2009-09-16.log

--- Day changed Wed Sep 16 2009
-!- nchaimov [n=cowtown@c-24-21-45-17.hsd1.wa.comcast.net] has quit [Read error: 104 (Connection reset by peer)]00:23
-!- nchaimov [n=cowtown@c-24-21-45-17.hsd1.wa.comcast.net] has joined #hplusroadmap00:23
flamoothttp://www.boners.com/grub/811686.html monster face00:53
-!- deviantlarva [n=deviantl@oline.itek.norut.no] has joined #hplusroadmap02:08
deviantlarvahi, me, again^^02:08
-!- dira [n=chatzill@de2-as1165.alshamil.net.ae] has joined #hplusroadmap02:54
dirahttp://www.youtube.com/watch?v=RB_l7SY_ngI&feature=fvst02:58
-!- wrldpc2 [n=benny@ool-ad03fe34.dyn.optonline.net] has joined #hplusroadmap03:20
deviantlarvayeah, intersting03:37
deviantlarvahi btw dira 03:37
dirahey deviantlarva03:44
drazakkanzure: doI mind if I link to one of your books on heybryan on the diybio list?06:36
-!- QuantumG [n=qg@206.248.137.22] has quit [Remote closed the connection]07:31
-!- QuantumG [n=qg@nuclear.biodome.org] has joined #hplusroadmap07:32
-!- QuantumG [n=qg@nuclear.biodome.org] has quit [Remote closed the connection]07:47
-!- QuantumG [n=qg@nuclear.biodome.org] has joined #hplusroadmap07:47
-!- genehacker [i=genehack@wireless-128-62-149-238.public.utexas.edu] has joined #hplusroadmap07:58
kanzuredrazak: no08:35
dirakanzure , how is the 3D thing going ?08:36
kanzurei don't know what you are asking about08:37
diracollision detection to be exact08:38
kanzurefenn rewrote it to use occ's bounding box methods, seems to work ok for now08:38
diraso what is the next step ?automated engineering ?08:39
kanzurea few different projects08:42
kanzureone, to compute a number that represents the "compatibility" of a design using particular parts and components08:42
kanzuretwo, an algorithm for converting from point clouds to boundary representations or bezier splines-- you can see the progress in skdb/import_tools/surf.py08:42
kanzurethree, some unit tests for graphsynth, and eventually integrating the graphsynth Node class into an assembly graph representation for parts in skdb. you can see the graphsynth progress over at: http://adl.serveftp.org/skdb/thirdparty/graphsynth.py08:43
kanzureyou can see the code for point clouds to breps over here: http://adl.serveftp.org/skdb/import_tools/surf.py08:44
diraI see08:44
-!- genehacker2 [i=genehack@wireless-128-62-141-141.public.utexas.edu] has joined #hplusroadmap08:53
-!- genehacker [i=genehack@wireless-128-62-149-238.public.utexas.edu] has quit [Read error: 110 (Connection timed out)]08:56
-!- any21724810 [n=someone@75-120-28-93.dyn.centurytel.net] has joined #hplusroadmap09:15
-!- genehacker2 [i=genehack@wireless-128-62-141-141.public.utexas.edu] has quit [Read error: 110 (Connection timed out)]09:18
-!- katsmeow-afk [n=someone@75-121-61-243.dyn.centurytel.net] has quit [Read error: 110 (Connection timed out)]09:33
dirakanzure: I assume you've already seen this but just to make sure , http://ldd.lego.com/09:47
dira+ Microsoft robotic studio and Virtual Work Benches in general09:48
diraincluding , virtual chemistry lab or even virtual medical training kits09:52
-!- genehacker [i=genehack@wireless-128-62-172-108.public.utexas.edu] has joined #hplusroadmap09:54
-!- deviantlarva [n=deviantl@oline.itek.norut.no] has left #hplusroadmap []09:59
-!- any21625432 [n=someone@75-120-31-43.dyn.centurytel.net] has joined #hplusroadmap10:24
fenndira: ldd is actually pretty sophisticated, i'm impressed10:25
diraalthough such systems without proper HCI device are hard to work with10:25
fennoh i've seen much worse10:26
dira:) don't tell me you are mechanical engineer ?10:27
fennmeh10:27
fenni've done my fair share of 3d modeling10:28
-!- any68940998 [n=someone@75-120-31-43.dyn.centurytel.net] has joined #hplusroadmap10:29
dirathe worst type of modeling systems are mechanical systems , then civil engineering systems and after that architecture 10:29
diraI see10:29
fenndunno why you say they are the worst10:30
dirabased on details and oddness of particles , in modeling of building you rarely face particles inside each other , components are integrated and well defined , but in mechanical engineering , one particle might include hundreds of sub-particles with special materials and mechanical state10:32
fennthat's just because architects don't care about details for some reason10:32
fennif i turned in a general idea of a sketch for an engine, you'd be pissed10:33
fennbut for some reason that's okay with buildings10:33
diraI'm talking about Civil engineering too, they don't need details too , at least as deep as mechanical sketches10:34
fennwell that's just crap10:34
diraI've been into modeling (more like simulation) of physical systems , applying Quantum mechanics  (simplified) to system of particles , since formula and situation is well defined that is easy too, so I guess , mechanical modeling is the worst case ;)10:37
diraunless you give me an example...10:37
fennbiological modeling is harder10:39
diraah , never thought of that, good example 10:39
diraI've seen some medical training kit simulating surgery ! 10:40
-!- genehacker [i=genehack@wireless-128-62-172-108.public.utexas.edu] has quit [Read error: 110 (Connection timed out)]10:40
fennsure but i think they just start off with a data set10:41
fennthey don't actually model the growth process10:41
diraI cannot believe in their accuracy but even thinking about modeling hearth pumping and all those breathing effects is hard !10:42
-!- any21724810 [n=someone@75-120-28-93.dyn.centurytel.net] has quit [Read error: 110 (Connection timed out)]10:42
dira*heart10:42
diraI just get into an idea , is it possible to create HCI able to simulate resistance  or tension of  virtual items ? like when you reach wall in virtual world , it stops you from going further ?10:46
-!- any21625432 [n=someone@75-120-31-43.dyn.centurytel.net] has quit [Read error: 110 (Connection timed out)]10:46
-!- any99506369 [n=someone@75-120-44-77.dyn.centurytel.net] has joined #hplusroadmap10:47
dirabecause that's critical in virtual engineering using HCI while working with particles ,  when you want to pickup a brick in real world , your fingers keep closing until they sense the resistance (touching surface of brick).10:51
fennyes that's called a haptic interface10:51
* dira is googling the phrase10:52
fennthere's even a consumer device, the novint falcon10:52
dirauseful link ?10:52
fenni don't know much about their use in CAD and 3d modeling10:53
fennit's mostly just "VR" whatever that meanss10:53
-!- genehacker [i=genehack@wireless-128-62-49-92.public.utexas.edu] has joined #hplusroadmap11:03
-!- any68940998 [n=someone@75-120-31-43.dyn.centurytel.net] has quit [Read error: 110 (Connection timed out)]11:04
fenngenehacker: check this out http://adl.serveftp.org/lab/fenn/pngs/lego_digital_designer11:54
genehackercool11:59
genehackernow get a robot to assemble them12:00
genehackerhttp://adl.serveftp.org/lab/fenn/pngs/lego_digital_designer12:00
genehackeroops12:00
genehackerhttp://fora.tv/2009/05/30/Rodney_Brooks_Remaking_Manufacturing_With_Robotics12:00
fenntoo bad it's proprietary12:00
fennwhy do people keep linking to that?12:00
fenndamn, worm gears don't work12:02
dirathat logo software even generates catalog indicating each action step by step  to create the total image12:07
diraso it is easy to create one robot ;)12:07
genehacker????12:08
kanzurethere's something called 'parametric action representation' or something which could potentially be used for step-by-step task generation, however i'm not so sure yet12:09
diraI'm sure it does the same job12:09
dirathey use it in their mindstorm website for robotic competitions12:10
dirayou just need to build robot to able to execute each operation like rotating , pushing and ... 12:12
dirayou can even use LDD to build that robot :p12:12
-!- kardan_ [n=kardan@p54BE62B6.dip.t-dialin.net] has joined #hplusroadmap12:19
diraa little off topic : http://en.wikipedia.org/wiki/Microsoft_Robotics_Developer_Studio12:29
dirasupport lego/NXT mindstorm too , beside its physix engine to support physical simulation based on defined models 12:29
-!- genehacker [i=genehack@wireless-128-62-49-92.public.utexas.edu] has quit []12:30
kanzureso?12:32
kanzureit's microsoft..12:33
diraeh , M$'s12:33
-!- kardan| [n=kardan@p54BE42CB.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)]12:36
any99506369plausable deniability inherent in ordering cognitive inhancement pills online "sorry, i don't remember ordering those, and i never got them"13:09
-!- wrldpc2 [n=benny@ool-ad03fe34.dyn.optonline.net] has quit []14:05
-!- wrldpc2 [n=benny@ool-ad03fe34.dyn.optonline.net] has joined #hplusroadmap14:11
kanzurehere's the transcript from paredis' talk today at UT:14:50
kanzurehttp://adl.serveftp.org/lab/talks/2009-09-16-paredis-utexas.html14:50
kanzurefenn: http://www.modelica.org/libraries/Modelica/releases/3.1/ModelicaStandardLibrary_v3.1_build4.zip/at_download/downloadableFile/at_download/downloadableFile14:56
kanzureline 2126 and onward show some equations for a linear spring damper14:58
kanzureer sorry I guess you'd need to know which file I was looking at15:00
-!- mason_l [n=x@202-89-188-136.static.dsl.amnet.net.au] has joined #hplusroadmap15:03
-!- mason-l [n=x@202-89-188-136.static.dsl.amnet.net.au] has quit [Read error: 110 (Connection timed out)]15:06
kanzure"PyMbs is a Python library for modelling holonomic multibody systems. It is able to generate simulation code for Python, Modelica and Matlab."15:07
kanzurehttp://sourceforge.net/projects/pymbs/15:07
kanzuresimulink, scicos, ptolemy, modelica15:11
-!- jonathan__ [n=jonathan@dhcp-146-6-214-4.icmb.utexas.edu] has joined #hplusroadmap15:13
jonathan__ping15:14
kanzurehello15:14
kanzurejonathan__: 15:14
kanzurehttp://adl.serveftp.org/lab/talks/2009-09-16-paredis-utexas.html15:14
kanzurejust got done with that talk today15:14
kanzurehope you liked yesterday's rothemund transcript15:14
jonathan__"The "diagram" is not a model."   well that's good15:16
kanzurethis is "buzzword soup"15:16
kanzureour lab already implements most of this15:16
kanzurejust without the buzzword soup15:17
kanzurealso there was another talk yesterday from schlumberger .. http://adl.serveftp.org/lab/talks/2009-09-15-schlumberger.html15:19
kanzurefenn: is it useful to have the components of simple linkages in skdb even though they don't actually correspond to anything physical other than basic chunks of materials?15:22
kanzurejonathan__: what's up?15:26
jonathan__workin on the robotics module to OO it15:27
kanzurebtw, yaml supports lists15:27
kanzureor "sequences"15:28
kanzuredo you have any experience with modelica?15:28
jonathan__never heard of it15:29
kanzurei've been ignoring it for a while15:29
jonathan__looks like vhdl kind of15:29
jonathan__these days I dont want to use something unless 10,000 people are also using it.  otherwise it's too niche15:30
kanzureunfortunately that's hard to tell on the net15:31
jonathan__yaml has a lot of complex power15:31
jonathan__I like the references the best15:32
kanzureyou mean pointers?15:32
jonathan__whatever they call the &thing15:32
kanzureah yeah15:32
jonathan__and the *thing15:32
kanzureyes it's nice15:32
kanzurei wish it would work across multiple documents though15:32
kanzurethat would absolutely blow my mind15:32
jonathan__a bio guy sent me his java framework for a specific bio robot program.  He used xml.  totally unreadable and yucky15:33
kanzure(at the moment it just crashes if it can't find what's being referenced)15:33
kanzureyep15:33
kanzurethe other day I wrote an xml2yaml app thingy15:33
kanzurehowever, it sucks15:33
kanzurehttp://adl.serveftp.org/skdb/import_tools/del_repo/repo_to_yaml.py15:33
jonathan__I dunno what's up with these CS guys, they always create these abnormally complex languages15:34
kanzureit doesn't construct lists/sequences, doesn't use pointers, etc.15:34
kanzureyes15:34
kanzureUML seems to be a case of that15:34
kanzurethe guy who presented today is a big fan of UML, 15:34
kanzurebut he's doing the same thing that ADL is doing15:34
kanzureexcept we're somehow doing it without as much bloat15:34
kanzurealthough graphsynth, sadly, also uses bloatware/xml to represent graphs15:34
jonathan__no one in industry uses UML.   bring it up in a design meeting and people will laugh or stare blankly.  it's useless15:34
kanzurewell that's what I thought15:35
kanzurebut apparently there are still people using it?15:35
kanzurelike, for code generation? wtf15:35
jonathan__well, for meaningless code like widgets, ok15:35
kanzureheh :)15:35
kanzurei've almost always used code generation for gui stuff15:35
kanzurebut that's about it15:35
kanzureexcept experimenting with SWIG once or twice. don't kill me :(15:35
jonathan__or database query languages etc.  it is a pain to write that stuff by hand15:35
kanzurecheck this out:15:36
kanzurehttp://adl.serveftp.org/lab/sysml/15:36
kanzurethis was apparently at a conference that matt campbell (the lab pi) went to the other week in san diego15:36
kanzureit's pretty bad15:36
jonathan__uh15:36
jonathan__basically the rule is, if you want people to use i, make it easy.   & thats not easy15:37
kanzureapparently they use all these graphical diagram editors and don't like writing code any more15:37
kanzurebut anyway. i agree it's a terrible idea.15:37
jonathan__plenty of academia is geared towards "invent super cryptic thing so that I can be the expert and publish it" etc15:38
kanzureoh that's the trick is it? 15:38
kanzurewhy didn't i think of that15:38
jonathan__if they like graphical, why not use labview's language already15:38
kanzurecan you show me some of what your code is capable (or not capable) of?15:38
kanzureyou're doing it in perl, right?15:38
kanzureif it was python i'd ask for a session demo15:39
jonathan__right now it's all broken since I am moving to OO15:39
kanzurecamelpoop?15:39
jonathan__I will post a video once it is done15:39
jonathan__the manufacturer tech support isnt being quick about providing low level device commands either so right now I'm only doing movement15:40
kanzurewhat's your connection interface?15:40
jonathan__the commands are string commands for their s/w, or active-x for their newer s/w   - most people use visual basic to do similar things but only specific robot tasks15:41
jonathan__thus none of their code is re-usable beyond a specific lab problem15:41
kanzureisn't there something for windows called "hyperterminal" that can help you listen for the strings that are being sent?15:42
jonathan__i.e. spend months creating s/w framework and all it does is 1 experiment15:42
kanzureit's been so long since i last used windows. is it hyperterminal that i'm thinking of?15:42
jonathan__hyperterminal doesnt do that, but it's not so much a question of getting the commands, more like decoding them15:42
kanzureit comes pre-installed on windows98 or something15:42
kanzurei see15:42
kanzurewell i'd hope that they have this plaintext stream15:43
kanzurebut that's unlikely i guess15:43
jonathan__D1,OV420P30M0R15:43
kanzurehah15:43
jonathan__means, aspirate to well #715:43
kanzurewell that sucks15:43
jonathan__ha, I sent a very simple email to the tech support and they replied:  "what's perl?  I have not come across that before"15:44
kanzureyou should make them pay you the privledge of letting them give you answers to your questions that they counterquestion.15:44
kanzureer, wait15:44
kanzurethere's a joke somewhere in there.. sorry.15:44
jonathan__yea15:47
jonathan__except, they hold the proprietary docs, so got to be nice15:47
jonathan__I did reply something like "surprising since it is the #1 language in bioinformatics" or something, ha15:48
kanzuretrue that15:49
jonathan__it's been going back & forth for about 2 mos. now.   tech support is always this way.  first ask, then they say "do it the simple way".  then tell them that isnt powerful enough, then they say "try it the simple way again".. etc....  until finally they break15:49
kanzuretech support actually talks to you?15:49
kanzureare they under some sort of legal obligation?15:49
kanzurehow does this work15:49
jonathan__ha15:51
jonathan__the customer (ut lab) pays a lot of $$$ for support contract15:51
jonathan__support contract says stuff like:  "must respond to all email within 72 hours" etc15:51
jonathan__which means, they can reply "Got your email will discuss with the internal team" or etc of course15:52
kanzurehah! "learn perl regular expressions within 72 hours, go!"15:52
kanzurebwahah15:52
jonathan__then they assign a tracking ID to the issue for reference, it goes in their database, and on their side if they have ID's open too long, their managers complain to them15:52
jonathan__etc15:52
jonathan__i.e. support issue that is open >2 weeks will send automatic email to the regional boss who will hammer the support team "why isnt this resolved???" etc15:53
jonathan__which of course leads to the habit of support personell who artificially close the support ticket with stuff like, "this should help you, let me know if it doesnt, I am closing this ticket now" etc15:54
jonathan__kind of dismissive15:54
jonathan__it depends on the company15:54
jonathan__I think all undergrads should do at least 6 mos as tech support somewhere.  great problem solving skills.15:54
kanzurehm, that would depend actually15:55
jonathan__"IT" is the worst of course.   IT guys never give support tickets or follow up.  lame.15:55
kanzurei bet in many cases they would lose brain cells15:55
kanzurebecause a lot of tech support call centers just have a knowledge management system15:55
fenni learned regex in << 72 hours15:55
kanzureand you go through the routines15:55
kanzurefenn: wtf?15:55
jonathan__there's a lot of levels to tech support15:55
jonathan__front line, sure, follow the internal databse15:56
jonathan__closer to the tier1 customers, each problem gets more unique and there is skill involved15:56
kanzurei hope so. that would renew some faith in commercial tech support.15:56
jonathan__forces the basics...    too many postdocs go for these crazy solutions to problems like "I know!  We can use UML and XML and a graphical blah blah blah"...    ugh, just get it to work command line first15:57
jonathan__the question is:  "which regex".   sysv ok simple.  gnu, more stuff & more annoying.  perl, uhhh I doubt i15:57
fenni actually liked the whole SysML idea once it was explained properly15:58
jonathan__perl re is ridiculously powerful/complex15:58
kanzurethe implementation is totally wrong15:58
fenni still don't think it _needs_ to be graphical as the UML people seem to think15:58
kanzurea lot of it is very similar to matt's work15:58
jonathan__it needs a graphical paperclip guy like microsoft word15:58
fennkanzure: do you know how to run a different python with bpython? 2.5 vs 2.615:59
kanzure#bpython will15:59
kanzurei don't15:59
-!- dira [n=chatzill@de2-as1165.alshamil.net.ae] has quit ["ChatZilla 0.9.85 [Firefox 3.5.2/20090729225027]"]16:14
drazakI swear to god the next person that emails me to start a pissing contest, I'm going to smite16:18
-!- QuantumG [n=qg@nuclear.biodome.org] has quit [Remote closed the connection]16:19
-!- QuantumG [n=qg@nuclear.biodome.org] has joined #hplusroadmap16:20
jonathan__I think I'm going to refactor the diybio faq16:48
jonathan__split up into smaller sections16:48
jonathan__ha, see the biosci spam?  "MicroRNA Clunkers Program. Economic Crisis affecting your microRNA research?   Exiqon is here to help.  Trade in your clunker. "16:55
kanzurejonathan__: that's ok but last time you totally mangled it :)17:13
kanzurehttp://openwetware.org/wiki/DIYbio/FAQ17:13
kanzureok new sections seem fine to me17:14
jonathan__considering everyone seems to want to "do stuff" its funny there is a lack of faq on "doing stuff"17:24
kanzureyeah :(17:24
jonathan__even simple stuff like computational i.e. blast17:25
kanzurebioperl and biopython are really, really easy to use17:25
kanzurehm i wish i had more time to explore the parametric action representation research17:29
kanzurethey have these methods of automatically generating service manuals for different machines17:29
kanzureapparently the airforce has funded much of this research17:29
kanzurein particular for the maintenance of jets17:29
kanzurewhich is pretty much what skdb is going towards17:30
kanzurelemme grab some pretty diagrams17:30
kanzurehm this is taking a while to upload17:31
kanzurehttp://adl.serveftp.org/papers/instructions.zip17:49
kanzureok added the images17:51
kanzureadl.serveftp.org/papers/instructions/17:51
kanzureer, http://adl.serveftp.org/papers/instructions/17:51
kanzurefor instance: http://adl.serveftp.org/papers/instructions/Virtual%20humans%20for%20validating%20maintenance%20procedures.pdf.1.png17:52
kanzureof course, 3D isn't that important to me17:53
jonathan__looks like ikea17:53
kanzurebut in skdb you already have the models so it's not a big deal17:53
-!- genehacker [i=genehack@wireless-128-62-43-62.public.utexas.edu] has joined #hplusroadmap17:54
kanzurejonathan__: isn't ikea a car company?17:54
kanzureoh you mean for furniture17:54
kanzuredunno what's going on with the skeletons though17:55
kanzurekind of gruesome if you ask me17:56
jonathan__omg17:58
genehackerikea a car company?17:58
jonathan__the dissertation in that directory even includes an ikea as a case study17:58
genehackerhuh?17:58
genehackerwhat did I mis?17:59
kanzuregenehacker: http://adl.serveftp.org/papers/instructions/17:59
jonathan__the premise of using pictures is what, do avoid translation?17:59
kanzurei don't want to use pictures18:00
kanzurei like text instructions too18:00
jonathan__I dont really find the pictures too useful, I prefer a recipe18:00
kanzurei think text, video and pictorial are all useful18:00
kanzurebut "only pictorial" is totally useless18:00
jonathan__the pictures can be generated from a good recipe18:00
kanzureyes18:00
kanzurei agree18:00
kanzurethe best paper by far in that directory was Automating maintenance instructions study18:00
genehackeryup it's to avoid translation18:01
kanzurewell they suck.18:01
kanzurejonathan is right.18:01
genehackertrue18:01
genehackerbut it sucks not having a robot to assemble stuff too18:01
kanzurei don't think you know what we are talking about18:01
genehackerI know what you are talking about18:03
jonathan__the robot is useless unless it can operate on something.  so it is chicken and egg18:03
genehackerautomatic insturction generation right18:03
jonathan__robot can be replaced by an undergrad18:03
kanzureyou mean me? :(18:04
genehackerI'm an undergrad18:04
genehackerso have you been able to generate instructions?18:08
kanzureyes but fenn hates it18:08
drazakI just posted the most ghetto gel electrophoresis chamber idea, ever18:09
genehackerlet me guess, it uses gello?18:10
drazakno, tupperwear18:10
drazakand you can use agarose or whatever18:10
genehackerok18:10
genehackerhahaha18:10
drazakoops, I forgot to say that you have to remove the tape after casting18:10
drazakor well18:10
drazaker, oh well18:11
drazakI'm stuffed full of sushi18:11
jonathan__who here is a mech eng18:11
genehackerI am18:11
kanzureme18:11
drazakI've taken mech eng18:11
jonathan__if so why tupperware.  pls post how to pour a mold which results in a nice chamber18:11
drazakjonathan__: you use the tupperwear as your mold18:12
jonathan__I want a 1-piece chamber poured into some cheap disposable (clay or something) mold18:12
fennlego18:12
genehackerhttp://reprap.erikdebruijn.nl/files/research/NASA%20as%20user%20innovator%20-%20WPMicrogravityFDM.pdf18:12
jonathan__bu hao lego18:12
fenni used to make casting boxes out of lego for plaster and wax molds18:12
genehackernasa put a stratasys in vomit comet and it worked18:12
genehackerwhat did you cast?18:13
drazakjonathan__: cut rectangles off each end, cover with tape, pour gel, let it set, take off tape, put gel in tupperwear in bigger tupperwear, in larger tupperwear attach electrodes, add buffer, add voltage, boom18:13
genehackerjonathan cline right?18:13
jonathan__the material for the cast is not important18:13
genehackerI could just machine one for you18:13
jonathan__finding the pourable acrylic is the important part18:13
genehackerpolyester resin might work too18:14
genehackera gel box mold is a fairly simple structure18:14
jonathan__exactly, simple18:14
drazakjonathan__: I'm not saying to make a gel box, I'm saying use the tupperware as the gel box18:15
drazak:D18:15
drazakit's ghetto18:15
jonathan__yes, I know 18:15
jonathan__if you need a really simple 220v power source let me know18:16
genehackerare there any standard gel box sizes?18:17
genehackeralso what could we use a gel electrophoresis setup for?18:18
genehackerbesides gene "finger printing"18:20
jonathan__it is used for everryyyyything18:21
genehackerI really thing we need to do something more advanced than gel boxes if we want to do true diybio18:21
jonathan__define "true diybio"18:22
drazaksure18:22
jonathan__you mean genetic engineering?18:22
drazakwhich is why I proposed the simplest way18:22
drazaktakes 10 minutes to build18:22
genehackergive me an example of everything first18:22
jonathan__nothing in biology can be measured except by amplifying (pcr) or by running on gel if already amplified18:23
genehackermeasuring what?18:23
jonathan__everything in the lab here eventually hits a gel to verify "it is what it is"18:23
jonathan__genes of course18:23
jonathan__or proteins18:23
fennso you don't like fluorescent labeling?18:24
genehackerwhat about the chemicals to run the gel aren't those going to be hard to obtain18:24
fennor radioisotope uptake18:24
drazakjonathan__: proteins you do on a PAGE, and those are verticle18:24
jonathan__too expensive for common stuff18:24
drazakgenehacker: nah, the only hard thing might be EtBr, but eh18:24
jonathan__page is still a matrix run on high voltage18:24
jonathan__and they are vertical only for convenience18:24
genehackerso can you seperate out proteins?18:24
genehackerusing gel eletrophoresis?18:25
drazakgenehacker: any time you want to seperate something by size you run a gel18:25
drazakgenehacker: or my charge18:25
drazakgenehacker: or by shape18:25
drazakor by shape and charge if you do a 2d gel18:26
genehackerawesome18:26
jonathan__capillary electrophoresis is out of reach of diy for the long term I think18:26
drazakyeah18:26
genehackercould we seperate Taq polymerase out?18:27
jonathan__"gel" = agar18:27
jonathan__agar is too thick for proteins, the proteins dont move18:27
genehackerso what do we use?18:27
drazakpolyacrylmide18:28
drazak2-20%18:28
genehackeroh cool18:28
jonathan__that's the point of all the different reagents...18:28
genehackerthat's availale at home depot18:28
genehackerplant section18:28
genehackercomes in granules though, so it might be hard to use18:29
genehackerI think that's what they are18:29
genehackeryup18:30
drazakyeah, it was mentioned on the group that you can use agar instead of purified agarose18:31
drazakif you're worried about DNAase or RNAase contamination, you can treet them with DETC18:31
genehackerso could we use gel electrophoresis to somehow test  for diseases?18:33
drazakit's done all the time :)18:33
drazakwell I guess they use qrt-pcr more than GE now, but you can do GE, pcr some samples, run them on the gel, if you get a band the right size then you have a gene, if it's a viral gene then you had a virus18:34
genehackerhow? and how do we get the chemicals?18:34
genehackerthe problem is getting the chemicals18:35
drazak99% of this stuff you can buy as a private citizen from invitorgen or sigmalaldritch18:36
genehackerand delivered to my house?18:36
genehackerhow much?18:36
drazakit depends on what you're buying18:37
genehackerisn't this stuff expensive(bad for stuff that is expended)?18:37
drazakyou can buy primers for 12 bucks a primer, and primers are good for atleast a couple hundred pcrs, taq polymerase, you only need a little, it's maybe 100 bucks for 200 reactions, primermix is cheap, DETC is cheapish, and you can get agar at the hardware store18:39
drazakyou need a thermocycler or to use hot water baths18:39
genehackerwow primers are cheaper than I though18:40
drazakyeah18:40
drazakit's 6$ for the forward, 6$ for the reverse18:40
drazakit's more at sigma, but I'll get you the name of the palce that we order from that's so cheap18:40
genehackerso taq or any use ful polymerase is going to be expensive18:41
drazakthink about how many reactions you get out of it though18:42
drazakit's only a couple bucks per reaction18:43
genehackerso just as an example, I want to determine if I have a certain virus18:43
genehackerhow would I do that?18:43
genehackerthat's comparable to the cost of pharmaceuticals18:44
-!- jonathan__ [n=jonathan@dhcp-146-6-214-4.icmb.utexas.edu] has quit []18:45
genehackeryou mentioned that this could be done with gel electrophoresis18:46
drazakok, so you need a sample from blood, if it's an rna virus you isolate rna, dna virus isolate dna, if you isolated rna you need to do reverse transcriptase, so then you have your dna, determine your concentration of dna, and then mix 80ng of dna (2 microlitres if you had 40ng/ul), .5ul taq, 2ul primer, 7.5ul primer mix, 13ul water, do the pcr, then take 15ul of your sample and mix it with 3ul of loading dye on some parafilm, then load it into the gel (whic18:46
drazakwhere'd I get cut off?18:46
genehacker(whic18:47
drazakh you made and put in the gel box and covered with buffer) and then run the  gel for an hour18:47
genehackerso how do you isolate DNA?18:47
drazakuse a dna isolation kit, it's usually chaper than buying the reagents18:48
drazakor rna isolation kit18:48
genehackerand how is concentration determined18:48
drazaknanodrop or EtBr dilutions18:48
genehackerhow do DNA and RNA isolation kits work?18:49
drazakgenerally they have a salt spin column18:49
genehackerwhich is?18:49
drazakso you take your sample homogenize it, put it in the spin column and spin for whatever time it says, then your run some buffers through to clean it up, do dnase digestion if you're isolating rna, then run some more buffers, then elute into water18:50
drazakyou need a microfuge18:50
drazakall dna isolation methods need one though18:50
drazakor a highspeed centrifuge18:51
genehackernot all18:52
genehackerthere is microfluidic dna isolation18:52
genehackerhttp://research.ncku.edu.tw/re/news/e/20090109/images/090105065937DihqBc.jpg18:54
genehackercheck that out18:54
drazakall traditional methods18:54
genehackerwe need not be traditionalists18:55
drazaklisten18:55
drazakfor getting started18:55
drazakdiybio needs to get it's head out of it's ass and stop trying to reinvent the wheel18:55
genehackerhow so?18:57
drazakmicrofluidics=out of our reach18:58
genehackernot so18:58
drazaksynbio=we don't have the infrastructure for it18:58
genehackerI wouldn't call it trying to reinvent the wheel18:58
drazakit is18:59
genehackerI'd call it trying to do something we can't18:59
drazakif we can't do the traditional methods we certainly can't manage to do the enw methods that 2 labs in the whole world are using18:59
genehackerwe can do shrinky dink microfluidics18:59
genehackerwhat infrastructure is needed for synbio?18:59
drazakeverything19:00
drazakwe don't have the reagents, we don't have the equipment, and most people don't have the knowledge19:00
genehackerwe don't have the reagents?19:01
drazakmaybe kay aull does19:01
genehackerwe don't have the reagents for much of this diybio stuff19:01
drazakbut 90% of diybio people are all theoretical19:01
drazak95% of diybio have no wetwork experience19:01
genehackercorrect19:02
genehackerhell I hardly have any19:02
drazakit kinda pisses me off when people think it's so easy just to do everything19:02
drazak"oh yeah you just pippete stuff into tubes"19:02
drazakwell pippetting .5uL isn't the most easy thing in the world, thanks19:02
genehackeroh wow19:03
genehackerI found a bunch of sources for pyrex wafers19:03
fenn0.5ul with a p2 is pretty easy19:23
drazakfenn: well yeah, but theres a technique to it, you can't put it on the side, you need to get it in the bottom of the tube with the rest of it, you need to make sure you don't pick up extra on the outside of the tip19:27
drazakwith that small ammount having extra on the outside of your tip will mess up your ratio easily19:28
fennyeah it's not like dumping in a liter into a swimming pool19:28
drazakexactly19:28
drazakmost people pippetting for the first time are just like19:28
drazakSPRAY19:28
drazaknot gently getting the liquid with the rest19:28
drazakof course if you're doing 600ul, you can just screw around with it :P19:29
drazakbut there is a technique to it19:31
drazakpippeting is a skill19:31
kanzurelike not forgetting what you just pippetted19:31
fennthat's a tough one19:31
drazakdefinitely19:32
drazakthat's what we need a smartbench for19:32
drazak:)19:32
kanzurehow hard could it be to turn on a webcam?19:32
kanzurethey haven't done anything with that project in a year19:33
drazakthat was always my thought19:33
kanzurehttp://diybio.org/projects/19:33
drazakcamera+computer19:33
kanzurethere's also this weird ass shot of mackenzie19:33
drazakthere's programs to add tags to objects19:33
kanzurenever understood that either19:33
drazakmac is lulzy19:33
kanzureopencv or other image process library stuff comes later.. just use a webcam for starters.19:33
-!- genehacker [i=genehack@wireless-128-62-43-62.public.utexas.edu] has quit [Read error: 110 (Connection timed out)]19:34
kanzurealso, voltage sensors would be neat19:34
kanzurelike playing "operation"19:34
fennheh AR augmented pipettor19:34
fenndon't make me PCR reaction you19:34
-!- any99506369 is now known as katsmeow19:34
drazakrofl19:37
fennbut seriously i'd love to have protocol instructions overlaid on stuff on the lab bench19:40
kanzureoverlaid?19:41
kanzureoh, AR/VR19:41
fennliterally arrows going from one tube to another19:41
fenn"pipette 5ul from here to here"19:41
kanzureit wouldn't be too hard to get "color coded object labeling" going on19:41
kanzurewhere you have it know the identity of an object by a color19:41
fennpinouts for breadboarding too.. i'm always looking back and forth at the datasheet19:41
kanzurecounting tiny pins?19:41
fennseems i can only remember one pin at a time19:42
fennelectronics chips19:42
kanzureyes i know19:42
kanzurei too have spent many hours counting pins19:42
fenni think you could train a computer to recognize a chip just by its shape19:42
fennthen add some fiducial code (what are they called again?)19:43
drazaknah, too many share the same package19:43
fennthe square barcode thingies19:43
drazakif it has resolution to read those19:43
fennactually i dont think you need that, you could just say "this chips is this number"19:44
drazakyeah19:44
fennactually you just need to know the orientation of the breadboard19:46
kanzureopenmodelica tutorial: http://www.ida.liu.se/labs/pelab/modelica/OpenModelica/Documents/ModelicaTutorialFritzson.pdf19:48
kanzurewhy is this a language?20:03
kanzure"keyword 'flow' indicates that (in this case) currents of connected pins sum to zero, example: flow Current i;" (pg 36)20:04
kanzure"Connections between connectors (ports) are realized as equations in modelica" (pg 37) 20:05
kanzureah20:05
kanzureconnect(pin1,pin2) corresponds to "pin1.i + pin2.i = 0" 20:05
kanzureso it's just setting up the constraints of the problem20:06
kanzureopenmodelica notebook seems to be their gui: http://images.google.com/images?q=openmodelica%20notebook&oe=utf-8&rls=org.debian:en-US:unofficial&client=iceweasel-a&um=1&ie=UTF-8&sa=N&hl=en&tab=wi20:08
kanzureon pg 44 they show a simple dc model.. they connect the "p" variable of an inductor to the "n" variable of an electromotive force (emf). any ideas as to what that might mean?20:10
drazakwhat is openmodelica?20:10
kanzurewhat is the "n" variable? 20:10
kanzuredrazak: modelica is a modeling language apparently. openmodelica is an open source library and compiler for the language.20:10
kanzurebut again i'm not sure why this is a language20:10
drazak'modeling language' like, 3d modeling?20:11
kanzurewhat does "compiling" it do?20:11
kanzureno, this is for parts and components being connected together20:11
kanzurekind of like in skdb20:11
fenni think p and n are like "effort" and "flow" (not necessarily p=effort but you get the idea)20:11
drazakoh right20:11
fennthere are 4 variables20:11
fenn(this i remember from the paredis slides)20:11
kanzurethere was a page in here before pg 40 that had a nice, more extensive diagram that slightly matches matt's usual "efforts and flows" lookup table20:12
fennoh, duh. "previous" and "next"20:12
fennor maybe positive/negative20:12
fenni don't really like that :(20:13
fennpolarity only matters for diodes20:13
fennand what do you do with a transistor20:13
fennkanzure: shouldn't you be reading about quantum mechanics?20:14
-!- genehacker [i=genehack@wireless-128-62-43-62.public.utexas.edu] has joined #hplusroadmap20:15
kanzureit's okay, i took some drugs so i'll be up forever20:15
drazakneed that adderall20:16
kanzurefenn: it wouldn't be a stretch to go fetch openmodelica's model of a screw and dump it into the skdb screw package20:17
fennthere are a lot of things about screws that they won't be concerned with20:18
kanzurethese do seem to be simplified models, yes20:18
fenni'm all for a "generate openmodelica" of course20:18
kanzurehow about import openmodelica?20:18
fenneh?20:18
kanzurewell there seems to be some connectivity information that is recoverable, although it may not be useful20:18
fennthere's a mapping from each skdb class to modelica class20:18
kanzuresure.20:19
fenni need a better way to integrate packages that may not necessarily be installed20:19
kanzureto integrate? what?20:19
fennlike try: import openmodelica; except ImportError: blah20:20
fennbut less sucky20:20
kanzurei actually like that20:20
kanzurei mean, it's worked so far20:20
fenni don't. it screws up the flow of the code20:20
fennif i didn't have to worry about it, things would be in different places (much more logically arranged)20:21
fenni shouldn't have to turn everything inside out because of a package dependency20:21
kanzureoh you mean when debugging something?20:21
fennno i mean the structure of the code, the order words appear on the screen20:21
kanzurethe import statements seem to be at the top usually20:22
fennyeah but if i'm doing try: import then i have to put all the code for that plugin inside the try statement20:23
fennunless i can do "if using_modelica:"20:23
fennbut i dont think that always works20:23
kanzurewhen does it not work?20:23
kanzurethat's the "c preprocessor" way :p20:24
fennhmm20:24
fennwell it turned out to be a bit more involved with OCC20:24
kanzureskdb.geom should fail with the "occ" issues on your box20:25
fennand then again with igraph, python-graph20:25
kanzureon your laptop i mean20:25
fennright, that's how i've been testing (halfassedly)20:25
fennyou could also do something like OCC.py: raise ImportError20:26
genehackerkanzure have you tried modafinil?20:26
fenni don't get why modelica doesn't use icons20:26
fennwhat's the point of graphical programming without icons?20:26
kanzuregenehacker: i forget. i can check my medical records if you want.20:26
kanzurethere's such a long list it's ridiculous.20:27
fennoh, nevermind, i guess they do use icons20:27
kanzurefenn: yeah the "notebook" thingy.20:27
fennit's sysml that doesn't use icons20:27
kanzurebut they use diagrams instead?20:27
fennthere's this labview-ish thing on p. 15/1620:27
fennmujahideen20:31
-!- katsmeow is now known as katsmeow-afk20:33
* ybit needs adrafanil20:36
* ybit is sleepy20:36
* ybit prefers jobs which don't suck the life out of you20:37
fennkanzure: http://www.flickr.com/photos/oddwick/3016657700/in/set-72157608736702937/20:37
ybiti have quite a few comments on the pythonocc install that i'm looking at20:37
ybitfirst off, make sure the person has python2.5-dev20:37
ybittook me forever to figure out why gcc test was failing and thus scons wouldn't compile (still won't compile, getting there...)20:38
ybit wrapper/SWIG/linux_darwin/BRep*20:38
fenndoes it have to be 2.5?20:38
ybitthat's the dir20:38
ybitdoubt it, i'm using 2.5 though20:38
fennso just python-dev then20:39
ybityeah20:39
ybitwrapper/SWIG/win32/LocOpe*20:39
ybittake note of the directory, it's different than the dir structure which you were using when the install notes were written20:40
fennare you following compile_pythonocc?20:40
kanzurewhich version did you checkout?20:40
* fenn loses track of how many mangled copies kanzure has littered around the internets20:40
ybitpython-gccxml doesn't exist20:40
fennyou shouldn't need it20:40
fennthat's for rebuilding the swig files20:41
ybithttp://pastebin.com/f64b005dc20:41
ybitthat's where i'm at currently, but i'm so flippin tired right now, not even sure if i'm going to bother fixing it tonight20:42
fennwell i updated the file that's being displayed on the wiki (why isn't it under version control hmm)20:43
fennybit: first of all, you shouldn't even bother with scons20:43
fennuse setup.py20:43
fennit seems to me like scons is a half finished project (and maybe setup.py is 7/8's finished)20:44
ybitspent the entire afternoon going over business details with my grandfather and other family members, looks like it will be pushed to the side for the time being and i'll continue learning welding and whatever metal work my dad and job teaches me + ev @ work20:44
fennwho is ev?20:44
ybitoh, electric vehicles20:44
ybitguess i will have to refer to Phreedom as evgeny now 20:45
fennhuh?20:45
ybiti called Phreedom ev for short for quite awhile20:45
katsmeow-afkybit : 1960's early 70's Ford had a hydraulic boost cylinder on the end of the pitman arm rather than a hydraulic boost steering gearbox , let me put in an order for 2 now20:45
fennwell, you could try calling him Phreedom20:46
ybitkatsmeow-afk: i'll keep an eye out if the business goes back into business :|20:46
katsmeow-afkk :-)20:46
fennshockingly brilliant idea, i know20:46
katsmeow-afkoff to postoffice, bbl20:46
ybiti could, but then that's a brilliant idea left alone20:46
ybitsetup.py, here we come, and then bedtime20:47
* fenn cries himself into chocolate induced insomnia20:47
ybitbtw, nobody told me about steel toe boots20:47
ybiti have to buy me a pair if i'm going to do some welding20:47
fennyou don't need steel toed boots to weld20:47
fennjust not sandals or running shoes20:47
ybiti just kind of figured that's what i would need and i asked a few guys today who have welded and they all suggested it20:47
ybiti only have running shoes and cheap dress-up boots20:48
fennthey probably wear carhart overalls too20:48
ybiti've yet to be meet anyone in a long time that's worn those :P20:48
ybits/that's worn/that wears/ (for katsmeow-afk's sanity)20:49
fennthat's is posessive20:49
ybitoh right20:49
fennno, wait20:50
ybiti'm sleepy?20:50
ybiti think you're right20:50
ybiti'm going to say you are20:50
fennyou is -> you'se20:50
ybitoh right, so did you look at the pastebin fenn, do i need some of that stuff? it looks like a bunch occ libs are missing20:51
fennthe "checking ... no"  stuff is fine, but you have to fix the environment paths to find the .i files20:52
fennthey moved them into src/wrapper/ i think20:52
fenni updated this page, you have to hit reload to see changes http://adl.serveftp.org/dokuwiki/pythonocc20:53
fennanyway it's for setup.py instead of scons20:53
fennseems to work a lot better20:53
fennsorry it's not totally checked and polished etc20:54
kanzureChlorophyll derivatives as visual pigments for super vision in the red20:54
kanzurehttp://adl.serveftp.org/papers/Chlorophyll%20derivatives%20as%20visual%20pigments%20for%20super%20vision%20in%20the%20red.pdf20:54
fennkanzure: that's ridiculous20:55
genehackerno this is:20:55
genehackerhttp://www.physorg.com/news172225206.html20:56
fennyeah20:57
ybiti'm almost certain this will break your system fenn 20:58
ybitsudo apt-get install libopencascade-visualization-dev libopencascade-dev20:58
ybitsudo apt-get install libopencascade-dev scons python-psyco swig checkinstall x11proto-core-dev libx11-dev20:58
fennwell, the second line isn't really necessary20:58
ybitbecause visualization-dev depends on the foundation-dev which can't be installed at the same libopencascade-dev is installed20:58
fennyar20:59
ybitand if libopencascade-dev is install, libopencascade-wok is removed20:59
ybitopencascade-wok20:59
ybitnot lib*20:59
fennguh20:59
fennfirst off, remove *opencascade*20:59
ybitit's almost better to just manually install occ20:59
fennthen enable opennovation repo's and install libopencascade-visualization-dev21:00
fennlibopencascade* should all be 6.3.021:00
fennyou need 6.3.0 for pythonocc (don't ask me why)21:00
kanzureprobably because the wrappers were generated on 6.3.021:01
kanzurethe wrappers/headers21:01
fenni would really like to figure out how this whole thing works some day21:01
kanzurehere's how it works:21:02
kanzureyou give ".h" files to swig21:02
fenni get the general idea21:02
kanzureswig takes these and..21:02
kanzureoh.21:02
fennit's the specific bits that don't really add up21:02
fennlike where are all the .i files hiding?21:02
fennand where are the transformations described?21:02
kanzurei think they may have created those by hand21:02
fennno21:02
ybitguess i should also edit /etc/preference to prefer the jaunty mirror over the unstable for occlibs21:03
fenni think you should forget about /etc/apt/preference21:04
ybits21:04
fennwhen you meddle too much it tends not to work as desired21:04
fennor you could learn the hard way, whatever21:04
fenni thought you were using debian?21:05
ybitfenn: i am21:10
ybitpython-gccxml is still nowhere to be found if you want to go ahead and remove that from the instructions21:24
ybitwill have to fix the following tomorrow21:27
ybit/home/heath/builds/pythonOCC/src/wrapper/SWIG/linux_darwin/Standard_wrap.cpp:1: warning: The C++ parser does not support -dy, option ignored21:27
ybit/home/heath/builds/pythonOCC/src/wrapper/SWIG/linux_darwin/Standard_wrap.cpp:3037:32: error: Standard_Failure.hxx: No such file or directory21:28
ybit/usr/include/opencascade/Standard_Failure.hxx21:29
ybitthe libs aren't linked correctly as you mentioned21:29
ybitoff to bed21:29
-!- davidnunez__ [n=davidnun@209.6.203.217] has quit []21:38
fennit's really unfortunate that compiling stuff is procedural knowledge and not declarative knowledge21:52
-!- Phreedom [n=quassel@195.216.211.175] has quit [Read error: 110 (Connection timed out)]22:08
genehackerkanzure don't you have some papers on microcontact printing22:42
kanzureprobably?22:42
kanzuredo you mean imprinting?22:43
genehackeryeah or stamp lithography22:44
kanzureyeah somewhere22:48
genehackerok22:49
genehackerhttp://www.rsc.org/Publishing/ChemTech/Volume/2007/03/hot_wax_and_peel.asp22:52
genehackerexcellent22:52
genehackera way to purify dna 22:52
-!- genehacker2 [i=genehack@wireless-128-62-43-62.public.utexas.edu] has joined #hplusroadmap22:59
-!- genehacker [i=genehack@wireless-128-62-43-62.public.utexas.edu] has quit [Read error: 110 (Connection timed out)]23:13
genehacker2WTF?23:58
genehacker2nvm23:59

Generated by irclog2html.py 2.15.0.dev0 by Marius Gedminas - find it at mg.pov.lt!