--- Log opened Fri Oct 29 00:00:08 2010 00:12 -!- hundred-ideas [~100ideas@c-24-4-6-179.hsd1.ca.comcast.net] has quit [Quit: hundred-ideas] 00:22 < kanzure> more "take on the machine" video stuff http://www.vimby.com/video/sponsor/us/all/detail/10908 00:23 < kanzure> light-related fashion items? http://www.romanillumination.com/company.htm 00:24 < kanzure> $150/yard haha no thanks 00:35 -!- memorex [~durp@c-68-34-207-23.hsd1.tn.comcast.net] has joined #hplusroadmap 00:40 -!- phreedom_ [~quassel@109.254.17.41] has quit [Quit: No Ping reply in 180 seconds.] 00:40 -!- phreedom [~quassel@109.254.17.41] has joined #hplusroadmap 01:09 -!- archels [~neuralnet@541EEC68.cm-5-7d.dynamic.ziggo.nl] has quit [Ping timeout: 252 seconds] 01:19 -!- amaruk [~freeze@p54B29881.dip0.t-ipconnect.de] has joined #hplusroadmap 01:19 -!- archels [~neuralnet@541EEC68.cm-5-7d.dynamic.ziggo.nl] has joined #hplusroadmap 01:19 -!- amaruk [~freeze@p54B29881.dip0.t-ipconnect.de] has left #hplusroadmap [] 01:30 -!- memorex [~durp@c-68-34-207-23.hsd1.tn.comcast.net] has quit [Quit: memorex] 01:44 -!- killall-9 [~paulc@diana.null.ro] has joined #hplusroadmap 01:46 -!- memorex [~durp@c-68-34-207-23.hsd1.tn.comcast.net] has joined #hplusroadmap 01:51 < fenn> reprap for tissue engineering? 02:15 -!- jennifer2 [~jennifer@c-67-170-193-255.hsd1.ca.comcast.net] has quit [Ping timeout: 245 seconds] 02:29 -!- memorex [~durp@c-68-34-207-23.hsd1.tn.comcast.net] has quit [Read error: Connection reset by peer] 02:30 -!- memorex [~durp@c-68-34-207-23.hsd1.tn.comcast.net] has joined #hplusroadmap 02:55 < archels> http://www.kickstarter.com/projects/86514518/diy-arduino-eeg 02:55 < archels> ^ re-inventing the wheel for the 42nd time 02:56 < Utopiah> this could be a feature of kickstarter "How is your project different from this other project?" (here http://openeeg.sourceforge.net maybe) 02:57 < archels> yes, OpenEEG is the first that comes to mind. You can just re-use the analog board and use your own digital board (e.g. Arduino). 02:57 -!- memorex_ [~durp@c-68-34-207-23.hsd1.tn.comcast.net] has joined #hplusroadmap 02:59 -!- memorex [~durp@c-68-34-207-23.hsd1.tn.comcast.net] has quit [Ping timeout: 240 seconds] 02:59 -!- memorex_ is now known as memorex 03:00 < Utopiah> could share the information a lot better but then it could also quickly become less about inefficiency (read "personal fun") and maybe more political (read how you spread the money) 03:01 < Utopiah> I mean if their answer is "well we are not really different, just new" then people would be less willing to giving money but if they say "well we could work with them, set up a network on the topic" then it means money will be shared thus allocated 03:01 < archels> OpenEEG is about as open as open hardware can get, imho 03:02 < archels> not to mention the dozens of other homebrew EEG designs with (partially) open hardware & software. 03:02 < Utopiah> but since they started sth new, either they didn't know about it (thus telling them would help) or they knew but think they could do sth better, at least different 03:03 < archels> It's impossible they didn't know about it, unless they are Google-handicapped. 03:04 < archels> It looks like their goal is cheap & simple (I read the word "art" somewhere). 03:04 < archels> Which is all good and well, but why they need $1k for that is beyond me. 03:05 < Utopiah> to make marketing (read "knowledge diffusion") to get more money of course 03:05 -!- memorex [~durp@c-68-34-207-23.hsd1.tn.comcast.net] has quit [Quit: memorex] 03:06 < archels> heh. Their little circuit would probably be $10 cost price at qty 10~20. 03:13 < Utopiah> they do talk about OpenEEG in the cmment btw 03:14 < Utopiah> and to "4. What exactly is the $1000 for? Can you outline what sort of development and testing you're planning?" they replied "4. Now that we have a working prototype, our next step is cleaning up the board. Through hole soldering so many pieces in close proximity is pretty messy and I'd like the soldering on the final product to be more intuitive" 03:15 < archels> Artsy types shouldn't be allowed to do electrical engineering, anyway. 03:15 < archels> The result is inevitably crap, like the Arduino. 03:18 < Utopiah> right but I dont think they aim at replacing medical brain imaging, rather "democratizing" for any type of purpose especially for people with nearly no EE skilles 03:20 < archels> Which would be fine if they just went ahead and did it, instead of asking for $1k for no apparent reason. 03:22 < Utopiah> you think KickStarter should display in the front page of each project the total of money asked for and blocks representing expected usages? 03:23 < archels> No, I'm just ranting about this particular project. 03:24 < archels> What's that old saying again, the only thing more dangerous than a hardware guy with a code patch is a programmer with a soldering iron? 03:44 -!- memorex [~durp@c-68-34-207-23.hsd1.tn.comcast.net] has joined #hplusroadmap 03:53 -!- marainein [~marainein@220-253-112-146.VIC.netspace.net.au] has quit [Quit: Ex-Chat] 03:54 < Utopiah> (Guns Germs and Steel by Jared Diamond http://video.google.ca/videoplay?docid=-4008293090480628280# 03:57 -!- streety [~Jonathan@host86-135-137-122.range86-135.btcentralplus.com] has joined #hplusroadmap 03:59 < killall-9> heh! still using batteries to avoid noise in amplifier - did that 20 years ago 03:59 < killall-9> much cheaper than a low noise power supply 04:00 -!- streety1 [~Jonathan@93.182.130.14] has quit [Ping timeout: 245 seconds] 04:01 < killall-9> of couse 20 years ago we didn't had ICs for that - and computers had like 64k of RAM ;-) 04:13 -!- memorex [~durp@c-68-34-207-23.hsd1.tn.comcast.net] has quit [Quit: memorex] 04:34 -!- archels [~neuralnet@541EEC68.cm-5-7d.dynamic.ziggo.nl] has quit [Ping timeout: 264 seconds] 04:38 -!- archels [~neuralnet@541EEC68.cm-5-7d.dynamic.ziggo.nl] has joined #hplusroadmap 05:54 -!- streety1 [~s0678364@cpat001.wlan.net.ed.ac.uk] has joined #hplusroadmap 06:08 -!- eridu [debian-tor@gateway/tor-sasl/eridu] has joined #hplusroadmap 06:12 -!- archels [~neuralnet@541EEC68.cm-5-7d.dynamic.ziggo.nl] has quit [Ping timeout: 250 seconds] 06:23 -!- archels [~neuralnet@541EEC68.cm-5-7d.dynamic.ziggo.nl] has joined #hplusroadmap 06:34 -!- eridu [debian-tor@gateway/tor-sasl/eridu] has quit [Ping timeout: 245 seconds] 06:34 -!- augur [~augur@208.58.6.161] has quit [Read error: Connection reset by peer] 06:42 -!- mheld [~mheld@pool-173-76-224-45.bstnma.fios.verizon.net] has joined #hplusroadmap 07:06 -!- Juul [~Juul@h55eb1609.dkkoost.sta.perspektivbredband.net] has joined #hplusroadmap 07:16 -!- streety1 [~s0678364@cpat001.wlan.net.ed.ac.uk] has quit [Ping timeout: 245 seconds] 07:23 < kanzure> http://rapcad.org/ 07:23 < kanzure> http://gitorious.org/openscad 07:28 -!- JayDugger [~duggerj@pool-173-74-76-197.dllstx.fios.verizon.net] has joined #hplusroadmap 07:28 < JayDugger> Good morning, everyone. 07:32 -!- augur [~augur@129.2.129.34] has joined #hplusroadmap 07:41 < kanzure> more videos up http://telexlr8.wordpress.com/2010/10/25/transvision-2010-october-22-24-2010/ 07:42 -!- Juul [~Juul@h55eb1609.dkkoost.sta.perspektivbredband.net] has quit [Quit: Leaving] 07:42 < kanzure> Call for Book Chapters: FOSS for Sustainable Development in Developing Countries http://webserv.ias.unu.edu/fel/sites/default/files/Call%20for%20Book%20Chapters-FOSS_Sdev.pdf 07:42 < kanzure> (deadline: nov. 5) 07:42 < kanzure> contact: Sam Kritikos 07:44 -!- wolfspraul [~wolfsprau@lucia.q-ag.de] has quit [Quit: leaving] 07:46 < kanzure> ian is adding a python intrepreter to openscad.. 08:03 -!- eridu [debian-tor@gateway/tor-sasl/eridu] has joined #hplusroadmap 08:25 -!- Juul [~Juul@h55eb1609.dkkoost.sta.perspektivbredband.net] has joined #hplusroadmap 08:26 -!- codeshepherd [~Deepan@122.166.152.152] has quit [Quit: codeshepherd] 08:27 -!- eridu [debian-tor@gateway/tor-sasl/eridu] has quit [Ping timeout: 245 seconds] 08:28 -!- mheld [~mheld@pool-173-76-224-45.bstnma.fios.verizon.net] has quit [Quit: mheld] 08:31 -!- Juul [~Juul@h55eb1609.dkkoost.sta.perspektivbredband.net] has quit [Quit: Leaving] 08:39 < kanzure> cool, ed uthman's image got on to the mit.edu front page http://www.mit.edu/ 08:43 -!- eridu [debian-tor@gateway/tor-sasl/eridu] has joined #hplusroadmap 08:46 -!- hundred-ideas [~100ideas@c-24-4-6-179.hsd1.ca.comcast.net] has joined #hplusroadmap 09:02 -!- mheld [~mheld@216.214.247.202] has joined #hplusroadmap 09:03 -!- klafka [~klafka@129.21.73.176] has joined #hplusroadmap 09:09 < kanzure> xcor/masten interview video http://www.dminuszero.com/post/1374633309/everyday-sci-fi 09:12 < kanzure> "At a recent event in San Francisco, NASA Ames Research Center Director Pete Worden introduced the Hundred Year Starship initiative, a project to embark on a one-way mission from Earth to Mars by 2030 and permanently settle the red planet." 09:12 < kanzure> QuantumG: aren't dennis and pete good friends or something? 09:17 -!- mheld [~mheld@216.214.247.202] has quit [Ping timeout: 240 seconds] 09:17 -!- mheld [~mheld@74.61.205.248] has joined #hplusroadmap 09:24 -!- augur [~augur@129.2.129.34] has quit [Remote host closed the connection] 09:24 -!- augur [~augur@129.2.129.34] has joined #hplusroadmap 09:41 < kanzure> http://beanstalkapp.com/ looks like an alternative to unfuddle/github 09:46 < archels> Has this been spammed here yet? http://www.eurekalert.org/pub_releases/2010-10/ciot-cic102610.php 09:47 < klafka> ooh kristof koch 09:57 < kanzure> "more than 800 people in the pilot stage of the so-called 1,000 Genomes Project, which aims to complete 2,500 sequences by the end of 2012 at a cost of $120 million." 09:57 < kanzure> wait, it's hosting how much? 09:57 < kanzure> *costing 10:00 -!- streety1 [~s0678364@cpat001.wlan.net.ed.ac.uk] has joined #hplusroadmap 10:07 -!- augur [~augur@129.2.129.34] has quit [Remote host closed the connection] 10:15 < klafka> what ever happened to the 5,000 genome 10:17 < fenn> i don't get the whole "hundred year starship" thing 10:17 < fenn> first of all, it doesn't take 100 years 10:18 < fenn> second, why can't they return? huh? too hard to shovel some ice into your nuclear rocket's tanks? 10:18 < fenn> third, nuclear rockets? 10:19 < fenn> "What the hell is NASA and the international space programme doing asking the public for handouts?" 10:22 -!- Juul [~Juul@h55eb1609.dkkoost.sta.perspektivbredband.net] has joined #hplusroadmap 10:25 -!- hundred-ideas [~100ideas@c-24-4-6-179.hsd1.ca.comcast.net] has quit [Read error: Connection reset by peer] 10:36 < kanzure> fenn: some people have been telling me to approach pete worden with some stuff. 10:37 < kanzure> turns out i've been facebook buds with pete for a few years now 10:38 < fenn> he is probably getting his inbox slammed right now what with all the media attention 10:39 < kanzure> so, dennis wants to melt down some moon rock/ore at the ames campus and demonstrate a simple machine tool 10:39 -!- archels [~neuralnet@541EEC68.cm-5-7d.dynamic.ziggo.nl] has quit [Ping timeout: 240 seconds] 10:39 < kanzure> it's not on the top of his priorities list but if someone showed up and said "i'll do that" he'd like that very much 10:40 < fenn> "No bucks, no Buck Rogers!" 10:40 < kanzure> oh please 10:40 < kanzure> since when 10:41 < fenn> yeah well supposedly they hired worden because of his reputation of getting things done cheaply 10:41 < kanzure> story time! tell me tell me 10:42 < fenn> uh, based on my five minute google search 10:42 < kanzure> you suck at stories :( 10:42 < fenn> http://www.thespacereview.com/article/612/1 10:42 < fenn> "Worden is an innovator and a risk-taker whose record shows that he can accomplish difficult goals without depending on massive funding." 10:42 < fenn> who is dennis? 10:43 < fenn> i'd love to melt down some moon rock 10:43 < kanzure> this would require a few trips to mountain view 10:43 < kanzure> dennis is http://google.com/search?q=dennis+skycorp 10:47 -!- archels [~neuralnet@541EEC68.cm-5-7d.dynamic.ziggo.nl] has joined #hplusroadmap 10:52 < fenn> it'd be cool to make some kind of SLS machine out of a solar concentrator and moon dust 11:02 < kanzure> he agrees 11:09 < fenn> how do you know dennis? and talking now? 11:09 -!- memorex [~durp@c-68-34-207-23.hsd1.tn.comcast.net] has joined #hplusroadmap 11:10 * fenn reads about solar pumped lasers 11:15 -!- Juul [~Juul@h55eb1609.dkkoost.sta.perspektivbredband.net] has quit [Ping timeout: 240 seconds] 11:16 -!- Juul [~Juul@h55eb1609.dkkoost.sta.perspektivbredband.net] has joined #hplusroadmap 11:20 -!- archels [~neuralnet@541EEC68.cm-5-7d.dynamic.ziggo.nl] has quit [Ping timeout: 265 seconds] 11:20 -!- killall-9 [~paulc@diana.null.ro] has quit [Quit: Leaving] 11:31 -!- archels [~neuralnet@541EEC68.cm-5-7d.dynamic.ziggo.nl] has joined #hplusroadmap 11:35 -!- memorex [~durp@c-68-34-207-23.hsd1.tn.comcast.net] has quit [Remote host closed the connection] 12:01 < kanzure> fenn: via QuantumG 12:27 -!- augur [~augur@129.2.129.34] has joined #hplusroadmap 12:27 -!- augur [~augur@129.2.129.34] has quit [Remote host closed the connection] 12:28 -!- augur [~augur@129.2.129.34] has joined #hplusroadmap 12:28 -!- streety1 [~s0678364@cpat001.wlan.net.ed.ac.uk] has quit [Quit: Leaving.] 12:39 < kanzure> jrayhawk: how would piny handle 2M to 2.3M repos? 12:51 -!- phreedom [~quassel@109.254.17.41] has quit [Quit: No Ping reply in 180 seconds.] 12:52 -!- phreedom [~quassel@109.254.17.41] has joined #hplusroadmap 12:55 -!- phreedom [~quassel@109.254.17.41] has quit [Client Quit] 12:55 -!- eridu [debian-tor@gateway/tor-sasl/eridu] has quit [Ping timeout: 245 seconds] 12:57 < kanzure> "john lewis on asteroidal resources" http://www.youtube.com/watch?v=wYaf2ZE6LvY 12:57 -!- eridu [debian-tor@gateway/tor-sasl/eridu] has joined #hplusroadmap 12:58 -!- phreedom [~quassel@109.254.17.41] has joined #hplusroadmap 12:59 < kanzure> jrayhawk: i meant million, not megabytes 13:04 -!- alystair [Alystair@bas1-toronto10-1279397903.dsl.bell.ca] has joined #hplusroadmap 13:05 -!- phreedom [~quassel@109.254.17.41] has quit [Quit: No Ping reply in 180 seconds.] 13:05 -!- phreedom [~quassel@109.254.17.41] has joined #hplusroadmap 13:09 -!- phreedom [~quassel@109.254.17.41] has quit [Client Quit] 13:13 -!- phreedom [~quassel@109.254.17.41] has joined #hplusroadmap 13:16 -!- phreedom [~quassel@109.254.17.41] has quit [Client Quit] 13:19 < jrayhawk> ooh, good question. I *think* everything is at least 32-bit. 13:19 -!- phreedom [~quassel@109.254.17.41] has joined #hplusroadmap 13:21 < kanzure> no that's not what i'm asking 13:21 < kanzure> scalability, etc. 13:22 < jrayhawk> It wouldn't be particularly performant, but it would be doable. If you wanted this to happen fairly immediately, there'd be issues to work out. 13:24 < kanzure> yeah.. elevenarms wasn't happy when i told him about piny/perl :P 13:24 < kanzure> so he was asking about scalability 13:24 < kanzure> hm. 13:26 -!- phreedom [~quassel@109.254.17.41] has quit [Quit: No Ping reply in 180 seconds.] 13:26 < jrayhawk> That said it's not like the storage architecture is obscure, so it's not like they can't yoink everything and do it better elsewhere. 13:26 < kanzure> this is stupid.. let me get him in here 13:27 -!- phreedom [~quassel@109.254.17.41] has joined #hplusroadmap 13:28 < jrayhawk> I'll be disappearing for 20 minutes or so. 13:28 -!- jblake [~jblake@pool-96-225-226-221.ptldor.fios.verizon.net] has joined #hplusroadmap 13:29 < kanzure> okay. 13:29 < kanzure> jblake: he'll be around in a few minutes 13:29 < kanzure> uh. 30. 13:30 -!- jburk [601a22bf@gateway/web/freenode/ip.96.26.34.191] has joined #hplusroadmap 13:31 -!- augur [~augur@129.2.129.34] has quit [Ping timeout: 240 seconds] 13:31 < kanzure> hi jburk 13:31 < jburk> howdy 13:31 < kanzure> any MITRE updates? :D 13:32 < jburk> i dont know very much about that 13:33 < kanzure> oops i'm thinking of jim bark 13:33 < kanzure> haha 13:33 < kanzure> ok, well, how about marshome.org updates 13:34 < jburk> np 13:34 < jburk> what do you want to know 13:34 < kanzure> just if there's anything new :) 13:35 < jburk> yeah a lot going on with it 13:35 < jburk> we have a site in CA which we are planning to build a mars analog research center 13:36 -!- phreedom_ [~quassel@109.254.17.41] has joined #hplusroadmap 13:36 < kanzure> jburk: know anything about the one in the attacama? 13:36 < jburk> a little 13:36 < kanzure> same idea? 13:37 -!- phreedom [~quassel@109.254.17.41] has quit [Ping timeout: 276 seconds] 13:37 < jburk> that one is more desert and has university involvement there. i think it will be multi-use site, whereas the one in CA is only marshome.org 13:37 < kanzure> so are you deploying equipment? 13:38 -!- mheld [~mheld@74.61.205.248] has quit [Ping timeout: 255 seconds] 13:38 < jburk> we are doing site planning now but yes in 2011 13:38 < kanzure> cool. i've been talking with some people interested in doing some automated manufacturing planning (for mars/moon stuff) ala 13:39 < kanzure> http://www.islandone.org/MMSG/aasm/ 13:39 < jburk> haha same here 13:39 < jburk> that will be a component of our stuff including teleoperation 13:40 -!- phreedom [~quassel@109.254.17.41] has joined #hplusroadmap 13:40 < jburk> teleoperation of robots to build/develop the base is one of the features of the marshome.org engineering plan for 1st human settlement 13:41 -!- phreedom_ [~quassel@109.254.17.41] has quit [Ping timeout: 240 seconds] 13:41 < jburk> and the analog research center will be structured around those same topics/themes 13:41 < jburk> myself i am also working on an iphone game 13:42 < jburk> think a game version (like farmville or pizza parlor) of building the base or teleoperating rovers on moon/mars 13:42 < kanzure> why teleoperation- what about autonomous manufacturing? 13:42 < jburk> right 13:42 < kanzure> yeah, there's been some community suggestions of mmo simulations 13:42 < kanzure> realistic simulations poised as a game or something 13:42 < jburk> the mars colony online one is still going on, howard deutsch 13:43 < kanzure> maybe using freecity (the sims clone) or something 13:43 < kanzure> links? 13:43 < jburk> hyperkatgames.com\ 13:43 < kanzure> i might be interested in doing some programming 13:43 < jburk> sorry http://www.hyperkat.com/MarsOnline.html 13:43 < kanzure> (or machining) 13:43 < jburk> yeah if you know objective c or the iphone sdk i could use some help LOL 13:45 < jburk> that goes for anybody else reading this 5 days from now LOL - email me at jburk at 43tech.com 13:45 < jblake> i know haskell and have written microcontroller code to manage CDMA radios, does that help 13:46 < jburk> well go download the xcode stuff and play around with it, email me in a few days :) 13:47 < kanzure> i don't see what mars online has to do with iphone apps 13:47 < jburk> it doesnt you mentioned mmo 13:47 < kanzure> also you should just make it javascript/html5 and not bother with platform-specific code 13:48 < kanzure> jburk: another 12min. sorry for the lag. 13:48 < kanzure> oops 13:48 < jburk> thats a good suggestion 13:48 < kanzure> jblake: another 12min. sorry for the lag. 13:49 < jrayhawk> now i am sad i do not have a secondary name starting with b 13:49 < jrayhawk> i feel all left out 13:49 < jblake> It's really tempting to see if I can throw together a simple city simulator in HTML5+JS 13:49 < kanzure> joe brutus rayhaawk 13:49 < jblake> jrayhawk: jbutthead 13:50 < jburk> oh man 13:50 < kanzure> jblake: didn't sim city go open source at the end of its life? 13:50 < jblake> oh, probably 13:50 < kanzure> JayDugger: can you /nick jdugger for a few minutes? 13:50 < jrayhawk> there are a bunch of existing open source sim-games 13:50 < jrayhawk> jbugger 13:51 < jrayhawk> most of which are more sophisticated than sim city anyway 13:51 < jburk> http://en.wikipedia.org/wiki/List_of_open-source_video_games 13:51 < kanzure> jrayhawk: the idea has been (for a while) to do a simulator with skdb data 13:52 < kanzure> "you just upgraded to vertical milling machine stations! woo" 13:52 < kanzure> or "error: you must first forge the metal" 13:53 < jrayhawk> Ah, so Civilization 6: Your Garage 13:53 < jrayhawk> oh no, a wandering barbarian 13:53 < jrayhawk> the barbarian is impressed by your technological advancement and agrees to become your flunky 13:54 < jburk> ever play buzz aldrins race into space 13:54 < kanzure> was that a real game? 13:54 < jburk> yeah 13:54 < jburk> dos era 13:55 < jburk> its online 13:55 < jblake> Feh, HTML5 canvas doesn't appear to have the kind of expressive power that cairo has. Dealing without clipping regions is just too painful to be worthwhile. 13:56 < jburk> WOW somebodys doing an open source of that http://sourceforge.net/projects/raceintospace/ 13:56 < jburk> i should just join that LOL 13:57 < jburk> "hey everybody, want to do an iphone version?" 13:57 * kanzure pokes elevenarms 13:59 < archels> completely off-topic http://iwdrm.tumblr.com/ 14:02 < kanzure> timothyschmidt: ping 14:08 < jblake> this is the most productive conversation i've had about piny in /weeks/ 14:08 < kanzure> i just called him, he's jumping in :/ 14:08 < kanzure> basically elevenarms' concern is about scalability for hundreds of thousands of repos 14:08 < kanzure> and that flatfile hosting might not do the trick here 14:09 < kanzure> he'll probably recommend nosql as a solution 14:09 -!- jburk [601a22bf@gateway/web/freenode/ip.96.26.34.191] has quit [Ping timeout: 265 seconds] 14:09 < jblake> there are some administrative concerns with that many repos; a few operations like repo creation have global side effects that you'd probably want to batch and consolidate 14:09 < kanzure> for instance, if we were using piny for github 14:09 -!- elevenarms [~elevenarm@71.22.75.110] has joined #hplusroadmap 14:10 < kanzure> hi elevenarms 14:10 < elevenarms> hello all 14:10 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has joined #hplusroadmap 14:10 < elevenarms> kanzure catch me up 14:10 < kanzure> so, you were suggesting a database solution for scalability 14:10 < kanzure> jblake and jrayhawk work on a project called piny, it's basically a project management solution for git hosting 14:10 < elevenarms> link? 14:10 < kanzure> https://secure.piny.be/cgit/piny-web.git/ 14:11 < kanzure> http://piny.be/piny-code/architecture/needed_user_facing_infrastructure/ 14:11 < jrayhawk> http://piny.be/piny-code/architecture/needed_user_facing_infrastructure/ has the best description of what it'll look like 14:11 < jrayhawk> to users, anyway. 14:11 < jblake> it's really depressing that the list of things i haven't done yet is the definitive "this is piny" webpage :-( 14:11 < kanzure> so anyway, 14:11 < kanzure> the idea is that a thingiverse/github clone would have to host a lot of repos 14:12 < elevenarms> I'm concerned with hosting thousands of git repos 14:12 < kanzure> and that this would run into scalability issues 14:12 < kanzure> ok even that 14:12 < jblake> What sorts of scalability issues are you thinking of? 14:12 < elevenarms> how/where will they be stored 14:13 < kanzure> access times, especially into revision history 14:13 < elevenarms> what would the directory structure look like 14:13 < elevenarms> perhaps http://github.com/blog/606-announcing-ernie-2-0-and-2-1 14:13 < jrayhawk> I think cgit is the only thing that attempts to do operations based on every repo; that might have to be modified extensively. 14:14 < streety> doesn't github blog extensively on their infrastructure? won't man of the same solutions work for thingiverse? 14:14 < kanzure> cgit: http://hjemli.net/git/cgit/about/ 14:14 < jblake> At large scales, we'd probably have the repos on a SAN exported over NFS to the frontend machines, and just organized by partial name if we're concerned about the size of dentries. 14:14 < kanzure> streety: we're not the developers of thingiverse. i'm building a better thingiverse 14:14 < jblake> Piny doesn't do anywhere *near* as much work as github does at the moment. 14:14 < kanzure> streety: there's a lot of things wrong with thingiverse, blah blah blah 14:15 < streety> fair enough but the same point applies 14:15 < kanzure> sure, but if joe and jules want to provide some development work towards that, then why not take it? 14:15 < kanzure> gitorious too 14:15 < kanzure> ok so anyway 14:16 < jblake> As it stands, I suspect we could probably host in the tens-of-thousands of git repos from a single machine; once a repo is created there's essentially no global work that needs to be done and only the most trivial local work to host it. 14:17 < jblake> Administrative tasks would start getting painful at that scale, but we have plans to remove the bits that would be causing problems anyway. 14:18 < jblake> (The largest limiting factors afaik would involve some steps that recreate large config files during repo creation/deletion) 14:18 < jrayhawk> passwd/shadow/group/gshadow et. al can be moved into a bdb 14:19 < jblake> And the apache/cgit/ikiwiki stuff should all be moving to .d style configs, anyway. 14:19 < kanzure> http://develop.github.com/ 14:20 < jblake> cgit or ssh would probably be the main consumers of cpu time at that scale, but those both scale horizontally. 14:20 < jrayhawk> I already moved cgit to a new model. 14:20 < jblake> Oh, excellent. 14:20 < elevenarms> I think the github api is the way to go 14:21 < elevenarms> if were talking foss hardware, solving the project metadata problem is a greater issue 14:21 < jblake> Introducing an API like that would probably cost a lot of CPU and make scalability significantly harder than what we have now. Do you have a big use case for API access to the repos? 14:22 < jrayhawk> Apache might be a bit of a problem, actually. We've got a fair amount of per-project configuration that I'm not sure Apache would be happy to traverse on every request. 14:22 < elevenarms> are you talking about adding an api to piny? 14:22 < jblake> It caches its configuration in memory, doesn't it? 14:22 < kanzure> jblake: elevenarms is talking about writing a website that uses the github api instead of hosting git stuff on its own 14:22 < jblake> Oh, OK. 14:23 < jrayhawk> I'm not interested in getting involved in walled-garden development. 14:23 < jrayhawk> I'm sure it wouldn't be hard to make something useful in that ecosystem, though, so have at :) 14:24 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has quit [Ping timeout: 276 seconds] 14:26 < kanzure> in terms of development time, it is probably faster to just use ready-made solutions, and then replace the bits with custom solutions 14:26 < kanzure> well, actually, it could go either way 14:27 < kanzure> jblake: why are you guys not using gitorious? 14:27 < jblake> It's git. You can just start hosting anywhere now and move around later. 14:27 < kanzure> http://gitorious.org/ 14:27 < kanzure> huh? 14:27 < kanzure> i don't think you understand what the github api lets you do 14:27 < kanzure> elevenarms is talking about using that api so that they solve the scalability issues for you 14:28 < kanzure> (which, they have, somehow) 14:28 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has joined #hplusroadmap 14:28 < kanzure> http://gitorious.org/gitorious 14:28 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has quit [Read error: Connection reset by peer] 14:29 < jblake> Well, a big claim here is that there isn't really a huge scalability problem inherent in just hosting ikiwikis and git repos; static file hosting has been a solved problem for ages. 14:29 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has joined #hplusroadmap 14:30 < jblake> There *is* a scalability problem with doing things like searching across repos and other "big" dynamic tasks. 14:30 < jblake> But piny isn't currently targetting that at all. It's interesting if you guys have a strong use case for that, though, because that will give us some good input on what direction piny wants to move in. 14:31 < kanzure> huh gitorious hosts all of their repos in /gitorious/repos and they store metadata stuff in mysql 14:31 < jrayhawk> They're very interested in categorizing and aggregating information from all repos, but that's the closest thing they've got to dynamic that I've heard so far. 14:32 < jrayhawk> And that's really just a minutely update rather than true dynamicism. 14:32 < jrayhawk> or, rather, something that is just as well served by being accurate to 60 seconds as it is by being fully dynamic. 14:33 < jblake> As long as the kinds of queries would be known ahead of time you could probably get away with post-updates that manage an index and arrange for updates after each commit. That'd get you better than 60 seconds without involving any dynamic work at all. 14:34 < kanzure> i'm not sure how many commits/sec that would handle though 14:34 < kanzure> i guess you'd hit that same problem elsewise 14:34 < kanzure> is gitorious using their own custom git daemon? 14:34 < kanzure> they say they store their repos in /gitorious/ and to clone their src it's git clone git://gitorious.org/gitorious/mainline.git gitorious 14:35 < kanzure> so clearly they are doing at least some funky url rewriting for git paths 14:35 < kanzure> well, i guess if you ran into scalability with commits/sec problems with the post-commit hooks, you can just use post-commit hooks to add work to a queue on some server, and then just work through the queue as fast as possible 14:36 < jblake> Yeah, managing large indexes like that is a relatively well-studied problem. 14:39 < jblake> Anyway, I'm mostly concerned with more precisely what you would want to aggregate over and what kinds of github/gitorious features (like the social network) y'all are interested in? Those are really a dominating concern when it comes to scalability. 14:40 < jrayhawk> It'd be nice if Apache had something similar to Exim's embedded Perl so we could make the configuration stuff fully dynamic. It's really all defined by just one variable, the name of the repo, and 9 lines of templated configuration based on that variable. 14:41 < jblake> We might just want to ditch apache at some point for a lighter server, anyway. 14:41 < jrayhawk> That'd be hard; I don't think anything else supports the PAM user and group nonsense. 14:41 < kanzure> jblake: having a list of friends (and comparing lists to find mutual friends) is definitely something worth doing, but updating a list of friends isn't hard 14:41 < jblake> Well, in comparison to what we're doing now, updating a list of friends would be literally the hardest thing piny does. 14:42 < kanzure> so user profile data would have to be stored somewhere 14:42 < kanzure> presumably ikiwiki has a user page?? 14:42 < jrayhawk> What would 'friends' do? 14:42 < kanzure> that's not quite the same thing 14:42 < kanzure> jrayhawk: appease people 14:43 < jrayhawk> No seriously. 14:43 < jblake> Joe: We could do a mod_piny that virtualizes the configuration over all possible repos without too much trouble. 14:43 < kanzure> it's kind of like tags for figuring out what activity feeds to aggregate and show the user 14:43 < jrayhawk> Oh, so it'd just be an RSS system. Hmm. That's interesting. 14:43 < jblake> Ah, "activity feeds", another hard thing we don't do. 14:43 < jblake> At least, per-user feeds. Per-repo is trivially easy. 14:43 < kanzure> well yeah, these things are p. simple in a "dynamic" website (well, simple to implement quickly.. perhaps not immediately scalable) 14:44 < kanzure> updating every single static feed page for each user subscribed to a particular project, just sounds like it's going all backwards 14:45 < kanzure> instead of pulling the feeds when requested 14:45 < jblake> Eh. Commits happen much less often than pageviews. 14:45 < kanzure> i.e. you can imagine 500k users on github that subscribe to 20 projects each; only 10% of those users might actually ping their RSS feeds, or even log in 14:46 -!- marainein [~marainein@220-253-38-27.VIC.netspace.net.au] has joined #hplusroadmap 14:47 < jblake> With feeds I suppose you get weird numbers like that. Eh. You'd still want to build the per-repo feeds statically and just aggregate on demand, which preserves locality of dynamism. 14:49 < jrayhawk> The user feeds are a little weird, though, because there's only a handful of things you can reliably track. 14:49 < jblake> We don't really track committers at all. I'm not sure where we'd do that, we'd probably have to introduce some kind of a database. Ugh. 14:50 < kanzure> on github you track (via your feed): your friends' actions (not their entire feeds), projects you own/commit to, projects you watch 14:50 < jrayhawk> Tracking committers is just stupid; I don't think that will ever be done. 14:50 < jblake> I think that's what he's asking for with "your friends' actions" though. 14:50 < jrayhawk> Even github can't really do that, they can just make a note of who pushes when. 14:50 < jblake> Oh, that's easier. 14:50 < kanzure> what can't they do? 14:51 < kanzure> *what can't github do? 14:51 < jrayhawk> Commits are orthogonal to user identity; at no point does a user need to be associated with Piny in order to commit; only to push. 14:51 < kanzure> can't you sign a commit with your public key? 14:52 < jblake> Sure, but that doesn't have anything to do with piny. 14:52 < jblake> You can do whatever you want to your commits. 14:52 < kanzure> well, github keeps a copy of your public key 14:52 < kanzure> it sometimes links commits to actual users 14:52 < jrayhawk> Github keeps a copy of your public *ssh* key. 14:52 < jrayhawk> For purposes of giving you a transport to push over. 14:53 < kanzure> actually maybe it's just doing a username string match ("hurr we seem to have a 'kanzure' let's link to him") 14:53 < jrayhawk> Unless they have a hook that rejects unsigned commits that I don't know about. 14:53 < jrayhawk> That'd be a good idea, though. 14:53 < kanzure> but they aren't actually doing that for feeds 14:53 < jblake> And I think it's only doing committer-to-user association heuristically, yeah. 14:53 < kanzure> for feeds you primarily watch kanzure/some_repo not "all some_repos" 14:54 < jblake> by which you mean "watch all pushes to the repo named kanzure/some_repo", yes? 14:54 < kanzure> http://github.com/kanzure unless you're me and you're crazy and just watch everything 14:55 < kanzure> sort of- it really just means "tell me anything about the github status feed of kanzure/some_repo" which might be commits, but might also be "so-and-so forked this project" 14:55 < kanzure> (github only knows about forks when you click "fork" of course..) 14:56 < jblake> Whatever. So the repo knows how to build a feed for itself, and then the feed you consume is just the sum of several such feeds. 14:56 < kanzure> i think the link i just dropped is fairly informative 14:56 < kanzure> yeah 14:56 -!- eridu [debian-tor@gateway/tor-sasl/eridu] has quit [Quit: Leaving] 14:57 < jrayhawk> ugh, web-based rss sure is awful 14:57 < jrayhawk> rss readers, that is 14:57 -!- elevenarms [~elevenarm@71.22.75.110] has quit [Ping timeout: 252 seconds] 14:57 < kanzure> google reader? (i haven't tried it) 14:57 < jblake> OK, so that's nice because we can do that sum dynamically and still have no side effects. That kind of aggregation wouldn't be hard to scale. Generating the feed for each repo - doing a per-commit feed is trivial, per-push isn't much harder. I don't think we're interested in a "fork button" at this point, are we Joe? 14:58 < kanzure> bah fine, quit with a timeout 14:58 < kanzure> i think a "fork" would be important for people who (1) have soemthing to contribute and (2) don't know how to do git clones on their own computer 14:58 < kanzure> (joe might remember some talk about that sort of use scenario from the other day) 14:59 < kanzure> gee i should write a spec.. thing. 14:59 < jblake> Even after forking you'd still have to clone it to actually do work. 14:59 < jrayhawk> The 'fork' button is something I'm fine with and wouldn't be very hard to implement. 14:59 < kanzure> "fake fork"- wouldn't complete until the user actually provides a new set of diffs or whatever :P 14:59 < kanzure> (i'm half joking; i fork a lot of crap on github without updating it, so it's just sitting around) 15:00 < jrayhawk> That's hard for us and storage is cheap, so whatever. 15:00 < kanzure> *shrug* 15:00 < jrayhawk> Anyway, the fork button would be something like newrepo, turn off denynonfastforwards, clone origin, force push new, delete clone. 15:01 < jblake> If we get desperate for storage we'll just start unifying the object stores. But that's not worth considering for another few million repos. 15:01 < jrayhawk> and turn back on denynonfastforwards 15:01 < jrayhawk> We already would need to sort out automated temporary clones in order to do the big dumb 'upload zip file' function. 15:01 < kanzure> for hardware, these repos are probably going to be large :/ especialy since most users are going to be committing dumb stuff 15:01 < jblake> If the repo doesn't have any commits you can push anything you want. We shouldn't need to mangle denynonfastforwards for that. 15:02 < jrayhawk> Oh, yeah, I guess that's true. 15:02 < kanzure> hey so uh writing down this stuff is kinda important 15:02 < kanzure> how do i turn this into a spec so i don't forget 15:02 < jrayhawk> Our default action is to have a dumb commit just so Ikiwiki has something useful to clone. 15:02 < kanzure> s/so i don't forget/so joe doesn't forget/ 15:02 < jblake> You can generate commits against piny-web if you want, and just send-patch them to us. 15:03 < kanzure> yes but if i was to do that what would i write 15:03 < jrayhawk> Or you can create a user and we can add you to piny-code 15:03 < jblake> Or you can just count on the fact that I think all three of us are archiving this conversation. 15:03 < kanzure> archives are a bad way to store specs 15:04 -!- elevenarms [~elevenarm@71.22.75.110] has joined #hplusroadmap 15:04 < jrayhawk> Not that we've really specced anything. 'graft awful web rss reader system to piny user database' seems to be it. 15:05 < kanzure> elevenarms: http://gnusha.org/logs/2010-10-29.log for stuff you missed 15:05 < jblake> Actually generating the RSS feeds is more important. 15:05 < jrayhawk> I'll go add 'fork button' and 'upload button' to user-facing infrastructure, though. 15:05 < kanzure> jrayhawk: also the "ability for site admin to update the front page and overall layout/themes" thing. 15:06 < jblake> We need per-repo feeds for commits and pushes, and per-user feeds for pushes, which means we need per-user information stores and some code to catch pushes and manipulate said store. 15:06 < jrayhawk> per-repo feeds are already done by Ikiwiki 15:06 < jrayhawk> Or by cgit 15:06 < kanzure> also the cgit/ikiwiki integration 15:06 < jblake> Are those feeds commits or pushes or both? 15:06 < kanzure> both 15:06 < jrayhawk> Commits. 15:06 < jrayhawk> Not pushes. 15:06 < kanzure> ok fine 15:07 < kanzure> but also stuff like "user x started watching project y" and "forked project y" 15:07 < jblake> Because it sounds like kanzure is expecting pushes if he wants it to be like github. 15:07 < jblake> ugh i have a fairly strong aversion to "started watching" 15:07 < jblake> really not looking forward to solving the social networking problem 15:08 < kanzure> it's a solved problem i thought :P 15:08 < jrayhawk> No, there are a lot of complicated social issues, like privacy. 15:08 < jrayhawk> The way to solve it is to fuck metametadata 15:08 < jrayhawk> If someone wants to provide that elsewhere, they can go ahead. 15:10 < kanzure> well, access controls on individual repos would be nice i guess 15:11 < kanzure> i.e., a list of allowed users to view a repo, public/private repos 15:11 < jblake> Got those for commits, but everything is always readable at the moment. 15:11 < jrayhawk> Yeah, that's already on the TODO list. 15:12 < jrayhawk> That's why, for instance, as you asked earlier, cgit is under https:// 15:18 -!- shepazu [~schepers@adsl-242-206-152.rmo.bellsouth.net] has quit [Quit: shepazu] 15:23 < kanzure> jrayhawk: adding me to piny-web, piny-code, piny-shared would be hawt. 15:24 < jrayhawk> You'll have to createuser@piny.be if you want to edit them directly. 15:24 < kanzure> what? 15:24 < kanzure> i don't have ssh access on piny.be 15:25 < jrayhawk> That's why you ssh createuser@piny.be 15:25 < kanzure> hrm. 15:25 < kanzure> zerosignal.nl? 15:25 < kanzure> Last login: Fri Oct 15 03:46:06 2010 from zerosignal.nl 15:25 < jrayhawk> ha ha whoops, that's a privacy mistake 15:26 < kanzure> am i able to set up my own authorized_key or does piny still need you to do that 15:26 < jrayhawk> readkeys, writekeys, appendkeys 15:26 < kanzure> yay 15:27 < jrayhawk> I do not recall whether I actually tested appendkeys. 15:28 < kanzure> a list of allowed commands upon login to pinyshell would be cool in the future too 15:28 < jrayhawk> oh yeah, help system. Good idea. 15:29 < kanzure> "man addaccess" fails 15:30 < jrayhawk> I just added you to those. 15:30 < kanzure> ls /srv/git/ works, but ls doesn't work 15:31 < jrayhawk> Hmm. I should turn tab completion off. 15:31 < jrayhawk> well, for arguments. 15:31 < jrayhawk> Command completion is still useful. 15:31 < jblake> There is a manpage for addaccess? 15:31 -!- klafka [~klafka@129.21.73.176] has quit [Quit: This computer has gone to sleep] 15:31 < jblake> I wonder why that didn't work for you. 15:31 < kanzure> "man addaccess" works on gnusha for me 15:31 < jrayhawk> He's on pinyshell, he can't man anything 15:31 < kanzure> probably because 'man' doesn't work 15:32 < jblake> Oh, feh. 15:32 < jblake> We should probably add man to the safe commands. 15:32 < kanzure> afraid i'll start readin' and lurnin'? 15:32 < jrayhawk> Ehhh. We'd need to restrict pagers as well. 15:32 < jblake> I think both less and more just use whatever your $SHELL is. 15:33 < jrayhawk> less has 'log to file', though. 15:33 < jblake> oh yeah, feh 15:33 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has quit [Ping timeout: 240 seconds] 15:33 < kanzure> what is the difference between piny-web, piny-shared and piny-code 15:33 < jblake> I've been wondering that myself. 15:34 < jrayhawk> Piny-web was supposed to be about the hosting itself, piny-code is the engine, and piny-shared is a repository full of static resources other repositories need. 15:34 < kanzure> uh, what? 15:34 < jblake> Anyway, man without a pager is still reasonably useful. 15:34 < kanzure> sounds like piny-web should be web/ in piny-code.git 15:34 < jrayhawk> I have not done much to piny-web, I need to refactor how piny-shared works. 15:35 < jrayhawk> Well, piny-code is an orthogonal concept to piny-web. Anyone can deploy piny-code. 15:35 < kanzure> then why does piny-code and piny-web both have architecture/* 15:36 < jrayhawk> That was a throwback to when I was documenting piny-code in piny-web because I am a bad person. 15:36 < kanzure> i still don't see how piny-web shouldn't be in piny-code 15:36 < kanzure> piny-web would be the project website describing piny-code, right? 15:37 < jrayhawk> 'piny' is an overloaded concept that I might need to split. 15:37 < kanzure> piny.be is an instance of the piny codebase 15:37 < jrayhawk> At the moment piny is both the hosting service at http://piny.be/ and the engine driving the hosting service, piny-code.git 15:37 < kanzure> http://piny.be/ looks like it's piny-web.git 15:38 < jrayhawk> That's correct. 15:38 < kanzure> right.. so when you deploy piny-code.git it would make sense to me for it to be a copy of piny.be 15:38 < kanzure> and then you just set it up however you want for whatever you're hosting 15:39 < jblake> piny-code is easily amenable to being packaged as something a distro might ship. piny-web isn't. 15:39 < kanzure> huh? wouldn't you just put into the .deb package-creation-script to ignore web/ in piny-code or something 15:40 < jblake> So you could apt-get install piny-tools for example and get the interesting bits of our infrastructure without completely taking over your system. 15:40 < jrayhawk> But piny-web is specific to my hosting service. That's the point. 15:40 < jrayhawk> Not that it matters much at this point since I haven't put anything into it. 15:41 < kanzure> it just seems a little weird for piny to not be hosting piny. 15:41 < jrayhawk> Piny is two concepts. 15:42 < jrayhawk> Perhaps I should rename the engine to lolgit 15:43 < jrayhawk> Anyway, you're free to put whatever wherever and we can argue about it if I manage to care enough to complain at you. 15:43 < kanzure> three repos just doesn't make sense to me :P 15:44 < kanzure> anyway, how about piny-shared? 15:44 < jrayhawk> if it makes you feel better we can redirect / to piny-code/ until I feel like developing piny.be-the-hosting-service more thoroughly. 15:45 < jrayhawk> Piny-shared is a bunch of static files Ikiwiki and cgit currently need to operate. I may or may not continue using it depending on some other stuff I need to decide. 15:46 < jrayhawk> If you were to want to make a pretty layout or something you could shove it in piny-code. 15:46 < jrayhawk> pretty CSS layout 15:47 < kanzure> maybe you could give me a tour of piny-code/ and why there's an etc/ and usr/ (i've looked at them and know what's in there, but not why) 15:47 < kanzure> like why usr/src/ instead of just src/ 15:48 < jblake> joe is bad at organizing repos 15:48 < jblake> usr used to be even worse before i attacked it 15:48 < kanzure> great. 15:49 < jblake> (it was originally a partial filesystem snapshot, but has been gradually migrating into real packages) 15:49 < kanzure> sigh 15:50 < jrayhawk> Buildable packages are in /usr/src, anything that hasn't made its way into the buildable packages (because Jules it too lazy to work out how to use debconf to do everything we want, for the most part) lives outside of /usr/src 15:51 < jblake> Some of the stuff in etc, for example, is config files which belong to packages I don't control, so it's sort of a pain to make policy-compliant packages for them. 15:51 < kanzure> this is all awful 15:51 < jrayhawk> haha 15:51 < jrayhawk> you can git mv usr/src src if it makes you feel better 15:52 < jblake> srv/templates probably needs to get packaged to live somewhere in /usr/share but I haven't really decided who should own it, yet. 15:52 < jrayhawk> We should really relocate all the /srv stuff anyway. 15:52 < kanzure> i'm not really sure how to reorganize this productively. i'd probably move libpiny and pinyadmin to the top-level 15:52 < jblake> Basically everything else in that repo should probably get moved to a doc directory now that we supposedly will support those properly. 15:53 < jblake> Yeah, toplevel with pinyadmin and libpiny and doc and maybe one or two other debian packages is sort of where I'd like to head in the long term. 15:56 < jblake> It also probably needs some smarts to build webpages from the generated manpages in pinyadmin, which is going to be all kinds of ugly and I'm none too excited about. 16:01 -!- streety [~Jonathan@host86-135-137-122.range86-135.btcentralplus.com] has left #hplusroadmap [] 16:02 < kanzure> jblake, jrayhawk: ok i pushed 16:02 < kanzure> http://piny.be/piny-code/architecture/needed_user_facing_infrastructure/ 16:03 < kanzure> this file is awful 16:03 < kanzure> hrm. 16:03 < jblake> I agree. I hate that file. 16:03 < jrayhawk> Still more graceful than actual html tables! 16:04 < kanzure> these tables are stupid 16:04 < jrayhawk> But yes, basically dumb. 16:04 < jblake> It's ugly and a constant reminder of my failures of motivation. 16:04 < kanzure> uhuh 16:04 < kanzure> what's with having to have .mdwn to all of these files 16:04 < kanzure> what's wrong with plaintext 16:05 < jrayhawk> Markdown is more flexible once you get used to it. 16:05 < kanzure> i'm also not happy about issues/ 16:05 < kanzure> jrayhawk: what distributed bug system does fossil-scm use? 16:05 < jrayhawk> fossil 16:06 < jrayhawk> fossil is entirely self-contained for ease of deployability. 16:07 < jblake> hence we can't easily use it 16:07 < kanzure> too bad bugseverywhere doesn't believe in releases 16:07 < jrayhawk> I can whip issues into better shape for you if you'd like. 16:07 < kanzure> i should submit that as a bug to them 16:07 < kanzure> jrayhawk: i don't trust you with repos any more 16:08 < kanzure> at least, structuring them well 16:08 < jrayhawk> Well, you don't really know what Ikiwiki's bug tracking looks like at this point, so someone should probably give you an example. 16:08 < kanzure> hmm 16:08 < kanzure> okay fine 16:08 < jblake> Can you get away with showing his Cas's tracker? 16:09 < jblake> Eh, I guess ikiwiki's may be a better example, anyway. 16:09 < jrayhawk> No, cas's is definitely more sophisticated than ikiwiki.git's 16:09 < jrayhawk> Does PSAS do something trackery? 16:09 < jblake> I doubt it. Andrew would know better than I. 16:10 < jrayhawk> eh, I may as well just do it here. 16:11 < kanzure> i wish the "clear" command was hooked up to a shocker so that every time i needlessly clear my terminal i get shocked 16:11 < jblake> control-L is faster, by the way 16:12 < kanzure> you're not helping 16:12 < jblake> sure i am; this way it doesn't clutter your .bash_history! 16:12 -!- mheld [~mheld@pool-173-76-224-45.bstnma.fios.verizon.net] has joined #hplusroadmap 16:12 < jblake> also contributes less to carpal tunnel 16:13 < kanzure> what is ikiwiki's built-in bug tracker and why should you use it over BE? 16:13 < kanzure> and why aren't the todo/features/wishlists/issues in a bug tracker 16:13 < jrayhawk> I'M WORKING ON IT JEEZ OKAY GOD YOU ARE SUCH A SLAVEDRIVER 16:14 < jblake> i think while joe is being productive i'm going to go play a game involving zombie cowboys 16:20 -!- jblake [~jblake@pool-96-225-226-221.ptldor.fios.verizon.net] has quit [Ping timeout: 265 seconds] 16:25 -!- wolfspraul [~wolfsprau@lucia.q-ag.de] has joined #hplusroadmap 16:31 < kanzure> hi wolfspraul 16:31 < kanzure> wolfspraul: i was wondering, how did you finance ben nanonote? 16:33 < kanzure> "how to build your own UAV for $300" http://spectrum.ieee.org/automaton/robotics/diy/robots-podcast-diy-drones 16:33 < kanzure> this article doesn't have anything new in it, but the title is relevant IMHO http://blog.databazaar.com/2010/10/will-the-open-source-hardware-movement-bring-you-a-free-printer.html 16:33 < kanzure> (namely, there has never been a free printer, ever) 16:35 < Utopiah> echo "There is not free meal." | sed -e "s/meal/printer/" 16:36 < kanzure> no, i mean the gutenberg printer was supposed to be this era of free press and other BS, but it was actually always a proprietary, licensed machine 16:37 < Utopiah> political tool 16:37 -!- elevenarms [~elevenarm@71.22.75.110] has quit [Quit: elevenarms] 16:40 < kanzure> jrayhawk: what's your workflow for piny development? do you write code, make the deb package, uninstall previous .deb, install latest, then test? 16:40 < kanzure> or are you just making changes on your filesystem with symlinks 16:41 < jrayhawk> write code, make deb package, install deb package on dev.piny.be, test on dev.piny.be 16:41 < jrayhawk> If you createuser@dev.piny.be I can switch your shell over to something less restrictive so you can mess around. 16:42 < kanzure> ssh wants a password for createuser 16:42 < jrayhawk> Any new commands need to be symlinked into /srv/rbin and enabled in sudoers.d. 16:42 < jrayhawk> oh, that's interesting. 16:42 < jrayhawk> I guess I disabled that for dev. 16:51 -!- phreedom [~quassel@109.254.17.41] has quit [Remote host closed the connection] 16:51 -!- phreedom [~quassel@109.254.17.41] has joined #hplusroadmap 16:53 < kanzure> oh man, bugseverywhere is in worse of a state than i thought 16:53 < kanzure> git clone git://gitorious.org/be/be 16:53 < kanzure> try: "be list" on itself 16:54 < jrayhawk> http://piny.be/piny-code/ and http://piny.be/piny-code/issues/ should be a little more interesting, now. 17:00 < kanzure> blerg but i don't wanna merge 17:01 < jrayhawk> You have some unmerged issues? 17:02 < jrayhawk> if so, you can apt-get install moreutils and for i in list of files; cat ../templates/issues.mdwn $i | sponge $i; done 17:02 < QuantumG> sounds like you need to see a psychologist. 17:02 < kanzure> jrayhawk: no, unmerged unrelated other changes. it's ok, i'll live :) 17:03 < kanzure> what is tag/ 17:03 < jrayhawk> Pages that are autocreated whenever someone uses a [[!taglink]] directive. 17:04 < jrayhawk> So, for instance, I can point rss2email at http://piny.be/piny-code/tag/jrayhawk/index.rss to follow my bugs. 17:04 < kanzure> what is the difference between srv/templates/ and templates/ 17:05 < jrayhawk> /srv/templates/ is part of what needs to be installed on piny. It'll get moved somewhere more explicable once we refactor the install paths. 17:06 < kanzure> i grepped through everything and nowhere is srv/templates/ mentioned 17:06 < kanzure> so it seems to be unreferenced 17:07 < jrayhawk> ./usr/src/libpiny/share/ikiwiki.setup:templatedir => "/srv/templates", # TODO: user-customizable templates 17:08 < kanzure> yes but that's a reference to the installed path, not the path in the git repository 17:08 < kanzure> but, grep -r has failed me anyway 17:09 < jrayhawk> Yeah, that's one of the things that needs manual management outside of dpkg. Everything else is documented in http://piny.be/piny-code/issues/debian_packaging_concerns/ 17:09 < kanzure> so is there an ikiwiki interface to the issues now, or what? 17:09 < kanzure> presumably you were showing me something 17:14 < kanzure> i guess there's stuff like https://secure.piny.be/repos/piny-code/ikiwiki.cgi?page=issues%2Fwmd&do=edit 17:14 < kanzure> but it seems to be just "edit a section of this document" thingy? 17:15 < ybit2> bleep blop 17:15 < kanzure> hi ybit2 17:15 < kanzure> did you pick up that vertical cnc machine center? 17:17 -!- memorex [~durp@c-68-34-207-23.hsd1.tn.comcast.net] has joined #hplusroadmap 17:29 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has joined #hplusroadmap 17:32 < kanzure> jrayhawk: pruned the tree a bit https://secure.piny.be/cgit/piny-code.git/tree/ 17:32 < jrayhawk> So I see. 17:32 < kanzure> i suppose tag/ could be moved 17:32 < kanzure> but it seems useful to ikiwiki in general 17:33 < kanzure> (same with template/) 17:33 < jrayhawk> Yeah, we should probably move Ikiwiki over to doc/ anyway. 17:33 < kanzure> me or you 17:33 < jrayhawk> Well, I suppose Jules should implement that as a pinyconfig thing first. 17:33 < kanzure> the best answer of all. 17:35 < kanzure> so where is ikiwiki getting favicon.ico (and others) from? http://piny.be/piny-code/ 17:36 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has quit [Ping timeout: 272 seconds] 17:39 < jrayhawk> It isn't. You could define one in /srv/templates/page.tmpl if you wanted, though just adding http://ikiwiki.info/plugins/favicon/ is a better idea, I guess. 17:41 < kanzure> http://piny.be/piny-code/favicon.ico seems to exist to me. 17:42 < jrayhawk> oh, that's an underlay item. Hmm. I did a reckless ikiwiki upgrade I need to clean some stuff up for. 17:43 < jrayhawk> One of Ikiwiki's more irritating features is this concept of an 'underlay' that Joey just assumes people will want in every Ikiwiki. 17:43 < jrayhawk> One of the hopes was that I could use piny-shared as a way of avoiding that, but there's a lot of manual overhead and I need to decide what to do with it. 17:44 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has joined #hplusroadmap 17:56 < kanzure> jrayhawk: testing piny-tools, libpiny and pinyadmin seems kind of painful 17:56 < kanzure> isn't this what metasploit (et al.) was created for? 17:56 < jrayhawk> Yeah, we should really have regression testing. 17:56 < kanzure> i guess metasploit is too extreme for this 17:57 < kanzure> i meant: "a simple vm image to boot up for a developer to test a new installation on" 17:57 < kanzure> although, it's fairly annoying to have to rebuild the .deb 17:57 < kanzure> hrm. 17:59 < jrayhawk> It'd be easiest to grab the entirity of dev, I guess. Let me go see what's going on with createuser. 17:59 < kanzure> even with dev.piny.be the development workflow still sucks 18:01 < jrayhawk> We are developing debian packages. 18:02 < jrayhawk> Actually ensuring those packages are usable is not an optional part of the workflow. 18:05 < kanzure> but just for testing even simple changes? 18:05 < kanzure> i.e. the set of simple changes that can't be tested in isolation of the other piny infrastructure 18:06 < jrayhawk> Perhaps you are irrationally terrified of building and installing packages? 18:06 -!- memorex [~durp@c-68-34-207-23.hsd1.tn.comcast.net] has quit [Quit: memorex] 18:06 < kanzure> quite possibly 18:07 < jrayhawk> the commands you'd want would be 'dpkg-buildpackage -tc' and 'debi' 18:07 < jrayhawk> err, sudo debi 18:09 < jrayhawk> well, prepended by 'cd piny-code/libpiny' 18:09 < jrayhawk> or piny-code/pinyadmin 18:12 < kanzure> so what if multiple people are testing wacky shit out on dev.piny.be at the same time? 18:13 < jrayhawk> At that point, 17:59 < jrayhawk> It'd be easiest to grab the entirity of dev, I guess. 18:14 < jrayhawk> Hmm. I guess an openssh-server upgrade broke PAM. Lame. Anyway, createuser is working on dev.piny.be if you want to make an account I can enable the shell on. 18:15 < kanzure> jrayhawk: ok done 18:15 < jrayhawk> Should be better, now. 18:16 < jrayhawk> ugh. using git mv instead of Ikiwiki's move means all the paths in the markdown and directives are broken. 18:16 < kanzure> didn't i fix that? 18:16 < kanzure> (manually) 18:16 < jrayhawk> http://piny.be/piny-code/docs/architecture/ 18:16 < kanzure> oh, i didn't see any in there (i tried grep "\[\[" for links) 18:17 < jrayhawk> https://secure.piny.be/cgit/piny-code.git/tree/docs/architecture.mdwn 18:17 < kanzure> what's the propagate-your-authorized-keys ssh command thingy? 18:18 < jrayhawk> writekeys, though you should have a real shell now, so 'mkdir .ssh; cat > .ssh/authorized_keys' should work similarly well. 18:18 * kanzure defaults to writekeys 18:18 < kanzure> well, yes, but there's actually an ssh script provided by openssh 18:18 < kanzure> the name escapes me 18:18 < jrayhawk> oh 18:19 < jrayhawk> ssh-copy-id i guess 18:19 < kanzure> aha 18:20 < jrayhawk> huh, i never knew about that. 18:20 < kanzure> /home looks awful with ikiwiki everywhere :P 18:22 < jrayhawk> oh, I just added you to sudo; you'll need to relogin for that to work 18:24 < jrayhawk> You should avoid running Ikiwiki directly under any circumstances because it is incredibly easy to screw up with it and break things in subtle ways. 'rebuildrepo' is the command you want. 18:24 < jrayhawk> Other than that I think you're good to break stuff as you like. 18:24 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has quit [Ping timeout: 276 seconds] 18:36 -!- eridu [debian-tor@gateway/tor-sasl/eridu] has joined #hplusroadmap 18:40 -!- marainein [~marainein@220-253-38-27.VIC.netspace.net.au] has quit [Read error: Operation timed out] 18:50 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has joined #hplusroadmap 18:55 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has quit [Ping timeout: 252 seconds] 19:01 -!- flamt [~root@bas5-barrie18-2925347894.dsl.bell.ca] has joined #hplusroadmap 19:11 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has joined #hplusroadmap 19:14 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has quit [Read error: Operation timed out] 19:16 -!- Juul [~Juul@h55eb1609.dkkoost.sta.perspektivbredband.net] has quit [Quit: Leaving] 19:54 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has joined #hplusroadmap 20:07 -!- augur [~augur@208-59-167-26.c3-0.slvr-ubr1.lnh-slvr.md.cable.rcn.com] has quit [Ping timeout: 245 seconds] 20:10 < kanzure> fenn: are you still interested in going after vehicleforge, in spite of what we heard during the webinar? 20:29 -!- jblake [~jblake@pool-96-225-226-221.ptldor.fios.verizon.net] has joined #hplusroadmap 20:33 -!- flamt [~root@bas5-barrie18-2925347894.dsl.bell.ca] has quit [Ping timeout: 276 seconds] 20:38 -!- jblake [~jblake@pool-96-225-226-221.ptldor.fios.verizon.net] has left #hplusroadmap [] 20:51 < fenn> what in particular was heard in the webinar? besides the whole soul-sucking aspect, which we knew from the start 20:53 < kanzure> that it's just a repository for whatever language-format the other DARPA performers (who responded to META) come up with 20:53 < kanzure> not actually CAD or stuff like that. 20:55 < kanzure> i don't see this as an opportunity to do what we want to do in a reasonable way 20:56 < kanzure> "reasonable" is the operative word there 20:56 < kanzure> so i'm much less inclined to want to be at the helm of a VF project 20:57 < kanzure> still, todd's contact is still moving forward and has redhat involved and a few others, it's going to end up being a large group of people 20:57 < kanzure> so being involved in that way is still possible, if that's of any interest 21:01 * fenn sleeps on it 21:02 < kanzure> did you createuser yourself on piny.be? 21:07 < kanzure> jrayhawk: if on a homepage hosted by piny (for the entire server) the user logs in and gets an aggregated feed of what's up, how is that not 'dynamic'? 21:08 < kanzure> (i think this 'dynamic' term needs better definition) 21:09 < jrayhawk> We haven't decided how to do that part yet. 21:09 < kanzure> is piny-web, in your case, that "front page" ? 21:09 < jrayhawk> The dynamicity is fairly trivial implementation detail once we have all the data set up that we need. 21:10 < kanzure> yes but i thought you were bloodlustingly avoiding dynamic pages 21:10 < kanzure> bloodlustfully 21:10 < jrayhawk> With the user case it might be inevitable. Depend on how a number of other things work out. 21:11 < jrayhawk> s/Depend/Depends 21:12 < kanzure> okay. 21:22 -!- ybit [~quassel@unaffiliated/ybit] has joined #hplusroadmap 21:23 -!- eridu [debian-tor@gateway/tor-sasl/eridu] has quit [Remote host closed the connection] 21:28 -!- eridu [debian-tor@gateway/tor-sasl/eridu] has joined #hplusroadmap 21:30 < kanzure> http://github.com/mojombo/egitd 21:30 < kanzure> apparently they don't use that anymore 21:37 < fenn> "Is Robin Hanson a humanoid computer sent from the future to convince humans to agree to upload our future brains onto computers? Do the future computers need to devour our minds because they have learned all the knowable facts in the universe but been unable to logically deduce love? Or are they hoping to gain strategic insight from our brains for the future man vs computer war? Is Bryan Caplan our John Conner?" 21:38 < kanzure> that's not funny, robin is genuinely looney 21:40 < kanzure> jrayhawk: how would i go about moving ikiwiki into docs/ (or some customizable folder)? 21:41 < kanzure> i'm guessing piny_ikiwikidestdir or piny_ikiwikisrcdir in piny-code/libpiny/lib/Piny/Config.pm 21:41 < kanzure> but it might be something more involved 21:41 -!- eridu [debian-tor@gateway/tor-sasl/eridu] has quit [Ping timeout: 245 seconds] 21:41 < kanzure> wait, that's definitely wrong 21:42 < kanzure> i should just break piny-code.git and have jules fix it by solving it 21:44 < kanzure> i also think wikipath should be a per-repo configurable. 21:45 < jrayhawk> Yeah, that's on the TODO list. 21:47 < kanzure> really i just don't want to see tags/ templates/ and *.mdwm all the time in a development repo :P 21:47 < kanzure> s/tags/tag 21:48 < kanzure> oh well. maybe i'll figure out something else to do 21:48 < jrayhawk> You can get it minimally working by arranging for srcdir to be changed in the generated setup file. 21:49 < jrayhawk> There should probably also be a means of cleaning up the destdir and the leftover .ikiwiki directory, but that's less essential. 21:50 < kanzure> eh i feel there's a few things missing from http://piny.be/piny-code/docs/architecture/needed_user_facing_infrastructure/ 21:50 < kanzure> based on what we've been talking about 21:52 < kanzure> you mentioned ikiwiki/cgit integration? or was that me 21:52 < jrayhawk> I might have, but I don't remember the context. 21:53 < kanzure> for some reason "Page History" in ikiwiki links over to cgit 21:53 < kanzure> on a related note, cgit doesn't seem to do --follow for history.. i.e.: https://secure.piny.be/cgit/piny-code.git/log/docs/architecture/needed_user_facing_infrastructure.mdwn 21:54 < kanzure> i would imagine that http://piny.be/piny-code/libpiny/ should still be under the ikiwiki interface too.. 21:56 < jrayhawk> Jules gets upset when autoindex adds markdown files. 21:56 < jrayhawk> You can duke that out with him. 21:56 < kanzure> adds markdown files? 21:57 < kanzure> i don't want to add markdown files 21:58 < kanzure> i guess i can edit apache's default index page for that sort of subdirectory 21:59 < kanzure> this is stupid. 22:08 < kanzure> ah, .htaccess is slightly less of a stupid solution 22:09 -!- shepazu [~schepers@c-98-218-227-232.hsd1.dc.comcast.net] has joined #hplusroadmap 22:10 < kanzure> jrayhawk: what happens if i commit .htaccess at the moment? and should it be overwritten server-side/disallowed? 22:24 < JayDugger> Good morning, everyone. 22:30 < JayDugger> There also exists a board game of developing space named Rocket Flight. http://www.boardgamegeek.com/boardgame/5525/rocket-flight 22:31 < JayDugger> Followed by a better successor, High Trader, which seems stuck in permanent private playtest at Ad Astra Games. 22:32 < JayDugger> Rocket Flight has Santa Claus Machines near the top of its tech tree. 22:37 < jrayhawk> There are no AllowOverrides 22:40 -!- flamt [~root@bas5-barrie18-2925347894.dsl.bell.ca] has joined #hplusroadmap 22:45 -!- marainein [~marainein@220-253-49-99.VIC.netspace.net.au] has joined #hplusroadmap 23:09 < JayDugger> kanzure, have you met Robin Hanson in person? 23:12 -!- codeshepherd [~Deepan@122.166.152.152] has joined #hplusroadmap 23:14 -!- memorex [~durp@c-68-34-207-23.hsd1.tn.comcast.net] has joined #hplusroadmap 23:24 -!- ybit [~quassel@unaffiliated/ybit] has quit [Remote host closed the connection] 23:24 -!- ybit2 [~ybit@unaffiliated/ybit] has quit [Remote host closed the connection] 23:27 -!- wolfspraul [~wolfsprau@lucia.q-ag.de] has quit [Read error: Connection reset by peer] 23:28 -!- wolfspraul [~wolfsprau@lucia.q-ag.de] has joined #hplusroadmap 23:37 -!- augur [~augur@208.58.6.161] has joined #hplusroadmap 23:44 -!- augur [~augur@208.58.6.161] has quit [Remote host closed the connection] 23:44 -!- augur [~augur@208.58.6.161] has joined #hplusroadmap --- Log closed Sat Oct 30 00:00:07 2010