2008-09-02.log

--- Day changed Tue Sep 02 2008
fenni think people like tabs because windows doesn't give much in the way of window management00:00
kanzureeven my vertical scrollbar hack isn't all that helpful00:01
kanzurestill have to scroll00:01
kanzurehow do you keep a bajillion things going at once00:02
newgenomeWOAH00:02
kanzureand still keep it easily addressable00:02
newgenomejust looked up the figures for graphene paper00:02
kanzureyeah?00:02
newgenomeit has a tensile strength of 35000 gigapascals00:02
fennthat doesnt sound right00:02
newgenomecarbon nanotubes have a tensile strength of 62000 gpa00:03
newgenomespider silk is 1200 gpa00:03
kanzurelink00:03
newgenomeno wait it's all in megapascals00:04
newgenomesorry got it wrong there00:04
fenn*bonk*00:04
newgenomefacepalm00:04
newgenomeall measurements above are in Mpa00:04
newgenomeso this graphene paper stuff is half the strength of single molecule carbon nanotubes00:05
fennthat figures00:05
fennit's basically the same stuff, but with rough edges00:05
newgenomeyep00:06
newgenomeand it is easy to make00:06
kanzurehttp://heybryan.org/graphene.html00:06
kanzureoh00:06
kanzurelet me export this to a wiki for you00:06
newgenomebut it gets messed up with water00:07
newgenometensile modulus00:07
newgenomethat is what I put00:08
newgenomehttp://ttp.northwestern.edu/abstracts/viewabs.php?id=31600:08
newgenomeI don't know if that is any different00:08
kanzurehttp://heybryan.org/mediawiki/index.php/Graphene00:08
kanzurehave fun00:08
kanzureoh00:08
kanzureI already had stuff on it00:08
kanzurehuh00:08
fenntensile modulus is stiffness00:09
kanzureFeel free to edit and dump notes.00:09
newgenomeso it isn't strong enough for building incredible strong things00:09
fenngraphene paper is not the same thing, it's like the difference between wood and paper00:11
newgenomeyeah I figured that out00:11
newgenomethere really needs to be a way to synthesize huge sheets of single molecule graphene00:12
kanzuremelt graphene down00:12
kanzurelet it cool on a surface00:12
newgenomeI don't think you can melt graphite00:12
newgenomemost single molecule graphene is made by CVD00:13
newgenomeI believe00:14
fennwhy do you want huge sheets of it?00:14
fennhmm interesting, monocrystaline silicon has a tensile strength of 7GPa vs 5.5GPa for carbon fiber (62GPa for carbon nanotube)00:15
kanzurefenn: I think the biobricks registry might be useless 00:16
kanzurewhen it comes to using the xml data00:16
newgenomeso you can somehow weld the sheets together to make a container that holds metallic hydrogen00:16
kanzureso far I haven't "cracked" it00:16
fennwhat do you mean "cracked" it? its all right there00:16
newgenomeor make a graphene balloon and pump it up to 6000 psi00:16
kanzure   <START>1008</START>    <END>1062</END>    <SCORE>0.0</SCORE>    <ORIENTATION></ORIENTATION>    <PHASE>-</PHASE>     <NOTE>BBa_R0062</NOTE>00:17
kanzurethat tells me nothing00:17
kanzurethat's the same data they use for all of the parts00:17
kanzurethe same variables I mean00:17
kanzureso what's distinguishing these parts in the XML data ?00:17
kanzurenot much ..00:17
kanzureit's all the semantic descriptions of the parts more than anything else00:17
fennthe XML data is just annotations00:17
kanzureannotations re: the sequences, right?00:17
fennshowing where the promoters, gene products, etc are00:18
fennit doesnt say what the part is supposed to do, that's on the wiki page00:18
fennthere's no formal way of describing how to use the part00:18
kanzurekinda like skdb in general00:19
kanzurehow do you describe turtles ?00:19
kanzureturtle societies especially00:19
fennskdb is supposed to have a formal description of how the parts fit together00:20
fennthat's why we have all these classes and units and stuff00:20
kanzureright00:20
fennotherwise its just a bunch of text like appropedia00:20
kanzureso then what would it take to work for partsregistry.org ?00:21
fennyou'd have to read each of the descriptions and stuff them into a formal category with parameters and hyperlinks etc00:21
fennand also write some code describing how the different categories interact00:21
fennsame thing as skdb really00:21
fennturtle = class = category00:22
kanzureright00:22
kanzurethey already have them in categories sorta00:24
kanzurehttp://partsregistry.org/cgi/partsdb/pgroup.cgi?pgroup=Signalling00:24
kanzureCell-Cell signalling section00:24
kanzurehttp://partsregistry.org/Part_Types00:24
kanzuredoes that count ?00:24
kanzureit's kind of hard to make something reasonable with that methinks00:24
kanzureit really requires external knowledge being hammered into the system I think00:24
kanzuregoing through the bio literature and finding a new gene to play with using the same parts00:24
fennthose categories (part_types) are useful but there's just not enough information on how they interact00:25
kanzurethere are some comments / descriptions that mention a relationship with another part00:26
kanzurebut it's not formalized00:26
fennright00:26
fenn"produces fluorescence in either the absence of TetR or in the presence of the inducer aTC." could be easily formalized00:26
kanzureokay, so let's fork it00:27
kanzurewhat mechanisms will we add in to let those formalizations to occur00:27
kanzurein the git-orized version of the db.00:27
kanzuremaybe each new type of relationship between two bricks00:27
kanzurewill be a separate set of files00:27
fennthose are like the skdb classes00:28
kanzurekind of like when you're doing mysql db tables you add in a new table to relate the unique id of some other table (the key) with some new property that you want to sort by00:28
fennwhereas each part is an skdb project00:28
kanzurethe classes are the relationships?00:28
fennno, the classes have methods which define the relationships00:28
kanzureand where in the proc do you run/call the methods?00:28
kanzurejust checking.00:28
fennwhen testing if the parts will work together the way you want?00:29
kanzureI'm very confused00:29
kanzureI agree that each part is an skdb project00:29
fennsay signaller A makes ampicillin, and signaller B makes tetracycline, then measurement unit C that responds to tet will "work" with B but not with A00:29
kanzureokay00:30
kanzureA, B, C = 3 skdb projects00:30
fennso when you're looking around for stuff that works with C, you get a list of things that make tetracycline00:30
fennright00:30
kanzurethat kind of makes sese00:30
kanzuresense00:30
kanzurewould it run through the list of all objects and run the make() method and if it produces 'tet' then it keeps it on the list ?00:30
kanzureor is there something else you're thinking of00:31
kanzurelike would that be in the metadata spec of the part ?00:31
fenn'makes tet' would be in the metadata00:31
fennuh, that's turtle layer number 100:31
fennwhat its supposed to do00:31
kanzurethis sounds like a good story00:32
kanzure"Turtles and Python"00:32
kanzuresecond ed. is "Turtles in Python"00:32
fennturtle was slow and stupid, but python knew how to tie himself in knots00:32
fennone day python ate turtle and became so bloated he couldn't tie himself in knots anymore00:33
fennthe end00:33
kanzurethen a gnu appeared and set everything right00:33
kanzurehrm00:33
kanzuremaybe also a camel00:33
kanzureanyway00:33
fennnooo00:33
kanzurepythons eat camels?00:33
fennpythons just ignore camels and hope they go away and die already, or maybe spit and hiss at them00:34
kanzureso in the metadata turtle representation layers thingies00:34
kanzurewhat's layer 2 specified at ? is it just another set of metadata that you could download, or would it be just extra stuff hidden in the package00:35
kanzurehm, I like hidden extra stuff00:35
fennlayer 2 is like, more introspection00:35
fennlooking at inherited traits and building new models of what is going to happen00:36
fenni guess00:36
fennforget i said anything about layer 1 2 300:36
kanzurek, that helps00:36
fennnominal output products: tet, GFP00:36
kanzureokay, so we can have a simple output specification for these biobricks00:37
fennactual output products: tet, GFP, NADP+, ADP, amino soup, tRNA00:37
kanzureand input spec00:37
newgenomehmmm00:37
newgenomethis gives me an idea00:38
kanzureI'd be worried if it didn't00:38
newgenomehow fast can gfp be produced00:38
fennmegatons per millisecond00:38
newgenomeand can it be turned off00:38
fennnevar!00:38
kanzurepolymerase per second will tell you sort of00:38
kanzurePoPs00:38
fennnewgenome: are you thinking about some kind of optical signalling mechanism?00:38
newgenomeas in could I make a cellular automata00:39
fennlike a blinking led00:39
kanzurekind of like Andy's lab's bacterial photography?00:39
kanzureyeah00:39
newgenomeyeah00:39
kanzurethey're working on cellular automata00:39
kanzurethat's what we were doing in the lab00:39
kanzuresort of00:39
kanzureturing patterns aren't quite cellular automata but it's kinda close enough00:39
newgenomeexcept blinking all over00:39
newgenomecool00:39
newgenomewait turing patterns?00:39
kanzureyes, but there's a discrete basis to it00:39
newgenomehow the heck are you getting turing patterns00:39
kanzureI was going to try to implement "LiquidWars"00:40
kanzurego play LiquidWars for about 2 seconds and you'll get it00:40
kanzure2D surface with patterns00:40
kanzureit's not cellular automata rules, I know00:40
newgenomehttp://www.bitstorm.org/gameoflife/00:40
newgenomei want to implement this sort of00:40
kanzureI know what CA is. ;-)00:40
newgenomegood00:41
newgenomesomeone is working on literally cellular automata?00:41
newgenomeare they using luciferin?00:41
newgenomedoes it have a faster response?00:42
kanzureNot in Andy's lab .. all in vitro00:42
kanzureanyway00:42
kanzurenonprotein00:42
-!- ybit [n=h@unaffiliated/ybit] has quit [Read error: 104 (Connection reset by peer)]00:42
fennyou have sequence binding fluorophores right?00:42
kanzureriight00:42
kanzureright00:43
fennwell something like that00:43
-!- ybit [n=h@unaffiliated/ybit] has joined #hplusroadmap00:43
kanzurefenn: okay, so I'll rip the data set shortly and just automatically generate relationships based off of which category it's in, kinda randomly assume all of the data, and then the parts within each category will have some sort of relationship generated for starters (we don't need to be precise at this point)00:43
kanzureand then we can just use that to try to build graphs, no ?00:43
kanzureI don't know what the selection function will be though00:43
kanzureonce all of that data is filled out (even if half-assed / autogenerated)00:43
fennsounds good enough00:44
kanzureno, I mean, let's say you're at your graph app thingy00:44
kanzurehttp://bloodgate.com/graph-easy/00:44
kanzurelet's say you're here again00:44
kanzureactually I might like  http://bloodgate.com/graph-demo/more 00:44
kanzurehttp://bloodgate.com/graph-demo/ more00:44
fennok, what about something like stupidfunclub.com?00:44
kanzureis it flash?00:45
fennwhere instead of trying to see the whole graph at once, you tell a 'story' about what you want to do, one step at a time00:45
fennyes it's flash00:45
kanzureok, now I see it00:45
fenndoesnt have to be flash though00:45
kanzurehad to switch browsers00:45
kanzurehm00:45
kanzureokay, I like that00:45
fennotherwise you just get a big hairball00:45
kanzureentirely doable00:45
kanzureright00:46
kanzurebut then what about parallel processes that you want to represent or something00:46
kanzuredon't know how to say that00:46
fennmultiple branches in the story00:46
kanzureyou don't branch on a conditional?00:46
kanzurethat's important.00:46
fennhuh?00:46
kanzureimagine an rpg with a tree for the storyline00:46
kanzurethose are conditional branches00:47
kanzurenot "always running concurrently"00:47
kanzurejust an important difference to consider.00:47
fennsure00:47
fenni dont think of this story as being a chronological sequence00:47
fennmore like a circuit diagram00:47
kanzuresignaler -> teh signal -> receptor00:47
fennbut the point is, each step along the circuit you have a set of options for what to add next00:47
kanzureodd, that's not how I design my electrical schematics00:47
kanzuresometimes it is I guess00:48
kanzureusually it's just me hearing somebody yelling at me "you forgot some damn resistors, that thing will fry like an egg"00:48
fennyou could even work backwards from the end point, or work both directions if it's a loop00:48
fennfwiw graph-easy doesn't show up in konqueror00:49
kanzureblah00:49
kanzuregraph-demo?00:49
kanzurewahoo00:50
fennoh, wait, its just words with > between them00:50
fennnevermind00:50
kanzurehm?00:50
* kanzure checks.00:50
kanzureyeah, that's an odd character blitting error 00:51
* kanzure wonders if this could be combined with collabedit.com just be to ridiculous .. ok, maybe not.00:51
kanzures/could/should/00:51
fennew00:52
kanzurecollaborative mind maps00:52
newgenomehttp://www.physorg.com/news139500643.html00:54
newgenomecheck this out00:54
newgenomemeat anyone?00:54
kanzureDepends, how complex is the synthesis of this polymer?00:55
kanzurefenn: ok, so the storymaker interface here .. how should this be implemented?00:56
kanzureI know how to make a storymaker in general,00:56
kanzurebut in the context of graph-demo I'm not so sure00:56
kanzuremaybe a couple of pages of a 'wizard' ?00:56
kanzurea wizard to guide the pythons and turtles, you see00:56
fennthe idea is to end up with a circuit diagram, right? so you draw it out as you go00:59
kanzurewhere do you see your options for the next unit thingy00:59
fennat each vertex in the graph, you have all the edges from that vertex available to choose from00:59
kanzureor your options for the relationship thingy00:59
fennlike some kind of context menu or something00:59
kanzureoh00:59
kanzureew00:59
fennok, another frame00:59
kanzureew ew00:59
fennwhat?00:59
kanzureyou mean a listbox maybe?00:59
fennlittle icons with words on them doesn't help make good decisions00:59
kanzurethis is true01:00
fennyou want to see what your options are and how they work01:00
fennum, bad phrasing01:00
fennyou want to see probably some pictures, explanation of how it works, testing readiness level, cost, etc01:00
kanzureall in one little window? yikes01:01
kanzureinterface nightmare01:01
kanzurethat should be stuff the user types into a cli and stuff pops up01:01
fennnot really, basically it would just show you the wiki page for that skdb project01:01
kanzuresorry ?01:01
kanzureso at the bottom of the screen, have it dynamically load the project you're about to make an edge to?01:01
fennright01:01
kanzurethe wiki page of the project01:01
kanzureok, sounds good, but I still don't see how they're going to see their options of possible new vertices to choose from01:02
kanzurelistbox at the side of the widget/panel in which they are 'drawing' ?01:02
newgenomeman this is way over my head01:02
fennso on the left side you have little icons with names stuck to them, and on the right you have the detailed view of whatever you just selected or focused on01:02
kanzurefenn: so you don't have an overall view of what you've written in the story so far?01:02
fennno, of course not01:02
kanzureso you just go linearly ?01:02
fennif you knew what you were doing you wouldnt need this program01:03
kanzureah01:03
fennuh, you have a diagram of what you've picked so far, yes01:03
kanzureokay01:03
kanzureoh01:03
kanzureleft side -> little icons with names stuck to them, check01:03
kanzureright side you have a detailed view of what you've just selected01:03
kanzurebut then where is your diagram of what you've picked so far?01:03
kanzuremaybe just a small little horizontal diagram at the top ?01:04
kanzurehrm01:04
fennit's being built up in the window with little icons01:04
kanzureoh01:04
fennon the left, same window01:04
kanzurethen where do they see their list of options for the new vertices to pick from ?01:04
fenneach icon has edges going out of it01:04
kanzurebah01:04
kanzuretime for you to do a mockup01:04
fennwhen you're at the bleeding edge, there's multiple edges to pick from01:04
fennthey lead to 'ghost' icons that are the options available to you01:05
kanzureuhm, so let's say that there's 10k options for the next part01:05
kanzure10k ghost icons, really ?01:05
fennwell, maybe not01:05
fennthe other option was a list01:05
fennyou could click on an edge and get a list of things that fit that edge in the usual description window01:06
fennok mockup time01:07
kanzureheh01:07
fennkolourpaint wheee01:07
kanzureyep, same here01:07
fennit always looks so much better in your head01:07
kanzurebtw, if there's such a "click an edge" feature, it's easy to implement in code at least for something like01:07
kanzureexamine_edge.py file.diagram edge-id-here01:07
kanzureand then it could spit out the information or whatever01:07
kanzurethat's simple enough.01:07
kanzure(of course, we're talking about a browser thingy)01:07
kanzure((but we both know it's not restricted to a browser thingy))01:07
kanzureI'm almost tempted to go cold turkey and not use a graphical interface at all ;-)01:08
kanzuremaybe that's the way to go anyway01:09
kanzurehack out the list of commands for the shell01:09
kanzurethen wrap those up into the web app01:09
bkerodo it01:11
bkeroGet the output to be through an irc bot.01:11
kanzurearen't irc bots typically terrible?01:11
bkeroYes, but they're an effective output mechanism.01:11
kanzureYou want to do engineering in a chat room with an irc bot ?01:12
kanzureheh01:12
bkeroyes01:12
newgenomehey01:12
kanzureyou're crazy :-)01:12
bkero!run command3.pl01:12
* kanzure adds it to his todo list01:12
newgenomejust found a cool biobrick01:12
bkero< Bot>  Hits: 0, Misses: 4, Anomolies: 8192, Entropy: 93%01:12
kanzurebkero: no, I'm thinking of trying combinations/permutations in a chat room with collective community experience or something01:12
newgenomehttp://partsregistry.org/wiki/index.php?title=Part:BBa_I74211101:13
newgenomelimonene synthase01:13
newgenomelimonene is a chemical that can dissolve polystyrene01:13
newgenomeit's also flammable01:13
bkeroCan't gasoline also?01:14
newgenomeand smell's like lemons01:14
newgenomegood point01:14
newgenomeyou could make napalm with it01:14
bkeroI've done that01:14
bkeroStyrofoam + gasoline = gooey01:14
bkeroIt absorbs about 1000:101:14
kanzurebkero: entropy? :)01:15
bkeroYes01:15
bkeroLots of entropy in your theoretical data. :P01:15
newgenomebactonapalm01:16
kanzureDoesn't make sense ..01:16
bkeroIt was just sample output :P01:16
kanzureokay01:16
kanzurejust had to know if there was something behind it01:16
kanzureyou know, just in case01:16
bkeroNope01:16
bkeroYou could have calculated entropy as an output01:16
fennblah.. god this is ugly01:17
fennhttp://fennetic.net/pub/irc/skdb-mockup/select-icon.png01:17
kanzurehehe01:17
fennhttp://fennetic.net/pub/irc/skdb-mockup/select-edge.png01:17
kanzureokay01:17
kanzureinstead of clicking, let's just do shift + left/right arrow01:18
kanzureand then no-shift means switching the nodes01:18
kanzurewhen you right-arrow over to something that doesn't exist, you're auto-transported to the list of options01:18
kanzureto the right01:18
kanzurewhere right arrow goes down, placing the blocks on the page01:18
kanzuresimple enough of course01:18
fenncan do both01:18
kanzureclicking lines in js just sounds hard01:19
fennkeyboard and mouse interface is common in gui widgets01:19
kanzureI don't think I know of a way to click on lines in javascript like that01:19
kanzureunless you're pixel perfect01:19
fenni dont know any javascript so i probably wont be learning it just for this01:19
kanzureor "snap on" action01:19
kanzuredon't bother01:19
kanzureI feel impure even mentioning it01:19
kanzurebut I might do this mockup.01:20
fennyou dont have to be pixel perfect, the line can light up based on which element the cursor is nearest01:20
* kanzure wonders if the bloodgate script has something for that01:20
fennlike in electrical cad01:20
fenni need to do something like this already for emc01:21
kanzuregui/js thingy?01:21
fennno, a graph/list thingy01:21
kanzurehuh?01:21
kanzurenot your mockup?01:21
fennconnecting boxes that can only be connected in certain ways01:21
fennits basically the same idea01:22
fennbut totally different context01:22
kanzurethat's what the "validator" that we mentioned a long time ago should be doing01:22
kanzurewell, not just "validator"01:22
fennthe validator is what gives you the list of options01:22
kanzurebut also the "find a compatable next part" idea01:22
kanzure*compatible01:22
-!- newgenome [n=chatzill@resnet-47-165.dorm.utexas.edu] has quit ["ChatZilla 0.9.83 [Firefox 3.0.1/2008070208]"]01:23
kanzurethe trick is that we have a data set with the biobricks stuff, sort of01:23
kanzureor at least we can auto generate the data set01:23
kanzuresince they are already somewhat staggard01:23
kanzurewe could do the random rule stuff and have it be bad meta data01:23
kanzureuntil they yell at us and are friendly enough to fork it for a while and then merge it when it's acceptable :)01:23
fennuntil they laboriously type all their sloppy plain english metadata into our system, you mean01:24
kanzurethey are welcome to come up with their own metadata system, I think01:24
kanzurearen't they?01:24
fennindeed, but they havent01:24
kanzureslackers01:24
fennsometimes its easier to let someone else tell you what to think (set up the metadata scheme)01:24
fennthen filling out all the details is just grunt work01:25
kanzuresome poor undergrad01:25
fennor even just kids on the diybio list01:25
fennshrug01:25
kanzurewhere do they have access to that info?01:25
kanzureI don't think it's all in the descriptions all the time01:25
fennhmm01:25
kanzureit's usually in the literature, which kids typically don't have access to01:25
kanzurebut01:26
kanzurefinding those papers and giving them out isn't hard either01:26
kanzureokay, 8 am classes01:26
kanzureI should sleep01:26
fenngn801:26
bkeroick 8am classes01:27
kanzureick 8 am calculus class01:28
kanzureick 8 am calculus class oh shit don't have the book homework due thursday01:28
kanzurethat sort of thing01:28
bkerog2library01:28
kanzurelike everyone else?01:28
bkeroMebbe01:29
bkeroGet there early in the morning, copy the pages, return the books, and figure out how to do the problems by wikipedia01:29
kanzureoh, the content isn't the problem01:29
kanzureit's the questions01:29
kanzureso yes01:29
bkeroor find a torrent on textbooktorrents.com?01:29
-!- Depucelato1 [n=Depucela@c-24-118-116-62.hsd1.mn.comcast.net] has joined #hplusroadmap01:31
kanzure'You are not currently enrolled in any courses.'01:32
kanzurehahah01:32
kanzurebeautiful ..01:32
kanzurefairly popular book, it has been torrented 01:33
kanzureaha01:36
kanzuredemonoid hurrah01:36
bkeroDemonoid :P01:41
bkeroThere you go01:41
bkeroNow you can be even lazier.  Thank you internet.01:41
kanzurethe torrent isn't starting though01:41
kanzurehm01:41
bkeroTried textbook torrents?01:42
kanzureregistration is currently disabled01:43
kanzurehas been for a while IIRC01:43
bkeroI have an account01:43
bkeroWhats the book called?01:43
kanzurehttp://www.demonoid.com/files/details/1552394/15625924/01:46
Depucelato1There's a private tracker for textbook torrents?01:46
bkeroDidn't used to be private.01:47
Depucelato1you're talking about Demonoid or one specifically for textbooks?01:47
Depucelato1Because I'm in partickler sick of paying for latest edition textbooks01:47
kanzureDepucelato1: textbooktorrents01:47
Depucelato1I've taken to melting the binding off of my textbooks and putting them on the scanner/copier at work to scan and OCR01:48
Depucelato1which is great for timed open-book texts where I supposedly wouldn't be fast enough to browse through the book for answers01:49
kanzuredamnit, wrong book01:55
bkerohttp://labs.mozilla.com/2008/08/introducing-ubiquity/01:58
kanzure"doing something" requires an art to construct the task02:04
kanzureit doesn't just magically exist02:04
kanzureso once one person puts in the effort, it should be much less effort to just copy/paste the file02:04
kanzurejust like my idea of having Austin2008.zip for all of the events around town02:04
kanzureargh02:15
kanzurehow could this be ?02:15
kanzurethere's no question #85 in this section02:15
kanzureso I'm pretty sure it's the wrong book02:15
kanzurebut amazon does not know of the right book's existence02:15
fennwow ubiquity is like surfraw v2.002:24
kanzureyou have it running?02:25
fennno i'm watching the demo02:25
fennit runs in the browser though.. not sure how hard it would be to fix that02:25
-!- Depucelato1 [n=Depucela@c-24-118-116-62.hsd1.mn.comcast.net] has quit ["Leaving."]02:34
fennubiquity is really quite fun to play with04:22
-!- nsh [n=nsh@eduroam-52.uta.fi] has joined #hplusroadmap04:57
bkeroYes it is :)05:22
bkerofor i in *; do mencoder $i -ovc copy -oac copy -forceidx -of avi -o $i.new && mkvmerge -o `echo $i|sed -e '{ s/\.avi/\.mkv/ }'` $i.new && rm $i.new && rm $i; done05:22
-!- Phreedom [n=freedom@ip-194-50-167-184.mir.dn.ua] has quit [Read error: 110 (Connection timed out)]05:46
nshEGREEK06:24
-!- faceface [n=dbolser@bioinformatics.org] has quit [Remote closed the connection]06:44
nsh"In the past two years, a laboratory in France and another in the United States independently reconstructed a functioning HERV-K retrovirus from pieces found in the human genome. This summer, both showed that the gene sequences of some of those viruses bear the characteristic fingerprints of APOBEC3, a human enzyme that mutated them into submission."06:59
nshcolour me dubious...06:59
nsh--http://www.washingtonpost.com/wp-dyn/content/article/2008/08/31/AR2008083101759_pf.html07:02
kanzurehuh07:45
willPow3r...07:49
* kanzure just woke up and isn't processing effectively07:49
-!- elias` [n=me@unaffiliated/elias/x-342423] has quit [Read error: 131 (Connection reset by peer)]08:43
-!- elias` [n=me@unaffiliated/elias/x-342423] has joined #hplusroadmap08:47
willPow3ri hate when that happens08:49
kanzurewaking up?08:50
willPow3rwaking up is nice, but the need for sleep isn't08:51
* kanzure vanishes08:52
-!- Phreedom [n=freedom@ip-194-50-167-184.mir.dn.ua] has joined #hplusroadmap09:32
-!- jm|earth [n=jm@p57B9E010.dip.t-dialin.net] has joined #hplusroadmap10:47
-!- nsh [n=nsh@eduroam-52.uta.fi] has quit [Success]11:32
-!- Phreedom_ [n=freedom@ip-194-50-167-184.mir.dn.ua] has joined #hplusroadmap11:53
-!- Phreedom [n=freedom@ip-194-50-167-184.mir.dn.ua] has quit [Read error: 110 (Connection timed out)]12:11
-!- Nade [n=lixasd@cpc2-clif5-0-0-cust516.nott.cable.ntl.com] has quit []12:29
-!- ybit [n=h@unaffiliated/ybit] has quit [Read error: 104 (Connection reset by peer)]12:46
-!- ybit [n=h@unaffiliated/ybit] has joined #hplusroadmap12:49
kanzuresomebody tempt me to use ubiquity12:59
bkerokanzure: It's got context-sensitive verbs.13:00
bkeroYou can use it to embed google maps into anything :P13:01
kanzurealso, the skdb app grapher thingy, is that just for structure? or is it for process? if I wanted to mesh a 3d object and call it a part, is the viewer thingy showing that I have an object "touching" other things? or what?13:01
kanzurebkero: so, it's just greasemonkey userscripts all over again?13:03
bkerokanzure: Sort of.  It's more like quicksilver for firefox.13:03
bkeroHave you ever used quicksilver?13:03
-!- ybit [n=h@unaffiliated/ybit] has quit [Remote closed the connection]13:12
-!- ybit [n=h@unaffiliated/ybit] has joined #hplusroadmap13:13
kanzureI don't remember.13:33
kanzurequickfox13:33
kanzurenot quicksilver13:33
kanzureoh crap, 1 pm appointment with an automated professor13:33
kanzureprofbot or something13:33
-!- pk [i=pk@cash-3b-130.res.umass.edu] has joined #hplusroadmap14:09
pkhi14:10
willPow3rpk14:12
pki want google chrome14:16
willPow3rchrome is going to compete with firefox, not IE14:19
willPow3rso keep that in mind when you decide to switch14:20
pkI have great add-ons in firefox14:25
pkchrome would have to blow firefox away for me to switch14:26
pkit's released now15:06
pkthe beta is out15:06
pki can't get it from google though, must be getting hammered badly15:06
-!- ybit [n=h@unaffiliated/ybit] has quit [Connection timed out]15:12
kanzurehahah15:29
kanzurefirefox has lots of problems15:29
kanzureanyway15:29
kanzurehahah15:29
kanzureI met with the doc behind the automated design lab today15:29
kanzureso here's what he's doing15:31
kanzuregiven a bag of parts, make a graph to represent the design and their interconnections and so on15:31
kanzurethen, he wants to simulate the functionality of these designs15:32
kanzurefor example -- I didn't tell him this -- but a good example would be biobrick kinetics15:32
kanzure*cough*15:32
kanzureand then he wants to see what manipulations of the parameters when generating those designs (from 'seeds') produce good stuff, or whatever15:32
kanzureyou could manually optimize all of the parameters of a spaceship if you had to, for example15:33
kanzureanyway,15:33
kanzurefor figuring out which parameters are interesting and the values that are interesting, he wants to use not genetic algorithms but simple neural networks with their own graph representation15:33
kanzurethen use graph rewriting 'grammar rules' to find a suitable amount of complexity in the 'neural net' (or hidden markov model, more appropriately?) 15:33
kanzurethat can deal with the problem space that you are designing within15:34
kanzure'grammar rules' are the things that can be applied to a 'neural network graph' and add a new edge into it as a permutation15:34
kanzureor whatever15:34
kanzureanyway, I piped up at that point and suggested that you should learn which of these 'grammar rules' start to make for interesting parameter-variable-values in the solver/evaluator of the designs that you tree-generate15:35
kanzureturns out that that's the project that he was going to pitch me15:35
kanzurebe careful what you say, you might be assigned to actually doing it15:35
* kanzure is wondering about computational complexity classes and whether or not graph rewriting grammar rules are more appropriate for certain classes than others, and whether or not manufacturing problems can be reduced to comp sci15:36
kanzuresurely 'traveling salesman and routing' is a good example, but that's not the same context.15:36
kanzureanyway, he mentioned there might be money involved15:36
kanzurenow, which biobricks would be particularly good for using in such a system15:38
kanzuresince we have chemical kinetics of some of them, I'm sure we do15:38
kanzurethen we can just use the random connections between the parts of each 'category' on partsregistry.org15:40
kanzurethat we were talking about last night15:40
kanzureplug in some kinetic simulation stuff, and then generate a full tree of all of the possible designs and whatever, then optimize that by coming up with either the 'neural network' or 'markov model thingy' so that the designs can be better selected "somehow" -- not quite sure what the framework is going to look like ...15:41
kanzurethe trick is finding the designs that work best, so I'm pretty sure kinetics is a good place15:41
kanzurethis guy apologized *to me* for using C#15:48
kanzurehttp://clm.utexas.edu/MM-MM08talk-med/15:52
* kanzure is not so sure how cross-lateral these 'graph grammar rules' are going to be across the computational complexity domain15:54
-!- Nade [n=lixasd@cpc2-clif5-0-0-cust516.nott.cable.ntl.com] has joined #hplusroadmap16:07
pkis this guy talking to UT austin donors or something?16:12
kanzurenot sure16:18
kanzuremight be students?16:18
-!- biopunk [n=p@h224n2c1o261.bredband.skanova.com] has joined #hplusroadmap16:19
-!- jm|earth [n=jm@p57B9E010.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)]17:48
-!- ybit [n=h@unaffiliated/ybit] has joined #hplusroadmap17:55
-!- Depucelator [n=Depucela@c-24-118-116-62.hsd1.mn.comcast.net] has joined #hplusroadmap18:43
bkerohttp://www.engadget.com/2008/09/02/mythbusters-rfid-hacking-episode-canned-by-credit-card-company-l/19:24
-!- kanzure_ [i=bbishop@66.112.232.116] has quit [Read error: 104 (Connection reset by peer)]19:28
-!- kanzure_ [n=bbishop@66.112.232.116] has joined #hplusroadmap19:29
-!- kanzure__ [i=bbishop@66.112.232.116] has joined #hplusroadmap19:36
ybithttp://www.ted.com/index.php/talks/paul_rothemund_details_dna_folding.html19:39
-!- kanzure_ [n=bbishop@66.112.232.116] has quit [Read error: 104 (Connection reset by peer)]19:40
-!- kanzure_ [i=bbishop@66.112.232.116] has joined #hplusroadmap19:45
-!- kanzure__ [i=bbishop@66.112.232.116] has quit [Remote closed the connection]19:45
pknice, looks like a good one19:50
pkI'll watch after dinner19:50
-!- biopunk [n=p@h224n2c1o261.bredband.skanova.com] has quit []19:51
bkeroYay ted19:53
Nadei love ted20:42
Nadei've watched like nearly every vid on that site20:42
pkoh my21:09
pkthat video is awesome21:09
kanzureaughghghghgh21:09
kanzureHulk crush!21:09
pkwatching all those smileys floating around is surreal21:09
pkthanks for sharing21:10
kanzureokay21:10
kanzureso21:10
kanzuredr. campbell has me paid for doing his research21:10
kanzuredr. mauk has his cerebellum simulator in fucking visualbasic21:10
kanzurewants that over to C/C++ via OpenMP21:10
kanzurehe's doing his own fortran version, wtf21:10
kanzurethought that was dead21:10
kanzureoverall great people21:10
kanzurealso went to the ASME meeting, the Engineers for a Sustainable World meeting (building a biodiesel reactor over at the pickle research center off campus, also poop bacteria for Mexico), and the Robotics and Automation Society with free 24/7 lab access in a trailer right off of the engineering 'square' of buildings21:11
kanzurewho the fuck codes a cerebellum in visual basic21:11
-!- ybit [n=h@unaffiliated/ybit] has quit []21:13
-!- ybit [n=h@unaffiliated/ybit] has joined #hplusroadmap21:15
* kanzure turns into hulk21:20
kanzure 21:23
kanzurePrevious Claims of siRNA Therapeutic Effects Called into21:23
kanzure Question by Report in Human Gene Therapy 21:23
pkhuman gene therapy is some kind of publication?21:28
pkyup.21:29
kanzurecrap21:30
kanzurewhere'd I send something about that yellow drum bot21:30
kanzureanybody know which mailing list?21:30
-!- Nofaris [n=Nofaris@adsl-75-42-77-105.dsl.scrm01.sbcglobal.net] has joined #hplusroadmap21:32
kanzurehttp://heybryan.org/~bbishop/docs/mauksimzip.zip <-- visual basic hurray?22:06
-!- pk [i=pk@cash-3b-130.res.umass.edu] has quit []22:48
proctokanzure: dunno if that's something you're interested in, by my research project this semester will be on qualitative spatial reasoners22:59
kanzurewhat does this mean23:00
kanzurequalitative?23:00
kanzurereasoners?23:00
proctowell, reasoners use logics and calculi to infer conclusions from premises23:01
kanzureeww23:01
kanzure:)23:01
proctoprolog is an example of a language and a particular logic23:02
kanzureright23:02
proctowell, qualitative spatial reasoning has to do with spatial things23:02
proctoso, for example you say "there's a bum near the mcdonald's on main st"23:03
proctowhat does "near" mean?23:03
proctoit's a qualitative spatial operator23:03
proctothere are ways to treat such statements in a formal mannel23:03
proctoI plan on using this to augment my gargoyle system23:03
proctoor imagine saying something like this to a system: "I'm on a building on the 3rd floor. out the windown I can see X, Y and Z. Where am I?"23:04
proctothings like that23:05
proctothe specific area I will be working on, I'm not sure yet23:05
kanzurereverse reasoning to figure out where an agent is, would certainly be interesting23:07
kanzurethat's worthwhile methinks23:07
kanzureto reconstruct overall location from clues23:08
proctothat's one application23:08
proctoin general, you don't always have perfect spatial information23:08
kanzureright23:08
kanzurerarely do I :-)23:08
proctoespecially in unfamiliar locations23:08
kanzuretoday I found myself staring at a 12 by 9 foot neural tissue slice on a wall23:09
kanzureit was beautiful23:09
kanzurebut I had no clue where I was.23:09
kanzurebeautiful in the evil way23:09
proctohehe23:10
proctowith my system implemented, you can show it a photo, and then ask for the nearest bathroom23:10
procto:>23:10
kanzurecell phone pics23:10
proctofully implemented. which probably won't quite happen very soon23:11
proctorealistically23:11
proctothing is, no one ever told it where the bathrooms in the building are explicitely23:11
proctoit probably scraped it from a site fro a small symposium, where one of pages said the bathroom was "Across the hall"23:12
proctothat's the kind of thing I want23:12
kanzure"download the poopfactor, dial 222 now."23:13
kanzureI thought of a silly wearable computer idea23:24
kanzureif you cross your hands to your alternate shoulders, there's a good place to strap on one half a keyboard on each outer edge of the arm23:25
proctotoo tiring23:28
kanzureyeah :(23:28
kanzureokay, how about the old card trick23:29
kanzurebut with giant keyboards23:29
kanzurecard under the sleeve23:29

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