean people would buy your not-bullshit indicators?
16:27 < fenn> reputation can't be purchased
16:27 < fenn> i guess you could game the system by hiring lots of "real" people embedded in real social networks
16:27 < kanzure> how is that possible?
16:27 < kanzure> i mean how is it possible that it can't be purchased?
16:28 < fenn> well, if you care about having an accurate measure, you un-friend people who "like" coca-cola or whatever
16:28 < maaku> well that's what prediction markets are, right? you buy/sell bs/not-bs shares to get an idea of whether something is trustworthy
16:28 < fenn> companies started offering "free stuff" if you liked them on facebook
16:29 < fenn> maaku: that's a little more subtle
16:30 < fenn> so, bitcoinwiki.org is known to be a scam, but due to limited information flow it tricks people into thinking it's a real site
16:31 < fenn> there's nothing stopping people from saying "hey, that's a scam, don't use it"
16:31 < fenn> but what if the owner thinks it's not a scam, and his friends put their support into it
16:31 < fenn> who do you believe?
16:31 < kanzure> nobody, everyone is equally awful and i hate all of them
16:31 < kanzure> or, rather, everyone has some minimum level of awfulness going on there
16:32 < fenn> yes, there's a non-zero noise in the measurement
16:32 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has joined ##hplusroadmap
16:32 < fenn> as always you have to set a noise floor
16:33 < maaku> so a prediction market helps here because it lets people on the inside anonymously profit from knowing it is a scam, but in doing so revealing that fact to the world
16:33 < fenn> doesn't that just incentivize more scams?
16:34 < fenn> also how does this work in practice, who would buy/sell predictions?
16:35 < kanzure> and if that method works, then why does gmaxwell spend a non-trivial amount of his time manually debating marketing scams
16:35 < fenn> also i dont get why it has anything to do with buying and selling
16:36 < maaku> kanzure: I'm not convinced the method works (if we're talking about truthcoin specifically. prediction markets do work in practice but are highly regulated)
16:36 < kanzure> nah not truthcoin
16:36 < fenn> can you name an example of a real prediction market?
16:37 < fenn> is it the same as betting on an outcome? (for the things i'm talking about, there usually will be no final day of judgement)
16:38 < maaku> InTrade (now defunct for regulatory reasons)
16:38 < maaku> fenn: yes
16:38 < kanzure> pagerank is like reputation i guess, and even pagerank links can be bought and sold
16:38 < kanzure> so you would have to argue pretty hard that there exists a way to do reputation that is impervious to purchase
16:39 < kanzure> s/purchase/trade
16:39 < fenn> yes pagerank is like reputation
16:39 < maaku> InTrade famously did better than any poll in predicting the 2008 presidential election outcome
16:39 < fenn> afaik there is no way to "downvote" a link
16:40 < fenn> also pagerank has issues with determining who is a real person
16:40 < kanzure> so will your reputation system
16:40 < fenn> yes but it will have ways to deal with sybil attacks
16:40 < fenn> (that's the whole point after all)
16:41 < kanzure> ?
16:41 < fenn> web pages don't represent people, so pagerank can't judge a page by whether it's a person or not
16:42 < fenn> glug
16:42 < kanzure> that's not what i asked
16:42 < kanzure> what is your magic sybil resistance method
16:42 < fenn> whether nodes are in your social network
16:43 < fenn> if all your friends are morons you will have a lot of noise in your measurement
16:44 < kanzure> social networks have never been immune to sybil attacks
16:44 < fenn> because people are morons
16:44 < fenn> "oh a hot girl friended me wow"
16:45 -!- yashgaroth [~ffffff@2606:6000:cb85:6a00:50a4:f569:684f:fe1b] has joined ##hplusroadmap
16:46 < kanzure> yeah that's not resistance
16:46 < kanzure> or immunity
16:47 < fenn> well so far key-signing parties have failed to take off
16:48 < kanzure> what about them?
16:48 < fenn> it's sort of like a turing test
16:48 < fenn> there's a cost to participate, so one person can't spam a million key signing parties at once
16:49 < fenn> but it turns out the cost is too high for anyone in practice
16:49 < kragen> fenn: except among Debian Developers
16:49 < fenn> (the cost is the time and effort it takes to show up)
16:49 < kanzure> you can't say "wah this other thing isn't working, so here's a broken design"
16:50 < fenn> look just because the social aspect doesn't work for morons doesn't mean the system shouldn't exist at all
16:50 < kanzure> sybil attacks don't attack morons, they attack everyone
16:52 < fenn> if you see spamspamspameggsandspam.com was recommended by bob's friend spambot9999 you might want to unfriend bob
16:52 < fenn> or un-trust or whatever the verb is
16:53 < fenn> facebook has this set up so bob feels shamed; i dunno if that is correct
16:53 -!- Qfwfq [~WashIrvin@unaffiliated/washirving] has quit [Ping timeout: 255 seconds]
16:53 < fenn> "unfriend" operation is considered hurtful
16:54 < kanzure> yeah, that's not immunity
16:54 < kragen> FB doesn't tell you when someone unfriends you
16:54 < fenn> but like, /part from a noisy channel isn't judging the people in the channel
16:54 < kanzure> and also, you can totally buy reputation in that system
16:55 < fenn> spambot9999 paid bob how much?
16:55 < kanzure> does it matter?
16:55 < fenn> yes
16:55 < kanzure> why?
16:55 < fenn> because it limits the number of nodes a spammer can compromise
16:56 < fenn> if bob can be bought for a free soda then he's easily compromised
16:56 < fenn> (fwiw the LIBOR exchange rate could be bribed with day-old sushi)
16:56 < kanzure> this just sounds so awful
16:56 < kragen> pagerank itself is relatively hard to compromise
16:56 < kanzure> haha
16:57 < kragen> what killed it was that people stopped building the graph it analyzed
16:57 < fenn> on the other hand, if bob is detected as a sell-out he loses credibility with all his friends, not just with the detector
16:57 < kragen> because it's easier to search for things on google than to search for them in your blog archive
16:57 < kanzure> that requires his friends to do active maintenance
16:57 < kanzure> your system also sucks because it assumes that sybil attacks aren't graphs
16:57 < fenn> people still link to stuff
16:57 < kragen> not as much as they used to
16:58 < kragen> and a lot of the linking is nofollow
16:58 < kanzure> and if they do link to things it's hidden away in proprietary systems
16:58 < kragen> pagerank does have the huge advantage that it converges, assuming ergodicity
16:58 < kanzure> yes lots of nofollow these days
16:58 < kragen> ergodicity is a very weak assumption
16:58 < fenn> .wik ergodicity
16:58 < yoleaux> "In mathematics, the term ergodic is used to describe a dynamical system which, broadly speaking, has the same behavior averaged over time as averaged over the space of all the system's states (phase space). In physics the term is used to imply that a system satisfies the ergodic hypothesis of thermodynamics." — http://en.wikipedia.org/wiki/Ergodicity
16:58 < fenn> ow my brain
16:59 < kragen> but Google is now using systems with negative edge weights
16:59 -!- Qfwfq [~WashIrvin@unaffiliated/washirving] has joined ##hplusroadmap
16:59 < kragen> which means that you can get "negative PageRank" and then link to sites that you want to demote
16:59 < fenn> how do you do that
16:59 < kragen> by e.g. spamming Gmail users
16:59 < fenn> but doesn't that end up promoting the thing you want to demote?
17:00 < kanzure> gmail filters spam
17:00 < kragen> right, and penalizes the spammy domains
17:00 < kragen> apparently also in the google search index
17:00 < kragen> it doesn't promote the thing you link to; it penalizes it for blackhat SEO
17:00 < fenn> but what i gmail detects your trick and realizes it's an attempt to demote a valid page, ad infinitum
17:00 < fenn> flickering intentionality
17:01 < kragen> systems like that tend to have multiple possible attractors
17:01 < kragen> e.g. maybe you trust Barack Obama and distrust Rush Limbaugh, or maybe vice versa; either state is stable
17:01 < kragen> because of the negative edge weights
17:02 < kragen> PageRank avoids that problem
17:02 < kragen> as does the Advogato reputation system
17:02 < fenn> i'm kinda lost now
17:03 < kragen> (unless, in the case of PageRank, your graph is non-ergodic.  But that's vanishingly unlikely)
17:03 < kragen> sorry
17:03 < fenn> say i send 1000 emails saying "visit kragen's blog!"
17:03 < fenn> 1000 people visit your blog
17:03 < fenn> now i send 10,000 emails, but gmail detects it as spam and ... what?
17:04 < fenn> kragen's blog gets blacklisted from google?
17:04 < kanzure> pagerank does not mean blacklists
17:04 < kragen> right.  or at least gets a spam penalty applied to its pagerank
17:04 < kragen> you can do even better though
17:04 < kragen> you can send 10,000 emails saying "visit fenn's blog!"
17:05 < fenn> but that would lower my blog's popularity
17:05 < kragen> and then once you've verified that your blog has been classified as spam, you can use it to demote my blog whenever you like
17:05 < kragen> you can also send those emails from the domain of your blog, which makes it a more effective tactic
17:06 < kanzure> nobody cares about a spam blog's popularity, fenn
17:06 < kanzure> that's the whole point
17:06 < kanzure> they are trivial to create
17:06 < fenn> sure
17:06 < fenn> you are both yellow and start with k, i might be getting you both mixed up
17:07 < fenn> ok so penalizing is free
17:07 < fenn> why hasn't the internet imploded?
17:07 < kanzure> hasn't it?
17:08 < fenn> i'm not sure, i don't really use google anymore because of the stupid features like "did you mean... "
17:09 -!- nmz787_i [~nmccorkx@192.55.54.40] has quit [Ping timeout: 250 seconds]
17:09 < fenn> it seems the internet not working is more of a technological problem than a game theoretical one
17:10 < fenn> link rot, site closure, forking
17:11 < fenn> slashdot style content moderation systems have managed to hold back the sheer tide of stupidity, for a while at least
17:12 < kragen> I think that's largely a game-theoretical problem
17:12 < kragen> you could reduce the cost of preventing link rot
17:13 < fenn> kanzure do you think people purchase karma on reddit/slashdot/hn/lesswrong?
17:13 < kanzure> YES
17:13 < kanzure> digg even put a price on it
17:13 < kragen> but unless you reduce it to zero you still need people who take the time to actually do it
17:14 < kragen> which is largely a game-theoretical problem
17:14 < fenn> it only takes one person to fix the link, and it doesn't actually have to be the webmaster
17:14 < kragen> that's right
17:15 < fenn> IF we had a web of trust, it would be simple enough to follow the fixed link provided by a trusted person
17:15 < kragen> but you have to do that without turning the entire internet into linkspam or YouTube comments
17:15 < kragen> yeah
17:16 < kragen> Wikipedia does it without a web of trust, but only by deleting unpopular pages
17:16 < fenn> i don't really get wikipedia's deletion policy
17:16 < kragen> well, that's basically it
17:16 < kragen> since WP doesn't have a web of trust, only popular pgaes get maintained
17:17 < fenn> what prevents hordes of b-tards from colonizing wikipedia then?
17:17 < kanzure> http://en.wikipedia.org/wiki/Goodhart's_law
17:17 < kragen> uncyclopedia and encyclopedia dramatica
17:17 < kanzure> ""When a measure becomes a target, it ceases to be a good measure.""
17:17 < kragen> I mean basically there aren't that many b-tards
17:18 < kanzure> i would guestimate at least 100k
17:18 < kragen> yeah, but that's not enough to overwhelm Wikipedia's vandal patrol
17:18 < kanzure> really? hm
17:18 < kragen> especially when it's more fun to write shit on https://encyclopediadramatica.se/Girlvinyl instead
17:19 < fenn> kanzure: that's not universal; it only applies to systems that can be gamed
17:19 < fenn> i'm thinking about people trying to lose weight or gain height or something
17:19 < kanzure> your reputation network can be gamed
17:20 < kragen> gaining height does get gamed
17:20 < fenn> yeah people chop their legs apart and stuff, it's gross
17:20 < kragen> think about dwarfs taking HGH and Chinese femur lengthening surgery
17:20 < fenn> i don't have any problem with HGH
17:20 < kanzure> why is gaining height a gaming of height gaining
17:20 < kanzure> fenn is it possible that the universe disagrees with your beliefs about reputation
17:21 < fenn> because height is attractive for evolutionary reasons having to do with the ability to eat a lot of food
17:21 < fenn> ... or something
17:21 < kragen> it's a gaming of people's "tall = high status" wiring
17:21 -!- rak[1] is now known as dog
17:21 < kragen> similar to dieting, liposuction, makeup, plastic surgery, clothing...
17:21 < fenn> yeah we should get rid of all that stuff :P
17:21 < kragen> piercing, tattooing, bodybuilding...
17:22 < kragen> really? why?
17:22 -!- dog is now known as Guest75225
17:22 < fenn> genetic modification is simpler and more effective
17:22 -!- Guest75225 is now known as raise-the-hunt
17:22 < kragen> really?
17:23 < fenn> if people want to be tall, is chopping their legs apart the answer?
17:23 -!- raise-the-hunt is now known as rak[1]
17:23 -!- rak[1] is now known as rk[1]
17:24 < fenn> some time in the far future, spam will cease to exist because the bots have gotten so good they actually contribute in a useful way
17:24 < kragen> it's probably less dangerous than gene therapy and stem cell therapy
17:25 < fenn> so here's my issue with goodhart's law: if you game the "height" measure, you really are taller
17:26 < fenn> if you do steroids, you really are stronger
17:26 < kanzure> and if you spam more, you really are spamming. who cares?
17:26 < fenn> at some point it ceases to be gaming
17:26 < streety> what's the underlying desire behind wanting to be taller?
17:26 < kanzure> er, does that matter?
17:26 < kragen> streety: discrimination against short people
17:26 < kragen> yeah.  it's only a problem if you're using height as a proxy for something else, like maturity or social class
17:27 < fenn> but there is no such thing as social class
17:27 < kragen> I mean if strength is really what people value in a fuckbuddy (if that's what you're trying to game)
17:27 < kanzure> fenn: so now it sounds like your reputation system is really just some communication system to see what your friends said
17:27 < kanzure> and nothing about reputation at all
17:27 < fenn> yes that's all it ever was
17:27 < fenn> what is "reputation"
17:27 < kragen> a girlfriend of mine told me that before she met me she thought that people had to be physically fit to be good in bed
17:28 < kanzure> you're the one building a reputation system, argh
17:28 -!- strangewarp_ [~strangewa@c-76-25-206-3.hsd1.co.comcast.net] has joined ##hplusroadmap
17:29 < fenn> reputation is a mushy hodgepodge combination of signals, trying to measure some hidden variable like "trustworthiness" that maybe doesn't even exist
17:29 < kanzure> that sounds dumb
17:29 < fenn> me i'd prefer to measure the signals directly, but people find that too difficult
17:29 -!- helleshin [~talinck@66-161-138-110.ubr1.dyn.lebanon-oh.fuse.net] has joined ##hplusroadmap
17:30 < fenn> 1 out of 5500 people said this website was a scam
17:30 < fenn> hurr
17:30 < fenn> .wik fitness
17:30 < yoleaux> "Disambiguation: Fitness" — http://en.wikipedia.org/wiki/Fitness
17:31 -!- strangewarp [~strangewa@c-76-25-206-3.hsd1.co.comcast.net] has quit [Ping timeout: 256 seconds]
17:31 < fenn> evolutionary biologists get into fits over the meaning of "fitness"
17:31 -!- hehelleshin [~talinck@66-161-138-110.ubr1.dyn.lebanon-oh.fuse.net] has quit [Ping timeout: 264 seconds]
17:31 < kragen> really?
17:32 < fenn> "fitness contributes to survival and reproduction"
17:32 < kragen> I thought they were the only ones who didn't
17:32 < fenn> ok but what about random plagues that kill everyone except the mutant previously considered "unfit"
17:33 < kanzure> biologists get angry when you say fitness, they like to correct you with "survival of the fit, not the fittest"
17:33 < kanzure> *fittest
17:33 < fenn> suddenly fitness means something different than it did before the plague
17:33 < kanzure> no
17:34 < kanzure> the concept of fitness recognizes that
17:34 < kanzure> nobody goes around applying fitness math though
17:35 < fenn> you can have two mutually exclusive plagues where, depending on a random variable, resistance to one or the other is selected for
17:35 < fenn> in fact you can actually build this with a vial of bacteria and two different antibiotics
17:36 < kanzure> selection doesn't work like that
17:37 < fenn> ok sometimes there are tradeoffs and sometimes there aren't
17:37 < kanzure> what did this have to do with your reputation system, again?
17:37 < fenn> uh, now i'm not making sense
17:38 < fenn> reputation is one of those things that is meaningless out of context
17:38 < fenn> like reproductive fitness
17:40 -!- ebowden [~ebowden@147.69.147.27] has joined ##hplusroadmap
17:42 < fenn> streety: in ancient times, better hunters ate more food and were taller, so it was a quick shortcut to determining who to follow, who to mate, etc
17:42 < kragen> [ev psych just so story]
17:43 < fenn> height is still correlated with IQ to a small degree
17:43 < kragen> it's a large degree if the population you sample from includes people who were malnourished as children
17:44 < fenn> "...  or that both height and intelligence may be affected by adverse early environmental exposures."
17:45 < kanzure> i don't know what this is http://aflowlib.org/apps.php
17:45 < kanzure> "Library of online structural ab initio calculations (308975 Calculations)"
17:46 < fenn> or more importantly for our evo-psycho story, that "common genetic factors can influence both hight and intelligence"
17:46 < kanzure> "sponsored by Department of Homeland Security - Domestic Nuclear Detection Office ."
17:46 < kragen> "AFLOWLIB.ORG: a distributed materials properties repository from high-throughput ab initio calculations"
17:48 < fenn> it doesn't really matter why height is desirable, it just is
17:49 < kanzure> http://en.wikipedia.org/wiki/Campbell%27s_law
17:49 < kanzure> "The more any quantitative social indicator (or even some qualitative indicator) is used for social decision-making, the more subject it will be to corruption pressures and the more apt it will be to distort and corrupt the social processes it is intended to monitor."
17:49 < jrayhawk> pygmyism is what happens in environments of scarcity; the large take too many calories to mate with
17:49 < jrayhawk> so it can go both ways
17:49 < jrayhawk> but yes, usually tall is a giant advantage
17:49 < kanzure> why is the evo psych stuff even necessary, why not just say "tall people tended to be tall and do tall people things"
17:50 < jrayhawk> because it adds a general associative fitness factor
17:50 < jrayhawk> or, adds to a general ssociative fitness factor, rather
17:50 < fenn> i don't see why it's necessarily "evo psych" - it sounds like run of the mill sexual selection to me
17:50 -!- juri_ [~juri@vpn166.sdf.org] has quit [Ping timeout: 255 seconds]
17:51 < kanzure> .wik kerckhoff's principle
17:51 < yoleaux> "In cryptography, Kerckhoffs's principle (also called Kerckhoffs's desiderata, Kerckhoffs's assumption, axiom, or law) was stated by Auguste Kerckhoffs in the 19th century: A cryptosystem should be secure even if everything about the system, except the key, is public knowledge." — http://en.wikipedia.org/wiki/Kerckhoff%27s_principle
17:51 -!- juri_ [~juri@vpn166.sdf.org] has joined ##hplusroadmap
17:52 -!- TheShadowFog [~TheShadow@2601:8:3e80:c:7d71:2e6c:7049:6035] has quit [Quit: Leaving]
17:52 < fenn> i love the examples on this page http://en.wikipedia.org/wiki/Perverse_incentive
17:53 -!- TheShadowFog [~TheShadow@2601:8:3e80:c:7d71:2e6c:7049:6035] has joined ##hplusroadmap
17:54 < kragen> once we're talking about tallness influencing things other than sexual attractiveness, we've gone beyond run of the mill sexual selection into evpsycho
17:54 < jrayhawk> i am an evpsycho, it is true
17:55 < kanzure> "Before the Anatomy Act of 1832, executed criminals were the only legal source of bodies for hospitals to use for surgeon training. Due to high demand from chronic shortage of legal cadavers, "resurrection men" resorted to illegal means to obtain bodies, such as digging up corpses from graveyards or even murder. In 1828, William Burke and William Hare murdered 16 people and sold the bodies. Thomas Williams and John Bishop, part of a group of ...
17:55 < kanzure> ... body snatchers known as the London Burkers, committed murder for the purpose of selling the victim's body in 1831."
17:55 < fenn> what if tallness is correlated with running speed, is that "ev psych"?
17:55 < kanzure> the only fitting punishment is to execute them and sell their bodies to science as well
17:55 < kragen> fenn: no, that's just kinematics :)
17:55 < kanzure> "illegal means to obtain bodies"
17:56 < kanzure> man, why don't serial killers sell the bodies
17:56 < kanzure> .wik anatomy act of 1832
17:56 < yoleaux> "The Anatomy Act 1832 (2 & 3 Will. IV c.75) was an Act of Parliament in the United Kingdom that gave freer licence to doctors, teachers of anatomy and bona fide medical students to dissect donated bodies. It was enacted in response to public revulsion at the illegal trade in corpses." — http://en.wikipedia.org/wiki/Anatomy_Act_of_1832
17:56 < fenn> i'm really not following how sexual attractiveness is not psychology?
17:56 < kragen> well, Burke and Hare were serial killers who sell the bodies
17:56 < kragen> fenn: it is!
17:56 < kanzure> why do you need psychology for sexual attraction?
17:57 < fenn> because humans are not animalsss!!!
17:57 < kanzure> "The Act was repealed by the Anatomy Act 1984, which was, in turn, repealed by the Human Tissue Act 2004. Access to corpses for the purposes of medical science is now regulated by the Human Tissue Authority."
17:57 < kragen> sexual attraction is a psychological phenomenon, no?
17:57 < kanzure> .wik human tissue authority
17:57 < yoleaux> "The Human Tissue Authority (HTA) is an executive non-departmental public body of the Department of Health. It regulates the removal, storage, use and disposal of human bodies, organs and tissue for a number of scheduled purposes such as research, transplantation, and education and training." — http://en.wikipedia.org/wiki/Human_Tissue_Authority
17:57 < kragen> .ety repeal
17:57 < yoleaux> repeal (v.): "late 14c., from Anglo-French repeler, Old French rapeler "call back, call in, call after, revoke" (Modern French rappeler), from re- "back" (see re-) + apeler "to call" (see appeal (v.)). Related: Repealed; repealing." — http://etymonline.com/index.php?term=repeal
17:58 < jrayhawk> i am confused, what does this have to do with a theoretical reputation system
17:58 < fenn> because tallness correlates with trustworthiness, apparently
17:58 < jrayhawk> oh, okay, yeah
17:58 < fenn> actually i just made that up
17:58 < jrayhawk> is someone arguing against that in principle, or in practice?
17:59 < fenn> kanzure is saying that we can never trust anything anyone says because someone can game the system
17:59 -!- juri_ [~juri@vpn166.sdf.org] has quit [Ping timeout: 240 seconds]
17:59 < kanzure> trust is a terrible idea
17:59 < kragen> ha
17:59 < kragen> jrayhawk: basically the question is how to make a useful reputation system
17:59 < jrayhawk> don't conflate imperfection with uselessness
18:00 < kragen> in the face of Goodhart's Law
18:00 < kragen> fenn posited height as a good measure that is still a target
18:00 < kragen> since it's hard to game
18:00 < jrayhawk> tell that to the chinese folks getting bone grafts
18:01 < jrayhawk> oh, i guess it's breaking and rehealing incrementally, not grafting
18:01 < fenn> that's what i said, but then i realized they got all the advantages of being tall
18:01 < jrayhawk> christ, so complicated
18:01 < kragen> both fenn and I brought that up too
18:01 < kragen> 01:20 < fenn> yeah people chop their legs apart and stuff, it's gross
18:01 < kragen> 01:20 < kragen> think about dwarfs taking HGH and Chinese femur lengthening surgery
18:01 < kanzure> "wow for some reason the people in here have the same thoughts!" is not surprising
18:01 < kragen> 41 minutes ago
18:01 < kragen> so the question is how to do useful reputation systems
18:02 < kanzure> or if you can survive without one just fine
18:02 < kanzure> or if reputation is a bad idea
18:02 < jrayhawk> PGP seems pretty good, but gmaxwell pointed out that there really aren't social incentives to report or act on reports of bad players in the trust web
18:02 < jrayhawk> there's no "anti-signing"
18:02 -!- rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has quit [Ping timeout: 264 seconds]
18:03 < fenn> pgp just says "these keys really belong to this person" right?
18:03 < jrayhawk> ayup
18:04 < kanzure> .w reputation
18:04 < yoleaux> kanzure: Sorry, that command (.w) crashed.
18:04 < fenn> but there's a cost involved in setting up your connections
18:04 < jrayhawk> If and only if you want to do it right.
18:04 < jrayhawk> There's no cost to doing it wrong.
18:04 < fenn> it takes a nonzero effort and technical knowledge
18:05 < kanzure> "the beliefs or opinions that are generally held about someone or something"
18:05 < jrayhawk> And no incentives to punish those who do it wrong.
18:05 < fenn> you can trust people with pgp keys not to be spammers, because so far it's not worth spammers' time to try to game the pgp trust network
18:06 < kragen> .ud reputation
18:06 < fenn> even though it wasn't set up for that purpose at all
18:06 < jrayhawk> not yet, anyway
18:08 < jrayhawk> once a spammer gets into the strong set, how to you incentivize trust path revocations?
18:08 < fenn> shame?
18:09 < jrayhawk> One person's word against another.
18:09 < fenn> out of band verification
18:09 < kragen> .wik reputation
18:09 < yoleaux> "Reputation of a social entity (a person, a social group, an organization) is an opinion about that entity, typically a result of social evaluation on a set of criteria. It is important in business, education, online communities, and many other fields." — http://en.wikipedia.org/wiki/Reputation
18:09 < kragen> jrayhawk: is that even necessary?
18:09 < kanzure> if reputation really mattered then you wouldn't see an infinite parade of scams on bitcointalk.org
18:10 < fenn> i haven't read up on how the pgp network works, but "strong set" sounds fishy in principle
18:10 < kragen> kanzure: why not?
18:10 < kragen> I mean, what would happen to the infinite parade of scams if reputation really mattered?
18:10 < jrayhawk> it's just a simple term meaning "is a part of the largest network"
18:11 < fenn> "while islands of sets of keys that only sign each other in a disconnected group can and do exist, only one member of that group needs to exchange signatures with the strong set for that group to also become a part of the strong set"
18:11 < fenn> ok that's dumb
18:11 < fenn> that's not strong at all
18:11 < fenn> "look at me, i'm on the internet, i'm in the strong set!"
18:11 < jrayhawk> yeah, i agree it was poorly chosen
18:12 < jrayhawk> okay, so if you have a network of, say, a hundred spammers, all of them interconnected trusting eachother because they don't care about doing things right, and each of them gets three signatures with the rest of the WoT, how can shame scale with that?
18:13 < jrayhawk> to get rid of one of them, you're looking at dealing with 300 of them all at once
18:13 < jrayhawk> do you think you can kill trust connections faster than they can make them?
18:14 < kragen> jrayhawk: trust metrics like Advogato's and PageRank are robust against that kind of thing
18:14 < jrayhawk> (I do not think you can kill most trust connections at all, mostly because PGP is about identity and not reputation)
18:14 < kanzure> pagerank is not robust against that
18:14 < kanzure> pagerank is getting things totally wrong all the time, what are you talking about
18:15 < fenn> http://www.advogato.org/trust-metric.html scroll down to last figure
18:15 < jrayhawk> PGP is somewhat extensible, so if we write a toolchain that isn't complete garbage we can probably graft a reputation system onto it.
18:15 < kragen> there is one parameter in the PageRank algorithm from the paper which makes it potentially not robust against that, kanzure, it's true
18:15 < kragen> but you can set that parameter quite low
18:15 < kragen> that's the "random jump" probability, I forget what they called it
18:16 < kragen> c
18:16 < kragen> well, 1-c
18:17 -!- juri_ [~juri@vpn166.sdf.org] has joined ##hplusroadmap
18:17 < kragen> but if you disregard 1-c, then the PageRank entering the 300-spammer subgraph is independent of how many spammers are in it
18:18 < kragen> that is, it's the same if it's just one spamming node or a whole interconnected subgraph of them
18:18 < kragen> the Advogato metric is maybe more interesting
18:18 < jrayhawk> thank you for pointing it out to me
18:19 < fenn> i dont actually understand how you're supposed to figure out which nodes are "good" and which are "bad"
18:19 < kragen> sure
18:19 < kragen> fenn: typically you start by deifying yourself
18:20 < fenn> i notice in the diagram the "bad" nodes only have one link to the supersink
18:20 < fenn> but if they were really spammers they'd have lots of links to the thing they're spamming
18:21 < kragen> yeah, Advogato's metric worked out well in practice but it wasn't subject to the level of attack that PageRank was
18:23 < kragen> so on a different topic, I calculated that Iceland's geothermal resource is a few thousand times bigger than Ghawar
18:24 < kragen> fossilized heat, in a sense, diffusing up through the crust on a timescale of billions of years
18:25 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has quit [Ping timeout: 272 seconds]
18:26 -!- rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has joined ##hplusroadmap
18:27 < fenn> the jetstream is pretty powerful too
18:27 < kragen> is it? how powerful?
18:27 < fenn> enough to make it cold as fuck here :)
18:27 < kragen> not very powerful then?
18:28 < kragen> I think Iceland's not that atypical in quantity, just in how close the heat is to the surface.  I think you could do self-sustaining robotic hot-dry rock heat mining.
18:28 < kragen> pretty much anywhere.
18:29 < kragen> and, at least for now, you could do it without attracting attention.
18:30 < fenn> "There are two major scientific articles about jet stream power. Archer & Caldeira[36] claim that the jet streams can generate the total power of 1700 TW, and that the climatic impact will be negligible. Miller, Gans, & Kleidon[37] claim that the jet streams can generate the total power of only 7.5 TW, and that the climatic impact will be catastrophic."
18:31 < kragen> 2011 world marketed energy consumption is 17.5 terawatts
18:31 -!- Vutral [ss@mirbsd/special/Vutral] has joined ##hplusroadmap
18:31 < fenn> that's quite a difference in estimates
18:31 < kanzure> "large hadron migrator" https://github.com/soundcloud/lhm
18:31 < kragen> so either way it's a significant amount of power
18:31 < kragen> but it's also not thousands of times bigger than Ghawar except at the high end
18:32 < fenn> one is power and one is energy, not the same thing
18:32 -!- justanotheruser [~Justan@unaffiliated/justanotheruser] has quit [Ping timeout: 255 seconds]
18:33 < kragen> yes, it depends on what timescale you are contemplating mining the energy over, or if you're just planning to extract sustainably
18:34 < kragen> https://en.wikipedia.org/wiki/Enhanced_geothermal_system#EGS_potential_in_the_United_States says 13000 ZJ with 200 ZJ extractable in the US
18:34 < kragen> 200 ZJ / 50 years is 130 terawatts
18:36 < fenn> 50 years is not very long
18:36 < fenn> that's less than the operating life of a nuclear plant
18:37 < kragen> should be long enough to start launching a Dyson swarm
18:37 < fenn> lack of energy is not the problem there
18:38 < fenn> do you know about space tethers?
18:38 < fenn> the rotating kind
18:39 < kragen> natch
18:39 < kanzure> jrayhawk: maybe you could pay for link removal or something with a fraud proof
18:39 < kanzure> jrayhawk: so if a fraud proof ever happens then the other links should be automagically reverted when shown the fraud proof or something
18:40 < kanzure> jrayhawk: even if that person isn't online or taking actions
18:40 < kanzure> andytoshi would have some okay ideas for how to construct proofs of fraud or something
18:41 < fenn> kragen: solar power can be efficiently converted to orbital kinetic energy and transferred to suborbital flying objects http://tethers.com/MXTethers2.html http://tethers.com/EDTethers.html
18:41 < kragen> fenn: I think the problem there is the lack of self-reproducing asteroid-mining robots
18:41 < kanzure> a prediction market about reputation fraud would probably suffer from the ability of the fraudsters to make bets in or against their favor
18:42 < kragen> interesting, didn't know that
18:42 < fenn> this uses currently existing materials too, no wacky carbon nanotubes
18:42 < jrayhawk> well, that's why you only provide as much incentive as their is value in the reputation
18:42 < jrayhawk> s/their/there/
18:43 < fenn> oh there's a video somewhere that shows how this works
18:43 < jrayhawk> i would be entirely willing to associate a few mBTC with every reputation claim i make
18:44 < kanzure> a fraudster can always tank his reputation and cash out against the bonds or something
18:44 < jrayhawk> Which is basically how it works in real life, too.
18:45 < kanzure> well that's useless
18:45 < fenn> kragen: http://youtu.be/mPx1Nq80jm8
18:45 < jrayhawk> Empirically not useless.
18:45 < jrayhawk> Just not perfect.
18:46 < kanzure> the point was to have a system that is more useful than not existing at all
18:46 < jrayhawk> well, the difficulty is that the lack of formalization means there's a lot more guesswork in real life
18:46 < kanzure> enough to justify or cover the costs of development, maintenance and use
18:46 < jrayhawk> like, what if we could automatically confirm every claim on a resume?
18:47 -!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap
18:47 < kanzure> what does that even mean
18:47 < kragen> fenn: I recommended someone else to watch Emacs screencasts today, so I am watching this video you linked as my penance
18:48 < jrayhawk> like if i go up to every previous boss i have had and told them to sign each of my claims for what i achieved
18:50 < kragen> fenn: I don't know that orbital momentum exchange tethers really solve the problem
18:50 < kragen> they cut it in half
18:50 < jrayhawk> hiring a treasurer on the strength of a confirmable resume rather than a guessworkable resume doesn't mean they can't still run off to the bahamas, but it is at least improved priors
18:50 < kragen> but getting to orbit in the first place is already pretty hard
18:50 < kanzure> jrayhawk: or you can use a system where your treasurer never has the private keys to sign away your treasury anyway
18:50 < fenn> kragen: the rocket equation says that if you halve the delta V required you square the payload capacity (i'm probably getting it wrong)
18:51 < jrayhawk> yeah, well, you have to trust everybody with everything sooner or later
18:51 < kragen> self-reproducing geothermal heat-mining robots are already halfway to being self-reproducing asteroid-mining robots
18:51 < jrayhawk> but yes, hopefully you don't have to trust somebody with everything.
18:51 < kanzure> do we really have to call those astrochickens still
18:51 < kanzure> .wik astrochicken
18:51 < yoleaux> "Astrochicken is the name given to a thought experiment expounded by theoretical physicist Freeman Dyson. In his book Disturbing the Universe (1979), Dyson contemplated how humanity could build a small, self-replicating automaton that could explore space more efficiently than a manned craft could." — http://en.wikipedia.org/wiki/Astrochicken
18:52 < fenn> the video spends too much time on the rocket launch and not enough on how tethers work
18:52 < kragen> that is true
18:52 < fenn> it shows boosting from LEO to GTO, but it can also work for boosting suborbital to LEO
18:53 < fenn> or you can do both, or do stuff like pick up bags of dirt from the moon and use them to power your tether launch system
18:53 < jrayhawk> i guess a bitcoin-like formalized structural incentive system for reputation would pretty much turn society into a unified paperclip maximizer.
18:53 < jrayhawk> that sounds like fun.
18:53 < kragen> "See also: Grey goo"
18:53 < jrayhawk> bitcoin is like halfway there
18:53 < kragen> jrayhawk: gold coinage is halfway there
18:54 < kanzure> uh?
18:54 < fenn> kragen a boost from a tether is the difference between riding to orbit on spaceshiptwo vs a falcon-9
18:54 < jrayhawk> gold coinage is halfway to bitcoinage, maybe
18:54 < fenn> in economic terms that's $200,000 per person vs $20,000,000
18:55 < kragen> kanzure: 2500 years ago we started coining gold into coins, and ever since then society has suffered massive paperclip-maximizing problems
18:56 < kragen> actually maybe you could go back further to heterotrophy, but gold coinage has really exacerbated the situation
18:56 < jrayhawk> Yeah, basically true.
18:58 < jrayhawk> Obviously it's subject to manipulation on account of physical control and has actual instrumental value, so it's a bit more distorted than BTC, but it is a big part of the progression.
18:58 < fenn> damn those carbohydrate maximizers!
18:59 < kragen> fenn: I see what you mean
19:00 < jrayhawk> anyway, re: reputation: proportional response to disreputable acts seems really hard
19:00 < fenn> once you have cheap transport to orbit, using orbital solar power is way cheaper
19:01 < fenn> it sounds dumb when i say it like that
19:01 < kragen> it might be true
19:01 < kragen> how do you get the energy to earth?  boron and aluminum meteorites?
19:01 < kanzure> jrayhawk: also the disreputable act is out-of-band anyway, so that makes it inaccessible or something
19:01 < kragen> or do you just not bother?
19:02 < fenn> o'neill did a lot of calculations on using phased array masers to transmit power from GEO to stationary antenna arrays
19:03 < fenn> it turned out to be efficient and impossible to use as a weapon
19:03 < kragen> yeah, but I don't think you can make rectennas much more power-dense than sunlight, or much cheaper than photovoltaic
19:03 < kragen> photovoltaic on the ground, I mean
19:03 < fenn> how could it not be cheaper than photovoltaics, it's just wires
19:03 < kragen> I haven't done the math, though, so I could be wrong
19:03 < kragen> photovoltaic is really, really cheap
19:03 < kragen> I mean it's reduced and doped sand with wires on it
19:04 < fenn> ok well tune your frequency to whatever power density makes sense
19:04 < fenn> vaporize new jersey for all i care
19:04 < kragen> with the rectenna you avoid having to reduce and dope the sand but you don't avoid the wires
19:04 < kragen> and the usual proposal is to use microwaves, so the wires have to be fairly dense
19:04 < kragen> I mean, densely spread
19:04 < kragen> a few centimeters apart
19:05 < fenn> 1 meter apart
19:05 < fenn> for birds
19:05 < fenn> microwaves would cook birds
19:05 < kragen> not if the power density is low enough
19:05 < fenn> well, not really it turns out
19:06 < fenn> right, the power density is about 1kW/m^2 or equivalent to sunlight, but if you're a bird that's enough to make you not want to fly around
19:06 < kragen> could be, yeah
19:06 < fenn> with a longer wavelength the bird doesn't absorb any heat at all
19:06 < kragen> why do the existing rectenna proposals use microwaves?
19:07 < fenn> because we lack a suitable vocabulary for specifying wavelength ranges
19:08 < fenn> the reason they set the power density at 1kW/m^2 was to allay the fears of the public
19:08 < fenn> you can have higher power densities with long wave
19:08 < fenn> i dont remember the maximum
19:09 < kragen> I'm skeptical
19:09 < fenn> of what exactly
19:11 < kragen> that the reason is that "we lack a suitable vocabulary for specifying wavelength ranges"
19:11 < kragen> I suspect it's Airy-spot diameter at geosynchronous focal lengths
19:12 < kragen> or something else similarly difficult to overcome
19:12 -!- pete4242 [~smuxi@boole.london.hackspace.org.uk] has joined ##hplusroadmap
19:13 < fenn> oh, that too
19:13 < fenn> it's either the birds or new jersey
19:14 < kragen> I think the boron and aluminum meteorites are more feasible
19:15 < delinquentme> gun kata
19:15 < delinquentme> https://www.youtube.com/watch?v=A2KJHysK6k8
19:15 < delinquentme> bc
19:15 < kanzure> .title
19:15 < yoleaux> Equilibrium Gun Kata Compilation - YouTube
19:15 < fenn> awful movie
19:15 < fenn> it's "the matrix" without all the robots and philosophy
19:17 < fenn> kragen: what do you do with the meteorites?
19:17 < delinquentme> fenn, better movie suggestion?
19:18 < fenn> oh another powertransfer mechanism is a launch loop or similar dynamic compression system
19:18 < jrayhawk> i quite liked Revolver
19:19 < jrayhawk> original UK cut if possible, but the U.S. cut was tolerable
19:19 < fenn> delinquentme: if you're into kung fu, i liked Azumi and House of Flying Daggers
19:20 < kragen> fenn: burn them in oxygen from the atmosphere
19:20 < kanzure> .title https://www.youtube.com/watch?v=bXcmG2uvgRY
19:20 < yoleaux> Fist of the North Star (1995 Live Action) - YouTube
19:21 < fenn> kragen: waste of energy.. the meteorites are worth more in orbit due to their kinetic energy
19:21 < kanzure> trailer: https://www.youtube.com/watch?v=4xRxlusrAT4
19:21 < kragen> really?
19:21 < fenn> .wa 0.5*kg*orbitalvelocity^2 in MJ
19:21 < yoleaux> fenn: Sorry, no result!
19:21 < kragen> I hadn't done that calculation before
19:21 < fenn> gasoline is 49MJ/kg
19:21 < kragen> yes
19:22 < kragen> LEO is 8km/s
19:22 < fenn> assuming low earth orbit is 8km/s that yields 32MJ/kg
19:22 < kragen> which is 32kJ/kg
19:22 < kragen> sorry, MJ
19:22 < kragen> as you said
19:23 < kragen> boron is 58.9MJ/kg if you burn it in oxygen
19:24 < kragen> so dropping it to earth only loses a bit over a third of the energy it had in orbit, the kinetic third
19:24 < kragen> aluminum is 31MJ/kg
19:25 < kragen> so you lose half
19:25 < kragen> not as efficient as rectennas but likely a lot more practical
19:25 < fenn> boron-11 is 1377109 MJ/kg if you burn it in hydrogen :P
19:25 < kragen> especially if you can just use the aluminum rather than burning it for heat
19:25 < kragen> heh, you mean with fusion?
19:25 < fenn> right
19:27 < kragen> burning boron is tricky.  aluminum less so.
19:28 < fenn> burning things is inefficient
19:28 < kragen> both have the advantage that they can survive a lot of abuse in contact with air
19:28 < kragen> only 60% inefficient though
19:28 < fenn> more than that i'd wager
19:28 < fenn> would you be boiling steam and turning turbines? or some kind of molten salt fuel cell?
19:29 < fenn> aluminum can generate hydrogen fairly easily
19:29 < fenn> just use electricity, sheesh
19:30 < kragen> boiling steam and turning turbines at a large scale is often about 38% efficient these days
19:30 < fenn> there was a lot of sci-fi about metastable helium isotopes as an energy storage medium
19:31 < kragen> you could, yes, probably also use aluminum as a plate in a battery
19:31 < fenn> no i mean, skip the aluminum and just transfer electricity via microwave
19:31 < kragen> like a molten-cryolite and carbon fuel cell, the same thing as an aluminum electrolysis pot in reverse
19:31 < kragen> I don't know that that's more efficient
19:31 < superkuh> What is the dielectric strength of a typical environment induced aluminum oxide coating? Can the thickness of it be controlled by any mechanism?
19:31 < superkuh> (on aluminum)
19:32 < kragen> superkuh: in many cases people anodize aluminum to thicken that coating
19:32 < fenn> mercury readily dissolves the aluminum oxide, and there are probably other catalysts
19:32 < kragen> does it really dissolve it? I've wondered how that works
19:32 < superkuh> Excellent. Thanks.
19:32 < fenn> there are aluminum alloys that don't form skin at all
19:32 < kragen> superkuh: lots of aluminum things are actually anodized aluminum
19:33 < kragen> in many cases people incorporate other crap into the coating to e.g. color it
19:33 < kragen> e.g. black
19:33 < fenn> also known as "sorta purple" or "sorta green"
19:33 < kragen> you can get those without even dyeing it though :)
19:33 < superkuh> I've been trying to find a low temperature curing alumina coating epoxy for about a year. I needed to coat a thin copper wire. A lot of the rest of the machine is already aluminum (the capacitor plates) and I know how to solder to it. Just growing it on an aluminum wire and then filing under oil might just work.
19:34 < superkuh> This is relevant to the discussion because I am building a demo dense plasma focus.
19:34 < fenn> superkuh: you could do a reduction reaction to plate the copper with aluminum, and then anodize it
19:34 < fenn> or sputter aluminum onto it and then anodize
19:35 < superkuh> Aluminum is already pretty conductive.
19:35 < kragen> fenn: how would you electroplate aluminum onto copper?  you'd need a water-free electrolyte, no?
19:36 < fenn> iodine and aluminum metal burns and covers stuff around it with aluminum metal
19:38 < kragen> you're saying to heat up AlI₂ in air?
19:39 < kragen> uh, I guess it's AlI₃ (beats me as to why)
19:39 < fenn> hum it might leave a film of aluminum iodide instead
19:39 < fenn> Al2I6
19:40 < kragen> yeah
19:46 < fenn> superkuh: i didnt really understand your problem, but you can use zinc-based solders to connect aluminum components (alumalloy)
19:46 -!- justanotheruser [~Justan@unaffiliated/justanotheruser] has joined ##hplusroadmap
19:52 < fenn> Procedure:
19:52 < fenn> Clean the area to be soldered
19:52 < fenn> Apply Stay Clean aluminum flux
19:53 < fenn> Heat until the flux becomes a nut brown color
19:53 < fenn> Apply the alloy
19:53 < kragen> superkuh: normal electrolytic capacitors actually involve growing aluminum oxide coatings I think
19:54 < heath> fenn just described my job from ~6 years ago
19:55 < fenn> ok heath now finish that plasma focus fusion device
19:55 -!- juri_ [~juri@vpn166.sdf.org] has quit [Ping timeout: 264 seconds]
19:55 < fenn> remember, we are CITIZEN CYBER SOLDIERS
19:56 < kanzure> wrong channel
19:56 < kanzure> no context
19:57 < fenn> amendment N freedom of context
19:59 < fenn> the state shall make no law restricting grammar or the use of intercontextuality
20:02 < kanzure> http://3.bp.blogspot.com/-362PrgdQUNQ/UH-urJVH_gI/AAAAAAAAJEQ/pFiWlfpY_ZI/s1600/dredd-3d-headshot.jpg
20:03 < kragen> fenn: https://www.youtube.com/watch?v=cZkAP-CQlhA sometimes maintaining microwave antennas can have unexpected difficulties
20:07 < fenn> now imagine all those are meteors falling from the sky
20:09 < fenn> due to the way momentum exchange tethers work, no energy is expended if the mass on the tether remains constant, so you could download the refined metal and upload raw ore for only the cost of suborbital launch
20:10 < fenn> you'd end up with a mass imbalance unless you were downloading oxygen
20:11 < kragen> yeah, burning extraterrestrial fuel will deplete atmospheric oxygen
20:11 < kragen> but that won't become a problem until world marketed energy consumption is several times greater than at present
20:11 < fenn> i mean on the tether, the masses going down will be less than the masses going up
20:12 < kragen> oh
20:12 < kragen> yeah
20:12 < kragen> or you just use ablative heat shielding to land the meteorites
20:12 < kragen> and a parachute or something
20:12 < fenn> but that's super wasteful
20:13 < kragen> I think the waste is very small
20:13 < kanzure> kragen: find me something clever to do with eyelash diffraction
20:13 < fenn> it's half the energy of your energy storage medium, it's not small!
20:13 < kragen> oh, I thought you meant the material that gets ablated
20:14 < kanzure> .title http://lav.io/2014/11/stupid-projects-from-the-stupid-hackathon/
20:14 < yoleaux> Stupid Projects From The Stupid Hackathon – Sam Lavigne
20:14 < kragen> yeah, that's wasteful of energy
20:15 < fenn> ipad on a face is a good idea
20:15 < kanzure> "Egg Timer by Pam Liou is an hourglass that counts down to menopause."
20:16 < fenn> also profanity65 is good
20:19 < fenn> it should dynamically generate compound profanities though
20:19 < kragen> I don't understand profanity65
20:19 < fenn> cuntpunter asspunter horsepunter etc
20:20 < kragen> are the signatures it generates valid?  if so, how?
20:20 < fenn> ar symbol_index = originalSymbols.indexOf(original_line[sym]);        new_line.push(profanity[symbol_index]);
20:20 < fenn> you have to have a copy of the profanity list on both ends of course
20:21 < kragen> oh, it doesn't interoperate with standard PGP?
20:21 < kragen> that's cool
20:22 < fenn> so "a" becomes "anal", "b" becomes "ass", "c" becomes "asshat"
20:22 < kragen> sure.  you could improve efficiency a bit but whatever
20:25 < fenn> oh you can encrypt too
20:34 < fenn> "You don't need cables, pipes, gas or copper wires. We can send it to you like a cell phone call—where you want it and when you want it, in real time."
20:37 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-butqbhoardxhwipg] has quit [Quit: Connection closed for inactivity]
20:37 -!- TheShadowFog [~TheShadow@2601:8:3e80:c:7d71:2e6c:7049:6035] has quit [Ping timeout: 258 seconds]
20:41 < superkuh> Soldering isn't the issue. I just want to minimize the anode size while copying the design from Leopoldo Soto and Christian Pavez. They do not include much detail about how they prevent flashover of the capacitor plates in their parallel plate design. The anode punctures the dielectric. This diagram should make it more clear, dfp_annotated_anode_hole.png
20:41 < superkuh> Er, oops.
20:42 < superkuh> http://superkuh.com/dfp_annotated_anode_hole.png
20:50 < jrayhawk> http://scholar.google.com/scholar?q=sybil+reputation so much to read
21:01 -!- rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has quit [Ping timeout: 250 seconds]
21:07 -!- juri_ [~juri@vpn166.sdf.org] has joined ##hplusroadmap
21:32 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has joined ##hplusroadmap
21:39 -!- gene_hacker [~chatzilla@c-50-137-46-240.hsd1.or.comcast.net] has joined ##hplusroadmap
21:40 -!- Vutral [ss@mirbsd/special/Vutral] has quit [Ping timeout: 244 seconds]
22:05 -!- Vutral [ss@mirbsd/special/Vutral] has joined ##hplusroadmap
22:11 -!- yashgaroth [~ffffff@2606:6000:cb85:6a00:50a4:f569:684f:fe1b] has quit [Quit: Leaving]
22:22 -!- ebowden [~ebowden@147.69.147.27] has quit [Remote host closed the connection]
23:10 -!- ebowden [~ebowden@147.69.147.27] has joined ##hplusroadmap
23:56 < kanzure> http://i.imgur.com/5iHr8fz.jpg
--- Log closed Tue Nov 18 00:00:54 2014