2009-05-06.log

--- Day changed Wed May 06 2009
kanzure-um.00:35
-!- genehacker_ is now known as genehacker01:48
kanzure-odd, the head-pain is happening at the same time tonight02:29
kanzure-hi kardan02:32
kardanhiho02:33
kardankanzure could you smell that I just was reading on your site .)02:33
kanzure-one of my monitors is always showing who is accessing my site.02:34
kardanthats spooky02:39
kardanim interested in building up osh groups and product distribution here in germony02:45
kanzure-what do you mean distribution?02:45
kardanselling .. promotion02:46
kanzure-to make profit?02:46
kardanto show people that open alternatives are better, and some day hopefully even cheaper02:47
kardanmy idea was - i know its not new - to put all 'profit' into osh bank/foundation to support new projects02:47
kardanwe already do this with fair trade drinks02:48
kardanyou surely know the ccc?02:48
kanzure-yes02:49
kanzure-I'm just very cautious of anyone talking about money 02:49
kardanright .. 02:49
kardani live without money and dont want any02:50
kardanbut i want to bring things forward you know02:50
kanzure-but the thing is that including money or something02:50
kanzure-makes people make stupid decisions02:50
kanzure-for instance, the open source hardware bank people are making some bad choices02:50
kanzure-and don't seem to be technically competent.02:50
kardanok lets talk about this and possible alternatives02:51
kanzure-alternatives for what though?02:51
kanzure-do you know about SKDB?02:51
kardanmoney free resource exchange for instance02:52
kardanno, whats SKDB?02:52
kanzure-an open source hardware repository and design tool02:52
kanzure-but think of it also as a 'kernel' of sorts to fablabs or fabratories02:52
kardanah, kind of core tools to access hardware02:54
kardana database of projects would be interesting02:55
kardanthere are many sites02:55
kanzure-yes, but they all suck02:55
kardanyea02:55
kardanand i resitate to collect them on some wiki page02:56
kardana real database would be nice02:56
kanzure-right.02:57
kanzure-somebody already started doing that02:57
kanzure-and it was ridiculous because they kept on ignoring me02:57
kanzure-and the work that we're doing in here02:57
kanzure-blah.02:57
kardanmh 'inner' conflicts .. lets overcome it02:57
kanzure-what?02:57
kardanexclusion ..02:58
kanzure-what?02:58
kardandoesnt matter .. just sounded like sb ignored 'you' with intention02:59
kanzure-maybe. or they might just be genuinely stupid.02:59
kardanso there is no good start to work on?03:00
kanzure-sorry?03:00
kardani mean a really open database which everybody can edit03:00
kanzure-right03:01
kardanwith criteria like title, project status, crew, url, description, maybe price ..03:01
kanzure-that's right03:01
kanzure-did you read the metadata todo file on OM?03:02
kanzure-http://heybryan.org/om.html03:02
kardanyou mean http://groups.google.com/group/openmanufacturing/browse_frm/thread/8465dc23eb48e332#03:03
kanzure-yes03:03
kanzure-it's kind of old, so some of it may no longer apply03:03
kanzure-but the majority of it does03:03
kardan"Simple script to handle python exception errors" thats great, i would like to think this bigger .. (unix wide)03:04
kanzure-right03:05
kanzure-do you know of package repositories like apt-get and yum or yast?03:05
kardansure03:06
kanzure-so yes. that, but for hardware.03:06
kardanhehe03:06
kardan> Tie-in of "open source hardware directories" with "one-click03:06
kardanordering" and sourcing of materials from suppliers 03:06
kanzure-right03:07
kardanthat depends on direct cooperation with developers who maintain material lists for their products03:07
kanzure-not necessarily03:08
kanzure-debian has "package maintainers"03:08
kanzure-so that's probably what will happen here03:08
kanzure-because not everybody is willing to package their projects up03:08
kanzure-but if they describe them sufficiently, maintainers can package them up into some somewhat usable form03:08
kardangreat. so which informations do we need beneath used materials03:10
kanzure-what?03:10
kardani mean a list of criteria 03:12
kanzure-for what03:12
kardanfor this db03:12
kanzure-criteria for what?03:12
kanzure-I mean, that sounds very vague03:12
kardanlike provides or depends03:14
kanzure-yeah, different 'hardware packages' need to specify their hardware dependencies03:14
kardanso users can navigate through list of products03:15
kanzure-sure, why not03:15
kardanwhich or not only shops but also interesting for developers03:15
kardans/or/are/03:16
kardanwe (brueckenschlaeger) are developing with riseup to improve crabgrass which is a open source social platform03:17
kanzure-what does "social platform" mean03:17
kardanthere are ideas on https://we.riseup.net/opensourcehardware03:17
kardanits a website to collaborate03:18
kardanmuch more powerful than a wiki03:18
kanzure-I don't think a website for collaboration makes since03:18
kanzure-most of my programming is done in a text editor03:18
kanzure-*not* a browser03:18
kardando you know gobby?03:19
kanzure-no03:19
kardanits a very simple real time collaboration text editor03:19
kanzure-doesn't look like a website to me03:19
kanzure-looks like a standalone app03:20
kardanyep03:20
kardanmaybe this makes sense03:20
kardanto develop documents as group in realtime03:20
kardanbut more interesting is the database ..03:23
kardani could set up one for testing03:24
kanzure-maybe you would be interested in packaging up some demos/prototypes?03:25
kardanif you mean to put some projects into the database and maintain their development, yes03:28
kanzure-so at first we're looking to do very simple projects03:28
kanzure-nothing complicated03:28
kanzure-something like "screws" or "fixtures" would be a good start03:28
kanzure-I was doing some IGES demos of different microcentrifuge tubes03:29
kanzure-at the moment I'm sitting around thinking about how to extract linear dynamic equations from the packages03:29
kanzure-between different connections, I mean03:29
kardanmaybe i didnt get it, you mean manuals with pictures and step by step instructions?03:35
kanzure-the step-by-step instructions are to be generated by computer03:36
kanzure-that's what I work on in the lab I'm doing research in03:36
kardanso you need a db interface to insert what?03:38
kanzure-sorry?03:38
kanzure-you insert the packages into the repository03:38
kardanone file per project/pkg03:39
kanzure-no, I think there will probably be multiple files03:39
kanzure-but they could be put into a tar file03:39
kanzure-so that ultimately it is indeed a single file03:40
kardancould you draw an example in which format?03:40
kardanbecause to create such a tar a db would be necessary03:41
kanzure-well, there's a metadata format that I'm working on,03:41
kanzure-but I think that CAD files like IGES or STEP would be the standard 03:41
kanzure-the metadata would refer to surfaces in the CAD files03:41
kanzure-(as well as dependencies, materials, etc.)03:41
kardanso we need a web interface to edit those date03:42
kanzure-not necessarily03:43
kanzure-the web interface can come later03:43
kardanor some other db interface03:44
kanzure-right, maybe some tools.03:44
kardan(as rails developer this is my first choice, because it automatically create web interfaces to edit db entries)03:45
kardanso if you have an example i could configure the fields and set it up03:45
kanzure-no, I don't have the linear dynamic equation thingy worked out yet03:46
kardani create the model like this http://www.tutorialspoint.com/ruby-on-rails/rails-migrations.htm04:20
kardankanzure what exactly would be my part in your plan?04:33
kanzure-package maintainer :)04:41
kanzure-or possibly somebody who knows more about linear dynamics than I do?04:41
genehackerlinear dynamics04:48
genehackerI'm taking dynamics04:49
kanzure-need a way to represent the dynamics of individual components04:50
kanzure-and then string them together04:50
kanzure-for a symbolic equation of multiple components in some design04:50
kanzure-one way is bond graphs. another way is linear graphs. but I'm thinking there might be some other way to do it.04:50
kanzure"Reception of Pictures from the Weather Satellites Using Homemade Equipment"05:23
kanzure"home-improvement? now it's becoming you-improvement."05:29
kanzure A Do-it-yourself Expert System for Shopfloor Use05:31
kanzureM Louha - Proc. 5th Int. Symp.‘Automated Systems Based on Human Skill’ …, 199505:31
kanzurehttp://www.lavoisier.fr/notice/frEWO3KLRAKRWLKO.html DIY labotomy05:35
genehackercall me back when they do spy sats05:43
genehackerlagrange method05:43
kanzurehm. optical fiber made out of PMMA holes + fructose ?07:10
genehackerFructose?07:59
fenn_i wish people wouldn't automatically assume 'SQL' or tables and columns when you say database11:17
-!- fenn_ is now known as fenn11:17
fennthe whole tables thing is so lame11:17
fennCommies all over the place, a great worldwide conspiracy to rob Americans of their hard-earned cash. Goddamn open-source software, hardware. Open source mustangs, open source wagons headin out West’, gonna turn this society upside down. Ain’t no hole in hell deep enough for ‘em.11:53
kardan_"Commies all over the place, a great worldwide conspiracy to rob Americans of their hard-earned cash" lol12:21
kardan_thats ridiculous, i really dont know what you are afraid of12:29
fenni thought it was pretty funny12:33
kardan_what did you read?12:34
kanzurekardan_: do you speak english?12:34
kardan_no i just pretend to12:35
kanzurebecause I've had trouble communicating with you12:35
kardan_seemed so :)12:35
fennit was a comment on this page in response to another comment; search for 'patent' for the original comment http://www.wired.com/gadgetlab/2009/03/open-source-har/12:35
fenni hate it when they dont have anchors for comments12:36
kanzurehrm. water-core Fresnel fiber.12:48
kanzurewhat's wrong with these? why isn't everybody building these for optical data transfer?12:48
kanzureit's a freaking water-core optical fiber ..12:48
kanzurebetter yet, air-core. yay.12:52
fennair has the wrong index of refraction12:53
fennyou have to have a higher index at the center12:53
fennwater absorbs IR12:53
fenni dont know how the spectrum compares to glass12:54
kanzuregimme a chance to read more12:54
kanzure"The fibers produced resemble ring-structured Bragg fibers, and transmission in the hollow core was observed for a range of wavelengths, including the infrared"12:55
kanzurehttp://adl.serveftp.org/papers/Hollow-core%20microstructured%20polymer%20optical%20fiber.pdf12:56
kanzurefenn: that image of the forest of CNTs with microspheres in that lipkowitz-paper was really neat.12:59
kanzure"SEM image of a MWCNT forest seeded with 330 nm polystryene spheres"13:00
fennyeah13:01
kanzure"Dew should not be confused with guttation, which is the process by which plants release excess water from the tips of their leaves."13:08
kanzurehttp://en.wikipedia.org/wiki/Air_well_(condenser)13:11
kanzurehttp://en.wikipedia.org/wiki/Guttation13:11
fennhah. chopped all those knobbly things off my bike tire, now it works13:53
kanzureerm, how long did you spend on that13:56
kanzuremaybe you shouldn't tell me13:56
fennabout half an hour14:00
fennin the process i found that i had two broken spokes14:01
kardan__ikanzure this is what i thought of: http://oshw.0xb5.org/14:10
fenntwo projects?14:14
fennoh this is a demo of the web based database editor14:14
kardan__ithe content doesnt matter14:14
fennthe 'depends' field is where things get tricky14:14
kardan__ieverybody may edit it atm14:14
fennit's not so easy as with software where you just say 'requires such n such library, end of story' and it's no big deal for anyone to download that library14:15
kardan__idepends should point to another project14:15
kardan__iyes14:15
fennwe've gone over this at length.. i'd love to share some files with you but my webserver's down14:16
kardan__iat least instructions are necessary how to include other peaces14:16
kardan__iexamples would be great. you can send them to kardan@brueckenschlaeger.de14:17
fenni've been hoping someone would come help write a web interface for a long time now14:20
fennbut the back-end is still barely even started14:21
fenni dont think a relational database is going to cut it14:22
fennespecially not with distributed development of the database 14:22
kardan__iwith such a db interface we can create files of every we like, i just need to write an exporter14:22
-!- kardan__i is now known as kardan14:23
kardanmh maybe its not best but i have no better solution yet14:23
fennok. at worst you could just show some text forms to edit the raw yaml14:23
fennand store the yaml in the database14:23
fennbut that is sorta pointless14:23
fennyou want to do something like 'select from screws where thread = M8' right?14:24
kardansort of. especially i think of a material table which enables much more14:25
fennyeah. it's hard to know where to begin14:26
kanzure"microplasma powder deposition" <- heh.14:26
fennkanzure has scraped some material databases14:26
fenni'm not really sure of the utility though since we don't know where to actually get most of them14:27
kardanfor instance a motherboard could have some materials as provides and a computer would list it a s needed material / depend14:27
kanzureexcept for mindat14:27
fenni mean it wouldn't be hard to go through mcmaster.com or onlinemetals.com and make a reasonably complete materials selection14:27
kanzuresoo14:28
kanzureI think I might leave out symbolic equations14:28
fennoh, that's what we were calling a 'package'14:28
kanzureif I do, it's less awesome14:28
kanzurebut doable14:28
fennkanzure: i think mathematical models are useful but they odnt always exist so it shouldnt be a requirement14:28
kanzurebut where would they go?14:29
fennit's not very hard to implement though, see my screw thread force calculation for example14:29
kanzureif they are available, for, say, a gear?14:29
fennclass method14:29
kanzurethat doesn't explain anything14:29
fenninherit methods from more generic classes for more automagical solution14:29
fennbut you never get to 100% magic14:29
fennyou have to read the manual at some point14:29
fenni havent read albert's paper14:30
kanzureI don't recommend it14:30
kanzureso, in the case of a-design, the equations were being constructed by taking "through" and "across" variables14:31
fennmy idea of a gear train is it has some function that describes the relationship between input and output positions and forces14:31
kanzureerm14:31
fennyeah, except without the stupid names like 'through' and 'across' :P14:31
kanzurein the "control point" data structure, basically campbell had some variables that defined the derivatives and integrals for going in certain directions14:31
kanzurewell, I was hoping reading his source code would help explain things14:31
kanzurehaven't got around to that today yet14:31
fennreading the source code almost never helps explain things14:31
fennwelp i wish i had read this stuff, now i'm going to go make a fool of myself yes? :)14:32
kanzurenot really14:32
fenncampbell called if functional parameter right?14:32
kanzureyes14:32
fennnot control point14:32
kanzureyes14:32
fennok so, what is an fp?14:33
kanzureit is like what I am calling 'control point' heh'14:33
fennis it just some numbers that plug into super-duper universal modeling equations?14:33
kanzureexcept his FP thingy is tailored for something like 'bond graphs'14:33
kanzureerr14:33
kanzurenot really14:33
fenni dont know what a bond graph is either14:33
kanzureso in the FP data structure there is a sub list called MG-change14:33
kanzureand this MG-change structure has a few arrays: through[n], across[n]14:34
kanzureeach value in the arrays is an equation14:34
kanzurewhere n = number of ports or interfaces14:34
kanzuresomehow these equations are assembled together by doing certain path traversals14:34
kanzurethrough the design.14:34
fennok14:35
fennso they just use the same equations for everything, which is what i said14:35
kanzurein this paper, http://heybryan.org/books/papers/Guiding%20conceptual%20design%20through%20behavioral%20reasoning%20-%20Welch%20-%20Dixon%20-%201994.pdf14:35
kanzurethey use some sort of 'directional derivative' (pg 4)14:35
kanzurewhich is the stuff about either directly proportional or inversely proportional14:35
kanzure(or no proportional relation)14:36
kanzurethe equations though don't really seem to be per-component/class, but instead per-port.14:36
kanzurewhich is weird.14:36
kanzureI've been asking campbell why he thinks this bond graph stuff is necessary, and why can't it just work by resolving the units14:37
fennit shouldnt be proportional vs inversely proportional though; should be function vs its inverse14:37
kanzurewell in the equations I mean, er14:37
fenninverse like f(inverse(x)) = x14:37
kanzureok, I don't have it14:37
kanzurebut one of the examples was something that calculated the theta/angle of a dial14:38
kanzurein a given design generated by the a-design software14:38
kanzureand in this equation for calculating theta,14:38
fennthere were lots of those..14:38
kanzurethere was a (1 / spring-constant)14:38
fennright i remember that14:38
kanzureas opposed to (1 * spring-constant)14:38
kanzureso 1/x is inversely proportional14:38
kanzure1*x is a proportion14:38
kanzure* so 1/x is an inverse proportion14:38
kanzureand they some how then assemble (1/x)(y)(1/z) together by some path traversal through the design graph? or something?14:39
kanzurebut I don't know if that makes sense. I haven't really ever done that when I've done designs ..14:39
fennman your server is so slow14:41
fenni think it was this http://heybryan.org/books/papers/campbell-dissertation-diagram-thingy.png14:41
kanzureaha, yes14:41
kanzureer, so, I think it woudl be faster if I set up ProxyPass on one of the servers on the campus14:41
kanzuresince I have like 10 MB/sec in and out of the university14:41
fennthe spring constant is inverse because it's in parallel with the lever14:42
kanzureif it was serial, it would not be inverse?14:42
fennwell, that's just my gut reaction14:43
kanzureparallel versus serial is a common theme that campbell was ranting about yesterday when I was talking with him14:43
kanzure*was a common theme14:43
fenni dont really know what the spring is doing14:43
fenni suppose i should get down there14:44
kanzureok, give me a ring if you end up in Mexico or something14:45
-!- any72684526 is now known as katsmeow-afk15:37
-!- any61810568 is now known as katsmeow-afk15:44
-!- any81413435 is now known as katsmeow-afk15:49
-!- any38463417 is now known as katsmeow-afk15:57
-!- any83825325 is now known as katsmeow-afk16:06
kardanhi #h+19:35
kanzurehello19:38
kardan_i think i will collect projects in the database later. maybe it needs some security features first.20:02
kanzure-in adesign/tests/coordtests/oldtests/test.lisp in the first few lines there's an example of MG-change written out.23:46

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