--- Day changed Fri Nov 14 2008
00:37 < bkero> Brains
00:47 < fenn> what graduate program would ADL be associated with?
00:49 < fenn> i'm guessing either ECE - Manufacturing systems engineering; operations research & industrial eng; or mechanical engineering 
00:52 < kanzure_> Not industrial, as it turns out.
00:52 < kanzure_> Industrial and operations is upstairs.
00:53 < kanzure_> Manufacturing systems, hrm.
00:53 < kanzure_> well, ADL is within the MAD lab, "Manufacturing and Design"
00:53 < kanzure_> uhm. http://www.me.utexas.edu/ should have something.
00:54 < fenn> Manufacturing Systems Engineering
00:54 < fenn> This area emphasizes the application of computers, information sciences, and information systems to the development of equipment and software systems for manufacturing.
00:55 < fenn> mad lab website looks like it hasnt been touched in 5 years
00:55 < kanzure_> it's been a decade prolly.
00:55 < kanzure_> um, mechanical is a safe bet.
00:56 < kanzure_> I haven't figured out what the MSE stuff is (in terms of people/labs) yet
00:56 < kanzure_> maybe it's the same people. don't know.
00:56 < kanzure_> http://www.me.utexas.edu/~bryant/mfg/
00:57 < kanzure_> ah, it no longer exists anyway.
00:57 < fenn> too bad
00:58 < fenn> i wonder if they are going to make me take calculus again
00:58 < kanzure_> Did you fail it?>
00:59 < fenn> no, but i've forgotten all of it
00:59 < fenn> from sheer disuse
00:59 < kanzure_> Doesn't sheer require calculus?
00:59 < kanzure_> or am I thinking of shear?
00:59 < fenn> i'm sure lots of engineering crap requires calculus, but it's totally irrelevant for what i want to do
01:00 < fenn> since i'm not doing engineering per se, more like meta-engineering
01:00 < fenn> and besides, that's what computers are for
01:01 < kanzure_> the multivariable calc grad/TA-dude basically just uses mathematica most of the time, except when lecturing/teaching/helping evidently.
01:01 < fenn> of course, why would you do it by hand
01:01 < fenn> i dont do long division by hand either
01:01 < kanzure_> because you're in a class :(
01:03 < kanzure_> actually I think they have graduate curriculums on the site.
01:04 < fenn> i dont see that anywhere
01:05 < fenn> looks like the other MS's are just plain ME
01:06 < kanzure_> yes, but there's "subspecializations" thingies.
01:06 < fenn> i dont even want to be a grad student, this is stupid
01:06 < kanzure_> heh
01:06 < fenn> i just want someone to pay me to work on skdb
01:06 < fenn> how hard is that
01:07 < kanzure_> if there's a way to be on the payroll without being enrolled as a student, I'd be the first one to sign up
01:11 < kanzure_> it's worth investigating. 
01:12 < kanzure_> every lab seems to have their nonstudent person. in Ellington's lab it was Zack, in ACTlab it's just people hanging out, etc.
01:12 < fenn> it would really simplify things and allow me to begin immediately
01:12 < fenn> instead of waiting until.. august?
01:12 < kanzure_> for what?
01:12 < fenn> i might be dead by august ffs!
01:12 < fenn> for the academic year to start
01:14 < fenn> looks like there is no engineering GRE, yay!
01:14 < kanzure_> some of the grad students talk about 'quals'.
01:15 < kanzure_> that might be for phd stuff though
01:15 < fenn> i think that's like finals
01:16 < fenn> ah, quals is required before you start your dissertation (whatever that is)
01:44 < fenn> jesus christ why is everyone talking about the venus project
01:45 < kanzure_> it's not even that good.
01:45 < kanzure_> So it was funny today when I showed campbell approppedia
01:45 < kanzure_> appropedia
01:45 < kanzure_> and then when we went into his office for a few seconds to check his machine for something,
01:45 < kanzure_> in the corner it popped up, somebody emailed him "you have to check out appropedia"
01:45 < kanzure_> he hadn't heard of it before I told him about it 
01:45 < fenn> whee
01:45 < fenn> who was it?
01:45 < kanzure_> I don't know. Didn't catch it.
01:46 < fenn> your tachikoma agent?
01:46 < kanzure_> No, she was at home watching porn. :(
01:46 < fenn> lazy robots
01:47  * fenn glares around the room with x-ray-laser-eyes
01:47 < kanzure_> ?
01:47 < kanzure_> GITS ref?
01:47 < fenn> bot-on-bot violence
01:48 < fenn> nevermind
01:48 < fenn> i dont think i can keep up with paul's emails
01:49 < kanzure_> He may be more of a Markov bot than me.
01:49 < fenn> they're essay-length but unlike eric hunting they're mostly recycled material
01:49 < kanzure_> He has specific sections in his emails that repeat.
01:50 < kanzure_> Don't know how he keeps track of the recyclable elements.
01:50 < kanzure_> heheh.
01:50 < kanzure_> It would be fun to write a Paul email imitator app.
01:50 < kanzure_> it would be easy to write a few regexps to extract those recyclables, and then just generate emails based off of input keywords of a message
01:50 < kanzure_> I'm certain this would work.
01:50 < fenn> or you could run a paragraph-length markov bot on his website or a collection of previous posts
01:51 < kanzure_> is that equivalent?
01:51 < fenn> less work, more diversity
01:51 < fenn> unless you mean to use the regexp to keep quotes intact
01:51 < fenn> (he uses lots of quotes)
01:52 < kanzure_> yeah, I was thinking I'd just look for links
01:52 < kanzure_> since he usually has quotes nearby
01:52 < fenn> ok that's enough gossip.. perhaps i'll go spend the last of my money on ice cream
01:52 < kanzure_> last?
01:53 < fenn> borrowed money doesnt count in my book
01:53 < fenn> and i'm having some bank issues so i cant get at it anyway
01:53 < kanzure_> bank vanish?
01:53 < fenn> they are just being unreasonably slow
01:54 < fenn> i was trying to close my account, but the transfer is taking forever. they'll probably be closed for the weekend by the time it finishes
03:27 < gene> hey Kanzure want to get into the business of selling GMOs?
03:27 < kanzure_> Why do you ask?
03:28 < gene> how else would we get money to do research
03:28 < gene> on biohacking
03:28 < kanzure_> Don't worry about the money.
03:28 < kanzure_> I have that covered.
03:29 < gene> while at the same time making a useful gene more accessible to the general public
03:29 < gene> oh yeah
03:29 < kanzure_> I was going to write some inventory software to manage inventories of plasmids in biobrick labs for accessibility. It would allow people to order plasmids.
03:30 < gene> you aren't for profit
03:30 < gene> from grocery stores?
03:30 < kanzure_> No, from people who have plasmids.
03:31 < gene> btw, anyway to hide a gene in an organism's genome where it sits unused and easily cleaved with RE's
03:31 < kanzure_> What do you mean by hide?
03:31 < gene> Easter egg
03:31 < gene> bio easter egg
03:31 < gene> Hmmmm....
03:31 < kanzure_> ?
03:31 < kanzure_> Do you mean not expressed?
03:32 < gene> yup
03:32 < kanzure_> That's genetic regulatory networks stuff.
03:32 < kanzure_> Not a matter of where to place it.
03:32 < kanzure_> So you put something like a lac operon in front of it.
03:32 < kanzure_> And a few primers and so on.
03:32 < gene> uh huh
03:33 < gene> I'm saying sell a GMO with unexpressed plasmids in it
03:33 < gene> that the FDA won't find
03:34 < gene> so what does it take to modify the avian genome
03:35 < gene> pretty much the same as modding mammalian genome right?
03:35 < kanzure_> It's just DNA..
03:35 < drazak> kanzure_: how do you do rapid gel electroporesis? eg. a couple hours from sample+pcr+gel?
03:36 < kanzure_> Uh? It was always a couple hours.
03:36 < kanzure_> Are you using some ridiculously lengthy method?
03:36 < fenn> lotsa volts, a fan to cool off the gel
03:36 < gene> what're you trying to do drazak?
03:36 < kanzure_> crank up teh volts :)
03:36 < fenn> slower gels have better resolution
03:36 < gene> don't you mean current?
03:36 < fenn> same thing
03:36 < fenn> gel is a fixed resistance
03:37 < fenn> you're going after current per unit cross sectional area
03:38 < drazak> kanzure_: ah ok, I had been given poor info than, sorry
03:38 < kanzure_> The FDA doesn't sequence genomes.
03:38 < gene> exactly
03:38 < kanzure_> drazak: Maybe. What source?
03:38 < kanzure_> gene: So you wouldn't have to hide it.
03:38 < drazak> kanzure_: a teacher :P
03:38 < kanzure_> drazak: Ah, the worst.
03:38 < kanzure_> gene: Besides, you could just, you know, avoid the FDA.
03:39 < fenn> you could just keep it in a freezer
03:39 < gene> not if you want to sell it here
03:39 < kanzure_> Sell it?
03:39 < gene> in the US
03:39 < gene> yeah
03:39 < gene> sell it
03:39 < kanzure_> If you start selling it, I start replicating it for free
03:39 < gene> yeah
03:39 < gene> I know
03:39 < gene> but investors don't know that
03:39 < fenn> what are you selling exactly?
03:40 < gene> something that could be rendered irreproducible
03:40 < gene> probably
03:40 < fenn> fat chance
03:40 < kanzure_>   ?
03:40 < fenn> RIAA and MPAA failed, with hundreds of millions of $ invested
03:40 < gene> to avoid environmental contamination
03:40 < gene> people will still buy it
03:41 < gene> even if it's free
03:41 < gene> really
03:41 < fenn> so why not sell insurance
03:41 < gene> they sell plants and animals
03:41 < fenn> or newspapers
03:41 < fenn> or linux distro's :)
03:42 < gene> can you distribute useful plasmids via a newspaper?
03:42 < kanzure_> What do you consider useful to be?
03:42 < fenn> erm uh you're being deliberately obtuse
03:42 < fenn> i'm saying your business plan sucks
03:42 < kanzure_> yay
03:42 < gene> 1. easy to seperate out
03:43 < gene> 2. useful for gene modification operations( restriction enzymes,heat tolerant polymerases and such)
03:44 < fenn> you can find these plasmids in any bio lab, the problem is not availability but rather artificial restrictions like patents
03:44 < kanzure_> and MTAs.
03:44 < gene> that's why I want to hide them in the genome of thing I am selling
03:44 < fenn> mail transfer agent? :)
03:44 < kanzure_> fenn: material transfer agreements. see my recent (~past 2 weeks) rants on inventory for plasmid mailing stuff, software thingies.
03:45 < fenn> rants where?
03:45 < gene> Mechanical Tyrant Agency?
03:45 < fenn> is this like "i promise not to give my lab supplies to terrorists"
03:45 < kanzure_> Material Transfer Agreements are required by universities before you can send biological agents.
03:45 < kanzure_> biological agents by mail.
03:45 < fenn> oo e. coli and taq polymerase.. scary
03:46 < kanzure_> right, it's bullshit stuff.
03:46 < fenn> someone could put e. coli in a bomb and contaminate the water supply!
03:46 < gene> HE HAS POTASSIUM PERCHLORATE HE MUST BE A TERRORIST
03:46 < fenn> or a welder
03:46 < fenn> weldor?
03:47 < fenn> or likes to make his socks really bright white
03:47 < gene> a welder is on gnomegland pecurity's watch list
03:48 < fenn> gnomegland?
03:48 < gene> shh....
03:48 < gene> you know what I mean
03:48 < gene> don't type it
03:48 < kanzure_> You're just acting paranoid because you think it's cool to be paranoid though.
03:49 < gene> indeed
03:49 < kanzure_> ..
03:49 < fenn> most paranoid people are that way
03:49 < fenn> at least the ones that talk about it
03:49 < gene> I USE A 1 GIG ENCRYPTION KEY
03:49  * fenn mutters something about "security culture"
03:49 < gene> TRY TO CRACK THAT FEDS!
03:50 < gene> anyway I've considered making potassium perchlorate before
03:50 < fenn> back when i was a terrorist... we had training workshops on how to avoid the fbi, but they were very straightforward and more about knowing the extent of the law than undertaking unreasonable technical solutions
03:51 < gene> I wish I could know my rank on the terror watchlist
03:51 < fenn> you could ask
03:51 < fenn> i think it would be a fun game to play
03:51 < fenn> kinda like "chicken"
03:52 < gene> actually it would be more like an RPG
03:52 < fenn> no, a mixed reality media
03:53 < fenn> crap. whatever that ilovebees thing was called
03:53 < gene> read about how to making bombs increase probability of being terrorist by 10 point
03:53 < gene> hmm...
03:53 < gene> wonder what they look for anyway
03:54 < gene> so how about blowing up the whitehouse
03:54 < gene> you know the inflatable one
03:54 < gene> it's been in it's box ever since I bought it
03:55 < fenn> no no, i live in the whitehouse, you can blow up the shack out back though
03:58 < gene> this topic isn't useful
03:59 < fenn> fine, be that way
03:59 < kanzure_> Why does he leave?
04:00 < fenn> ChatZilla
04:00 < kanzure_> So what?
04:00 < kanzure_> Oh right, firefox can't be on for more than 20 minutes ;-)
04:01 < fenn> do you think campbell would be interested in this? http://fennetic.net/cadwiki/index.php?FrontPage#vb206cb3
04:02  * kanzure_ types in the url ..
04:02 < kanzure_> is the anchor important?
04:02 < fenn> it's notes to myself on features i'd want in a cad-ish/skdb-ish application (before i came up with the idea of a technology distribution)
04:02 < fenn> the anchor is "goal oriented flow"
04:03 < fenn> just go to fennetic.net/cadwiki
04:04 < fenn> feel free to read the rest though
04:05 < kanzure_> user interface and data struct stuff to bottom.
04:05 < kanzure_> yeah, okay, linking me to that anchor makes more sense.
04:10 < kanzure_> I'm trying to figure out how to best convey the style that he tends to like. Talking search algorithms is good.
04:10 < kanzure_> Goals / user preference modeling is supposedly what I'm presently doing, which I guess I can make up some strange connections to goalism
04:11 < fenn> it's not really about goals, that was just sort of a buzzword/placeholder
04:11 < kanzure_> it's about 'requirements' no?
04:12 < fenn> right
04:12 < fenn> sufficiently vague requirements
04:12 < fenn> not specified to the last detail like in a drafting program
04:14 < fenn> anyway the point is to build all this optimization stuff into the cad program and run the optimizations in the background/on the server automatically, so you dont waste time switching around between tools
04:15 < kanzure_> right, I'm coding something up like that at the moment.
04:15 < fenn> lisp interpreter has a neat feature where you start typing in a statement and it is already evaluating it, so when you press return the answer comes back instantly
04:16 < fenn> oh well, this is "in the ballpark" at least
04:16 < kanzure_> sort of. I'm trying to figure you out a home run.
04:16 < kanzure_> oh, talking about hierarchical bayesian networks/probability would score some points btw.
04:17 < kanzure_> if you happen to actually know that area :/
04:17 < fenn> well, not really
04:18 < fenn> i dont see what bayesian networks have to do with process planning?
04:18 < kanzure_> no biggie, it was just a one-off comment he made to me recently.
04:18 < kanzure_> me either.
04:18 < fenn> its not like you can just make up data
04:20 < kanzure_> It's hard for me to figure out the "what's next" for ADL because once you have the giant cad system, and the toolchain from the tree search to the 3D modeling software and so on, that's about it for that sort of toolchain, plus or minus improvements/bugfixes..
04:20 < fenn> "Generalizations of Bayesian networks that can represent and solve decision problems under uncertainty are called influence diagrams." this sounds vaguely related
04:20 < kanzure_> huh.
04:21 < fenn> but i think he just has a hard-on for graphs :)
04:21 < kanzure_> also significant invested time in his codebase.
04:22 < fenn> hm yeah and i dont really get what it's supposed to do
04:22 < kanzure_> have you played around with it?
04:22 < kanzure_> graphsynth, I mean.
04:22 < fenn> no
04:22 < fenn> only the design repo stuff
04:22 < kanzure_> it compiles on linux, but it has some errors, though mostly works.
04:23 < kanzure_> so what the software does is take the substitution rules, given lefthand sides and righthand sides, and then goes through your given graph and keeps track of the number of different ways that the substitution could be applied, and then executes that replacement
04:23 < fenn> with mono? (linux)
04:23 < kanzure_> actually the website is rather informative for each of the sections, it just doesn't help to explain the GUI
04:23 < kanzure_> yeah
04:24 < kanzure_> monodevelop.
04:25 < kanzure_> One student has been writing a DLL extension to graphsynth for gear optimization. He has a set of substitution rules that take a user's "requirements" (in the form of a 'function structure' graph) and by various substitutions (+additions sometimes) there's this overall gear system, and then sprinkle in the optimization, and in the end it's a .gxml file that supposedly I'm reading in to glxgears.
04:26 < kanzure_> there's like four guys in the lab, for what it's worth.
04:28 < gene> http://burrtools.sourceforge.net/
04:28 < gene> are you pondering what I am pondering?
04:28  * fenn grumbles about the lack of a readme
04:28 < kanzure_> gene: That's like my supermetal app.
04:29 < gene> what's supermetal?
04:29 < gene> is harder than dragonforce?
04:29 < kanzure_> http://heybryan.org/shots/2008-11-02_supermetal_2.png
04:30 < gene> so what does it do?
04:30 < fenn> oh this was on johnny quest (cyberspace version)
04:31 < gene> huh?
04:31 < kanzure_> Given a set of primitives, allows you to randomly generate shapes and then do cool things with them. Export to 3D file format, for instance.
04:31 < gene> like?
04:31 < kanzure_> See the screenshot.
04:31 < fenn> they dug up a bunch of smashed pieces of a statue, scanned them in, and then let the computer assemble the 3d models
04:32 < gene> wait a minute really?
04:32 < gene> just like in snow crash?
04:32 < kanzure_> It's a common scene in investigative TV shows.
04:32 < kanzure_> "Broken bottle! But it's okay, we can painstakingly reconstruct it and extract fingerprints!"
04:33 < kanzure_> or in the latest Batman, given sharpnel => reconstruct the original bullet.
04:33 < gene> how does supermetal do that?
04:33 < fenn> these crime scene scenarios are rather close to what we need to do with magical-engineering-app
04:33 < gene> you are pondering what I am pondering then
04:33 < kanzure_> Supermetal is going in the opposite direction, but you just reverse things around and do collision detection and some quick math for checking whether or not basic constraints are satisfied.
04:34 < kanzure_> fenn: ?
04:34 < fenn> kanzure_: interfaces
04:34 < fenn> mcmaster catalog = smashed pieces
04:34 < fenn> edges of glass = product specifications
04:34 < gene> so supermetal breaks a part into pieces?
04:34 < fenn> combine them into something useful
04:34 < kanzure_> gene: No, supermetal assembles from primitives.
04:34 < gene> assembles what?
04:35 < fenn> stuff
04:35 < gene> random stuff?
04:35 < kanzure_> All stuff.
04:35 < fenn> icosahedrons from the looks of it
04:35 < gene> like?
04:35 < kanzure_> Say there's N possible primitives.
04:35 < kanzure_> and P set of points per primitive.
04:35 < kanzure_> Then you can loop through each one and make the graph.
04:36 < kanzure_> Now, to rate them, and score the constructions, you do some mathy thingies. Like whether or not you like there to be GIANT GAPING HOLES in the structures.
04:36 < gene> did you make it yourself?
04:36 < fenn> giant gaping holes can be good or bad or not matter
04:36 < kanzure_> Supermetal? yeah
04:36 < kanzure_> fenn: that's user-dependento f course.
04:36 < gene> and the gui
04:36 < gene> for it
04:36 < kanzure_> There is no GUI except the 3D visualization.
04:37 < gene> which is in?
04:37 < kanzure_> Huh?
04:37 < gene> what?
04:37 < gene> is it in
04:37 < fenn> but kanzure_ how do you generalize the concept of "giant gaping holes", and how many of these descriptors are there?
04:37 < kanzure_> a computer
04:37 < kanzure_> fenn: plugins.
04:38 < kanzure_> plugins for new varchecks etc.
04:38 < fenn> varcheck = what?
04:38 < gene> huh?
04:38 < kanzure_> distance between points/edges for instance?
04:38 < gene> I have know idea what u r talking about
04:39 < fenn> so varcheck is just a function
04:39 < kanzure_> What did you think I was talking about?
04:39 < fenn> gene: please limit the drooling idiot abbreviations
04:39 < gene> ok
04:39 < fenn> thanks
04:40 < fenn> kanzure_: would you say this is basically a genetic algorithm?
04:40 < fenn> varcheck ~= fitness function
04:41 < gene> you made a GA kanzure?
04:41 < fenn> unless you mean to calculate every possibility
04:41 < kanzure_> varcheck =~ scoring function.
04:41 < gene> how does it score it?
04:41 < kanzure_> The scoring function is not yet implemented because I don't care.
04:41 < kanzure_> It doesn't because I haven't implemented any of the checks.
04:41 < gene> how do your gene's work
04:41 < fenn> uh oh, did i say a naughty word?
04:41 < kanzure_> I guess if you wanted to you could check the distance between tips of objects on it.
04:42 < gene> so it's not really a GA?
04:42 < kanzure_> I didn't say it was a GA.
04:42 < kanzure_> It could be a GA though if you wanted.
04:42 < kanzure_> The genes would be the sequence of primitives or whatever.
04:42 < gene> I don't get what it does
04:42  * fenn feels like this is going nowhere
04:43 < kanzure_> There's a set of N primitives; you assemble it by saying primitives #1, 34141, 341, 5, 1, 39, 394
04:43 < kanzure_> So the genes for that one would be that list of primitives
04:43 < kanzure_> It's slightly more involved than that though because I implemented points of tangency
04:43 < gene> what's input what's output?
04:44 < kanzure_> input: number of 'nodes' of the graph or whatever. 
04:44 < kanzure_> output: 3D object.
04:45 < gene> ok
04:48 < gene> can it convert graphs to real objects
04:48 < gene> like from the ADL?
04:49 < kanzure_> Yep. I'd have to write a gxml -> a data format thingy for the app but that's ok. That's just file format conversion stuffs.
04:49 < fenn> the design repo doesnt contain any 3d data
04:49 < fenn> so you'd have a bunch of spheres stuck together to represent a camera, i guess
04:50 < kanzure_> hold on a sec.
04:50 < fenn> unless i'm missing something?
04:50 < kanzure_> So, the gxml data is the graph stuff. Like those 'CFGs' (component function graphs) that I generated with graphviz
04:50 < kanzure_> (grxml is the substitution rule stuff)
04:50 < kanzure_> The design repo, and all of those .repo files, do not contain gxml. :(
04:50 < kanzure_> They have files within them though, that's why I was so annoyed last month.
04:50 < fenn> gxml describes what?
04:51 < kanzure_> a graph.
04:51 < fenn> a graph that represents what
04:51 < kanzure_> It either describes a 'function structure' (black box) or 'CFG'
04:51 < kanzure_> CFG would be like "axe -> hammer"
04:51 < gene> so how would you make that meat slice gun real?
04:51 < kanzure_> Hold on.
04:51 < kanzure_> Some of the .repo files do not have those images, some do, etc.
04:52 < fenn> i'm not interested in images
04:52 < kanzure_> Some have 'CDD' (ConceptDraw files). Some have 'functionCAD' files. But CDD is proprietary and not easily converted to anything usable, as we know.
04:52 < kanzure_> Also, the database sucks; there's no way to extract the 'function structure' information about the connectivity.
04:52 < kanzure_> Specifically, a subfunction of an 'artifact' (which is kinda like in the CFG layer)
04:52 < kanzure_> points to another artifact; not to a 'port' on that artifact (like what subfunction the other artifact has that is relevant)
04:53 < fenn> unfortunate, that
04:53 < kanzure_> So you see stuff like "laserpointer -> table" but not that it's the solid interface that matters, or not that it's the laser-pointing part.
04:53 < gene> can't you figure out things like component is wedge, gear, bearing, or plane or such?
04:53 < kanzure_> yeah, so it's basically somewhat unusable data
04:53 < gene> uh wow
04:53 < kanzure_> the design repo doesn't contain 3D information, that's correct
04:54 < gene> matlab represents a missile guidance system as a graph
04:54 < gene> and a car as a graph
04:54 < kanzure_> yeah, electronics is very commonly a graph
04:54 < kanzure_> yep.
04:54 < fenn> control system, not the actual "how to build it" info
04:54 < kanzure_> "Function structures" are popular.
04:54 < kanzure_> right, the 'how to build it' info is something lacking in this stuff. 
04:55 < fenn> diagrams (graphs) for control systems is the only way to do it
04:55 < kanzure_> conversion to gcode and such is one way, but I want the origami assembly folding stuff too.. anyway.
04:55 < kanzure_> as for the meat slice gun thingy,
04:55 < gene> you know knives are simple machines
04:55 < kanzure_> It's very easy to make it real *if* the individual parts have 3D representations in the repo
04:55 < gene> the first machines made by man were knives
04:55 < fenn> so are rocks
04:56 < fenn> and sticks
04:56 < kanzure_> and dung.
04:56 < fenn> and hooting noises
04:56 < kanzure_> Is it a monkey?
04:56  * fenn bangs on chest (visual state indicator)
04:57 < gene> why not use all that ADL stuff to make stone age 3d structures
04:57 < fenn> indeed
04:57 < gene> rome wasn't built in a day
04:57 < fenn> this is why i'm interested in "love"
04:57 < fenn> quelsolaar.com
04:57 < gene> I don't know what you mean
04:57 < gene> meat
04:58 < gene> ok
04:58 < fenn> graph grammars for building primited 3d structures
04:58 < fenn> primitive*
04:58 < fenn> (oh did i mention they're virtual?)
04:59 < fenn> yuck IDE's :)
05:00 < kanzure_> It's all IDEs. :(
05:00 < kanzure_> oh
05:00 < kanzure_> GUIs.
05:00 < kanzure_> Brainfart.
05:01 < fenn> so i have monodevelop and the source, what do i do?
05:01 < kanzure_> fenn: you could pitch campbell my automated reverse engineering of the patent db stuff.
05:01 < kanzure_> fenn: Alt+P + enter should build it.
05:01 < fenn> i dont think that's feasible or valuable
05:01 < fenn> i think i need to open a file first
05:02 < kanzure_> oh, open up the sln file
05:02 < kanzure_> GraphSynth.sln somewhere.
05:02 < fenn> yeah i did that and nothing happened
05:02 < kanzure_> try the .suo ?
05:02 < fenn> probably not liking being run over ssh/X
05:03 < kanzure_> where.. are you?
05:03 < fenn> oh i see my screen is too small
05:04 < gene> A = [1 2; 3 4];
05:04 < fenn> Assembly '/home/fenn/sandbox/GraphSynth.v.1.9.src/GraphSynthFiles/Representation/bin/Debug/Representation.dll' doesn't have an entry point.
05:04 < gene> crap
05:04 < gene> this isn't matlab
05:04 < kanzure_> fenn: try building the whole project
05:05 < kanzure_> fenn: and then run the Application_UI_thingy
05:07 < gene> you still have that john doe account kanzure?
05:07 < kanzure_> Nope.
05:07 < gene> damn
05:07 < gene> guess I'll have to download ubuntu my self then
05:07 < gene> wiat
05:07 < gene> so that's why the server went down?
05:07 < kanzure_> Yes.
05:08 < gene> that sucks
05:08 < gene> is it still down?
05:08 < gene> did you get throttled?
05:08 < kanzure_> Go check?
05:08 < kanzure_> http://heybryan.org/
05:08 < kanzure_> I don't know about the throttling. That's a new bug to me.
05:09 < kanzure_> fenn: Load up something from the input/ dir, then apply one of the things in the rules/ dir. Then you should get some stuff.
05:10 < gene> Ponder this: there will come a day when there will be no betas
05:10 < fenn> i have nfc what i am doing
05:10 < kanzure_> ..
05:10 < kanzure_> is it doing anything?
05:11 < fenn> it opened an xml file in firefox :(
05:11 < kanzure_> how did that happen?
05:11 < fenn> i did file -> open swirlsomething.gxml
05:11 < kanzure_> and then *bam* firefox?
05:11 < fenn> ya
05:12 < fenn> doesnt help that half the stuff is off the screen
05:12 < kanzure_> ?
05:12 < kanzure_> you can resize the window.
05:12 < fenn> no, there's a minimum pane size
05:12 < kanzure_> there's a menu item that lets you resort the nodes in the graph, if you see any graph. 
05:13 < fenn> i got rid of all the toolbars and now i can see one line of shell output
05:13 < kanzure_> I don't know why it opens firefox. Do you have some weird file clicker setup?
05:13 < kanzure_> shell output? 
05:13 < fenn> gah
05:13 < kanzure_> Did you get graphsynth running?
05:14 < fenn> no
05:15 < fenn> i am still at step 1: http://fennetic.net/pub/irc/monocrap.png
05:16 < kanzure_> yeah, I'm surprised opening up the sln file didn't work
05:17 < kanzure_> when you installed monodevelop do you think it installed the compilers?
05:17 < kanzure_> sln is weird. on the left in that list of directories I don't see the one that starts with the word 'Application'
05:17 < fenn> monodevelop installed like 50 packages (100MB or so)
05:17 < fenn> it better fucking have installed some compilers
05:17 < kanzure_> okay. good bet that it was there.
05:19 < fenn> "solution" means "binary" more or less?
05:19 < kanzure_> What?
05:19 < fenn> there's a lot of terminology i've never seen before
05:21 < fenn> i guess bytecode would be more appropriate
05:21 < kanzure_> bleh? I don't remember seeing that option.
05:26 < kanzure_> Try restarting monodevelop. The sln file is the project file and should be openable.
05:28 < fenn> when i open the sln it just acts like i opened the .suo
05:28 < kanzure_> is there other GraphSynth.<something> files?
05:28 < fenn> no
05:29 < kanzure_> what's the full dir path that you find the sln in?
05:29 < fenn> /home/fenn/sandbox/GraphSynth.v.1.9.src/GraphSynthFiles/GraphSynth.sln
05:29 < kanzure_> blah.
05:29 < kanzure_> Same for me.
05:30  * fenn googles on "doesn't have an entry point."
05:30 < kanzure_> it means that you're trying to run the wrong thing.
05:30 < kanzure_> the "Application_Search_UI" thing is what you want to run.
05:31 < kanzure_> but it's for some reason not included in your left-hand list thingy in that screenshot..
05:31 < fenn> because the files tab is selected
05:31 < fenn> it's there in the solutions tab
05:31 < kanzure_> oh good. try selecting it and right click and go to run or something like that.
05:32 < kanzure_> also, if you go into GraphSynthFiles/the one that isn't rules or input or output/Release/bin/GraphSynth.exe
05:32 < kanzure_> or whatever it is. just go see if there's a full exe.
05:32 < fenn> aha it worked
05:33 < kanzure_> it?
05:33 < fenn> right click on App_search_UI
05:33 < kanzure_> yay.
05:34  * fenn makes train derailment noises
05:35 < fenn> crashed
05:35 < fenn> wow that was only like twenty seconds
05:35 < kanzure_> did it give an exception about XML and input or output?
05:35 < fenn> Unhandled Exception: System.NullReferenceException: Object reference not set to an instance of an object
05:35 < kanzure_> you might have to replace "\\" in the files with "/" via the 'search 'tab. It's one of those windows-things.
05:35 < kanzure_> oh.
05:35 < fenn> it opened the file fine
05:35 < kanzure_> but then crashed?
05:35 < fenn> all the nodes were clumped together so i tried to pick them apart
05:36 < kanzure_> yeah?
05:36 < fenn> then it crashed
05:36 < kanzure_> I fixed this earlier today.
05:36 < kanzure_> um.
05:36 < kanzure_> basically the problem is that the node name is null.
05:37 < kanzure_> so there's a part of the system that reads in the nodes or something. 
05:37 < kanzure_> it will be in the XmlIO files 
05:37 < kanzure_> you want to do something like if(!(some var.Length==0) in front of a for loop that checks for arc names.
05:37 < kanzure_> yay for being terribly vague. 
05:38 < kanzure_> oh, it's in loadXml in the designGraphXmlIO.cs file.
05:38 < fenn> is there a way to kill just the app when it's misbehaving and not monodevelop?
05:39 < kanzure_> it should be executing it separately, so killall -9 graphsynth should work (check ps -A or top)
05:39 < kanzure_> but I've never had problems, I can always just close it via the 'x' button on the window.
05:40 < fenn> i should have been able to figure that out
05:40 < fenn> problem is "about graphsynth" has no 'x' button
05:40 < kanzure_> oh
05:40 < kanzure_> yes, well,
05:40 < kanzure_> on KDE what I do about that is I send it to another desktop
05:41 < kanzure_> :)
05:41 < fenn> hmm then it drags me with it to that desktop
05:41 < kanzure_> can you go back?
05:42 < fenn> i can send it to the bottom of the stack but then the pulldown menus dont work in the app
05:42 < kanzure_> can you send it to another desktop, get pulled, and the ngo back to the other desktop? :/
05:43 < fenn> er... yes sometimes
05:43 < fenn> i didnt have this problem the first time (when it wanted to create a config file)
05:43 < fenn> there's a text input box on "about graphsynth" that wants my attention
05:44 < kanzure_> text input? I haven't noticed that.
05:44 < kanzure_> you know what, this whole 'about' thing should be disableableble.
05:44 < kanzure_> I haven't bothered figuring out how yet.
05:45 < fenn> foreach (arc a in arcs)
05:45 < fenn> is that the right place to put if(!(some var.Length==0)
05:46 < kanzure_> right
05:46 < kanzure_> below it
05:46 < kanzure_> it should be looking for a.name 
05:46 < kanzure_> is that true?
05:46 < kanzure_> a.To.name
05:46 < kanzure_> the reason why a.To.name and a.From.name fails is because the From and/or To object is not instantiated.
05:47 < kanzure_> so, a good check is whether or not a.To && a.From == null or not..
05:50  * fenn clucks disapprovingly
05:50 < kanzure_> ?
05:52 < fenn> look before you leap coding
05:52 < kanzure_> I'm not looking at the code at the moment, but I seem to recall something about look-ahead in the graph
05:52 < kanzure_> is that what you mean?
05:53 < fenn> no, i'm referring to the "if some var == null" scattered throughout every function
05:53 < kanzure_> every?
05:53 < kanzure_> is it already there?
05:54 < fenn> i'm not sure.. still figuring out what this stuff does
05:54 < fenn> what kind of IDE doesnt have code folding
05:54 < kanzure_> a bad one.
05:54 < kanzure_> kwrite does.
05:55  * kanzure_ quickly learned kwrite's f11 
05:56 < bkero> fenn: xcode :(
05:56  * fenn opens the damn file in kate
05:56 < fenn> ah much better
06:18 < gene> I just thought of something that needs to be made
06:19 < gene> a program that scans forums for conspiracy theories and debunks them when it finds them
06:19 < bkero> I believe snopes has that covered
06:19 < fenn> any decent conspiracy theory can't be debunked
06:20 < gene> I know
06:20 < gene> but automically looks up snopes
06:20 < fenn> heh so i finally get the damn thing to compile and it still crashes
06:20 < gene> cars compile yet they still crash
06:21 < gene> any way
06:21 < bkero> I think an autotroll would be pretty amusing
06:21 < gene> I am trying to figure out how to make a program to troll a forum so I don't have to
06:21 < gene> yes it would
06:22 < gene> it would generate epic lulz
06:22 < gene> now what would be really epic
06:22 < gene> if you start generating a whole bunch of random garbage in a UFO forum
06:22 < fenn> bkero: if an autotroll trolls in the forest and nobody is around to hear, is it a troll?
06:23 < gene> and see if they believe it
06:25 < gene> http://www.autotroll.com/index.html
06:25 < gene> there already is an autotroll
06:28 < gene> but it doesn't autotroll forums
07:45 < gene> KANZURE I FIGURED OUT HOW TO MAKE A SPIRAL IN SOLIDWORKS
07:46 < gene> it's so easy
07:46 < gene> FACEPALM
08:30 < willPow3r> dark chocolate m&ms are fucking gross
08:31 < willPow3r> dont ever eat these disgusting nuggets of nastiness.
09:04 < gene> I hear clone meat is tasty
09:04 < gene> you should try it
10:57 < kanzure_> gene - Stop claiming you learn new things and then not explaining.
11:03 < kanzure_> fenn: why is paul still going on about his triples?
11:10 < kanzure_> "Remember, if immortality is outlawed then only outlaws will be immortal."
12:21 < kanzure_> haha, classroom emailing lists suck. "Last night, one of my friends attempted suicide, so if somebody could take notes for me today that would be great. Thanks. I'm going to hang out with the others that were there. See ya." 
12:27 < UtopiahGHML> http://www.nataliedee.com/090208/you-are-free-to-learn-how-to-do-stuff-at-any-time.jpg
13:25 < kanzure_> fenn: Did you ever send it to me?
14:13 < kanzure_> How did Bram Moolenaar get his name on the vim startup screen?
14:14 < kanzure_> I thought vim was some horrible monstrosity of a few hundred contributors.
16:03 < UtopiahGHML> Intelligence in Wikipedia
16:03 < UtopiahGHML> by googletechtalks
16:03 < UtopiahGHML> I can't access it, anybody else can? it was supposed to be the last video posted :/
16:07 < fenn> kanzure_: no, i didnt know what to write
16:40 < fenn> http://www.nataliedee.com/090208/you-are-free-to-learn-how-to-do-stuff-at-any-time.jpg
17:01 < kanzure__> fenn: the cad stuff is fine to write about. you'll get bonus points if you mention investigating the graphsynth codebase or something I guess.
17:05 < kanzure__> New neurophysiology notes. http://heybryan.org/school/buildingbrains/?C=M;O=D
17:14 < kanzure__> fenn: for paul's licensing BS, http://groups.google.com/group/gnu.misc.discuss/browse_thread/thread/df4b4363d544f766?pli=1 , how about just a "casting" command to "cast" a package with a certain license. I guess a synonymous word would be 'signing', but whatever.
17:14 < kanzure__> castlicense -lLGPL my.deb
17:22 < fenn> "a 
17:22 < kanzure__> This was the idea of metatags on pages anyway..
17:22 < fenn>  user contributing an extension to a GPL'd work" is either contributing under GPL or violating the terms of the license, so it's their problem
17:23 < fenn> the squeak "wah. ignore" strategy seems like the wrong way to go
17:24 < kanzure__> all configure-make-makeinstall packages have a LICENSE file anyway. These are usually copies verbatim, so if Paul's so concerned then he should just do a diff/compare match and write a few notes with a markup language.
17:24 < kanzure__> fenn: hm?
17:24 < kanzure__> That the squeak community said ignore?
17:24 < fenn> i guess
17:26 < kanzure__> http://home.comcast.net/~arid_shadow/interface.png <- Paul's still linking to this. I don't think he understands.
17:27 < kanzure__> "Human Readable Text" getting the largest input field and a CAD file thingy for uploading .. it's odd, he seems to know technical details, but then he goes off promoting this sort of thing.
17:27 < kanzure__> only so much you can do with another implementation of Wikipedia ..
17:28 < kanzure__> (I know, I know, "but semantic mediawiki allows querying!" (yes, only if you know the ontology beforehand, which you don't. That's why you need to smash a few layers together.)
17:30 < fenn> i hate paul's license management scenario
17:30 < fenn> it's like DRM coming from your grandmother
17:30 < kanzure__> Paul gets very offensive with licenses. To the point where I think he's bullying me/SKDB/ADL almost.
17:30 < kanzure__> "Oh, you're not hip with GPL flashing over the place. You're evil until that's fixed."
17:30 < fenn> unfortunately that's the way it is
17:31 < kanzure__> http://susaro.com/wp-content/uploads/2008/11/draft_consciousness_rpwl.pdf <- "Hard problem of consciousness solved", says Richard Loosemore. The abstract is all that matters, it's just some ontological dodging.
17:31 < kanzure__> fenn: Not true. I am living proof that Paul is false.
17:32 < fenn> being a license nazi from the start is much less effort overall than realizing you goofed three years later and have to delete half your work
17:33 < fenn> when you're investigating large numbers of libraries and software to integrate, tracking down the license information is a big chore
17:34 < fenn> so yes, you should have a big red GPL stamp on the front page
17:34 < fenn> (if that's true)
17:34 < fenn> (which it isn't, in your case)
17:35 < kanzure__> And yet here I am, with content in the repositories (granted, bad content (in the sense that the data sucks anyway)). Maybe Paul's theory is that it was the grace of God that got me this far, and not my copy button.
17:36 < fenn> no, he's saying you will get bit in the ass if you continue with this methodology
17:36 < fenn> sharing books on your server is one thing.. building a career on illegal information is a dumb idea
17:37 < fenn> or maybe that's what i'm saying
17:37 < fenn> paul is saying he doesnt want to deal with it if it's "tainted"
17:39 < fenn> it sucks that we have "evil by default" as the de-facto standard
17:39 < fenn> i wonder if skdb/wahtever shouldn't have a human readable field, just to train people to not use it
17:40 < kanzure__> what do you mean by that?
17:40 < kanzure__> Where does human input come into play then? just wondering what you're talking about.
17:41 < fenn> if it looks like a wiki, people are going to assume it's a wiki and then just ignore "all that complicated metadata stuff"
17:41 < kanzure__> Right.
17:41 < fenn> and then we end up with appropedia and no metadata
17:43 < kanzure__> before /me forgets, I should get Hod Lipson's robot work to do robot ecologies for unstable mining operations, i.e. evolving robots that build better versions for the bad mining conditions.
17:47 < kanzure__> erm
17:47 < kanzure__> so I'm wondering what actual data we're supposed to have.
17:48 < kanzure__> Graph interconnectivity is relatively useless because of (1) how incredibly intense port connecting information from part to part would have to be, and (2) how we don't have that.
17:48 < fenn> many product datasheets have Real Interface Specifications(tm)
17:48 < fenn> especially if they come from a huge multinational corporation
17:49 < fenn> most consumer crap doesnt come with a datasheet though
17:50 < fenn> so, electronics components in general are pretty good about this
17:50 < fenn> there's probably even a digital format already somewhere that i dont know about
17:50 < fenn> then there's stuff like NEMA
17:51 < kanzure__> NEMA?
17:51 < kanzure__> .org
17:51 < kanzure__> ah.
17:51 < fenn> for electrical stuff like motors and electrical outlets
17:51 < kanzure__> Yes, the electronics industry is pretty good about this.
17:51 < fenn> because they have to design complex things
17:51 < kanzure__> Verilog/RTL/VHLI stuff is a good example of good design thingies.
17:51 < kanzure__> VHLI is wrong.
17:51 < kanzure__> What am I thinking of?
17:52 < fenn> VHDL
17:52 < fenn> i'm not really sure that's the same realm
17:52 < fenn> those are more like "network topology" descriptions
17:52 < kanzure__> Algebraic languages for descriptions of the logic to be implemented
17:52 < kanzure__> yeah
17:52 < fenn> there's no actual geometry or real physical units
17:52 < kanzure__> It was my understanding that at some point they have to convert to that sort of in-house understanding for how they are going to implement, say, a transistor
17:53 < fenn> yes
17:53 < fenn> "synthesis"
17:53 < fenn> ~= CAM
17:53 < kanzure__> and then they go off to the computational physics gurus in the university 
17:53 < kanzure__> and they figure out how transistors work with various constraints, then take that information into account, avoid overlap scenarios, etc.
17:54 < kanzure__> But that's because of the very specific research of the physicists.
17:54 < fenn> not unless they are developing a transistor technology
17:54 < fenn> most chips are 90% computer generated from design rules
17:54 < kanzure__> Where do you think those rules come from? The vacuum?
17:54 < fenn> with humans to tweak the layout, like on a circuit board
17:54 < fenn> i'm saying they dont have to write new rules every time
17:54 < kanzure__> 'course.
17:56 < fenn> "Sorry, but I didn't know the patient needed heart  medication because I could not access a copy of the patient medication  records since the hospital DRM system payment account was temporarily  overdrawn." all too true
17:59 < fenn> haha "Our current  copyright system is now in some ways like the American South before the  Civil War where black people (or now "creative works") were often  assumed to be a runaway slave unless they had their papers of freedom  with them
17:59 < kanzure__> "minimum autodesign-capacity sufficiency: can the system sufficiently design the means to keep a small family (man, woman, pet and/or child) alive?"
18:00 < kanzure__> fenn: That's not Paul, is it?
18:01 < fenn> yes and he goes on for another page or so in that vein
18:01 < kanzure__> oops, I must have not clicked 'read more'
18:02 < kanzure__> Is this on squeak? it failed to load for me
18:02 < fenn> http://groups.google.com/group/gnu.misc.discuss/msg/11c0e6a52bc480e1?
18:02 < fenn> near the bottom
18:04  * kanzure__ read it.]
18:05 < kanzure__> Paul fails to provide me a meta-meta-meta-infinity license license license manager.
18:06 < fenn> not necessary
18:06 < kanzure__> just like licenses.
18:06 < fenn> license manager could bootstrap just as well as apt-get can bootstrap
18:07 < kanzure__> if you're worried about licenses, I guess you should bring it up with the chmod people.
18:07 < fenn> i would certainly like to stop paying rent on this house i live in, for which the parasite who "owns" it reciprocates nothing
18:08 < fenn> but then i would have to worry about dudes with guns coming, and having to schlepp all my stuff around
18:09 < fenn> anyway the point is for use as a *tool* not a cantankerous burden that you have to deal with
18:09 < fenn> the legal system is enough of a burden as is
18:09 < fenn> the tool is for avoiding unforseen obstacles, and reducing the amount of effort for investigating beforehand
18:11 < kanzure__> this is boring.
18:11 < fenn> why should i support unptnt?
18:11 < fenn> they look like double-alpha-secret capitalists to me
18:12 < kanzure__> I figure if they end up having anything of value, they might slap up the AJAX-graph-stuff and feed us data.
18:12 < kanzure__> I'm supposed to be doing some 'preference modeling of what a designer wants to see'. 
18:12 < fenn> we dont need any more crap data
18:12 < kanzure__> and this requires a lot of data.
18:12 < fenn> wtf is preference modeling
18:12 < kanzure__> no, this isn't for crap data, this is for user input on what they like / don't like in terms of designs that they see
18:12 < kanzure__> well, "I hate this graph. Show me another!"
18:13 < kanzure__> "yessir. how about something practically the same?" "off with your head!"
18:13 < fenn> you can do that with randomly generated or GA generated graphs
18:13 < kanzure__> guess what I'm doing.
18:13 < kanzure__> heh'
18:13 < fenn> i had this program for the mac plus called "blind watchmaker"
18:13 < fenn> which was essentially that
18:13 < kanzure__> the substitution rules are randomly generated at the moment. so I have 2.9 GB of 1,000,000 rules (I know, stupid)
18:13 < kanzure__> stupid format too, all XMLy and lots of overhead
18:13 < fenn> you could "evolve" weird animals/plants or whatever (they were just IFS fractals i think)
18:14 < kanzure__> *it's stupid
18:14 < fenn> gzip
18:14 < kanzure__> on the fly uncompression of the whole thing when you scan through the directory of rules to apply to a certain graph?
18:14 < kanzure__> decompression. whatever.
18:14 < fenn> ya
18:14 < kanzure__> That sounds even more computationally intense.
18:15 < fenn> well, why are you using 1000000 rules?
18:15 < kanzure__> because I wanted to emphasize that they were randomly generated
18:15 < fenn> dude, read NKS
18:15 < kanzure__> we have a set of 150 useful rules that the last undergrad made.
18:15 < fenn> you dont need a lot of rules
18:15 < kanzure__> hm?
18:16 < kanzure__> cellular automata don't do full recursion of the graphs fwiw.
18:16 < fenn> he has a chapter on substitution systems
18:16 < fenn> even a section on substitution networks
18:16 < kanzure__> oh, okay. that's a good excuse for me to finally read NKS.
18:17 < fenn> but the point is that the same characteristics appear in the system's behavior as with other simple sets of rules
18:17 < fenn> and also that more complex sets of rules dont yield more complex behavior
18:19 < kanzure__> well the idea is also to try to figure out what to map user preferences to, so if you have only 10 axiomatic rules, that might not be enough to figure out what the user likes, whereas the larger rules which are found in patterns in reverse engineered systems and such, might be the "points in conceptual space" or whatever, that mathematically would be good matches and thus make for the things that you should push to the top of the queue for
18:19 < kanzure__> gah, that's long-winded.
18:19 < fenn> "push to the top of the queue for..." cutoff
18:20 < fenn> oh that was the end
18:20 < kanzure__> "substitutions to make next"
18:20 < kanzure__> was the end.
18:20 < kanzure__> (in quotes too)
18:20 < fenn> oh well, i still dont get the whole substitution system concept anyway
18:21 < kanzure__> Do you want to?
18:21 < fenn> well i hope it actually contains value...
18:21 < kanzure__> did graphsynth start working?
18:21 < fenn> no
18:21 < fenn> it crashes a lot
18:22 < kanzure__> the way the graph substitutions work is that you're given basically a "dot file" for all practical purposes, and you have your substitution rule that looks for a certain subgraph in user provided input; given a match, the "right hand side" of the substitution rule is employed, thus replacing the subgraph with the 'implementation graph' (CFG)
18:22 < kanzure__> so if you see "import -> channel -> export" and you have a rule for "channel" to translate that to "GIANT METAL PIPE", then that's a valid branch to follow in the tree. Then you have "import -> GIANT METAL PIPE -> export"
18:23 < kanzure__> and the only things left to change are input-> and ->export, the CFG part being 'untouchable'.
18:23 < fenn> what is this subgraph supposed to look like?
18:23 < kanzure__> (unless there's a rule that manipulates CFGs too)
18:23 < kanzure__> well, that's what the substitution rule file contains.
18:23 < kanzure__> the left hand side has the subgraph to look for
18:24 < kanzure__> and if it matches the user input, or current graph in the tree of possibilities, then you know you have a match and a possible way to do yet another branch
18:24 < fenn> what about multiple possible solutions?
18:25 < kanzure__> those are accounted for, yeah
18:25 < fenn> "accounted for"?
18:25 < kanzure__> so if you see that it applies twice in the graph, then those are two branches to follow
18:25 < fenn> and you follow all the branches?
18:25 < kanzure__> well.
18:25 < fenn> seems computationally intractable
18:26 < kanzure__> graphsynth currently "takes the user's hand down a path", so it's stochastic, one-off.
18:26  * fenn grumbles about ghost diagrams
18:26 < kanzure__> my approach was "do the computationally intractable thingy, until you get input from the user (asynchronous) telling you to stop going down a particular type of path, unless you've exhausted options, in which case you should just consider the user to be wrong/incorrect/inconsistent in his preferences"
18:26 < kanzure__> fenn: grumbling about them to campbell would be fine methinks
18:27 < fenn> i'm still stuck with figuring out how to translate this paradigm into my own thoughts re: skdb
18:29 < fenn> it seems like a subgraph is an interface, and a rule is when two parts are connected, but i'm not sure
18:30 < kanzure__> oh, that would be a neat way to do it, but unfortunately not.
18:31 < kanzure__> okay, so the translation is that a 'graph' represents some system, like electrical circuits, 
18:31 < kanzure__> and the 'substitution rules' would be generated from valid ways of recognizing nodes on the graph (components I guess) and then offering possible ways of reconnecting them
18:31 < fenn> but if each node represents only a specific component, it would be very inflexible
18:31 < fenn> its like black boxing
18:31 < kanzure__> yes, it is like blackboxing.
18:31 < fenn> poop on that
18:32 < kanzure__> turtles, remember? :-/
18:32 < fenn> its drafting
18:32 < kanzure__> Not even drafting.
18:32 < fenn> yeah at least drafting conveys knowledge
18:33 < fenn> why is everyone doing it wrong
18:33 < fenn> gah
18:33 < fenn> fuck earth
18:33 < kanzure__> we need you!
18:34 < kanzure__> but really, it's obvious that there's more that could be done here.
18:34 < fenn> a NEMA 34 motor mount has four 3/8-16 tapped holes in a 3 inch square
18:34 < fenn> what is the basic unit in the graph, the NEMA34, or each of the bolt holes?
18:35 < fenn> and a square, and a 3 inch dimension
18:35 < fenn> i'm thinking the subgraph should be (square x 3/8-16 x 3")
18:35 < fenn> and who gives a fuck what it's called
18:36 < fenn> and if you only have M8 bolts, then you can fix the rule to allow a size range instead of just that specific thread size
18:38 < fenn> you should be able to generalize it down to however many turtles you wish
18:39 < fenn> i mean the motor could be welded on, or duct taped on, or held in place with a rabid bulldog
18:40 < kanzure__> the basic unit is the NMEA34
18:40 < kanzure__> in the current setup.
18:41 < fenn> it's not a specific motor though, right?
18:41 < fenn> so i could pick any nema34 motor from various manufacturers
18:43 < kanzure__> at the moment it's based off of labels (although more parameters could be added I guess)
18:43 < kanzure__> so if your 'rule set' has a rule that is specific to a specific motor, then yes it is?
18:44 < UtopiahGHML> can someone access the video of http://www.researchchannel.org/prog/displayevent.aspx?rID=27626&fID=345 ?
18:44 < fenn> UtopiahGHML: i'm going to guess no
18:44 < fenn> but i'll try, JUST FOR YOU
18:45 < UtopiahGHML> fenn: you are going to make me blush :-#
18:51 < fenn> UtopiahGHML: there is no video for that talk, hence the green freedesktop-ish logo
18:52  * kanzure__ goes off to chem.
18:54 < UtopiahGHML> "Runtime:00:56:30" so how could there be no video? it has also been published on YouTube but removed right after it (Ive seen it on the RSS feed)
18:54 < UtopiahGHML> :(
18:54 < fenn> maybe the dreaded copyright beast attacked in the night
18:55 < fenn> won't someone save us from this monster!
18:55 < fenn> interesting series though, like TED
18:55 < UtopiahGHML> yep, and the SEED ones too and the VideoLectures.net too
18:56 < fenn> videolectures.net is just droning MIT professors :)
18:57 < fenn> anyway did i mention i hate video
18:57 < kanzure__> Screw chem. I'm too hooked on 'golly', the CA visualizer.
18:57 < UtopiahGHML> they are just feaking lazy I asked for an RSS feeds years ago and they said "yeah yeah yeah we are working on it"...
18:58 < fenn> you need an Autoverse simulator!
18:59 < fenn> oh, you never read permutation city
18:59 < fenn> autoverse = CA universe simulator that approximates real-world chemistry
19:05 < kanzure__> fenn: but, if it means anything to you, the gear stuff can be exported to schematics/drafting for final results.
19:05 < kanzure__> (since I have the 3D model anyway)
19:05 < fenn> = "let a human figure it out"
19:05 < fenn> oh well its better than nothing
19:06 < kanzure__> ok, ok, also an export to gcode if emc has its end held up.
19:06 < kanzure__> *if it holds up its end
19:06 < fenn> you cant really cut gears with a milling machine
19:06 < fenn> you have to make a special cutter and then spin things at certain rates and positioned at various angles relative to each other
19:07 < fenn> it would be possible to do on a milling machine, but you cant just trace the outline with an endmill
19:07 < fenn> unless they're big fisher-price gears
19:09 < xp_prg> kanzure__ any progress?
19:12 < kanzure__> No. I've been doing other things.
19:12 < xp_prg> ok just wondering
19:18 < kanzure__> fenn: best if you send the email sooner rather than later, since directional pushes away from sloppy graphs might be possible.. maybe.
19:18  * kanzure__ sits around wondering what to do. watches CA.
19:19 < kanzure__> Actually, I know what to do, I'm just not doing it.
19:19 < marcel> kanzure find a girl to fuck!
19:19 < marcel> ^^
19:21 < kanzure__> basically diagnostics on the traversal algorithms for finding "placed" graphs within the trees that I want to see if the software will want to show me upfront
19:21 < kanzure__> but the fundamental formatting issues still have to be addressed. I mean, it stores some of the useful information but not quite "all of it", but to prove this I'd have to have "all", which I don't, because the format doesn't have "it" (supposedly)
19:21 < kanzure__> Boy this is confusing. :)
19:25 < kanzure__> http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1145699498/9#9
19:28 < kanzure__> http://www.freesoftwaremagazine.com/articles/making_open_hardware_possible?page=0%2C1 more licensing crap
19:47 < fenn> it's important to understand that article
19:48 < kanzure__> http://heybryan.org/shots/2008-11-14_emachineshop.png <- Lookie.
19:49 < kanzure__> I understand all of this licensing stuff, but that doesn't mean I want to.
19:49 < fenn> i feel the same way about cryptography
19:51 < kanzure__> So I like how they have that link to matweb.
19:52 < fenn> where?
19:52 < kanzure__> in the screenshot.
19:53 < kanzure__> Material window.
19:53 < kanzure__> Upper right.
19:53 < fenn> ah
19:53 < kanzure__> They seem to be using their own proprietary 3D part representation.
19:53 < kanzure__> dot "ems"
19:53 < fenn> rather useless link
19:54 < fenn> can you click on the material names and see specifications?
19:54 < kanzure__> right, what guarantee that they have the same exact materials inhouse?
19:54 < kanzure__> no, only those numbers on that display there
19:54 < kanzure__> but the "Compatible with .." message changes, which I like.
19:54 < fenn> oh, i guess there are more numbers and stuff if you scroll
19:55 < fenn> eh.. stainless is "compatible" with milling but it's a lot harder to do
19:55 < fenn> what is VR cost?
19:56 < kanzure__> I don't know. /me moves back to that computer
19:56 < kanzure__> variable relative cost
19:57 < fenn> derr.. is that like, dollars weighted for inflation or something?
19:58 < fenn> hey that's not your gear
19:58 < fenn> its a demo file
19:58 < kanzure__> http://www.emachineshop.com/users/index.htm
19:58 < kanzure__> it's like relative material cost I guess.. "this is generally cheaper" ? I don't know.
19:58 < kanzure__> this is just some commercial stuff from http://emachineshop.com/
19:59 < kanzure__> http://www.emachineshop.com/machines/metal_fabrication.htm
20:00 < kanzure__> I wonder where they are getting their material pricing information.
20:00 < fenn> their metal supplier probably
20:01 < fenn> i wonder how well emachineshop scales
20:02 < fenn> and data interoperability seems like a huge issue
20:03 < fenn> oh "simple just use our software from start to finish"
20:03 < fenn> bah
20:03 < fenn> 2.5D only
20:08 < kanzure__> nope, 3D visualization was in my screenshot
20:08 < kanzure__> they claim 100,000 units is an ok order
20:08 < kanzure__> maybe they're using onlinemetals.com heh'
20:09 < fenn> what was the "deal" with onlinemetals you were talking about?
20:09 < kanzure__> they want a backend for XML price quoting and order placement
20:10 < fenn> uh, writing software for them isnt much of a deal :(
20:10 < kanzure__> it is if I get to use it :)
20:10 < kanzure__> I mean, an open interface 
20:10 < kanzure__> for purchasing metals and such.
20:12 < kanzure__> 'The so-called "Free Dog" organization ( Freeedaug ) meets in the Boston area and discusses gEDA and related topics.'
20:15 < fenn> gEDA certainly toots its horn
20:16 < kanzure__> as it should.
20:16 < fenn> too bad it sucks :*
20:16 < kanzure__> oh?
20:16 < kanzure__> I've used it like, twice.
20:16 < kanzure__> so I'm not remembering how it sucks.
20:16 < fenn> its about as bad as any EDA software i suppose
20:20 < kanzure__> hm. so ngspice apparently takes a circuit and generates a simulator for that specific circuit.
20:20 < kanzure__> which is the "each component brings some equations to the table" approach or something
20:20 < fenn> i wouldnt say "generates a simulator"
20:21 < fenn> it's nodal analysis (aka "lumped circuit model")
20:21 < kanzure__> "The Spice Library. It takes a circuit description and builds a simulator from it." /ng-spice-rework-17/src/spicelib/
20:21 < kanzure__> hm.
20:22 < fenn> it's pretty good as far as simulators go though
20:22 < fenn> cant do digital electronics though
20:22 < kanzure__> oh really?
20:22 < kanzure__> that's odd. why?
20:22 < fenn> too complex
20:22 < fenn> and you dont have the specs anyway
20:23 < kanzure__> wtf, no results on sf.net for 'mechanical'
20:23 < kanzure__> oh, but what about parts libraries?
20:23 < kanzure__> and what's complex about it? 
20:23 < kanzure__> I used to run a simulator back in high school for electronics
20:23 < kanzure__> it was some commercial software package of course
20:23 < fenn> there are some projects to integrate digital 'black boxes' into a spice-ish environment, for example qucs or ktechlab
20:24 < fenn> hmm maybe not qucs
20:24 < kanzure__> http://ktechlab.org/
20:24 < fenn> gpsim?
20:25 < kanzure__> http://gpsim.sourceforge.net/
20:25 < fenn> not gpsim
20:25 < fenn> it's a hard problem
20:26 < fenn> this appears to do mixed signals; 
20:26 < fenn> http://ktechlab.org/wiki/index.php?title=Index:Screenshots
20:26 < kanzure__> I thought it would just be this messy hierarchical PDE (for a given circuit of some system)
20:26 < kanzure__> well, not for digital circuits
20:26 < kanzure__> what is the simulation supposed to do? timestep simulation?
20:27 < fenn> vs PDE? i dont know
20:27 < fenn> "Variable time step in Simulation"
20:28 < fenn> http://www.opencollector.org/summary.php
20:28 < fenn> lots of simulators there
20:29 < kanzure__> apt-cache show model-builder, <- what's going on there? hm.
20:31 < fenn> looks more or less like gnuplot
20:31  * kanzure__ grabs half a GB of simulators in apt
20:31 < kanzure__> five minutes to go. not bad.
20:33 < fenn> i like ktechlab, it's easy for beginners
20:33 < fenn> and since you should never trust a simulator anyway, that's about as far as it needs to go
20:43 < kanzure__> come to think of it, why do I have rules at the moment
20:43 < kanzure__> if they are randomly generated anyway, just say everything can connect to everything else,
20:43 < kanzure__> then just keep track of what connected things were a /bad/ idea.
20:43 < kanzure__> basically the same result ..
20:44 < fenn> most of them will be a bad idea
21:00 < kanzure__> 'Automatic reconstruction of 3D solid objects from 2D orthographic views'
21:00 < kanzure__> bwahah
21:03  * fenn is skeptical
21:50 < fenn> god i hate doing self promotion
21:51 < fenn> it just sounds so cheesy
21:51 < fenn> "And probably most important, I'm passionate about creating a new paradigm of personal and 
21:51 < fenn> open manufacturing"
22:13 < UtopiahGHML> fenn: what about the very cheesy "democratizing innovation"? eventually referring to work from Erik Von Hippel
22:17 < gene> you there kanzure?
22:22 < fenn> UtopiahGHML: i dont care about "democratizing" i care about empowering individuals
22:22 < fenn> the more empowered individuals the better
22:24 < UtopiahGHML> and work from Von Hippel isn't about that?
22:24 < fenn> i guess that is what he is saying
22:24 < fenn> i havent read it
22:24 < fenn> to me, "democracy" means majority rule
22:25 < UtopiahGHML> http://web.mit.edu/evhippel/www/democ1.htm
22:28 < UtopiahGHML> (even empowering those http://video.google.com/googleplayer.swf?docid=-7625142738072469708 ?)
23:12 < kanzure_> fenn: there's a certain amount of self-promotion that is legit and a certain amount that is bad-legit, like "most importantly I'm passionate .." <- (probably) bad. 
23:13 < kanzure_> I should send you teh email I sent Campbell
23:18 < kanzure_> fenn: Sent. His response was "When are you coming by?"
23:19 < kanzure_> so he's somewhat more of a phone guy I guess.
23:24 < fenn> ack
23:24 < fenn> did you already get and forward the email i just sent?
23:24 < kanzure_> No.
23:24 < kanzure_> I didn't get.
23:24 < kanzure_> when did you send?
23:25 < fenn> i'm confused, you just sent me an example? your own introductory email to campbell
23:25 < kanzure_> ah, just got it
23:25 < kanzure_> yeah, that's what I sent to him
23:25 < fenn> so " fenn: Sent." was in response to what?
23:26 < fenn> is he still at the lab?
23:27 < kanzure_> fenn: Sent was in respone to my "I should send you .."
23:28 < kanzure_> He's probably not, I sometimes catch him leaving and it's consistently at five.
23:28 < kanzure_> in fact, all of my professors tend to leave at five.
23:28 < fenn> hmph
23:28 < kanzure_> So anyway, two things stuck out at me
23:29 < kanzure_> first of all, campbell is like a comp sci guy somehow randomly in mechanical engineering, you see papers all out on his desk and they're from "European Journal of Graphical Network Theory" and so on :-p
23:29 < fenn> yes i gathered that much
23:29 < kanzure_> so when you emphasize "get something done", you should also throw in something that says "applied algorithms are cool"
23:29 < kanzure_> or something :p
23:30 < fenn> guh
23:30 < fenn> but that's totally false
23:30 < kanzure_> tree traversal is an algorithm.
23:30 < fenn> but it's a trivial part of the whole system
23:30 < kanzure_> heh.
23:30 < fenn> "for loops are super neato"
23:31 < fenn> well, yeah, for loops are necessary for an automated engineering system, i guess
23:31 < fenn> unless you use lisp
23:31 < kanzure_> I know, they are trivial, but 
23:31 < kanzure_> okay.
23:32 < kanzure_> so you need to sprinkle in some graduate-level goodness I guess, 
23:32 < kanzure_> worshiping algorithms might not be the way to do it
23:32 < fenn> yeah i wasnt feeling very inspired when i wrote that
23:32 < fenn> so it's just a glorified resume really
23:33 < fenn> i never took any comp-sci classes, or engineering classes
23:33 < fenn> so i have no idea what's "graduate level" in those fields
23:33 < kanzure_> I guess I haven't either, and the ones that I have taken (AP comp sci, me302) have been jokes
23:33 < kanzure_> bad jokes :(
23:33 < fenn> however i do know what is used in reality and what works
23:34 < fenn> so i dont see spouting off various tree traversal algorithms as being the right thing to do
23:34 < kanzure_> that's fine.
23:35 < kanzure_> oh, at the very end of the message is maybe the place to put "I have a bachelors degree." or a giant indicator saying "This is that graduate student you are looking for"
23:35 < kanzure_> I know it's in your resume, but whatever.
23:35 < fenn> haha but there's the catch
23:36 < kanzure_> you don't?
23:36 < kanzure_> oh that's right, you dropped didn't you?
23:36 < fenn> if i mention "graduate student" that files me immediately into the "i want to be a grad student" category
23:36 < fenn> no, i have a degree
23:36 < kanzure_> okay, then indicate that you could do "full time" (whatever that is)
23:36 < fenn> righto
23:41 < kanzure_> emphasis on VOICED is good .. a one-off link to the p2pfoundation.net blog post about it and quoting the paragraph, the one that talks about it, and then saying something like "I WANT TO DO THIS / kinda doing it anyway unless I have to go do manual labor."
23:41 < kanzure_> anyway, besides that I think it's ok
23:42 < kanzure_> we could do this a few ways. I could double check and see if he's most definitely looking for a full time person to work on the project, or you could just dump the email to him out of the blue.
23:45 < fenn> what else are they doing besides voiced and design repo?
23:46 < kanzure_> some stuff on preference modeling, trying to figure out how to get the best thing out of the possibility space of designs, the gear stuff, sheet metal folding rules for graphsynth
23:47 < fenn> huh. graphsynth does sheet metal?
23:47 < kanzure_> it's just the substitution rules, so if you come up with a set of rules for a given domain, yes
23:50 < fenn> have you actually talked to him about skdb?
23:50 < kanzure_> yes.
23:52 < kanzure_> so while he knows a bit about it, he's more interested in just doing stuff with the repository data more than anything, and since that data doesn't exist, there's not much that we can do with it
23:52 < kanzure_> so, if there was to be some sort of good file format already available with more usable variables, I could totally push that
23:52 < kanzure_> but all that we have is content of the format that graphsynth takes up, which leads to playing around with search trees more than anything else.
23:53 < fenn> right
23:53 < fenn> where did the data come from? (who entered it)
23:53 < fenn> (and what were they thinking!!!)
23:53 < kanzure_> mst.edu
23:54 < kanzure_> http://repository.designengineeringlab.org/
23:54 < fenn> Product armadillo armor was added to the repository.
23:54 < kanzure_> so we've talked with them, as I've mentioned, about possibly redoing the repository database format or whatever to capture the information that campbell is interested in (like the way that artifacts solve functions etc.)
23:55 < fenn> Product puffer fish was added to the repository.
23:55 < kanzure_> hrm
23:55 < fenn> Product two component regulatory system was added to the repository.
23:55 < kanzure_> wtf.
23:55 < kanzure_> is this recent?
23:55 < fenn> ya
23:55 < kanzure_> note that the repository stuff is GPL'ed :) so that's good. not the data though.
23:55 < fenn> meh
23:55 < fenn> its not hard to write a data entry form
23:56 < kanzure_> right, and their form sucks because you can get away with fucking up the db
23:56 < fenn> it takes a while to get good measurements of real world stuff though
23:56 < fenn> i would probably just make new data
23:56 < fenn> instead of trying to save the design repo stuff
23:57 < kanzure_> you might want to mention that in the email, if it's not too crowded.
23:57 < fenn> no, i dont want to tell him how all his stuff sucks in the intro email
23:57 < kanzure_> probably a good idea.
23:58 < fenn> so, on a tangent
23:58 < fenn> what do you think about having a virtual environment to play with virtual artifacts in?
23:58 < fenn> like the "love" environment
23:59 < kanzure_> if you say the word 'simulator' in the email, or mention how you worship Hod Lipson (do you?), that is bonus points
23:59 < kanzure_> love environment? I'm not familiar with that.
23:59 < fenn> hod lipson has a lot of cool stuff, ya
23:59 < fenn> he also suffers from academic head-up-the-assness
23:59 < fenn> oh well
23:59 < fenn> love = quelsolaar.com
23:59 < kanzure_> hod is a friend of campbell, just for the record
23:59  * kanzure_ checks