<@kanzure> it's not constantly replicating (is it?) 18:23 < fenn> hmm 18:23 < fenn> maybe 18:23 < fenn> if it has no other leaf nodes, yes 18:24 <@kanzure> it has to wait for resource accumulation, you realize 18:24 <@kanzure> so it can't constantly be replicating, it will hit run-time errors 18:24 < fenn> yes, that's part of replication 18:24 <@kanzure> okay 18:25 < fenn> it won't assemble the new unit until the parts are built 18:25 <@kanzure> so, 18:26 <@kanzure> (1) can skdb and autogenix do the package/functionality requirements we need to be able to specify? 18:26 <@kanzure> and (2) can we do a small proof of our 'fishing' method for some other design project? 18:26 <@kanzure> also, it looks like for each package and functionality we add, we have to include everything that it could possibly be used to do 18:26 <@kanzure> and mention its relationships to all existing functionality 18:27 < fenn> fishing the ring out of the swamp? 18:27 <@kanzure> right 18:27 < fenn> there arent any replicators 18:27 < fenn> or do you mean just fake it, and prove the algorithm works? 18:27 <@kanzure> the algorithm for fishing? 18:27 < fenn> s/prove/demonstrate/ 18:27 < fenn> yes 18:27 <@kanzure> yes, I suppose 18:28 <@kanzure> but with packages and functionality in there 18:28 <@kanzure> where *we* know a solution 18:28 <@kanzure> but we want the computer to find it 18:28 < fenn> re:1, skdb/autogenix only exist in our minds right now ;) 18:28 <@kanzure> yes, but I mean to say that we get to do functional specs for skdb/autogenix and make sure the file formats could possibly define all of this information that we need 18:29 <@kanzure> I guess this is going to be ad-hoc really 18:29 <@kanzure> since there's all sorts of variables we will want to include 18:29 < fenn> the design can change if it's needed 18:29 <@kanzure> perhaps we will make it abstract 18:29 < fenn> we can do .agx version 1, version 2, etc 18:29 <@kanzure> in the sense that we can "include package-required-for-electricity-variables" 18:29 <@kanzure> and then other variables can be further subunits which will each add functionality to the overall project 18:30 < epitron> http://kol.coldfront.net/thekolwiki/index.php/Time_sword 18:31 < fenn> what variables? diesel for your diesel generator? 18:31 <@kanzure> fenn: as well as flow and other information about that fluid 18:31 <@kanzure> there's tons of information that CRC keeps on such stuff 18:31 <@kanzure> but we can't know all of these variables from first principles 18:31 < fenn> true 18:31 <@kanzure> so when we want to include a new variable for a new functionality 18:31 <@kanzure> we should be able to just add a new 'library' addon thing 18:31 <@kanzure> or extension 18:32 <@kanzure> which would include subcomponents that would be called upon during simulation 18:32 <@kanzure> I think this is the 'virtual function' idea in C/C++ 18:32 <@kanzure> or template factory 18:32 <@kanzure> not sure. 18:32 < fenn> terminology update: streams are run-time capabilities/requirements, packages are build-time, packages can provide streams 18:33 <@kanzure> packages still vertices? 18:33 < fenn> yes 18:34 <@kanzure> packages deliver streams to the new unit? 18:34 < fenn> i guess i wasnt specific enough 18:34 <@kanzure> no, I think you were 18:34 < fenn> streams are things like materials, energy, information 18:34 < fenn> i havent figured out if they're vertices or edges yet 18:35 < fenn> there's going to be a lot of streams that have no package providing them, like sunlight, air 18:36 < fenn> the aluminum smelter relies at run-time on the clay processor to provide bauxite 18:36 < fenn> smelter, processor = packages 18:36 < fenn> clay, bauxite = streams 18:37 <@kanzure> streams sound kind of like vectors 18:37 < fenn> then you can have a factory object that makes packages 18:37 < fenn> a package that makes packages 18:38 <@kanzure> including itself? 18:38 < fenn> let's not go there :) 18:38 <@kanzure> isn't that what we wanted in the first place? 18:38 < fenn> yes of course 18:39 < fenn> but there are lots of intermediate packages that make other packages (i guess) 18:39 < fenn> maybe that's just poor refactoring on society's part 18:39 <@kanzure> that was my 'criss-crossing' example: where package X requires stream Y during build-time to reproduce package X in the new unit, but in normal operations, package Y would require stream X, or something 18:40 <@kanzure> erm 18:40 <@kanzure> criss-crossing: 18:40 <@kanzure> Note that the "sand-processing subsystem" need not be made out of sand. The "sand-processing subsystem" needs to be made out of X mineral where there exists in the design "X-processing subsystem" which, again, is not necessarily made out of X, but it has to be made out of something like "X" that eventually makes it full circle back to sand. 18:40 < fenn> ah a sub-loop 18:40 <@kanzure> so, package X (which processes Y) needs to be made out of stream L where stream L is 18:40 <@kanzure> I am having trouble variabilizing it, because it's a hard statement to make 18:41 < fenn> the loop is in stream-space, whereas the package tree is very linear and loop-free 18:41 < fenn> maybe draw a diagram 18:41 < epitron> TIME SWORD! 18:41 <@kanzure> the dependency loops are important, of course 18:41 <@kanzure> let's see 18:42 < fenn> the subtle knife 18:43 <@kanzure> (($mineralX)-processing-subsystem) can be made out of $mineralY, where there exists in the overall design a (($mineralY)-processing-subsystem), which either must immediately be made out of ($mineralX) or some (($mineralI-don't-care)-processing-subsystem) which eventually is made out of ($mineralX) 18:43 <@kanzure> where: 18:43 <@kanzure> you would probably say the mineralX-processing-subsystem is a package, yes? 18:44 <@kanzure> and then streams would be tapped during build-time by a certain package to manufacture a new package, for the replicated entity to have. 18:46 < fenn> i'm not convinced this is a problem we have to worry about 18:47 <@kanzure> how so? 18:47 < fenn> do you think a loop will kill the dependency resolution code? 18:47 <@kanzure> no? 18:48 <@kanzure> it's not a problem 18:48 <@kanzure> we're just formalizing what we need 18:50 <@kanzure> right? 18:54 < fenn> yah 18:58 < fenn> its funny, usually i'm trying to cut apart dependency loops, but for self-replication they're quite useful 18:59 <@kanzure> the more the better (in terms of design) 18:59 <@kanzure> but the more the worse, in terms of implementation 19:09 < fenn> grr do they really expect me to pay money for nasa publications? 19:11 <@kanzure> here's my attempt at defining a package: 19:11 <@kanzure> a package must (1) be made of a certain material, (2) has an input stream that gives it a material to make, (3) this input material is different than the material in #1, and (4) a package must make *some other package* with the input stream material. 19:12 < fenn> package must make a package? 19:12 <@kanzure> yep 19:13 <@kanzure> must make *another* package 19:13 < fenn> what about instruments 19:13 <@kanzure> and that other package might end up making the original package 19:13 <@kanzure> a package might be a fabricator, and that fabricator, by coincidence, can make extra stuff 19:13 <@kanzure> like an instrument 19:13 < fenn> i mean sensors, like an oscilloscope or something 19:13 < fenn> a spectrometer 19:14 < fenn> what does the spectrometer make? 19:14 <@kanzure> right, it makes nothing 19:14 <@kanzure> so these are your leaves 19:14 <@kanzure> leafs 19:14 < fenn> it makes information, which is necessary for replication 19:14 <@kanzure> interesting 19:14 < fenn> i feel funny calling data a package 19:15 <@kanzure> instrument == package? 19:15 < fenn> well sure, it's a collection of components that performs a function 19:16 < fenn> and there are control instruments that dont produce 'information' exactly 19:16 < fenn> the computer that's directing operations 19:17 <@kanzure> arrgh 19:17 <@kanzure> I just wrote down "Reverse Method of Making a Fabricator" 19:17 <@kanzure> (1) start with a fabricator 19:17 <@kanzure> 'erm, replicator 19:17 <@kanzure> (2) Break it down into its components. 19:17 < fenn> well gee with #1 we're already done :) 19:17 <@kanzure> (3) Make a machine to do the reverse. Done. 19:17 <@kanzure> heh' 19:17 <@kanzure> yeah :( 19:17 <@kanzure> man, relying on biology really helps though 19:18 <@kanzure> it gives you something to start with 19:18 <@kanzure> because if you want to be able to create enough for a human to live and grow 19:18 <@kanzure> then you can just require your materials for self-replication rely on that same stuff 19:18 < fenn> human is too complex, start with a minimal bacterial cell 19:18 <@kanzure> okay, what's agar made out of? 19:18 <@kanzure> unbranched polysaccharide? 19:18 < fenn> input streams: amino acids, sugar, nucleic acids, lipids, salts 19:19 <@kanzure> `Agar is a heterogeneous mixture of two classes of polysaccharide: agaropectin and agarose` 19:19 < fenn> now we need proteins to import each of those streams into the cell 19:19 < fenn> then there are assemblies of proteins that turn those streams into higher order molecules, like proteins and nucleic acids 19:20 <@kanzure> huh? 19:20 < fenn> nucleic acid polymers :) 19:20 <@kanzure> we don't need those proteins, do we? 19:20 <@kanzure> just make agar for the bacteria and that's it 19:20 < fenn> i'm talking about how the bacterium works 19:20 <@kanzure> hell, I am sure we can make bacteria that can eat moon rock 19:20 <@kanzure> that's hacking the bacterium 19:20 <@kanzure> not really necessary IMO 19:20 < fenn> no, as a model for how replicators work in general 19:20 <@kanzure> ah 19:21 <@kanzure> okay, so then the packages that we would define 19:21 <@kanzure> would they be genes and protein expressions ? 19:21 < fenn> take te bacteria apart, abstract and generalize the functions and structure, put it back together using other technologies 19:21 <@kanzure> remember the Minimal Cell Project? 19:21 < fenn> i've heard of it, not read much though 19:21 <@kanzure> http://heybryan.org/mediawiki/index.php/Minimal_cell_project 19:22 <@kanzure> lucky for you, I have a page 19:22 <@kanzure> we could try to model the whole damn cell, I guess 19:22 <@kanzure> and then figure out how to apply our terminology to it 19:22 < fenn> yeah 19:25 <@kanzure> isn't that kind of cheating, shouldn't we be able to come up with the model ourselves? 19:25 <@kanzure> and why can't we use von Neumann's work on theoretical self-replicators? 19:25 <@kanzure> surely there must be somebody who has been exploring the mathematics of this topic 19:25 < fenn> i was just trying to make sense of your 'reverse method' 19:26 <@kanzure> we're operating on many different levels here 19:26 <@kanzure> ah 19:26 <@kanzure> well, I see what you mean in that case 19:26 <@kanzure> the other reverse methods that I have been proposing are different of course 19:26 <@kanzure> such as starting with a fabricator arm that can move "completed components" to assemble the fabricator arm itself 19:27 <@kanzure> and then adding in the closed-form dependency loops to make all of the components on the shelf 19:27 <@kanzure> which is a reverse recursive definition method, but not quite the same thing as starting with a cell instead 19:27 < fenn> that's the reprap way, but unfortunately plastic goo isnt good for shit 19:27 < fenn> so you have to start with something that has potential 19:28 < fenn> and legos, have you ever tried to manufacture something using only legos? 19:28 <@kanzure> nope 19:28 <@kanzure> but 19:28 <@kanzure> I wonder if it would be possible to make a lego factory out of legos 19:28 < fenn> i saw a youtube where a 'factory' was assembling legos.. looked godawful inefficient 19:29 < fenn> the factory was made of legos too 19:29 <@kanzure> it surely had some electronics or someting 19:29 <@kanzure> *something 19:29 < fenn> yes there's an electronics brick 19:29 < fenn> i guess you could use control cams like in a screw machine (not that legos lend themselves to that method) 19:30 <@kanzure> Google isn't being kind to me - not much info on synthesis of agar 19:30 <@kanzure> or agarose for that matter 19:30 < fenn> agar is made of agarose :) 19:30 <@kanzure> but it occurs to me that there does exist such things as aerobic bacteria 19:30 < fenn> its refined from seaweed, fuchus i think 19:30 <@kanzure> hrm 19:30 <@kanzure> that's not good 19:30 < fenn> agarose is not a nutrient 19:30 <@kanzure> how does fuchus make it? 19:30 <@kanzure> ooh 19:30 < fenn> it's used for bacteria because they can't digest it 19:30 <@kanzure> well, we can just have giant water tanks pointed at the sun 19:30 <@kanzure> oh? 19:31 < fenn> so you get a nice smooth plate with bumps sitting on top 19:31 < fenn> instead of having the bacteria forming branching colonies inside the agar 19:31 < fenn> well, they do that too, but it doesnt break down at least 19:31 < fenn> if the agar broke down it would turn into a pool of slime 19:33 < fenn> lego factory: http://www.youtube.com/watch?v=GQ3AcPEPbH0 19:37 < fenn> with some odometry that whole thing could be replaced by a car with a gripper 19:40 < fenn> the minimal cell project has too much information processing in the form of molecules, it's like a factory made of cams and levers 19:41 < fenn> so it's a bad model to follow for macro-scale replicators 19:41 <@kanzure> take a look at this 19:41 <@kanzure> http://heybryan.org/mediawiki/index.php/Self-replication#Sipper_excerpt 19:42 <@kanzure> http://en.wikipedia.org/wiki/Langton's_loops 19:42 < fenn> sure but it relies on the cellular grid being just right 19:42 < fenn> the rules of the universe, if you will 19:45 < fenn> von neumann's design separates structure from information, which is desirable when dealing with a fuzzy non-digital universe 19:46 <@kanzure> http://ti.arc.nasa.gov/people/jlohn/bio.html <-- good guy to know 19:46 < fenn> so now we have DNA and proteins, instead of an RNA-based ribozyme mishmash 19:46 <@kanzure> "the rules of the universe, being just right" --> With any sufficient preparation ... 19:47 < fenn> neat stuff 19:48 <@kanzure> http://lslwww.epfl.ch/pages/embryonics/ interesting 19:50 < fenn> evolution is good at circuit-hiding :\ 19:51 <@kanzure> it doesn't look like anybody has actually focused on the mathematics of self-replication, away from cellular automata 19:51 <@kanzure> in fact, why the hell has cellular-automata been used in the first place 19:51 <@kanzure> this should have been graph based since the beginning 19:51 < fenn> because cellular automata is a math-based representation of 'the real world' 19:52 <@kanzure> so if we had a cellular automata configuration file that was a self-replicator 19:52 < fenn> also because, in math you can just say 'a = b' and you're done 19:52 <@kanzure> how would we use that to help us find a material implementation 19:52 < fenn> well, remember they came up with the information tape before DNA was discovered 19:53 <@kanzure> so? 19:53 < fenn> what if we hadn't discovered DNA? 19:53 <@kanzure> heh, we'd still be stuck on CAs. 19:54 < fenn> i guess they had paper tape already 19:54 < fenn> anyway, i dont know what the CA simulation is good for 19:54 <@kanzure> nothing as far as I can tell 19:54 <@kanzure> looks like we were going in the right direction with our definitions/graph-theory approach 19:55 <@kanzure> with the extendible simulator to generate posisble reactions and interactions between streams so that we can figure out if we have missed any optimizations 19:56 < fenn> btw have you ever seen animations of the vonneumann replicator? it's really cool to watch 20:01 <@kanzure> nope 20:01 < fenn> well too bad, you probably never will :P 20:03 <@kanzure> hm, this is interesting - 1. W.M.Stevens "NODES: An Environment for Simulating Kinematic Self-Replicating Machines" Proc. of the Ninth International Conference on the Simulation and Synthesis of Living Systems (ALIFE9) 39-44, 2004. 20:03 <@kanzure> the guy's web site - http://www.srm.org.uk/home.html 20:04 < fenn> yep found it 20:07 < fenn> looks quite biological 20:08 < fenn> havent any of these guys heard of a tape reel? :) 20:10 <@kanzure> http://www.landesbioscience.com/books/special/id/912 20:10 <@kanzure> $150. I may just have to go beg Freitas. 20:10 <@kanzure> Or maybe Max can lend me his copy, if he has it. 20:12 < fenn> it's on his website 20:13 < fenn> http://www.molecularassembler.com/KSRM.htm 20:13 < fenn> i need to read that too 20:13 <@kanzure> let's wiki it 20:14 < fenn> wiki what? the whole thing? :) 20:15 < fenn> gosh there are a lot of citations 20:16 <@kanzure> yep 20:26 <@kanzure> http://heybryan.org/mediawiki/index.php/KSRM 20:26 <@kanzure> my start to it 20:26 < fenn> this book must be rather heavy 20:27 <@kanzure> Freitas is a peculiar fellow. I saw him on video once, he had no clue as to how to run a presentation. He was reading straight from a transcript, looking down, was not energetic, very monotonous tone. But if he's really all that he's cracked up to be ... 21:18 <@kanzure> fenn: http://heybryan.org/mediawiki/index.php/KSRM/3 21:18 <@kanzure> it's getting all messed up 21:19 <@kanzure> http://heybryan.org/projects/ksrm/output.txt has the entire wiki code for the book, sort of 21:19 <@kanzure> but it's all out of order I think 21:19 <@kanzure> oh wait, nevermind 21:37 <@kanzure> hm, 21:37 <@kanzure> Freitas cites Winfree. 21:38 <@kanzure> 1175. Brent A. Ridley, B. Nivi, Joseph M. Jacobson, “All-inorganic field effect transistors fabricated by printing,” Science 286(22 October 1999):746-749; http://www.media.mit.edu/molecular/Science10-99.pdf 21:39 <@kanzure> 1987. Kinneret Keren, Rotem S. Berman, Evgeny Buchstab, Uri Sivan, Erez Braun, “DNA-templated carbon nanotube field-effect transistor,” Science 302(21 November 2003):1380-1382, 1310 (comment). See also: “DNA used to create self-assembling nano transistor,” ISRAEL21c, 23 November 2003; http://www.israel21c.org/bin/en.jsp?enPage=BlankPage&enDisplay=view&enDispWhat=object&enDispWho=Articles%5el557&enZone=Technology&enVersion= 22:03 <@kanzure> fenn: it occurs to me that we can use humans as the 'black box' for as long as we want for self-replication, slowly porting over certain functions to the actual hardware, until eventually the machine replicates on its own 22:05 < fenn> reading the printed FET paper.. i wonder if the nanocrystals could be aligned by an externally applied electric field 22:06 < fenn> kanzure: it's more than just 'humans' it's industrial civilization 22:06 <@kanzure> as long as we can move all of the functionality over to the human, we're good 22:06 < fenn> i dont know how to make aluminum from scratch :P 22:06 <@kanzure> what printed FET paper? 22:06 <@kanzure> are you reading my nanocrystals page? 22:07 < fenn> all inorganic field effect transistors fabricated by printing 22:07 < fenn> you just linked to it 22:09 <@kanzure> oh 22:09 <@kanzure> well 22:09 <@kanzure> more on semiconductor nanocrystals - http://heybryan.org/alternate_transistors.html 22:09 <@kanzure> i.e., printed FETs. 22:09 <@kanzure> but this isn't the DNA-based FETs. 22:10 < fenn> i cant load the israel21c.org url 22:11 <@kanzure> http://web.archive.org/web/*/http://www.israel21c.org/bin/en.jsp?enPage=BlankPage&enDisplay=view&enDispWhat=object&enDispWho=Articles%5el557&enZone=Technology&enVersion=0& 22:11 < fenn> did you ever read about qtc pills? 22:14 <@kanzure> "The DNA serves as a scaffold, a template that will determine where the carbon nanotubes will sit," Braun said. 22:14 <@kanzure> hmm 22:14 <@kanzure> nope 22:15 <@kanzure> looks like pressure-based switches? 22:15 < fenn> yes 22:15 < fenn> they're rubber-state switches 22:16 < fenn> well, variable resistors 22:17 < fenn> the fabrication sounds really low tech 22:17 < fenn> mix nickel powder and adhesive 22:18 <@kanzure> woah 22:18 <@kanzure> that's definitely worth it 22:18 <@kanzure> why haven't I seen more of this 22:18 <@kanzure> searching for 'qtc pills' doesn't show up much but a single company's website 22:18 <@kanzure> can you get me some other links or something ? 22:18 < fenn> it's fairly new, and there's only one company developing it (the inventor) 22:19 < fenn> hell if i know why it isnt more widespread 22:34 < fenn> popular article on the history of qtc http://www.sep.org.uk/catalyst/articles/catalyst_18_1_333.pdf 22:40 <@kanzure> hm 22:42 < fenn> i wonder if you could make a composite material that was both a qtc-resistor and a piezo actuator 22:42 < fenn> so a small electric field would change the conductivity 22:42 < fenn> it would probably have to be anisotropic, like a fiber embedded in silicone rubber 22:42 <@kanzure> how does a piezo actuator work? 22:43 < fenn> good question 22:43 <@kanzure> is it just a simple case of applying a current and making it move? 22:44 <@kanzure> I am pretty sure piezo actuator == piezo buzzer 22:44 <@kanzure> `http://www.sep.org.uk/catalyst/articles/catalyst_18_1_333.pdf` 22:44 <@kanzure> oops 22:44 <@kanzure> Inkjet printers: On some inkjet printers, particularly those made by Epson, piezoelectric crystals are used to control the flow of ink from the inkjet head to the paper. 22:44 < fenn> i think piezoelectric materials deform due to voltage, but there is a certain amount of charge that must be transfered (as if it were a capacitor) 22:45 <@kanzure> Piezoelectric motors: piezoelectric elements apply a directional force to an axle, causing it to rotate. Due to the extremely small distances involved, the piezo motor is viewed as a high-precision replacement for the stepper motor. 22:45 < fenn> yeah the piezo inkjets physically squirt the ink onto the page 22:45 <@kanzure> yes, piezoelectrics work just like that 22:45 < fenn> ugh, no, ignore the piezoelectric motor snippet 22:45 <@kanzure> but I have not figured out how that applies to piezo steppers / actuators 22:45 <@kanzure> oh? 22:45 <@kanzure> because that doesn't make sense 22:45 <@kanzure> if you apply a voltage 22:45 <@kanzure> it will deform and push the axil 22:45 <@kanzure> but then when you stop the voltage 22:45 <@kanzure> it should just move back to where it started 22:46 <@kanzure> which makes no sense :) 22:46 < fenn> you can cause rotational motion, but not _continuous_ rotation 22:46 < fenn> unless you have some ratchet mechanism or something 22:46 <@kanzure> I doubt they are using ratchets 22:46 <@kanzure> because you're using twice the energy 22:46 <@kanzure> that you need. 22:47 <@kanzure> where's Superkuh when you need him? heh' 22:47 <@kanzure> he's a self-made expert in piezos last time I checked 22:47 <@kanzure> was designing a piezo-based particle accelerator 22:47 < fenn> you could use a piezo actuator as the replacement for a piston in an engine (sorta) 22:47 < fenn> ah i bet he was doing something with fusion 22:48 < fenn> lithium deuteride crystals yes? 22:48 <@kanzure> so how would you cause rotational motion with piezos? 22:48 <@kanzure> let me chekc 22:48 <@kanzure> Ferroelectric/pyroelectric particle accelerators and research with homebrew methods. 22:48 <@kanzure> [context pressure ~=10^-3 Pa] Has anyone considered using pyroelectric crystal mediated acceleration of charged particles (spontaneous polarization of voltage oriented through the c-axis crystal faces (when heating the , +z surface has a negative potential and the -z surface has a positive potential, reverse for cooling) as the result of shifting ions in the crystal shape/lattice combined with funky dilute gas charge masking behav 22:48 <@kanzure> Particle (electrons or ions depending on the crystal face and heat/cool cycle) energies up in the 150keV+ range are very doable (or 200kev+ with dilute light gases and a bipolar setup). Plus with cylindrical crystals the beams are self focusing. A common choice for this is Lithium Tantalate, and making thin films and stacking this material is possible, but there are lots of tricks involved (and making crystal boules is a magnitud 22:49 < fenn> that's pyroelectric, not piezoelectric 22:49 <@kanzure> yeah .... oops. 22:49 <@kanzure> but 22:49 <@kanzure> he's the guy that introduced me to piezoelectricity 22:49 < fenn> i see 22:50 < fenn> well, anyway, there are piezoelectric polymers (kynar/polyvinylidene) 22:51 <@kanzure> hm, that's convenient 22:51 < fenn> but the most promising are ceramic 22:51 <@kanzure> but how does the rotational motion work that you were mentioning? 22:51 < fenn> a piston in a car engine goes up and down, so replace the piston with a piezo actuator 22:51 < fenn> the crank turns reciprocating motion to rotation 22:52 < fenn> there's more ways to do it 22:52 <@kanzure> huh 22:52 < fenn> like a clock wheel, hit each tooth with a hammer as it goes by 22:52 <@kanzure> so if you just do rapid voltage pulses 22:52 <@kanzure> you'd convert this into rotational motion with a crank 22:52 < fenn> i saw a 'wiggle motor' somewhere 22:53 < fenn> it's a piezo rod that wiggles its way around, using a gearing principle similar to harmonic drives to provide a decent output torque 22:56 < fenn> it was about 5mm wide 20mm long 22:56 < fenn> cylinder 22:56 <@kanzure> not much turning up on Google. 22:58 < fenn> http://www.newscaletech.com/squiggle_overview.html 23:01 <@kanzure> that's really convenient 23:02 <@kanzure> I wonder how we could make it. 23:02 <@kanzure> It looks like just a piezo + voltage + ... I don't know what. 23:02 <@kanzure> either a rachet or a crank 23:03 <@kanzure> from Superkuh's bookmarks - http://www.pollin.de/shop/shop.php?cf=detail.php&pg=NQ==&a=MDA5OTQ3OTk= giant piezo crystals 23:04 <@kanzure> http://www.chem.pacificu.edu/Johnson/JohnsonResearch/STM/PIEZO.HTM piezo tube 23:04 <@kanzure> http://www.colomar.com/Shavano/mini_piezo_tweeter.html DIY mini piezo tweeter 23:04 <@kanzure> Hm, in my trash I apparently have a Google search for pyroelectrical transistors - http://www.google.com/search?client=opera&rls=en&q=pyroelectric+transistor&sourceid=opera&ie=utf-8&oe=utf-8 23:05 < fenn> wow that's cheap (pollin.de) 23:07 <@kanzure> maybe we should stock up ;) 23:09 < fenn> the piezo tube looks like what the squiggle motor uses 23:10 < fenn> heh, go back on that page it says 'inchworm motor housing' for STM 23:11 <@kanzure> the piezo tube is for motion in x/y/z directions, not up/down, although ... if we could find a piezo that could change shape by say, 5 mm or something, that would be great 23:11 <@kanzure> or I guess it doesn't have to be too much 23:11 <@kanzure> it could just be a rubber band stretching system 23:11 < fenn> oh i guess 'inchworm motor' is the stick-slip thing i keep reading about 23:12 <@kanzure> http://en.wikipedia.org/wiki/Inchworm_motor 23:12 < fenn> why 5mm? what do you plan to do with it? 23:12 <@kanzure> well 23:12 <@kanzure> a 1 m piezo would be great 23:12 <@kanzure> as a piston or something :) 23:13 <@kanzure> http://www.physikinstrumente.com/en/news/fullnews.php?newsid=123&Wiki_iw I hate this website, but it's about the principles of the inchworm motor 23:13 <@kanzure> physikinstrumente spams all over the place 23:13 <@kanzure> tons of domain squatting 23:14 <@kanzure> if this piezo motor can scale from 1 Newton to up to thousands of Newtons that'd be great 23:15 < fenn> a motor might have force and speed but not at the same time 23:15 < fenn> but why does it matter? what are you going to use it for? 23:16 <@kanzure> huh? 23:16 <@kanzure> I was assuming you had a few ideas 23:16 <@kanzure> for using it in a replicator 23:16 < fenn> well, i never considered piezo's to be useful for producing mechanical power 23:16 < fenn> that might be a mistake on my part 23:17 <@kanzure> it looks fairly convenient 23:17 < fenn> did you notice the efficiency chart? looks like they are much less efficient when size goes up 23:18 <@kanzure> hrm 23:18 <@kanzure> qtc-resistor + piezo would be really useful though 23:18 < fenn> on the right side of here http://www.newscaletech.com/doc_downloads/SQUIGGLE_Overview_2-18-08.pdf 23:18 <@kanzure> combines a lot of parts 23:19 < fenn> yes, i was thinking mainly piezos for squeezing qtc's to control a more conventional motor 23:19 < fenn> like, a pancake motor or switched reluctance 23:20 < fenn> voltage is easy to get from a tiny computer chip 23:20 < fenn> anyway this all needs some empirical data 23:21 < fenn> i've never used either qtc's or piezo's 23:24 <@kanzure> the qtc research group should be responsive 23:24 <@kanzure> http://www.dur.ac.uk/psm.group/qtc.html 23:26 < fenn> i'd like to know if there are other metals besides nickel that form nano-spikes 23:27 <@kanzure> I need to take a break from this stuff and go figure out a circuit for Andy ... he wants me to come up with something that has an emergent property 23:27 <@kanzure> but this is pretty hard 23:27 <@kanzure> I am thinking about telling him about my 'evolvable DNA logic circuits' idea, but I think that's sidestepping his request 23:28 < fenn> something with an emergent property eh 23:28 <@kanzure> heh' 23:28 < fenn> cats and dogs and buttered toast! 23:28 <@kanzure> he cited the ring oscillator as the "classic example" 23:28 <@kanzure> indeed 23:29 <@kanzure> the oscillator example is really simple and takes up a huge class of possibilities 23:29 <@kanzure> so I can't do anything that oscillates 23:30 <@kanzure> I was thinking about trying to implement sine 23:30 <@kanzure> but that's analog, whereas this is discrete-state 23:30 < fenn> forget about emergence, come up with something that has 'a spiritual property' 23:31 < fenn> they're both meaningless words 23:32 <@kanzure> then what should I implement 23:32 <@kanzure> calculations are mostly meaningless in this context 23:32 <@kanzure> although the prospects are interesting for large-scale computing really 23:33 <@kanzure> because if you have a few million gates for a single circuit 23:33 <@kanzure> and want to do simultaneous computation all at once 23:33 <@kanzure> then you have them all extremely localized 23:33 <@kanzure> if you have it acting in a centrifuge then it would be interesting to store the result of the computation eventually, but I don't see how 23:33 <@kanzure> maybe a flip flop system 23:34 <@kanzure> but there's no guarantee of saturation or propagation of all of the same states to all of the flip flops 23:34 <@kanzure> hrm 23:35 <@kanzure> maybe a massive cellular automata implementation 23:35 <@kanzure> where the logic gates are the cells 23:36 <@kanzure> each logic gate has four neighbors, and they are all executed in parallel, so what would happen? 23:36 < fenn> *drool* http://sourceforge.net/projects/xandra 23:36 <@kanzure> awesome 23:38 < fenn> how do you connect the neighbors? through winfree's toehold addressing mechanism? 23:44 <@kanzure> yes 23:44 <@kanzure> and in this scenario, 23:44 <@kanzure> you don't have to be too specific either 23:45 <@kanzure> since you can have four neighbors for each 'cell' 23:45 <@kanzure> heh 23:45 <@kanzure> so that means I don't have to be as strict when selecting the toeholds and the recognition sequences 23:45 <@kanzure> sneaky, isn't it? 23:46 < fenn> screenshot from xandra showing a hexaglide actuator: http://imagebin.org/15164 23:48 < fenn> i dont get it. the reason you have to be strict when selecting toehold sequences is to reduce crosstalk 23:48 <@kanzure> yes 23:48 <@kanzure> well 23:48 <@kanzure> not only that 23:48 < fenn> crosstalk will bunch up your fabric 23:48 <@kanzure> but you also don't want one toe-hold to match for the wrong logic gate 23:48 <@kanzure> it's kind of like a lock-and-key method 23:48 <@kanzure> where the logic gate output is a key to the next one in the pathway 23:48 <@kanzure> so if your key unlocks the 4 neighboring cells .. 23:48 <@kanzure> where 'cells' are really just gates 23:50 < fenn> 'cell' as in cellular automata right? not bacterial cell 23:51 < fenn> blah my brain is hurting 23:52 <@kanzure> right 23:52 < fenn> another xandra screenshot http://imagebin.org/15165 23:52 <@kanzure> heh 23:53 <@kanzure> so what does it simulate 23:53 <@kanzure> just a 3D visualization? 23:53 < fenn> machine tool kinematics 23:53 <@kanzure> ooh 23:54 < fenn> windows only right now or i'd be playing with it 23:54 <@kanzure> wine? Day changed to 23 Mar 2008 00:04 <@kanzure> fenn: you wouldn't happen to remember how to use the X11 screenshot feature to take a snapshot of a particular program? 00:05 < fenn> i use ksnapshot 00:05 <@kanzure> I need to automate it. I think there was a program named xsd or something. 00:05 < fenn> try `man import` 00:06 <@kanzure> okay 00:06 < fenn> with import you have to click on the window.. 00:09 < fenn> xwd is what you're remembering 00:12 <@kanzure> yes 00:12 <@kanzure> thank you 00:12 <@kanzure> turns out xplanet ( http://xplanet.sf.net/ ) does image output itself 00:12 <@kanzure> so all is good 00:13 <@kanzure> a guy from the Moon Society wants me to write a quick script to show the earth from the moon at the given moment 00:13 <@kanzure> he said his web dev team has been at it for two years 00:13 <@kanzure> I laughed and set it up in 2 minutes :( 00:13 < fenn> two years! 00:14 <@kanzure> script kiddies do better than two years ... 00:18 < fenn> i feel like i've been doing JITT (just in time training) for a few years now 00:19 < fenn> exhibiting symptoms of 'jitt-stick' 00:19 < fenn> wikipedia overdose 00:24 <@kanzure> JITT indeed 00:24 <@kanzure> but what can be done about that 00:27 < fenn> mobile computing for exercise (faster removal of brain waste products, increased oxygen) 00:27 < fenn> might cause more harm than good though, you could walk into the street or something 02:16 -!- kanzure [n=bryan@cpe-70-113-54-112.austin.res.rr.com] has quit ["Leaving."] 05:06 [Users #hplusroadmap] 05:06 [ Enki-2] [ epitron] [ fenn] 05:06 -!- Irssi: #hplusroadmap: Total of 3 nicks [0 ops, 0 halfops, 0 voices, 3 normal] 05:07 < fenn> Light from the sun hitting lunar dust causes it to become charged through the photoelectric effect. The charged dust then repels itself and lifts off the surface of the Moon by electrostatic levitation. 05:07 < fenn> It is thought that the smallest particles are repelled up to kilometers high, and that the particles move in "fountains" as they charge and discharge. 05:23 -!- fenn_ [n=pz@adsl-76-251-81-249.dsl.bltnin.sbcglobal.net] has joined #hplusroadmap 05:23 [Users #hplusroadmap] 05:23 [ Enki-2] [ epitron] [ fenn] [ fenn_] 05:23 -!- Irssi: #hplusroadmap: Total of 4 nicks [0 ops, 0 halfops, 0 voices, 4 normal] 05:23 -!- [freenode-info] if you're at a conference and other people are having trouble connecting, please mention it to staff: http://freenode.net/faq.shtml#gettinghelp 05:23 -!- Channel #hplusroadmap created Sat Mar 22 15:44:12 2008 13:26 http://archives.seul.org/geda/dev/Nov-1999/msg00120.html 13:26 EDIF 2 format 13:26 oh 13:26 heh 13:26 good old ~rcary 13:26 http://www.rdrop.com/~cary/mirror/tools_htmlized.html 13:27 This guy is a god. 13:27 David Cary. 13:27 http://www.rdrop.com/~cary/html/idea_space.html <-- read this like a bible 13:31 http://www.geda.seul.org/wiki/geda:icarus_xnf 13:31 gEDA can do XNF. 13:32 heh i'm reading ~cary and he's slipping into orion's arm lingo 13:32 yes 13:33 I've emailed him over the years. He doesn't seem to exist any more. 13:35 reduced information diet 13:38 From #electronics 13:38 (2008-03-16 12:39:35) gordboy: i'd like to see store "disloyalty" cards. you pop in your card and it injects a virus into the system and causes all the wheels and dials to spin hopelessly out of control, while technicians with suspiciously german-sounding accents rush around shouting "gott und himmel. donner und blitzen" 13:39 screw it, I could have been done by now with a 'logic' grammar 13:39 just need something that can do "output1 = input1 xor input2 xor input3" 13:39 um, how bout C? 13:39 but it compiles into bytecode 13:39 what kind of bytecode? 13:40 I mean, the machine code for the specified architecture 13:40 oh 13:40 I wonder how to specify the architecture 13:40 and the specified architecture would thus be DNA 13:40 i think that's the hard part 13:41 Quick! to #gcc ! 13:43 maybe it would be technically smarter to 13:43 make an assembler 13:44 assembler assumes a linear sequence, whereas what you are building is inherently parallel execution 13:44 nopre 13:44 *nope 13:44 surprisingly not 13:44 assume one DNA molecule per cell 13:44 and the gene, the nucleotides on the molecule, *that's* the gate 13:44 :3 13:45 bwahah 13:45 so now I just need to look for a simple assembly language 13:45 I will _not_ implement x86 or MIPS. Screw that. 13:45 something's walking down the molecule though, recombining at various places 13:45 so? 13:46 you can have more than one of those 13:46 yeah, but there's only one gate 13:46 it's just asynchronous logic 13:46 it's a bottleneck 13:46 what is the 'something'? 13:46 your 'something'? 13:46 I thought it was RNA polymerase 13:46 or DNA polymerase 13:47 so, the state of the logic gate's inputs are blobs of protein stuck to the dna, and the 'flow' is whether the gene is transcribed.. got it 13:48 right, because when you transcribe the gene, you're making the "output" of the gate, which then becomes the input to the next gate 13:48 this is much simpler then :) 13:48 so I just need to write a DNA assembler. 13:49 so you need a library of transcription factors and inhibitors 13:49 that's your 'cache' 13:49 there are rules for making those library-items 13:49 so automatic generation of those is important 13:49 since you might have a circuit with a few thousand components 13:49 and you want your output going to the right place 13:49 has that been proven in a lab? 13:49 creating new transcription factors 13:50 yep, I'm pretty sure 13:50 http://www.dna.caltech.edu/DNAresearch_publications.html 13:50 see Construction of an in vitro bistable circuit from synthetic transcriptional switches. 13:51 it does not specifically imply that they have done this, but it is clear that they are using a mechanism so that logic does not 'jump' 13:51 also see Enzyme-Free Nucleic Acid Logic Circuits. 13:52 ah, shit: Protein Design is NP-Hard. 13:53 you could do a combinatorial approach, tying known-to-work blobs together 13:54 wait, why do I want an assembler 13:54 an assembler works off of an ISA 13:54 and the ISA is for using the hardware 13:54 but we are writing the hardware, no? 13:54 this is confusing 13:54 you want to generate the 'bitstream' i believe 13:54 ? 13:54 which is where your netlist gets turned into hardware state 13:55 in FPGA's there is some data which initializes the hardware into the state you want it to be (your circuit) 13:55 it's called a bitstream 13:55 VHDL/verilog describe the desired state in an abstract sense 13:56 this gets turned into something made of slightly more concrete elements (logic gates and digital devices like flipflops) 13:56 that's your xnf file 13:56 eh 13:56 let's modularize this 13:57 the basic output of this program should be a single logic gate 13:57 as in, the DNA 13:57 so, program xxdna should output DNA + information on "output specs" ... the parameter to xxdna should be (1) the logic gate needed and (2) the input-key that it should read (i.e., it can't just accept anything) 13:58 the space is limited obviously, so if you need 2^8 gates, you need something like 4 nucleotides for the 'key' 13:58 or something like that, 4^4 should be 2^8 13:58 anyway, the input parameter to xxdna (the second parameter) should actually be a file name 13:58 and in this file we can store all of the input keys already used 13:58 so it just uses the next one 13:59 it might be too hard to recognize DNA sequences, you will probably have to use combinations of proteins that bind together 13:59 and then it makes up its own output-key, which becomes the parameter for the next call to xxdna 13:59 nope, 13:59 as it turns out that's well studied 13:59 I need to go back and read the papers, but it looks pretty well studied 14:01 ah that's good 14:01 I need to go find the algorithm. 14:01 because if I can find the algorithm 14:01 this is a very simple perl scripot 14:02 *script 14:02 so you can just change the protein sequence to recognize a differen dna sequence? 14:02 no proteins involved 14:02 oh? it's RNA transcription factor? 14:02 all mRNA strands :) 14:02 yep 14:02 nice 14:02 protein folding is NP-hard, remember? hehe 14:03 i think that's a myth, but i digress 14:03 it's a paper on their site 14:03 heh 14:05 the papers talk about a "toetail" which is what we want to algorithmically change 14:05 but they do not mention where this is in the actual sequence of the gates and so on 14:06 they give 8 different DNA sequences, I need to go back and map these out to what each of them are supposed to do 14:06 I have the sequences at the end of the page over here -- http://heybryan.org/winfree.html 14:11 the output from the gate is tagged with the 'address' of the next gate 14:11 sort of, yes 14:14 is there a 'packet diagram' somewhere or do they just list the raw sequences and hope you can read AGTAGAACCATAACACAAGGGTTCTCAAGAA 14:14 packet diagram? 14:16 http://www.media.mit.edu/physics/publications/papers/04.10.sciam/barcode.jpg 14:17 ogh 14:17 heh 14:17 so in other words, a packet diagram of each of the DNA sequences 14:17 [this part should be changeable][READ ONLY!!] etc. 14:17 right 14:20 --> The sequences of all DNA molecules and expected RNA transcripts 14:20 were chosen to minimize the occurrence of alternative secondary 14:20 structures, checked by the Vienna group’s DNA and RNA folding 14:20 program (Flamm et al, 2000). All DNA oligonucleotides were 14:20 purchased (Integrated DNA Technologies, Coralville, IA). T21-nt is 14:20 hmm 14:20 I guess I need that program 14:21 flim flamm folding fun 14:22 Flamm's papers 14:22 http://www.tbi.univie.ac.at/~xtof/papers.html 14:22 Vienna RNA secondary structure server -- Hofacker 31 (13): 3429 ...The Vienna RNA package (12) is a free software package that implements a variety of ...... A. Weixlbaumer, A. Werner, C. Flamm, E. Westhof, and R. Schroeder ... 14:22 nar.oxfordjournals.org/ cgi/content/full/31/13/3429?ck=nck - Similar pages 14:22 (behind a paywall) 14:22 ah 14:22 RNAcofold 14:23 it doesn't matter so much if you can't get the papers (if you have the source) 14:23 http://www.tbi.univie.ac.at/~ivo/RNA/ 14:23 haha, bitches! 14:23 google ftw 14:24 actually :( http://en.wikipedia.org/wiki/List_of_RNA_structure_prediction_software 14:25 `Finally, we provide an algorithm to design sequences with a predefined structure (inverse folding).` 14:25 http://rna.tbi.univie.ac.at/cgi-bin/RNAinverse.cgi <-- Web interface :) 14:27 so now I am confused 14:27 the RNA inverse input is not a nucleotide sequence 14:27 therefore RNA folds too? 14:27 i think you want the other direction to figure out what those sequences do 14:28 you have to come up with a sequence in the first place 14:28 rna and dna can fold up, depending on their sequence 14:28 that web program takes a structure (a shape) and spits out a sequence that will give that shape 14:28 okay, then I want RNAup 14:28 The RNAup program for computing RNA-RNA interactions is now included. 14:28 yes 14:28 so, if we want the logic gate to go through with the operation 14:28 the two RNA sequences must bind 14:28 or actually 14:29 the RNA output sequence (from the last gate) 14:29 and relplot.pl 14:29 has to bind with the DNA 14:29 so this is RNA-DNA interaction 14:29 not RNA-RNA. 14:29 you need both? the mRNA will interact with itself 14:30 huh? 14:30 that's true 14:30 so this is to check RNA-RNA interactions 14:30 oh 14:30 the RNA output has to bring in a sequence 14:30 that will complete the DNA for the next logic gate 14:30 so the RNA has to be able to "bring/tug/port" an oligonucleotide sequence over to the next logic gate 14:31 hm 14:31 I wonder if we can make it so that 14:32 yeah, it has to be made so that when the oligonucleotide sequence is wrong (the wrong key) the RNA does not deposit the oligonucleotides into the logic gate 14:32 because that would 'block' the logic gate 14:34 one problem i can think of is sequence specificity.. a partly-mismatched 'address' will still bind somewhat 14:34 nope 14:34 so you're still dealing with analog 'fuzzy' logic somewhat 14:34 I got it 14:34 DNA repair mechanisms 14:34 you have only one strand of the DNA logic gate there 14:34 and the RNA transcribers bring in the complement 14:35 yeah, wait a second 14:35 how do they unbind 14:35 or, why doesn't the DNA repair mechanisms just fill in the gaps 14:35 I think it's at "run time" where everything happens 14:35 so that only at run time, when the logic gate is being transcribed, is it on or off 14:35 you lost me with the DNA repair stuff 14:35 well 14:36 I am now very confused 14:36 I was thinking that it may be using strand complements 14:36 for the on/off property 14:36 but that would mean that DNA repair mechanisms could just come in and add the right nucleotides and make it blocked 14:36 but I realize now that it has to be more "run time" than that 14:36 i.e., the actual magic happens during transcription 14:36 this is complex.. 14:36 yes 14:37 the dna structure exposes certain regions of DNA 14:37 no, it can't do that 14:37 because DNA repair guys would come around and fix it 14:37 in a double helix 14:37 think of this: AATTGGTTT[where you need your gate output to fit]TTCCCATTGTAC and then this is attached to a complete compliment 14:37 the DNA proteins would fix the gap 14:38 therefore this is not how it works. 14:38 everything's hunky dory, but 5% of the time (or whatever, depending on the binding strength of the DNA sequence) the double stranded structure comes apart and allows transcription factors to bind 14:38 AT pairs are stronger than GC pairs 14:38 hm? 14:38 it comes unzippered 14:39 it's all jumbling around in a soup of brownian motion 14:40 yes 14:40 so you can think of it like quantum mechanics, it's in suchnsuch state 20% of the time, 14:40 but I am having trouble parsing what you are saying 14:40 the DNA repair mechanisms only detect certain types of structural abnormalities, they won't bind to partially open sequences like when it comes a bit unzippered 14:41 when something 'binds' to a sequence, it's sometimes there and sometimes not there 14:41 oh 14:42 oh 14:42 also, 14:42 page 4 is good 14:42 Construction of an in vitro bistable circuit from synthetic transcriptional switches - bistable_switch2006.pdf 14:45 that's a lot of dense information 14:45 I need to diagram it or something 14:49 for some reason I keep on reading that paragraph over and over again, but it's not getting any better -- I think that they are giving the packet specification there, but I don't know how they figured out what sequence to put in it 14:49 their diagram sucks 14:51 oh boy the waveforms are plotted in hours 14:52 that's not good, is it? 14:54 this is in vitro, it could mean anything 14:57 http://en.wikipedia.org/wiki/List_of_RNA_structure_prediction_software#_note-book1 14:57 ? 14:58 i got confused.. just spent half of yesterday reading about the hitchhiker's guide 14:59 you've never read it? 14:59 i've read it, but i've never read _about_ it 14:59 ah 14:59 um, nevermind 15:02 you shouldn't describe your dna/rna transcription circuits as using 'genes' since genes usually produce a protein and this gets confusing 15:02 hrm 15:04 the ellington stuff on the other hand actually uses genes 15:05 here's the real Ellington stuff 15:05 http://ellingtonlab.org/mediawiki-1.10.0/index.php/Main_Page 15:06 hm 15:06 he has an amorphous computation guy 15:08 not a very useful webpage 15:08 right 15:09 i mean the transcription factors and gene products on this page are actually proteins http://heybryan.org/genetic-circuits.html 15:09 yes 15:10 you will have to interface with proteins at some point in order to do sensing 15:10 otherwise why bother running a biological computer? 15:10 that's okay 15:10 inhibition of circuits is possible 15:10 via the methods that were from genetic-circuits.html 15:10 i.e., a protein can prevent transcription 15:10 (inihibition) 15:11 dna.caltech.edu says -- 15:11 Software packages that we actively use for research, e.g. mfold, Vienna, Namot2, Tcl, RasMol..., as well as other software installed on our cluster, e.g. VMware, MATLAB, Mathematica, ... 15:18 huh 15:18 http://gcat.davidson.edu/GcatWiki/index.php/Logic_Gates_-_Emma_Garren 15:18 this might be good 15:25 also see Enzyme-free nucleic acid logic circuits - DNA_logic_circuits2006.pdf 15:25 yes i'm reading that 15:26 I like how they conveniently exclude DNA sequences. 15:35 ooh 15:35 look at the bottom right-hand corner of the last page 15:35 the materials are on sciencemag.com 15:35 which I do not have access to. 15:39 woiah 15:39 http://www.dna.caltech.edu/Papers/DNA_logic_circuits2006_supp.pdf 15:39 I got it 15:39 yeah, 18 pages 15:39 this is much easier to understand than the paper 15:40 page 16 for DNA. 15:41 Sequences were from: [S7] Lagos-Quintana, M., & Rauhut, R., & Yalcin, A., & Meyer, J., & Lendeckel, W., & Tuschl, T. (2002) Current Biology 12 735–739. 15:41 [S8] Lagos-Quintana, M., & Rauhut, R., & Meyer, J., & Borkhardt, A., & Tuschl, T. (2003) RNA 9, 175–179. 15:48 hm 15:48 it's interesting that they were using sequence optimization techniques 15:49 why not just generate a few megabyte file with a list of all possible sequences 15:49 for the lengths that they are working with 15:49 I am pretty sure they were using only the Flamm software packages. Maybe they will tell me if I ask nicely? 15:51 'sequence optimization' means getting rid of things that can go wrong, right? 15:51 yep 15:51 while also making sure the sequences meet requirements 15:52 what would you do with a bunch of lousy sequences that probably wont work? 15:52 huh? 15:52 you want a list of things that *will* work 15:52 it's a toolset 15:52 yes 15:52 so then xxDNA will just come in and select a few for the circuit that you are encoding into DNA. 15:52 sure, 1 2 3 15:52 or maybe 5 6 7 15:53 so save the effort and have a pre-generated list 15:53 generate once, use many 15:53 what are the limiting factors? 15:53 minimization of secondary structures in single-stranded species 15:53 as predicted by the minimum-free-energy (MFE) structure [S3] at 25◦ C using DNA parameters [S4]; 15:54 minimization of cross-talk between all single-stranded species, as measured by the ∆G◦ of association between pairs of 15:54 strands (estimated as intramolecular MFE for a ‘virtual’ strand linking the two sequences via 5 unpaired nucleotides); 15:54 i mean, you cant have a zillion different signals without paying for it somehow 15:54 right, so we'd constrain the generation-list to maybe length = 10 15:54 which would be 4^10 possibilities, and then you narrow it down from there 15:55 by obviously avoiding certain sequences for testing heh' 15:55 how did they calculate the 'delta-g of association' 15:55 for example 15:55 no clue 15:55 for example -- they say they don't want any 3 nt to be next to each other that are the same nt 15:55 so 15:55 that narrows it down significant 15:56 *significantly 15:56 I don't remember my number theory enough to be able to recall what that might look like 15:56 number theory? just do a brute force method, it's only 4^10 possibilities :) 15:56 that's quite a lot :) 15:57 only 1e12, and you only have to calulate the list once 15:57 and a partial list is still useful 15:57 oh 15:57 only E12 15:57 that's not bad 15:57 how much meta-information would be generated, after all? 15:57 only yes/no 15:58 and the actual sequence 15:58 so it's up to (number of bits required to express the strand + 1 bit for yes/no) * E12 15:58 not yes/no, you'd get a binding strength and a specificity value (standard deviation?) 15:59 you'd do the evaluation at runtime 15:59 and you'd just have to encode a "cut off point" for making the team or not ;) 15:59 *at generationtime (so runtime, but not runtime of xxdna) 16:00 right? am i thinking this right? (4^10 different sequences)^2 16:01 why's that? 16:01 to compare each sequence against every other sequence 16:01 oh shit 16:01 that's not good 16:01 and yes, you're right 16:01 4^10 is about 1e6 16:01 4^10 is E12 16:01 no 4^10*4^10 = e12 16:01 wait 16:01 okay 16:02 so 16:02 technically there would be subgroupings that can work together 16:02 but not with others 16:02 so for each sequence we would have to specify which ones they work with 16:02 maybe it would be optimal to do "groupings" 16:02 and then part of the "generation program" parameters would be "how many divisions do you want?" 16:02 and obviously you could do up to 4^10 divisions if you wanted to 16:03 or one division (making all results compatible) 16:03 *compatable 16:03 i think 'how many divisions' is the 10 base pairs 16:03 I disagree 16:03 think about it: the first and last might be exclusive 16:03 but the first one and #14141 is good 16:03 so just because first-and-last are mutually exclusive, does that mean #1 and #14141 is bad? not at all 16:04 you just need to have some 'rules' or 'classifications' 16:04 * fenn searches desperately for an analogy 16:04 kind of like ethnic stereotypes 16:04 no asians and jews in the same room 16:04 preferably something in a computer context 16:04 heh' 16:04 oh 16:05 either way, this idea sucks 16:05 lots of computation and simulation time is required 16:05 i guess the sequence would have 'factors' which are the sub-sequences 16:06 that would make the task much less daunting 16:06 but I think it requires a simulation of their interaction 16:06 like 0110 is 0011*2 (the 11 is a factor) 16:06 like, we can't just say "No AAA and TTTA" 16:07 so what you're really combining is the sub-sequences, not the individual bases 16:07 bleh 16:08 yeah ... 16:08 how did they do it, then? 16:08 did they just get lucky? 16:08 small search space 16:08 ooh 16:08 no 16:08 they only had 6 nt in the sequence 16:08 this is the inverse folding problem 16:09 so if you have specifications for a number of entities to inversely fold 16:09 you provide their stats and numbers 16:09 and then they come up with the sequences :) 16:10 are we talking about the same thing? 16:10 yes 16:10 see 16:11 i'm coming up with a 10 nt toehold sequence that should have low-cross talk 16:11 we were originally talking about generating a list of usable components 16:11 right 16:11 if we do it by generating nt seqs, then we have lots of processing to do 16:11 but we already know what we want the finished thing to be like 16:11 we don't care what its sequence is 16:11 as long as it interacts in the defined ways 16:11 this is basically why we have inverse folding =) 16:11 ok 16:12 searching bottom-up (from possible nt seqs) would be much more computationally intensive 16:12 isn't that how inverse folding works though? 16:13 there's energy minimization algorithms at work, I think 16:13 so that they constrain search space significantly 16:13 i thought the energy minimization referred to the self-assembly of the structure, not the algorithm 16:14 the structure assumes the minimal energy configuration 16:14 http://www.tbi.univie.ac.at/~ivo/RNA/RNAinverse.html 16:14 check out the options 16:14 oh 16:15 RNAinverse searches for seqs bruteforce 16:15 crap 16:17 http://freebirds.com/ 16:17 I'll be back later. 16:18 if only they delivered via UAV 16:19 we can arrange that 16:19 they are ran by college students with nothing better to do 16:22 EUREKA 16:22 we'll do it in solution 16:22 screw computation 16:22 we just need a way to get rid of all of the sequences that have *not* binded 16:22 erm, woops 16:22 get rid of the ones that have bound together 16:22 i.e., maybe they are too heavy 16:22 and cannot be electrically attracted 16:22 to a certain anode 16:22 and then we just do DNA sequencing for all of the strands that *do* work 16:23 * kanzure goes back to the shower ... 16:23 that's a lot of gels to run 16:33 this reminds me of DNA microarrays 16:36 anyway i'd rather a computer-based method as it is more easily replicated 16:37 easily run on parallel clusters 16:46 this is really cool http://www.dna.caltech.edu/Papers/DNAorigami-nature.pdf 17:30 also the prospect of an FPGA from self-assembling tile sets (DNA/CNT hybrid structures) 18:17 that's a straightforward path to a self-replicating computer 18:18 presumably there would be a way to read the configuration data from some DNA into the FPGA's electronic circuitry 18:20 -!- kanzure [n=bryan@cpe-70-113-54-112.austin.res.rr.com] 18:20 -!- ircname : purple 18:20 -!- channels : #gcc ##electronics #biology ##neuroscience #bioinformatics @#namcub #wrongplanet @##wikipeding 18:20 -!- server : irc.freenode.net [http://freenode.net/] 18:20 -!- away : Away 18:20 -!- : is identified to services 18:20 -!- End of WHOIS 18:25 What would take longer? Simulation on a supercomputer, or using a solution and just mix them? Maybe we can use the DNA microarray idea, yes. DNA microarrays use mRNA probe tips to do hybridization, so if you detect an energy change you know there's been some binding. But this would only test one probe to the other sequences that you have in all of the wells. 18:25 But this would only test one probe to the other sequences that you have in all of the wells. 18:35 if you're doing gel electrophoresis, the probes would all have to be different lengths or they won't separate 18:36 I prefer STM DNA sequencing. 18:36 assuming you could detect a 1% difference in length (tough) how do you go about setting up the combinations? 18:36 DNA synthesizer. 18:36 giant inkjet printer ;) 18:37 that does oligosynthesis chemistry. 18:37 crap STM sequencing was sci-fi last time i read anything about sequencing 18:37 heh 18:37 http://heybryan.org/mediawiki/index.php/DNA_sequencing 18:37 or sequencer or something 18:37 I have lots of notes on STM-DNA sequencing. 18:37 I also have a $100 STM design up there. 18:56 Huh. I think they are using a genetic algorithm to come up with the optimal sequences. 18:56 other side. Scores for each of these soft criteria were weighted and summed to obtain an overall score for the set of 18:56 sequences being designed. Sequence optimization proceeded by random descent to minimization of the overall score: 18:56 sequence mutations were made randomly (subject to satisfying the structural constraints) and accepted if the score 18:56 was reduced. If the final sequences were unsatisfactory, the scoring weights were adjusted, new initial sequences were 18:56 chosen, and optimization was attempted again. 18:56 Sounds like a GA to me. 18:57 agreed 18:58 Okay. I got most of the requirements for the simulation. I just need to go get ref 3, 5, and 6. 19:35 http://heybryan.org/music/Yngwie%20Malmsteen%20-%20I%20Am%20a%20Viking.mp3 19:35 Now just the Seeman ref. 21:46 well 21:46 that was interesting 21:46 Winfree offered me a spot. Day changed to 17 Mar 2008 17:52 Hey. 17:52 So Winfree replied to my email and said that I should try looking for an analytical method of coming up with the RNA sequences. I have come up with the following formalization, I want you to take a look at it. 17:54 In any set N with 4^N different possible combinations, there are X specific "lonely RNA"s. In set N+1, the set X of specific "lonely RNAs" is not necessarily the new X subset of set N+1. If there's a general pattern in **which** sequences are "lonely RNAs" per each iteration, then we can do it analytically. 19:22 I just got done unfolding 24 hand-written pages of notes (on my own ideas, not school stuff) from my pocket. They are all from the last month or so. 20:01 hands-down keyboard, keep it in your pocket 20:03 http://en.wikipedia.org/wiki/Image:Septambic_key_numbering.jpg 20:23 Know anything about NMR? 20:25 i know how it works 20:25 wikipedia is probably a better source of info though :) 21:33 ahh motherfucking irssi silently puts server notices like 'private messages are not allowed from unregistered users' into the server tab 21:33 heh 21:33 http://en.wikipedia.org/wiki/Image:Septambic_key_numbering.jpg keep it in your pocket 21:33 Forget my NMR question. I would need E308 molecules. 21:33 There's only E77 atoms in the universe. 21:34 d'eaux 23:35 I need to magically obtain a copy of matlab. Any suggestions? Day changed to 18 Mar 2008 00:41 iirc the demo is free 00:42 then there's the pirate bay 00:42 indeed 00:42 http://www.youtube.com/watch?v=vimz7sz9Fx4&feature=related Star One - Space Metal - High Moon. 00:42 or you could use Octave or Scilab 00:43 yep 00:43 I'm looking into Scilab. 01:18 http://shiriah.net/ayreon/Disc%201/07-Ayreon%20-%20Ride%20The%20Comet.mp3 01:18 Find your way home, little extremophiles 01:18 Find your way home, donors of life 01:18 You're on your own, little extremophiles 01:18 You're on your own, cleaving the skies 01:18 [Chorus]: 01:18 Carry out our dangerous task 01:18 Sail uncharted spheres 01:18 Live out our dreams, ride the comet 01:18 Journey on the Migrator trail 01:18 Cross the new frontiers 01:18 Pass on our genes, ride the comet 01:18 relevant. 02:04 a panspermian hymn? 02:04 -!- kanzure: No such nick/channel 17:20 -!- There is no such nick kanzure 18:49 There goes Arthur C. Clarke. Day changed to 19 Mar 2008 06:58 -!- fenn [n=pz@adsl-75-60-175-89.dsl.bltnin.sbcglobal.net] has quit [Read error: 110 (Connection timed out)] 16:11 -!- fenn [n=pz@adsl-76-251-87-108.dsl.bltnin.sbcglobal.net] has joined #hplusroadmap 16:11 [Users #hplusroadmap] 16:11 [ Enki-2] [ epitron] [ fenn] [ kanzure] 16:11 -!- Irssi: #hplusroadmap: Total of 4 nicks [0 ops, 0 halfops, 0 voices, 4 normal] 16:11 -!- Channel #hplusroadmap created Sat Mar 22 15:44:12 2008 16:12 -!- Irssi: Join to #hplusroadmap was synced in 28 secs 17:04 < kanzure> Hey fenn. 17:04 < kanzure> You get my email? 17:06 < fenn> yep 17:06 < kanzure> I don't think it's much of a problem 17:06 < kanzure> that it's a Hamiltonian cycle 17:06 < kanzure> but it just goes to show this will require lots of number crunching 17:07 < fenn> i looked it up and it didnt mean much to me 17:07 < kanzure> Oh, well, it's Traveling Salesman problem. 17:07 < kanzure> you have a guy that wants to go travel a path and wants to optimize his route 17:07 < kanzure> and hit up a few clients or whatever 17:07 < kanzure> there's an optimal path that gets all of the points 17:07 < kanzure> but this can't be solved in polynomial deterministic time or something 17:08 < kanzure> http://en.wikipedia.org/wiki/Traveling_Salesman 17:08 < kanzure> NP-hard ;) 17:08 < fenn> i hope the database isnt complex enough that it will take a computer very long 17:09 < kanzure> this depends on what we want to do 17:09 < kanzure> do we want to (1) hand-code all connections between packages? 17:10 < kanzure> or (2) let the computer figure out possible connections between packages? We can program in "matter/energy converters" - since all mechanical energy can become chemical energy, etc. 17:10 < kanzure> #1 means we will probably know it all so that we will know if something seems stupid 17:10 < kanzure> but by #1 we may miss stuff 17:10 < kanzure> by #2 we get lots of bloat, longer search time, and maybe not a 'reduced' graph (although it will give us the general idea) 17:11 < fenn> #2 seems very difficult 17:11 < kanzure> maybe not -- just code in modules for each type of variable to interact with each other 17:11 < fenn> not impossible, but crikey, it's like the holy grail of automated engineering 17:11 < kanzure> just like in C++ where you have to define how the + operator interacts with other classes for a certain object 17:11 < kanzure> haha 17:11 < kanzure> yes, I suppose that's true. 17:12 < fenn> now i've had people say stuff like that to me all the time, and it doesnt necessarily mean anything 17:12 < kanzure> the holy grail stuff? 17:12 < fenn> after all we have gigahertz computers with gigabytes of ram and terabytes of storage 17:12 < kanzure> that's true 17:12 < kanzure> and not all of it has to be 'simulated' with physics and so on 17:12 < kanzure> it's really just testing out how to put legos together 17:12 < fenn> and still using the same software from the 1980's for most CAD/CAE stuff 17:13 < kanzure> it's like having a giant bucket of legos 17:13 < kanzure> but these legos don't all connect 17:13 < kanzure> hrm, the lego analogy doesn't work I think 17:13 < kanzure> since they are not functional-legos 17:13 < fenn> its like trying to use your SCSI tape reader to play video on the VCR 17:13 < kanzure> hah 17:14 < kanzure> I don't think so. What happened to the dependency-loop ideas and so on ? 17:14 < fenn> and then record computer backup data onto a video tape 17:15 < fenn> well the interesting thing about a dependencey loop is it's sort of a mini exponential growth without self replication 17:16 < kanzure> Oh, Wikipedia is correcting me. It's not a Hamiltonian cycle. Just a closed path. 17:16 < kanzure> hm? 17:16 < fenn> "i'm fat because i eat so much, i eat because i'm depressed, i'm depressed because i'm fat. it's a vicious cycle that took years to perfect." -garfield the cat 17:16 < kanzure> closed cycle, actually (cycle means path, except the path ends/begins at same vertice) 17:16 < kanzure> I remember that one :) 17:18 < kanzure> so, the only reason why we need skdb and the solver is really just because we can't come up with a 2-module system that makes itself, really. A sand-processor that makes a glass-processor, and the glass-processor makes the sand processor (but this doesn't make sense). 17:18 < kanzure> and as you fill out the requirements, it is kind of like exponential growth 17:18 < kanzure> since the more you need, the more components you have to add 17:18 < kanzure> that's why I like starting with a part already -- an arm, say 17:18 < kanzure> and then we add components to the arm until it is able to self-replicate 17:18 < fenn> also i was thinking about symbiosis, where you have not one single point through which the loop passes, but more like several interconnected loops in parallel 17:19 < kanzure> like the character '8' ? 17:19 < fenn> more like a knitted headband 17:19 < kanzure> like oooooooooo 17:19 < kanzure> hm 17:19 < kanzure> xxxxxxxxxx ? 17:19 < kanzure> but those cross 17:22 < fenn> http://fennetic.net/pub/irc/directed_cyclic.png 17:23 < kanzure> that's hard to understand 17:23 < kanzure> could you color the cycles? 17:23 < fenn> a b c are the replicators 17:23 < kanzure> the entire design? 17:24 < fenn> alone they couldnt replicate 17:24 < kanzure> why is there two sets of a, b, c ? Showing the replication? 17:24 < fenn> just to connect the top and bottom 17:25 < kanzure> I'm sort of thinking there should be more symmetry here 17:27 < kanzure> The requirements: 17:27 < kanzure> A+B = C 17:27 < kanzure> A + C = B 17:27 < kanzure> B + C = A 17:27 < fenn> ok reload. the green are new copies of the red ABC, but their position in the graph is identical 17:28 < fenn> the purple is a leaf node, unrelated to replication (but requires b and c) 17:28 < kanzure> B + C looks required for replication 17:28 < kanzure> but not the purple one, right 17:29 < kanzure> aren't these supposed to be directed? 17:29 < fenn> yesmeh 17:30 < fenn> dia is a pain to use 17:30 < kanzure> this is dia? 17:30 < kanzure> haha 17:30 < fenn> no, that's kolourpaint 17:30 < kanzure> hehe 17:32 < fenn> i added numbers 17:32 < fenn> you might say that node 6 or node 7 is "the" replicator 17:33 < kanzure> What about 5? 17:35 < fenn> hmm how do you show in graph theory that you must approach a vertex from two edges at the same time? 17:35 < kanzure> that's interesting 17:35 < kanzure> draw arrows with matching colors? 17:35 < kanzure> lemme ask #math 17:36 < fenn> 7 should depend on 5 and 6 being present 17:37 < fenn> i added 9, which depends on 5 and 8 17:38 < fenn> i'm sure this could be simplified 17:38 < kanzure> a "required vertice" is interesting 17:38 < kanzure> required vertices 17:38 < kanzure> it's a functional graph, in a sense 17:38 < kanzure> so you just have to say that "vertice X does not exist without vertice Y" 17:38 < fenn> well, it might exist, but you can't get to it 17:39 < fenn> like, i might have a lathe and bar-stock, but no cutting bits 17:41 < fenn> ok poor example 17:43 < kanzure> maybe you need two graphs 17:44 < kanzure> one for the dependencies and then one for the vertice dependencies 17:44 < kanzure> or something like that 17:44 < fenn> earlier i was talking about different graphs for streams and packages (run-time and build-time, respectively) 17:45 < fenn> but you can also have a package that makes packages 17:45 < fenn> or streams that spontaneously degrade into something 17:49 < kanzure> hm 17:50 < kanzure> so then you have to figure out a way to specify the extra requirements for the graph then 18:02 < fenn> gosh math is really useless sometimes 18:02 < kanzure> reading up on Wikipedia? 18:03 < fenn> "the sum of all the vertex-degree is equal to twice the number of edges" 18:04 < kanzure> the vertex-degree is something like the number of connections 18:04 < kanzure> to a certain vertice 18:04 < fenn> yep 18:05 < fenn> since a line connecting the vertices has two end points, there are 2*lines number of endpoints 18:05 < fenn> big deal 18:06 < kanzure> haha 18:06 < kanzure> =) 18:20 < kanzure> fenn, what exactly are you looking into at the moment? 18:28 < fenn> strange loops and graph theory 18:31 < kanzure> mind pasting anything you find interesting ? 18:35 < fenn> In 1974 the Consumer Product Safety Commission recalled 80,000 of its own lapel buttons promoting toy safety. The buttons had paint with too much lead, sharp edges, and clips that could be broken off and swallowed. 18:37 < kanzure> huh? for graph theory? 18:37 < kanzure> hehe 18:38 < fenn> apparently irony is self-referential 18:41 < kanzure> didn't know? 19:09 < fenn> this looks about right: http://www.graphviz.org/Gallery/directed/fsm.html 19:09 < kanzure> neat. 19:11 < fenn> i really want something more interactive though 19:11 < kanzure> to what extent? 19:13 < fenn> to be able to poke at the graph and rearrange it.. as if the vertices were connected by springs 19:14 < kanzure> yes, I have wanted an interactive interface to graphviz for a while now 19:14 < kanzure> there's "freemind" out there on SF, it's sort of node-based but only for concept-maps, and it's all in java 19:14 < kanzure> so it's kind of disgusting. 19:16 < fenn> someone i know wrote a java app that did springy networks (FEA - he's an engineer) 19:16 < fenn> this should be pretty easy to write from scratch in whatever language 19:18 < kanzure> finite element analysis? 19:18 < kanzure> hm 19:19 < kanzure> but hacking graphviz is not so easy, last I checked 19:19 < kanzure> they had lots of literature that you have to sort through, you need to have connections to get the papers 19:19 < kanzure> maybe I'm wrong, it's worth investigating 19:20 < fenn> i need to write a similar application for emc2 actually 19:20 < fenn> graphical hal configurator 19:25 < fenn> this is neat http://www.kylescholz.com/projects/speaking/tae2006/music.fm/ 19:28 < kanzure> ooh 19:29 < kanzure> it looks pretty close 19:30 < fenn> the neat part is the search results are actually right 19:33 < fenn> you can drag the names around if you select them 19:38 < fenn> The main plastic we are using is polylactic acid. Anyone can make this by fermenting starch, so if you have a few tens of square meters of land to grow a starch crop you not only have a self-reproducing machine, you have a self-reproducing source of feedstock independent of the petrochemical industry. 19:41 < kanzure> reprap? 19:47 < fenn> yes 21:29 < kanzure> fenn: so what should the next steps be? any thoughts? 21:29 < kanzure> I think we need to formalize your graph-definition and make a program to generate expandable graphviz files that follow the format 21:34 < fenn> file format and parser 21:35 < fenn> needs to be text based, human and machine readable 21:35 < fenn> i'm thinking about yaml or slip-xml 21:41 < kanzure> maybe we can write the parser in flex/bison to convert to graphviz format for the data visualization 21:42 < kanzure> erm, I think I am cutting a few steps in the toolchain 21:43 < Enki-2> just use flex alone 21:43 < Enki-2> bison isn't needed if you know flex 21:46 < kanzure> right, but the file format would not necessarily be what we want to see a graph of 21:46 < Enki-2> i mean, flex can do translation of formats on its own, sans bison 21:47 < Enki-2> i've done complete compilers in flex 21:47 < kanzure> woah, you're not fenn 21:47 < kanzure> just realized this :) 21:47 < Enki-2> :P 22:18 < fenn> i dont know flex or bison, so i'm going to do a naive straightforward implementation :) 22:19 < kanzure> bison is extremely easy 22:19 < kanzure> http://dinosaur.compilertools.net/bison/bison_6.html 22:19 < kanzure> http://www.gnu.org/software/bison/ 22:19 < kanzure> http://cs.wwc.edu/~aabyan/464/Book/ 22:20 < kanzure> This one is really good - http://www.garshol.priv.no/download/text/bnf.html 22:20 < kanzure> short and to the point 22:20 < fenn> OT: just in the shower, thinking about cost of inventory space. A package's inventory complexity has to do with the number of unique items that need to be stored, their overall shape, and the size. Each physical object can be given metrics based on its bounding shape (i.e. cube cylinder sheet rope sphere amorphous) 22:20 < kanzure> inventory could be external for simple designs 22:21 < fenn> this cost comes out of human attention (our most precious commodity) unless there is an automated inventory system 22:21 < kanzure> Uh oh. News on http://utexas.edu/ (where I'm going next year) -- `BB&T gives $2 million for Ayn Rand researchPhilosophy Department to create new chair for study of novelist’s philosophy of Objectivism.` Oh boy. 22:22 < fenn> i take it that's bad? 22:22 < kanzure> Objectivisim is a word-hijack 22:22 < kanzure> it means lots of politics and capitalism 22:22 < kanzure> generally not good philosophy 22:22 < fenn> what does politics and capitalism have to do with philosophy? 22:22 < kanzure> Ayn Rand :( 22:23 < fenn> oh, objectivism is at least slightly different from objectivity 22:23 < kanzure> http://en.wikipedia.org/wiki/Objectivism_(Ayn_Rand) 22:23 < fenn> in terms of word structure 22:23 < kanzure> ` ... that the proper moral purpose of one's life is the pursuit of one's own happiness or "rational self-interest"; that the only social system consistent with this morality is full respect for individual rights, embodied in pure, consensual laissez-faire capitalism; ` 22:23 < kanzure> the Proper Moral Response 22:23 < kanzure> *Purpose 22:24 < fenn> if they can implement capitalism through a university philosophy department, more power to them :\ 22:24 < kanzure> heh 22:24 < fenn> as i see it, capitalism was demolished by corporations and the patent system 22:25 < fenn> or at least the prospect of it becoming a reality 22:26 < kanzure> fenn: in general, I am not communist, I am not capitalist, I am not anything like that at all -- I think that these old 'socioeconomic philosophies' need to be dismantled and replaced with a more intense, rigorous mathematical framework where we don't have bullshit "Social Contracts" etc. 22:26 < fenn> how can you have patents and call it capitalism? nonsense 22:26 < kanzure> these old ideas are *ancient* 22:26 < kanzure> technological development is accelerating 22:26 < kanzure> perhaps when we had a constant rate of development, capitalism/communism/Objectivism/Christianity/Republicanism worked, but the world just isn't the same 22:27 < kanzure> these are all old, stale ideas with a massive number of people still promoting them 22:27 < kanzure> and while there are good ideas within, 22:27 < kanzure> the culture is still largely one of identification and not of ... importing the good ideas ;) 22:27 < fenn> indeed 22:27 < fenn> let it be known i'm not a sports fan 22:27 < kanzure> hah =) 22:28 < fenn> got any opinion on this? http://www.debian.org/social_contract 22:29 < kanzure> oh-my-god I have access to scientific databases 22:29 < fenn> you mean journals? 22:30 < kanzure> yes 22:30 * kanzure hugs his UT login 22:30 < fenn> gratz. got ieee spectrum access? 22:30 < kanzure> fenn: Not anything off the top of my head. 22:30 < kanzure> fenn: yes 22:30 < fenn> sweet 22:30 < kanzure> fenn: I just opened up an article on pyroelectricity 22:31 < kanzure> Piezoelectricity and Pyroelectricity in Polymers 22:31 < kanzure> First legally accessed, paywalled scientific paper 22:31 < kanzure> this should go in my baby book or something 22:32 < fenn> today is something special.. i forget 22:32 < kanzure> there was chocolate today 22:32 < kanzure> so I'd venture a guess of easter 22:32 < fenn> aha 22:33 < fenn> right, we had jehova's witnesses 22:33 < fenn> (not for dinner) 22:33 < fenn> international bunny day - day of the replicator! 22:34 -!- fenn changed the topic of #hplusroadmap to: happy bunny day - day of the replicator! 22:36 < kanzure> epitron: I'll be adding a 50 MB zip file to the server later tonight. 22:38 < kanzure> woot, nature! 22:39 < kanzure> American Physical Society ! 22:54 < kanzure> fenn: I want to write a script to browse through Google Scholar. Any thoughts? I want the input to be a list of papers, and an output to be a nice package of PDFs. 22:55 < kanzure> Google does not provide a direct interface to a PDF for a title you enter: some of the links it returns may or may not be to a PDF. 22:55 < kanzure> so in the cases where this does not happen, you can go through whatever journal website there is and click on a link that says "PDF" 22:59 < fenn> if (grep ''): wget $1; else ... ? 22:59 < fenn> note that was pseudo-code 23:02 < kanzure> nah, 23:02 < kanzure> the url doesn't necessarily have .PDF in it 23:02 < kanzure> but you can scan for PDF next to the link 23:02 < kanzure> there's a way to process Google pages, they have a nice systematic format with div tags and so on 23:02 < kanzure> I need to go look at the HTTP libraries to see if there's some way to log in with cookies to a service 23:03 < kanzure> WWW:Mechanizer might be what I want. 23:03 < fenn> you need to get those cookies though, which might not be straightforward with a non-browser interface 23:04 < kanzure> ooh 23:04 < kanzure> WWW::Mechanize does something interesting though 23:04 < fenn> especially since scientific publishers arent known for their browser compatibility 23:04 < kanzure> use WWW::Mechanize; 23:04 < kanzure> use HTML::TokeParser; 23:04 < kanzure> my $agent = WWW::Mechanize->new(); 23:04 < kanzure> $agent->get("http://www.radiotimes.beeb.com/"); 23:04 < kanzure> $agent->follow("My Diary"); 23:04 < kanzure> $agent->field("email", $email); 23:04 < kanzure> $agent->click(); 23:04 < kanzure> so in other words it's like doing the physical actions 23:04 < kanzure> except there's no browser UI 23:10 < kanzure> Okay, first part done. Neat. 23:19 < kanzure> Hm, the rest is testing my knowledge of regexps. I want to extract the first link after [PDF]. 23:20 < fenn> $agent->follow("[PDF]") right? 23:21 < kanzure> if the name of the link was [PDF], yes 23:21 < kanzure> but PDF is not part of the link name 23:21 < kanzure> it's just 'near' it 23:21 < fenn> unless you want to do something like ~= \[PDF\].* 23:21 < kanzure> so I'm thinking of splitting up the string 23:21 < kanzure> exploding it just before [PDF] 23:21 < kanzure> and then after the next 23:21 < fenn> oh, 'follow' means click on the link 23:21 < kanzure> right 23:22 < kanzure> will your regexp work? 23:22 < kanzure> $1 would be my url, right? 23:22 < fenn> ~= '\[PDF\].*(.*)' $1 is the url $2 is the link text 23:23 < kanzure>

[PDF] Efficient exact stochastic simulation of chemical systems with many species and many channels - Species and Many Channels Michael A. Gibson* and Jehoshua ... 23:23 < kanzure> anyway, it gets nasty 23:23 < kanzure> so you can't assume it's precisely 23:24 < kanzure> hm, it's kind of smart with the "on-mouse-down" function 23:24 < kanzure> but that's no problem - that's what the follow() idea is, no? :) 23:24 < kanzure> or the "click" hehe 23:25 < fenn> it gives the url in the 23:25 < fenn> you want to ignore that? or what? 23:26 < kanzure> 23:26 < kanzure> erm 23:26 < kanzure> 23:26 < kanzure> where we want the xxx, and then after the yyy"> we can get the title 23:26 < fenn> ~= '\[PDF\].* do I have to say $var ~= ... ? 23:27 -!- Enki-2 [n=weechat@c-71-234-190-248.hsd1.ct.comcast.net] has quit [Read error: 104 (Connection reset by peer)] 23:27 < fenn> $var should be the whole document i think 23:27 < kanzure> ah, right 23:27 < kanzure> okay 23:27 < fenn> ~= loads all that crap into special variables 23:27 < fenn> did i mention i hate perl? :) 23:28 < kanzure> hehe 23:29 < fenn> oh joy they use single quotes in the url 23:31 < kanzure> hm? 23:34 -!- Enki-2 [n=weechat@c-71-234-190-248.hsd1.ct.comcast.net] has joined #hplusroadmap 23:35 < kanzure> $searchresults ~= '\[PDF\].*(.*)'; 23:35 < kanzure> parse error 23:36 < kanzure> hm 23:41 < fenn> ah i knew it would be something stupid 23:41 < fenn> =~ not ~= 23:43 < fenn> http://fennetic.net/pub/irc/test.pl 23:49 < kanzure> hahaha 23:49 < kanzure> this is awesome 23:50 < fenn> i wish they would have used a different operator 23:51 < kanzure> epitron, fenn -- Cosmic coincidences ;) http://heybryan.org/shots/2008-03-23_autoscholar.png 23:52 < fenn> who is eugen? 23:52 < kanzure> I've never talked about him? 23:52 < kanzure> http://eugen.leitl.org/ see for yourself 23:52 < kanzure> he's also mentioned here - http://heybryan.org/stats/2008-03-23-extropy-chat.html 23:52 < kanzure> he's also mentioned on http://heybryan.org/mediawiki/index.php/Roadmap 23:53 < kanzure> http://heybryan.org/mediawiki/index.php/Getting_up_to_speed too (I think) 23:53 < fenn> more like synthetic serendipity ;) 23:54 < kanzure> Eugen Leitl is sort of a polymath. 23:54 < kanzure> Good guy to know. 23:54 < fenn> many a time i've come across chat logs of me asking the same question i was googling for 23:54 < kanzure> I was talking on a mailing list one day about how you should just "do the transhumanism thing and be done with it, ignore politics - once you transcend, they can't stop you." He replied within 30 seconds and said "Yes, but now that they know what you are up to, they could stop you *now*." He and I have known each other ever since. 23:54 < kanzure> same here 23:55 < fenn> well, what do you do about that? 23:55 < kanzure> hide 23:55 < kanzure> but I have been a moron and am already on the net 23:55 < kanzure> in fact, there's a very significant portion of content generated by me on the net 23:55 < kanzure> heh 23:55 < fenn> hm.. secrecy doesn't work 23:55 < kanzure> right 23:56 < kanzure> it seems that the best way to go about it is to just be open as much as possible 23:56 < fenn> better to make yourself indispensable 23:56 < fenn> and dont get involved in shady criminal activities as a hobby 23:56 < fenn> because then there's a reasonable excuse for your mysterious disappearance 23:56 < kanzure> secretly, I was a script kiddie when I was 11 23:57 < fenn> there werent any scripts when i was 11 :( 23:57 < kanzure> old fart 23:57 < fenn> i'm not that old (25) 23:57 < kanzure> okay, so the script is mostly working 23:57 < kanzure> it gets the wrong URL though 23:58 < kanzure> it gets the stuff after a href=" but then doesn't stop at the first " 23:58 < kanzure> $searchresults =~ '\[PDF\].*(.*)'; 23:58 < kanzure> so I need those quotation marks to match exactly 23:58 < fenn> oh, probably greedy vs non-greedy matching 23:59 < kanzure> should I add /g to the end? 23:59 < kanzure> or /ig ? 23:59 < fenn> g is global? 23:59 < kanzure> oh 23:59 < kanzure> hrm 23:59 < fenn> (.*?) i think 23:59 < kanzure> oh 23:59 < kanzure> no, 23:59 < kanzure> =~ m/stuff/; Day changed to 24 Mar 2008 00:00 < fenn> not following you 00:00 < kanzure> nope, that didn't work 00:01 < kanzure> you're right 00:03 < kanzure> wget --output-file is not useful 00:04 < kanzure> it just routes the wget output to that file 00:04 < kanzure> not the downloaded file 00:05 < fenn> --output-document 00:06 < kanzure> Alright, neat. Now I need to figure out a way to fetch a document if there is no PDF available. 00:08 < kanzure> So, in the case where there is not a PDF available, there's usually a link that says "All $number versions >>" and then on the following page there's a list of possible places to get it, where we can do the PDF-search thing again, *or* click on a link that says " Get this article" which takes the user to a page where there's a Javascript "Go" button to get the fulltext from various databases (obviously, select the first one). But 00:09 < fenn> But (EOL) 00:11 < kanzure> eh 00:11 < kanzure> at this point it takes you to Nature, or some other database, and on that page the only thing you can hope to do is search for the link with a name including "PDF" and just get that URL. 00:11 < kanzure> this seems pretty simple in comparison 00:12 < kanzure> if javascript execution works 00:12 < fenn> does WWW::Mechanise run javascript? 00:12 < kanzure> asking #perl atm 00:13 < kanzure> ah, shit, no 00:13 < kanzure> http://search.cpan.org/dist/WWW-Mechanize/lib/WWW/Mechanize/FAQ.pod 00:14 < fenn> I don't use JavaScript myself, so I don't have an itch to scratch. <- wtf does he use WWW::Mechanize for then? 00:16 < kanzure> hahah 00:16 < fenn> read 'So what can I do?' - is the url in the html source somewhere? (even if its not standard) 00:17 < kanzure> yes, I figured I'd have to do that anyway 00:17 < kanzure> and it turns out it's pretty simple in this case 00:17 < kanzure> it's just a form call hehe 00:17 < fenn> is it the same from publisher to publisher? 00:18 < kanzure> http://p9003-www.lib.utexas.edu.ezproxy.lib.utexas.edu/sfx_local/cgi/core/sfxresolver.cgi?basic1&tmp_ctx_obj_id=1&service_id=110974981369254&request_id=4760844 <-- will redirect me to the first "full text" available option ... but not directly to the PDF. And service_id and request_id have to be extracted from the page (they are hidden form fields) 00:23 < fenn> i think 'socialized eng kdb' should be changed, because 'socialized' actually means something rather different 00:23 < fenn> socio-economic system in which property and the distribution of wealth are subject to control by the community 00:24 < fenn> but you mean, distributed development and distributed knowledge right? 00:24 < kanzure> right 00:24 < kanzure> socialized knowledge as in, "social knowledge" 00:24 < kanzure> social as in "community-known" 00:24 < kanzure> but I'm good for a change 00:27 < fenn> societal engineering knowledge db? 00:27 < kanzure> sure 00:29 < kanzure> hm, it turns out I need only search for either (1) PDFs or "Get this article" links on the Google Scholar search results page 00:31 < kanzure> my local "should be a programmer but isn't yet" buddy just sent me this: http://www.osnews.com/images/comics/wtfm.jpg 00:33 < kanzure> Bill: so like apt-get but for science journals 00:33 < kanzure> smart guy too 00:34 < fenn> heh. good for ui testing at least 00:35 < kanzure> Java.swing :( 00:35 < kanzure> I will never again use java swing 00:35 < kanzure> ever 00:36 < fenn> but what about the cool sketch-on-a-napkin skin? 00:36 < kanzure> hm? 00:36 < kanzure> xkcd ref? 00:36 < fenn> http://napkinlaf.sourceforge.net/ 00:37 < fenn> good for a laf at least :) 00:38 < kanzure> har har 01:00 < kanzure> fenn: does .* have to match anything? 01:00 < fenn> please restate the question in the form of an answer 01:00 < kanzure> hm 01:00 < fenn> .* means any character, any number of times 01:01 < kanzure> see, I want to now match for .*PDF.* 01:01 < kanzure> so that I get the URL to links that have 'PDF' in their name 01:01 < fenn> might belong to a url that is later in the file 01:01 < kanzure> hrm 01:01 < kanzure> okay 01:02 < kanzure> so I guess I can do (.*?) still 01:02 < fenn> so i think the second .* should be .*? 01:02 < kanzure> .*PDF.*? 01:02 < fenn> yes 01:02 < kanzure> I don't really need the second (.*?) 01:02 < kanzure> .*PDF.*? 01:02 < fenn> () just changes what gets stored in special variables 01:02 < kanzure> right 01:02 < kanzure> so I just need the href variable 01:03 < fenn> (in this example) 01:03 < fenn> uh, it might be a bit late to mention that there are probably lots of xml or html parsing tools 01:03 < kanzure> I was looking into them 01:04 < kanzure> the perl community is obsessed with LinkExtor 01:04 < kanzure> but it does not return the link name 01:04 < kanzure> and for some reason I didn't think to go more generalized with XML 01:05 < fenn> i think you need ? on all the .*'s 01:08 < kanzure> if () { /* stuff * / } else if ($searchresults =~ /Get this article/) { } <--- parse error at } else if () { ........ wtf? 01:10 < fenn> / get this article/ probably has a / in it somewhere 01:10 < fenn> you can use other characters to bound the regexp 01:11 < kanzure> ah, it was something else entirely 01:14 < kanzure> how ridiculous 01:14 < fenn> how Perl 01:14 < kanzure> my password, I have to pass it to wget with --proxy-pass 01:14 < kanzure> and it has a bang in it 01:14 < kanzure> for some reason 01:14 < kanzure> and this messes up bash 01:14 < kanzure> apparently I can't escape it 01:14 < kanzure> with \! 01:14 < fenn> !! i think 01:14 < kanzure> ugh 01:15 < fenn> no that's no it 01:15 < kanzure> hm 01:15 < kanzure> echo ! works though 01:15 < kanzure> oh 01:15 < fenn> are you doing double quotes? try single 01:16 < kanzure> ah 01:17 < kanzure> wget still says it is ambiguous though 01:17 < kanzure> wget --proxy-pass=hi! 01:17 < fenn> wget --proxy-pass='hi!' 01:17 < kanzure> right 01:17 < kanzure> even that 01:18 < fenn> it's --proxy-password 01:18 < kanzure> bleh 01:19 < kanzure> oh 01:19 < kanzure> I just realized something 01:19 < kanzure> Google pulls up numerous documents that are related to the title 01:19 < kanzure> so you have to do *exact* title searches 01:19 < kanzure> as in, intitle:"name" 01:19 < kanzure> where "name" is the name of the article 01:19 < kanzure> else you will get PDFs from other things 01:41 < kanzure> fenn: 01:41 < kanzure> http://search.cpan.org/dist/WWW-Mechanize/lib/WWW/Mechanize.pm#$mech->redirect_ok() 01:42 < kanzure> you think there's a way to figure out the 'current URL' that it is sitting at ? 01:42 < kanzure> ah, nevermind 01:57 < kanzure> fenn: Here's my regexp, II$r3 =~ '.*?PDF.*?'; 01:57 < kanzure> and here's what I need to match: Download PDF 01:57 < kanzure> for some reason, I am only getting back "/nature/" ... 02:16 < fenn> dunno 02:17 < fenn> $1 = "/nature/" ? 02:19 < fenn> also print $r3 and make sure it's what you expect 02:20 < kanzure> of course 02:20 < kanzure> #perl has convinced me to switch to LinkExtractor 02:20 < kanzure> http://search.cpan.org/~podmaster/HTML-LinkExtractor-0.13/LinkExtractor.pm 02:20 < kanzure> now I'm figuring out how to access the _TEXT variable 02:21 < kanzure> foreach my $x ($LX->links) { $x->_TEXT; } or something ... 02:22 < fenn> huh. linkextractor just dumps you back in the same situation 02:23 < fenn> there's no variable containing 'I am a LINK!!!' 02:23 < kanzure> _TEXT is it. 02:23 < kanzure> you just have to strip out the HTML. 02:24 < fenn> then why bother with linkextractor :) 02:24 < kanzure> blah, #perl is telling me to go back to the newbie tutorials 02:24 < kanzure> I just want to get the _TEXT variable :( 02:24 < kanzure> I thought it might be $x{_TEXT} but that wasn't it either 02:25 < kanzure> it's a hashref, I know that much 02:26 < fenn> $LX->links{_TEXT}? 02:26 < fenn> for my $Link( @{ $LX->links } ) { 02:26 < fenn> ## new modules are linked by /author/NAME/Dist 02:26 < fenn> if( $$Link{href}=~ m{^\/author\/\w+} ) { 02:26 < fenn> print $$Link{_TEXT}."\n"; 02:26 < fenn> change author to PDF 02:27 < kanzure> bleh 02:27 < kanzure> well, not quite 02:27 < kanzure> $$Link{href} is assuming there's a pdf in the URL 02:27 < fenn> i forget how to get an element from an array 02:28 < fenn> oh, change href to _TEXT then 02:28 < fenn> and then print the href :) 02:29 < kanzure> yep 02:29 < kanzure> hm, there's a few other parts of the code I have to fix 02:29 < kanzure> because nature.com is a bitch and provides relative links 02:29 < kanzure> but that's simple enough 02:52 < kanzure> fenn: It works. 02:52 < kanzure> fenn: Two things to make note of. (1) If Google does not provide a link to a PDF, then the name of the PDF has to be the search string you used. This gets kind of nasty. Maybe there's a perl-PDF library that I can use that can extract the name of the paper. That would be really useful. 02:53 < kanzure> (2) Doesn't yet do a list of inputs, but this is easy. 03:02 < fenn> know anything about webs of trust? 03:02 < fenn> is it possible to 'import' a web that's already existing? for example robots.net or advogato 03:04 < kanzure> don't know what you are talking about 03:04 < fenn> if software doesn't work it's no big deal 03:04 < fenn> but say you have a big dependency tree leading up to some hardware project, that costs money and time and effort 03:05 < fenn> so you'd like to know in advance the trustworthiness of the author (in the field of the invention) 03:06 < fenn> there exist already, open systems to gauge the amount of trust an individual has acquired 03:06 < kanzure> http://heybryan.org/projects/autoscholar/ 03:07 < fenn> i'd like to be able to jump-start the system by finding the small-world center nodes (people) and asking them to join, but that takes a lot of work 03:07 < fenn> and they wont necessarily contribute to our project and become center nodes 03:08 < fenn> so it'd be better if we could use the whole web of trust from elsewhere that already exists 03:08 < fenn> rather than trying to build community and trust all over again for every web site 03:09 < fenn> i get annoyed about having to make new logins all the time 03:09 < kanzure> have you heard of OpenID? 03:10 < fenn> nope 03:11 * fenn looks 03:14 < kanzure> oh shit 03:14 < kanzure> yeah, that's a good thing to know about 03:14 < kanzure> http://claimid.com/ is what I use 03:14 < kanzure> but there are others out there 03:14 < kanzure> probably a few that are more legit 03:29 < fenn> lawdy.. forget i said anything about trust networks mmkay 04:16 < fenn__> is it ironic that roderick long is an "ayn rand" scholar? http://en.wikipedia.org/wiki/Social_contract#Implicit_social_contract_theory_presupposes_its_conclusion 18:02 -!- Irssi: Starting query in freenode2 with kanzure 18:02 Hey 18:02 Thanks for that email, it reminded me to get into gear 18:02 let's make a database/repository for these projects, for "blackboxing" whole projects 18:02 we can begin by splitting up all sorts of projects into their units 18:03 and each package will include (1) some specs on what it does, (2) some variables on what can be changed, (3) an output in some "fabricator-friendly" format, and (4) a simulator to plug in to an overall program when you want to simulate your project 18:03 just like in APT or CPAN, you cannot specify functionality, which is unfortunate, but something that we just have to get over 18:03 also, I will be gone in 15 minutes, so think fast ;) 18:16 uh, hi 18:16 hello 18:17 what do you mean 'blackboxing'? like depends on other packages (rather than generic functionality) 18:17 based on my experience with python and hasattr(), specifying generic functionality is way cooler 18:28 what? 18:28 you'll have to tell me about that later 18:28 but basically you can't search for functionality with APT or even Google 18:28 you have to search for tags 18:28 or be able to talk with somebody who can give you a package name 18:28 bbl 18:54 Hey 18:56 in python if you pass an object to a function, it will run it, no questions asked. you can examine the object either by type (i.e. package name) or by its attributes. by attributes is much more flexible and powerful 18:59 if it's the wrong type of object, you'll get someting like 'error: "car" has no attribute "horsepower"' 19:00 when it tries to access car.horsepower 19:13 yes, but you can't encode the actual functionality spec 19:13 bbl 22:30 What I was basically saying is that APT works out not because it has some super-awesome way of finding software, but rather it's a very well known database of software. We just need the same for all sorts of other projects, with ways of managing different fileformats and specifying what a module contains, but not necessarily its entire specification, just enough to allow somebody who has already found it to find it again. 23:01 This seems to be the same thing as the intelligence problem. You can't really select for it, but if you see it appear, you know to run with it. I think the way Andy copes with this is via artificial biochemistries ... he has added fake nucleic acids to what he calls "uncoli" (his variant of ecoli), which boosts the bacteria from 1/2000 chance of mutation to something much higher, like 10% chance of mutation, which allows very rapid communi Day changed to 21 Mar 2008 15:29 hey 15:29 I found another "one of us" sort of guys 15:29 'epitron' on freenode 15:30 he has a favorable condition where he has this giant gaping air pocket in his skull, suitable for a brain implant 16:07 um, how did he figure that out? 16:07 MRI. 16:07 developed during fetal stage 16:08 http://chris.ill-logic.com/files/my_brain.zip 16:08 heh, "Chris Gahan - PROBLEM SOLVER" mentioned on his web page 16:08 my mom knew a girl who was apparently completely normal except that her brain was just a thin smooth balloon 16:09 interesting 16:09 there was an article about a man like that 16:09 he recently died, they investigated his brain, it was also a thin sliver 16:09 http://chris.ill-logic.com/wiki/CurriculumVitae 16:10 he happens to do engineering psych 16:10 so he's on the "self-modification" level 16:10 whereas I'd call you more self-replication ;) 16:12 sure 16:12 i have no tattoos 16:12 we need to figure out a way to make a "Getting up to speed" page 16:13 or at least a way to list the guys "like us". I'd list you, myself, BioMors, Superkuh, enki-2, epitron, and a few others that I may be forgetting 16:13 Maybe TheWOLPRO or endos127, but they seem to be too wrapped up in school to care 16:13 surely there are many others, we are just the ones on freenode 16:14 right 16:14 but I bet there are many others who would be interested in contributing 16:14 if they had, say, a map 16:14 1) define the problem. 2) solve the problem 16:14 http://heybryan.org/mediawiki/index.php/Getting_up_to_speed <-- I am going to dump some notes here. 16:16 i think solar concentrators can provide sufficient energy to do small scale foundry work 16:16 giant mirrors? 16:16 steel casting.. can't do that any other way except with huge amounts of electricity 16:17 arrays of small mirrors 16:17 Tesla used to use the earth for electricity and so on 16:17 it would be useful if we could do the same with giant asteroids 16:17 well, i dont know how to do that 16:17 neither do I. 16:17 giant solar cell arrays may have to be the solution 16:17 many square kilometers 16:18 not a big deal, IMO, if you have self-replication 16:18 oh, actually i'm wrong, you can melt steel with an oxygen concentrator 16:18 unless that needs steel foundry 16:18 oh, good 16:18 i was reading this and thinking about how to do it with a 'fresnel mirror' for both mirrors http://kmr.nada.kth.se/wiki/Main/PointFocus 16:19 *.kth.se is good 16:19 http://www.nada.kth.se/~asa/ <-- Anders Sandberg. 16:19 interesting 16:19 he used to run the Omega Point Society in Sweden. 16:19 the guy who wrote that paper is now head of the department(?) and doing research on "the semantic web" 16:20 Anders is a good guy to get on board, he does lots of free form fabrication stuff too 16:20 and used to do lots of TransVision conferences 16:20 saw him last week in Second Life. 17:25 hm 17:25 http://wiki.ill-logic.com/AugmentationNotAutomation 17:45 agreed, but why the title "Augmentation Not Automation"? 17:56 automation is subset of augmentation 18:13 huh. "learning depends on social interaction" is quite true but i had never noticed 19:00 yeah 19:00 which sucks 19:00 I would desperately like to prove otherwise 19:01 perhaps you could substitute randomness (weighted with a model of your interests of course) 19:02 but that's television 20:01 had to bring epitron up to speed http://heybryan.org/chats/epitron-2008-03-21.html 21:19 http://heybryan.org/shots/img_8032.jpg 22:27 hm, http://knome.org/ as a 23andme/decodeme clone 23:09 Found this hidden on my site -- http://heybryan.org/creating_communities.html how to make communities 23:13 yep that's why i mistakenly thought you were 22 23:13 hm? how's that? 23:14 scroll down to Throwaway identities 23:14 ooh 23:14 note the quotes 23:15 i dont like how a lot of those are essentially advocating spamming 23:15 :( 23:16 it's kind of weird, though, because it kind of worked for me: I published an email announcing my biohacking kit, and I had to spread it *myself* 23:16 I had to go push it on to the news websites 23:16 I had to push it to Make, to hackaday, etc., 23:16 the only way to not have to push is to go through AP and they do 'serious' stuff (supposedly) 23:16 sure 23:17 Hypermarketer website/tool- automated emails, rss feeds with "updates" on the outbreak, spam email from botnet, multiple fake accounts on all of the popular social networking or "web 2.0" sites, mediablitz, all within one hour, with continued postings throughout the internet to *take over* the smaller blogs if necessary. Goal: no trail to an original, single source. To really fool anybody who might be catching on, have multiple emails across the botnet submitted at the same moment with clocks in synch. 23:17 is that really something you want to be associated with? 23:17 not really 23:18 anyway it starts out good 23:18 I was just doing note taking, mind you 23:18 plus some brain storming ;) 23:18 a lot of people put basically zero effort into publicizing themselves though, understandably 23:19 oh, by the way 23:19 I have a question 23:19 nobody ever said 'in moderation' was easy 23:19 let's say that by some miracle I have figured out a way to hack DNA replication and can make ssDNA while inside a cell 23:19 and I know that the cell will degrade it immediately 23:19 however, 23:19 do you know of any 'molecular chaperones' that could bring it to the plasma membrane and push it through? 23:19 not DNA but there are lots of them for RNA 23:20 why would you want to make ss DNA? 23:20 ssDNA does logic gates 23:20 RNA logic gates are infeasible (apparently) 23:20 uh, i thought you needed dsDNA for your scheme 23:20 I have a plasma-membrane-bound SSBP 23:20 I thought it was ssDNA 23:21 http://heybryan.org/winfree.html 23:21 yeah, ssDNA 23:21 anyway, I have a SSBP that I can express on the outer surface 23:21 no, pretty sure the rna gets transcribed by regular RNA polymerase (which uses ds DNA) 23:21 which will bind to the ssDNA 23:22 oh 23:22 okay, so a dsDNA template is used for the switch itself 23:22 and then ssDNA for the activator for completing the promoter region or something 23:22 anyway, dsDNA then. 23:23 you want to truck ds DNA around? probably easier in a bacterial cell 23:23 they do that sort of thing for 'mating' 23:23 yepo 23:23 *yep 23:23 so that's what I am figuring. 23:24 I just need a dsDNA chaperone molecule to truck it to the membrane and then bind it to it (SSBP-- this is used in bacterial transformation apparently, as part of competence) 23:25 I want to encode the dsDNA logic gate strands into the genome, but that's another story -- I actually want the nucleotides to encode *which* logic to use for each gate, then have another part come in and swap in the actual gate (AND, OR, NOR, XOR, ...) while conserving the number of used gates (for various technicalities) 23:26 this is all so that I can evolve logic that has some 'emergent' property 23:26 since I can't come up with anything for Andy. heh' 23:26 so this would be like 'programmable' logic gates? 23:27 yes, I suppose that's true 23:27 is that how an FPGA works? 23:27 having a conserved number of gates, but you get to choose the pathways between them or something? 23:27 an fpga has a memory element that controls the function of the gate (and, or, etc) 23:27 you are very limited in choosing what gate connects to what other gates 23:28 its more like a grid than a brain 23:28 ah, so this is almost the same thing 23:29 except the memory is on the genome 23:29 wait, same thing 23:29 I am sure of it. 23:29 "have another part come in and swap in the actual gate" could you elaborate on what 'another part is' and how to 'swap in' the gate function? 23:30 like the DNA is the compressed version of the gate, and something comes along and rewrites the DNA macro? 23:30 expands the macro 23:31 do you know about introns and exons? 23:35 somewhat, not enough 23:37 as for the elaboration: Andy's lab has done this with nucleic acids. He has evolved unnatural amino acids and nucleic acids, plus the enzymes required for DNA replication with those new units; so if you have an encoding representing AND, OR, NOT, then instead of the amino acid you have the dsDNA come up. The dsDNA can be made by replicating a portion of the DNA molecule that specifies how to make the gates. (eof) 23:40 you cant make DNA with a ribosome though? 23:40 don't know 23:40 that would be useful 23:40 you could do lots of splicing with mRNA using introns/exons and then reverse-transcriptase it back to DNA 23:41 splicing inside of the cell? 23:41 yes 23:41 dna-splicing tools are generally implements of war 23:41 hrm 23:41 not something a cell would do in the course of everyday housekeeping 23:47 i think it's funny that people call bacteria 'species' Day changed to 22 Mar 2008 00:22 http://en.wikibooks.org/wiki/Circuit_Idea/How_do_We_Create_Sinusoidal_Oscillations%3F <-- somebody has put a lot of work into this article 00:24 i wonder who 00:25 wiki's are cool but i dont see any reason behind doing a lot of work and not putting your name on it somewhere 00:26 http://en.wikibooks.org/wiki/Circuit_Idea 00:26 this entire book is peculiar 00:26 it's right on 00:26 all about the social knowledge of electronic circuits 00:28 i'm glad something is working. the old electronics collaborative book was festering 00:28 here's the guy's website 00:28 http://www.circuit-fantasia.com/ 00:29 http://www.circuit-fantasia.com/circuit_stories/inventing_circuits/inventing_list_of_circuits.htm lots of content 00:29 *shudder* 00:29 make it stop! 01:19 black boxing = hiding : http://en.wikibooks.org/wiki/Circuit_Idea/Why_Circuit_Ideas_are_Hidden 01:19 -!- kanzure: No such nick/channel 01:19 -!- kanzure [n=bryan@cpe-70-113-54-112.austin.res.rr.com] 01:19 -!- was : purple 01:19 -!- server : irc.freenode.net [Sat Mar 22 04:33:19 2008] 01:19 -!- End of WHOWAS 12:57 http://heybryan.org/mediawiki/index.php/Socialized_engineering_knowledge_database 12:57 if you'd like to drop some notes. 13:22 huh, I think you are, in fact, the bottleneck - http://heybryan.org/mediawiki/index.php/Getting_up_to_speed 13:23 also, I noted your recent-changes on your wiki and will want to talk about amorphous fabrication + cells later, I am pretty sure cells are not the replicators that we need since they can't specifically manufacture in "solid-state" sense 13:24 although ultimately since we are biological, and we have constructed these machines, we know that it should be possible to make more biological systems to produce such machines as the ones we see before us, 13:24 but that's pretty hard in general, since it required extensive training of the humans 14:17 http://en.wikibooks.org/wiki/Circuit_Idea/Why_Circuit_Ideas_are_Hidden <- every time you say 'blackboxing' i think of this 14:17 i want open source hardware, and more than that, open knowledge 14:17 right 14:18 blackboxing just means "for now, we avoid this component and act like we know what it is" 14:18 and in many cases, the ideas are hidden to make the box/circuit 14:18 although hopefully we can change that 14:18 could we call it 'scrounging' or 'outsourcing' instead? 14:18 sure 14:19 http://heybryan.org/mediawiki/index.php/Blackbox 14:19 done 14:21 hextatic is actually aimed more at empowerment of the individual than turning into a swarm of self-replicators 14:21 it just happens that i can't think of a better starting point for a self replicator 14:23 we need to do a mega outline brain storming session for what is needed to make a self-replicator, once and for all 14:23 maybe it would be better to call it 'environment modification' 14:23 how's that? 14:23 because the goal of having a self-replicator is to do stuff with it 14:23 like make a floating sky palace or whatever 14:24 after all, it's not me that's replicating. i can do that already, just need to find a suitable female 14:25 and/or sheep and biotech lab :P 14:25 right 14:26 so I see what you mean 14:26 a tool to help extend the environments in which we (as von Neumann replicators already) can work in 14:26 or to help increase the number of environments in which we can setup and install ourselves 14:27 yes 14:27 it is interesting that even though we are biological, we have somehow been able to do solid state devices, even though our biology is quite 'fuzzy' 14:27 so that gives hope that we can do fuzzy biological systems that will make solid state devices 14:27 unless we want to do a self-replicator without an internal dependency on biology 14:28 i want to do a self-replicator without an internal dependence on biology 14:28 the dna tile thing was just a clever shortcut 14:28 hrm 14:28 i didn't anticipate it 14:28 the idea was 'computers are the most complex thing out there, so if we can make one of them, we can make anything' 14:28 turns out that's not true 14:29 if you do want to do it without biology, then what's with your "environment modification"-tool ? 14:29 i'm a selfish greedy gene 14:29 and so are the other 7 billion humans 14:30 if i want their help, it has to be relevant for them too 14:30 otherwise you have stuff like the artist christo - decorating islands in pink plastic 14:31 http://www.christojeanneclaude.net/si.shtml 14:32 also i feel like using biology is cheating. you dont really truely understand how it works until you've made it from scratch 14:32 so my previous method of getting started on a von neumann probe project is (in my opinion) still working pretty well: I defined what was needed and then was going to work backwards ("Humans would need these resources") 14:32 so let's do that same thing but without the biological component 14:33 and since this opens up too many doors at the same time 14:33 let's add in some constraints, namely mineral constraints 14:33 let's go find out what asteroids commonly have in them, and then figure out a way to do self-replication with the processing of those materials 14:33 let's not add artificial constraints 14:33 then we'd have to have an "everything fabricator" and an "everything tool" 14:33 there will be multiple ways to do it, why start with asteroids when we are on earth? 14:34 well, I just mean to propose that scenario 14:34 where resources are obviously limited 14:34 and you don't have to make a tool to process every possible mineral 14:34 there are asteroids full of nickel and platinum.. the resources aren't limited they're just different 14:35 i think we should try to make use of glass because the stuff is everywhere 14:35 and ceramics 14:36 but didn't you just say you didn't want to constrain the possibilities? ;) 14:36 i'm advocating their use, not saying everything else is off limits 14:38 perhaps it's possible to make silicone rubber from raw minerals? 14:38 sure, it would have to be 14:38 i mean easily, without an army of chemists and refineries :) 14:39 we have to do it first by hand before programming the machine to do it 14:42 correct 14:42 what would silicone rubber be used for? 14:42 and what other resources would you be considering for inclusion ? 14:42 insulation, gaskets and seals, lubrication 14:43 i'm also thinking about aluminum 14:43 what is the fabricator of this design? 14:43 do you notice a common theme? (i do) 14:43 sort of, looks mechanical or vacuum-tube based 14:43 well, maybe. but it all comes from clay-like and sandy soil 14:45 clay is aluminum magnesium silicate 14:46 aluminum would be much more 'expensive' to manufacture because it requires hydrochloric acid (and where do you get the chlorine?) 14:46 let's wiki it 14:46 http://heybryan.org/mediawiki/index.php/Clay-sand-soil_replicator 14:47 I just jotted down some notes 14:51 let's do something like this 14:51 http://en.wikipedia.org/wiki/Image:Advanced_Automation_for_Space_Missions_figure_5-29.gif 14:51 this is the one where arms pick up parts from a shelf 14:51 and then assemble the parts into the replicator 14:52 yeah looks really advanced ~~ 14:52 then we can add functionality where the arms can retrieve the parts or store the parts onboard 14:52 where the 'parts' are really just material resources 14:53 oh, as an experimental phase? instead of actually digging dirt you just give it glass and clay? 14:53 so if we have a 'number of parts' we can then go back and jot down all required machinery to make that particular part (to process the raw mineral-resources), and then if we want to do good then we have to optimize all of that machinery to use the same resources and be made from the same arm-manipulator 14:53 yeah 14:54 eventually we should just give it input such as a bucket of sand 14:54 and then further down the road we make it walk outside and search for the bucket of sand 14:54 (or whatever would be a good debug-test for searching for raw minerals) 15:05 i find myself reluctant to add carbon to the list for some reason 15:05 i think this is too complicated to do on a wiki 15:06 can it be done on paper? 15:06 there are lots of interconnections, and at different levels 15:06 here are my notes on the "contractionary design approach" - http://heybryan.org/mediawiki/index.php/Self-replication#The_.22contractory_phase.22_design_approach 15:06 yeah 15:06 maybe it needs a whole namespace instead of just a page 15:12 why disassembly? i dont think it's necessary for self replication, just makes for a more efficient ecology 15:12 no, dissassembly for *designing* it 15:12 the actual machine need not be able to dissassemble itself 15:13 in Hierarchy of self-reppers 15:14 ah, I was thinking about how cells seem to be self-destructive when they replicate 15:14 some species breed and then just die afterwards 15:15 they expend lots of energy in the process and sometimes cut their resources in half 15:15 which is something that can't be fixed, in general, unless they go get more resources etc. 15:15 wasteful 15:17 i read about an alien bacterial life form (possibly fictional?) that replicated by building miniature copies inside of the cell, when then burst open 15:17 i can see the sense in that 15:17 but afaik all earthian animals can lay eggs/do mitosis 15:18 spores, seeds, rhizomes 15:18 isn't there some insect that lays eggs inside her body and is eaten from the inside out for the larvae to survive? 15:18 not that i know of, but i'd be interested if you have a link 15:25 ping 15:25 i've not heard of this insect 15:26 no i've not heard of this insect 15:27 gah 15:27 http://69.180.166.50/old/_/files/operaessions_0.jpg 15:27 Superkuh gets kinda extreme. 15:41 so from the look of the http://heybryan.org/mediawiki/index.php/Clay-sand-soil_replicator page it looks like you're going in the opposite direction that I was proposing (which isn't bad, I'm just observing) 15:41 particularly, you're starting with basic minerals and then specifying what can be made from them 15:41 and then I guess you're just hoping that you started with the right mix of minerals to make machinery that can use all of them ? 15:44 ./join #hplusroadmap 16:00 I didn't mean to pull you away from work, I thought epitron was solid 16:01 hah 16:02 you thought i was solid? :) 16:02 what? 16:02 well sure, you seem to have enough background to be able to make yourself do work 16:23 -!- kanzure [n=bryan@cpe-70-113-54-112.austin.res.rr.com] 16:23 -!- ircname : purple 16:23 -!- channels : @#hplusroadmap #biology ##neuroscience #bioinformatics @#namcub #wrongplanet @##wikipeding 16:23 -!- server : irc.freenode.net [http://freenode.net/] 16:23 -!- : is identified to services 16:23 -!- End of WHOIS 16:23 a shill? 16:24 hm? 16:24 purple :) 17:48 -!- Irssi: Looking up irc.freenode.net 17:48 -!- Irssi: Connecting to irc.freenode.net [140.211.166.3] port 6667 17:48 -!- Irssi: Connection to irc.freenode.net established 17:48 !irc.freenode.net *** Looking up your hostname... 17:48 !irc.freenode.net *** Checking ident 17:48 !irc.freenode.net *** No identd (auth) response 17:48 !irc.freenode.net *** Found your hostname 17:48 -!- Welcome to the freenode IRC Network pz 17:48 -!- Your host is zelazny.freenode.net[zelazny.freenode.net/6667], running version hyperion-1.0.2b 17:48 !zelazny.freenode.net *** Your host is zelazny.freenode.net[zelazny.freenode.net/6667], running version hyperion-1.0.2b 17:48 -!- This server was created Sat May 19 03:15:54 UTC 2007 17:48 -!- zelazny.freenode.net hyperion-1.0.2b aAbBcCdDeEfFGhHiIjkKlLmMnNopPQrRsStTuUvVwWxXyYzZ01234569*@ bcdefFhiIklmnoPqstv 17:48 -!- IRCD=dancer CAPAB CHANTYPES=# EXCEPTS INVEX CHANMODES=bdeIq,k,lfJD,cgijLmnPQrRstz CHANLIMIT=#:20 PREFIX=(ov)@+ MAXLIST=bdeI:50 MODES=4 STATUSMSG=@ KNOCK NICKLEN=16 are supported by this server 17:48 -!- SAFELIST CASEMAPPING=ascii CHANNELLEN=30 TOPICLEN=450 KICKLEN=450 KEYLEN=23 USERLEN=10 HOSTLEN=63 SILENCE=50 are supported by this server 17:48 -!- There are 24761 listed and 19580 unlisted users on 27 servers 17:48 -!- 43 flagged staff members 17:48 -!- 19595 channels formed 17:48 -!- I have 3934 clients and 0 servers 17:48 -!- Current local users: 3934 Max: 4302 17:48 -!- Current global users: 44341 Max: 46091 17:48 -!- Highest connection count: 4304 (4302 clients) (650635 since server was (re)started) 17:48 -!- - zelazny.freenode.net Message of the Day - 17:48 -!- - zelazny.freenode.net Message of the Day - 17:48 -!- - Welcome to zelazny.freenode.net in Corvallis, Oregon, US! Thanks to 17:48 -!- - Oregon State University for sponsoring this server! 17:48 -!- - 17:48 -!- - ZELAZNY, ROGER JOSEPH [1937-1995]. Roger Zelazny was a 17:48 -!- - leading light in the New Wave of science fiction. He began 17:48 -!- - writing in 1962. Known for such works as This Immortal, 17:48 -!- - Lord of Light, The Dream Master and the Amber books, 17:48 -!- - Zelazny is best known for his idiomatic American 17:48 -!- - protagonists, embodying the mythological figures of the 17:48 -!- - Trickster and the Hero of a Thousand Faces. 17:48 -!- - 17:48 -!- - You're using freenode, a service of Peer-Directed Projects 17:48 -!- - Center (http://freenode.net/pdpc.shtml). 17:48 -!- - 17:48 -!- - By connecting to freenode you indicate that you have read 17:48 -!- - and agree to adhere to our policies and procedures as per 17:48 -!- - the website (http://freenode.net). We would like to remind 17:48 -!- - you that unauthorized public logging of channels on the 17:48 -!- - network is prohibited. Public channel logging should only 17:48 -!- - take place where the channel owner(s) has requested this 17:48 -!- - and users of the channel are all made aware (if you are 17:48 -!- - publically logging your channel, you may wish to keep a 17:48 -!- - notice in topic and perhaps as a on-join message). 17:48 -!- - 17:48 -!- - By registering your nickname with Nickserv you agree that you 17:48 -!- - are 13 years of age, or older. For more information about the 17:48 -!- - Children's Online Privacy Protection Act please see their 17:48 -!- - website at (http://www.coppa.org). 17:48 -!- - 17:48 -!- - freenode runs an open proxy scanner. Your use of the network 17:48 -!- - indicates your acceptance of this policy. For details on 17:48 -!- - freenode network policy, please take a look at our policy 17:48 -!- - page (http://freenode.net/policy.shtml). Thank you for using 17:48 -!- - the network! 17:48 -!- - 17:48 -!- - freenode is a service of Peer-Directed Projects Center, an 17:48 -!- - IRS 501(c)(3) not-for-profit organization. Our yearly 17:48 -!- - fundraiser will begin soon; if you'd like to donate early, 17:48 -!- - please see http://freenode.net/pdpc_donations.shtml for more 17:48 -!- - information. Thank you for using freenode! 17:48 -!- End of /MOTD command. 17:48 freenode-connect [freenode@freenode/bot/connect] requested CTCP VERSION from pz: 17:48 -NickServ(NickServ@services.)- This nickname is owned by someone else 17:48 -NickServ(NickServ@services.)- If this is your nickname, type /msg NickServ IDENTIFY 17:48 -!- Mode change [+i] for user pz 17:48 -!- Irssi: Unknown command: log show 17:48 -!- Irssi: Unknown command: log list 17:48 Logs: 17:48 1 /home/fenn/irclogs/hplusroadmap.log: ALL -autoopen 17:49 17:49 LOG OPEN [-noopen] [-autoopen] [-window] [-] [-targets ] [-colors] [] 17:49 LOG CLOSE | 17:49 LOG START | 17:49 LOG STOP | 17:49 17:49 -noopen: Create the entry to log list, but don't start logging 17:49 -autoopen: Automatically open this log file at startup 17:49 -: Targets are logged only in this server 17:49 -targets: Log only in specified channels/nicks (space separated list) 17:49 -window: Log output in the window. Active window is used by default, or 17:49 you can give the window's refnum in -targets. 17:49 : File name where to log, it is parsed with 17:49 strftime(), so %d=day, etc. see "man strftime" for 17:49 more info. 17:49 : Defaults to ALL 17:49 : ID number of log. 17:49 17:49 /SET log_create_mode - Specifies what file mode to use with 17:49 the created log files. Default is 0644. 17:49 17:49 All of these are parsed with strftime(): /SET log_timestamp - Specifies the time stamp format. 17:49 Default is "%H:%M ". 17:49 /SET log_open_string - Text written to log when it's opened /SET log_close_string - Text written to log when it's closed /SET log_day_changed - Text written to log when day changes 17:49 17:49 NOTE: Log files are locked after opened, so two Irssis can't accidentally try to write to the same log file. 17:49 17:49 Examples: 17:49 17:49 /LOG OPEN -targets cras ~/irclogs/cras.log MSGS 17:49 - Logs all messages from/to nick `cras' 17:49 17:49 /LOG OPEN -targets #linux ~/irclogs/linux/linux-%Y-%m-%d 17:49 - Logs all messages in channel #linux. Log is rotated daily, so 17:49 logs in 1. May 2000 goes to file "linux-2000-05-01", when the 17:49 day is changed, Irssi closes the log and starts logging to 17:49 "linux-2000-05-02" etc. 17:49 17:49 See also: SET LOG, WINDOW LOG 17:49 17:49 Irssi commands: 17:49 log close log open log start log stop --- Log closed Mon Mar 24 17:49:19 2008 --- Log opened Mon Mar 24 17:50:27 2008 17:50 -!- fenn [n=pz@adsl-76-251-84-177.dsl.bltnin.sbcglobal.net] has joined #hplusroadmap 17:50 -!- Topic for #hplusroadmap: happy bunny day - day of the replicator! 17:50 -!- Topic set by fenn [] [Sun Mar 23 22:34:25 2008] 17:50 [Users #hplusroadmap] 17:50 [ Enki-2] [ epitron] [ fenn] [ kanzure] 17:50 -!- Irssi: #hplusroadmap: Total of 4 nicks [0 ops, 0 halfops, 0 voices, 4 normal] 17:50 -!- Channel #hplusroadmap created Sat Mar 22 15:44:12 2008 17:50 -!- Irssi: Join to #hplusroadmap was synced in 0 secs 17:51 < fenn> meep 17:53 < kanzure> beep 17:57 < fenn> kanzure: have you seen n55? http://www.n55.dk/MANUALS/Manuals.html 17:58 < kanzure> No, hold on 17:58 < kanzure> fenn: I just found out about this: http://wiki.workatjelly.com/Jelly-Austin-March-25-2008 17:58 < kanzure> I think I'll go. 17:59 < fenn> constructive work in a bar? yeah, sure 18:00 < fenn> are you going to sneak in? :) 18:09 < kanzure> oh, a bar 18:09 < kanzure> shit 18:10 < kanzure> fenn: I don't get n55 18:10 < fenn> it's a technology distribution 18:10 < kanzure> is it? 18:10 < kanzure> it looks like a recursive self-defining manual 18:10 < fenn> aimed at the functionality required for living as a human on earth 18:11 < kanzure> hm 18:11 < fenn> it's also an art project, which helps in some ways and hurts in others 18:13 < fenn> i wish FACTORY was capable of making those damn plastic octahedron tanks 18:14 < fenn> apparently it's a lot easier to get them made in denmark 20:08 -!- fenn_ [n=pz@adsl-76-251-87-49.dsl.bltnin.sbcglobal.net] has joined #hplusroadmap 20:08 -!- Topic for #hplusroadmap: happy bunny day - day of the replicator! 20:08 -!- Topic set by fenn [] [Sun Mar 23 22:34:25 2008] 20:08 [Users #hplusroadmap] 20:08 [ Enki-2] [ epitron] [ fenn] [ fenn_] [ kanzure] 20:08 -!- Irssi: #hplusroadmap: Total of 5 nicks [0 ops, 0 halfops, 0 voices, 5 normal] 20:08 -!- Channel #hplusroadmap created Sat Mar 22 15:44:12 2008 20:08 -!- [freenode-info] if you need to send private messages, please register: http://freenode.net/faq.shtml#privmsg 20:08 -!- Irssi: Join to #hplusroadmap was synced in 12 secs