Conversation with #hplusroadmap at Sun 23 Mar 2008 09:35:47 AM CDT on kanzure@irc.freenode.net (irc)

(2008-03-23 15:15:25) fenn [n=pz@adsl-76-251-87-108.dsl.bltnin.sbcglobal.net] entered the room.
(2008-03-23 16:08:06) kanzure: Hey fenn.
(2008-03-23 16:08:08) kanzure: You get my email?
(2008-03-23 16:10:01) fenn: yep
(2008-03-23 16:10:10) kanzure: I don't think it's much of a problem
(2008-03-23 16:10:13) kanzure: that it's a Hamiltonian cycle
(2008-03-23 16:10:27) kanzure: but it just goes to show this will require lots of number crunching
(2008-03-23 16:10:38) fenn: i looked it up and it didnt mean much to me
(2008-03-23 16:10:44) kanzure: Oh, well, it's Traveling Salesman problem.
(2008-03-23 16:10:55) kanzure: you have a guy that wants to go travel a path and wants to optimize his route
(2008-03-23 16:11:04) kanzure: and hit up a few clients or whatever
(2008-03-23 16:11:11) kanzure: there's an optimal path that gets all of the points
(2008-03-23 16:11:24) kanzure: but this can't be solved in polynomial deterministic time or something
(2008-03-23 16:11:58) kanzure: http://en.wikipedia.org/wiki/Traveling_Salesman
(2008-03-23 16:12:07) kanzure: NP-hard ;)
(2008-03-23 16:12:12) fenn: i hope the database isnt complex enough that it will take a computer very long
(2008-03-23 16:13:16) kanzure: this depends on what we want to do
(2008-03-23 16:13:26) kanzure: do we want to (1) hand-code all connections between packages?
(2008-03-23 16:13:48) 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.
(2008-03-23 16:14:09) kanzure: #1 means we will probably know it all so that we will know if something seems stupid
(2008-03-23 16:14:12) kanzure: but by #1 we may miss stuff
(2008-03-23 16:14:29) 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)
(2008-03-23 16:14:41) fenn: #2 seems very difficult
(2008-03-23 16:14:53) kanzure: maybe not -- just code in modules for each type of variable to interact with each other
(2008-03-23 16:15:05) fenn: not impossible, but crikey, it's like the holy grail of automated engineering
(2008-03-23 16:15:12) kanzure: just like in C++ where you have to define how the + operator interacts with other classes for a certain object
(2008-03-23 16:15:16) kanzure: haha
(2008-03-23 16:15:20) kanzure: yes, I suppose that's true.
(2008-03-23 16:15:38) fenn: now i've had people say stuff like that to me all the time, and it doesnt necessarily mean anything
(2008-03-23 16:15:54) kanzure: the holy grail stuff?
(2008-03-23 16:16:08) fenn: after all we have gigahertz computers with gigabytes of ram and terabytes of storage
(2008-03-23 16:16:14) kanzure: that's true
(2008-03-23 16:16:20) kanzure: and not all of it has to be 'simulated' with physics and so on
(2008-03-23 16:16:26) kanzure: it's really just testing out how to put legos together
(2008-03-23 16:16:27) fenn: and still using the same software from the 1980's for most CAD/CAE stuff
(2008-03-23 16:16:34) kanzure: it's like having a giant bucket of legos
(2008-03-23 16:16:44) kanzure: but these legos don't all connect
(2008-03-23 16:17:01) kanzure: hrm, the lego analogy doesn't work I think
(2008-03-23 16:17:04) kanzure: since they are not functional-legos
(2008-03-23 16:17:08) fenn: its like trying to use your SCSI tape reader to play video on the VCR
(2008-03-23 16:17:14) kanzure: hah
(2008-03-23 16:17:39) kanzure: I don't think so. What happened to the dependency-loop ideas and so on ?
(2008-03-23 16:17:45) fenn: and then record computer backup data onto a video tape
(2008-03-23 16:18:43) fenn: well the interesting thing about a dependencey loop is it's sort of a mini exponential growth without self replication
(2008-03-23 16:19:41) kanzure: Oh, Wikipedia is correcting me. It's not a Hamiltonian cycle. Just a closed path.
(2008-03-23 16:19:47) kanzure: hm?
(2008-03-23 16:20:03) 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
(2008-03-23 16:20:08) kanzure: closed cycle, actually (cycle means path, except the path ends/begins at same vertice)
(2008-03-23 16:20:19) kanzure: I remember that one :)
(2008-03-23 16:21:41) 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).
(2008-03-23 16:21:57) kanzure: and as you fill out the requirements, it is kind of like exponential growth
(2008-03-23 16:22:05) kanzure: since the more you need, the more components you have to add
(2008-03-23 16:22:14) kanzure: that's why I like starting with a part already -- an arm, say
(2008-03-23 16:22:20) kanzure: and then we add components to the arm until it is able to self-replicate
(2008-03-23 16:22:27) 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
(2008-03-23 16:22:46) kanzure: like the character '8' ?
(2008-03-23 16:22:48) fenn: more like a knitted headband
(2008-03-23 16:22:51) kanzure: like oooooooooo
(2008-03-23 16:22:57) kanzure: hm
(2008-03-23 16:23:02) kanzure: xxxxxxxxxx ?
(2008-03-23 16:23:07) kanzure: but those cross
(2008-03-23 16:26:17) fenn: http://fennetic.net/pub/irc/directed_cyclic.png
(2008-03-23 16:27:00) kanzure: that's hard to understand
(2008-03-23 16:27:10) kanzure: could you color the cycles?
(2008-03-23 16:27:17) fenn: a b c are the replicators
(2008-03-23 16:27:24) kanzure: the entire design?
(2008-03-23 16:27:45) fenn: alone they couldnt replicate
(2008-03-23 16:28:20) kanzure: why is there two sets of a, b, c ? Showing the replication?
(2008-03-23 16:28:30) fenn: just to connect the top and bottom
(2008-03-23 16:28:48) kanzure: I'm sort of thinking there should be more symmetry here
(2008-03-23 16:31:02) kanzure: The requirements:
(2008-03-23 16:31:03) kanzure: A+B = C
(2008-03-23 16:31:06) kanzure: A + C = B
(2008-03-23 16:31:09) kanzure: B + C = A
(2008-03-23 16:31:14) fenn: ok reload. the green are new copies of the red ABC, but their position in the graph is identical
(2008-03-23 16:31:40) fenn: the purple is a leaf node, unrelated to replication (but requires b and c)
(2008-03-23 16:31:59) kanzure: B + C looks required for replication
(2008-03-23 16:32:14) kanzure: but not the purple one, right
(2008-03-23 16:33:25) kanzure: aren't these supposed to be directed?
(2008-03-23 16:33:31) fenn: yesmeh
(2008-03-23 16:33:53) fenn: dia is a pain to use
(2008-03-23 16:33:59) kanzure: this is dia?
(2008-03-23 16:33:59) kanzure: haha
(2008-03-23 16:34:05) fenn: no, that's kolourpaint
(2008-03-23 16:34:07) kanzure: hehe
(2008-03-23 16:36:01) fenn: i added numbers
(2008-03-23 16:36:28) fenn: you might say that node 6 or node 7 is "the" replicator
(2008-03-23 16:36:42) kanzure: What about 5?
(2008-03-23 16:38:43) fenn: hmm how do you show in graph theory that you must approach a vertex from two edges at the same time?
(2008-03-23 16:39:05) kanzure: that's interesting
(2008-03-23 16:39:17) kanzure: draw arrows with matching colors?
(2008-03-23 16:39:26) kanzure: lemme ask #math
(2008-03-23 16:40:08) fenn: 7 should depend on 5 and 6 being present
(2008-03-23 16:40:36) fenn: i added 9, which depends on 5 and 8
(2008-03-23 16:41:33) fenn: i'm sure this could be simplified
(2008-03-23 16:41:38) kanzure: a "required vertice" is interesting
(2008-03-23 16:41:42) kanzure: required vertices
(2008-03-23 16:41:52) kanzure: it's a functional graph, in a sense
(2008-03-23 16:42:02) kanzure: so you just have to say that "vertice X does not exist without vertice Y"
(2008-03-23 16:42:27) fenn: well, it might exist, but you can't get to it
(2008-03-23 16:43:29) fenn: like, i might have a lathe and bar-stock, but no cutting bits
(2008-03-23 16:44:49) fenn: ok poor example
(2008-03-23 16:47:25) kanzure: maybe you need two graphs
(2008-03-23 16:47:32) kanzure: one for the dependencies and then one for the vertice dependencies
(2008-03-23 16:47:33) kanzure: or something like that
(2008-03-23 16:48:15) fenn: earlier i was talking about different graphs for streams and packages (run-time and build-time, respectively)
(2008-03-23 16:48:38) fenn: but you can also have a package that makes packages
(2008-03-23 16:48:53) fenn: or streams that spontaneously degrade into something
(2008-03-23 16:53:29) kanzure: hm
(2008-03-23 16:53:41) kanzure: so then you have to figure out a way to specify the extra requirements for the graph then
(2008-03-23 17:06:11) fenn: gosh math is really useless sometimes
(2008-03-23 17:06:24) kanzure: reading up on Wikipedia?
(2008-03-23 17:06:40) fenn: "the sum of all the vertex-degree is equal to twice the number of edges"
(2008-03-23 17:07:49) kanzure: the vertex-degree is something like the number of connections
(2008-03-23 17:07:56) kanzure: to a certain vertice
(2008-03-23 17:08:29) fenn: yep
(2008-03-23 17:09:05) fenn: since a line connecting the vertices has two end points, there are 2*lines number of endpoints
(2008-03-23 17:09:18) fenn: big deal
(2008-03-23 17:09:34) kanzure: haha
(2008-03-23 17:09:35) kanzure: =)
(2008-03-23 17:23:46) kanzure: fenn, what exactly are you looking into at the moment?
(2008-03-23 17:31:42) fenn: strange loops and graph theory
(2008-03-23 17:35:03) kanzure: mind pasting anything you find interesting ?
(2008-03-23 17:38:39) 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.
(2008-03-23 17:41:19) kanzure: huh? for graph theory?
(2008-03-23 17:41:19) kanzure: hehe
(2008-03-23 17:42:05) fenn: apparently irony is self-referential
(2008-03-23 17:44:44) kanzure: didn't know?
(2008-03-23 18:12:47) fenn: this looks about right: http://www.graphviz.org/Gallery/directed/fsm.html
(2008-03-23 18:13:28) kanzure: neat.
(2008-03-23 18:15:18) fenn: i really want something more interactive though
(2008-03-23 18:15:26) kanzure: to what extent?
(2008-03-23 18:16:47) fenn: to be able to poke at the graph and rearrange it.. as if the vertices were connected by springs
(2008-03-23 18:18:05) kanzure: yes, I have wanted an interactive interface to graphviz for a while now
(2008-03-23 18:18:22) 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
(2008-03-23 18:18:26) kanzure: so it's kind of disgusting.
(2008-03-23 18:19:49) fenn: someone i know wrote a java app that did springy networks (FEA - he's an engineer)
(2008-03-23 18:20:06) fenn: this should be pretty easy to write from scratch in whatever language
(2008-03-23 18:22:26) kanzure: finite element analysis?
(2008-03-23 18:22:31) kanzure: hm
(2008-03-23 18:22:39) kanzure: but hacking graphviz is not so easy, last I checked
(2008-03-23 18:22:50) kanzure: they had lots of literature that you have to sort through, you need to have connections to get the papers
(2008-03-23 18:22:53) kanzure: maybe I'm wrong, it's worth investigating
(2008-03-23 18:23:48) fenn: i need to write a similar application for emc2 actually
(2008-03-23 18:23:54) fenn: graphical hal configurator
(2008-03-23 18:28:45) fenn: this is neat http://www.kylescholz.com/projects/speaking/tae2006/music.fm/
(2008-03-23 18:32:22) kanzure: ooh
(2008-03-23 18:32:34) kanzure: it looks pretty close
(2008-03-23 18:33:58) fenn: the neat part is the search results are actually right
(2008-03-23 18:36:38) fenn: you can drag the names around if you select them
(2008-03-23 18:42:28) 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.
(2008-03-23 18:45:12) kanzure: reprap?
(2008-03-23 18:50:50) fenn: yes
(2008-03-23 20:33:05) kanzure: fenn: so what should the next steps be? any thoughts?
(2008-03-23 20:33:22) kanzure: I think we need to formalize your graph-definition and make a program to generate expandable graphviz files that follow the format
(2008-03-23 20:38:13) fenn: file format and parser
(2008-03-23 20:38:54) fenn: needs to be text based, human and machine readable
(2008-03-23 20:39:04) fenn: i'm thinking about yaml or slip-xml
(2008-03-23 20:45:25) kanzure: maybe we can write the parser in flex/bison to convert to graphviz format for the data visualization
(2008-03-23 20:45:34) kanzure: erm, I think I am cutting a few steps in the toolchain
(2008-03-23 20:47:19) Enki-2: just use flex alone
(2008-03-23 20:47:29) Enki-2: bison isn't needed if you know flex
(2008-03-23 20:49:37) kanzure: right, but the file format would not necessarily be what we want to see a graph of
(2008-03-23 20:50:19) Enki-2: i mean, flex can do translation of formats on its own, sans bison
(2008-03-23 20:50:35) Enki-2: i've done complete compilers in flex
(2008-03-23 20:50:50) kanzure: woah, you're not fenn
(2008-03-23 20:50:53) kanzure: just realized this :)
(2008-03-23 20:51:17) Enki-2: :P
(2008-03-23 21:21:47) fenn: i dont know flex or bison, so i'm going to do a naive straightforward implementation :)
(2008-03-23 21:22:37) kanzure: bison is extremely easy
(2008-03-23 21:22:51) kanzure: http://dinosaur.compilertools.net/bison/bison_6.html
(2008-03-23 21:22:51) kanzure: http://www.gnu.org/software/bison/
(2008-03-23 21:22:51) kanzure: http://cs.wwc.edu/~aabyan/464/Book/
(2008-03-23 21:23:40) kanzure: This one is really good - http://www.garshol.priv.no/download/text/bnf.html
(2008-03-23 21:23:43) kanzure: short and to the point
(2008-03-23 21:23:59) 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)
(2008-03-23 21:24:31) kanzure: inventory could be external for simple designs
(2008-03-23 21:24:48) fenn: this cost comes out of human attention (our most precious commodity) unless there is an automated inventory system
(2008-03-23 21:25:07) 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.
(2008-03-23 21:25:35) fenn: i take it that's bad?
(2008-03-23 21:25:48) kanzure: Objectivisim is a word-hijack
(2008-03-23 21:25:54) kanzure: it means lots of politics and capitalism
(2008-03-23 21:26:00) kanzure: generally not good philosophy
(2008-03-23 21:26:11) fenn: what does politics and capitalism have to do with philosophy?
(2008-03-23 21:26:29) kanzure: Ayn Rand :(
(2008-03-23 21:26:47) fenn: oh, objectivism is at least slightly different from objectivity
(2008-03-23 21:26:56) kanzure: http://en.wikipedia.org/wiki/Objectivism_(Ayn_Rand)
(2008-03-23 21:26:59) fenn: in terms of word structure
(2008-03-23 21:27:22) 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; `
(2008-03-23 21:27:29) kanzure: the Proper Moral Response
(2008-03-23 21:27:31) kanzure: *Purpose
(2008-03-23 21:28:04) fenn: if they can implement capitalism through a university philosophy department, more power to them :\
(2008-03-23 21:28:09) kanzure: heh
(2008-03-23 21:28:29) fenn: as i see it, capitalism was demolished by corporations and the patent system
(2008-03-23 21:28:36) fenn: or at least the prospect of it becoming a reality
(2008-03-23 21:29:47) 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.
(2008-03-23 21:29:48) fenn: how can you have patents and call it capitalism? nonsense
(2008-03-23 21:29:55) kanzure: these old ideas are *ancient*
(2008-03-23 21:29:59) kanzure: technological development is accelerating
(2008-03-23 21:30:27) kanzure: perhaps when we had a constant rate of development, capitalism/communism/Objectivism/Christianity/Republicanism worked, but the world just isn't the same
(2008-03-23 21:30:37) kanzure: these are all old, stale ideas with a massive number of people still promoting them
(2008-03-23 21:30:46) kanzure: and while there are good ideas within,
(2008-03-23 21:30:58) kanzure: the culture is still largely one of identification and not of ... importing the good ideas ;)
(2008-03-23 21:31:10) fenn: indeed
(2008-03-23 21:31:19) fenn: let it be known i'm not a sports fan
(2008-03-23 21:31:23) kanzure: hah =)
(2008-03-23 21:32:15) fenn: got any opinion on this? http://www.debian.org/social_contract
(2008-03-23 21:33:07) kanzure: oh-my-god I have access to scientific databases
(2008-03-23 21:33:18) fenn: you mean journals?
(2008-03-23 21:33:44) kanzure: yes
(2008-03-23 21:33:50) ***kanzure hugs his UT login
(2008-03-23 21:34:04) fenn: gratz. got ieee spectrum access?
(2008-03-23 21:34:08) kanzure: fenn: Not anything off the top of my head.
(2008-03-23 21:34:09) kanzure: fenn: yes
(2008-03-23 21:34:12) fenn: sweet
(2008-03-23 21:34:16) kanzure: fenn: I just opened up an article on pyroelectricity
(2008-03-23 21:34:45) kanzure: Piezoelectricity and Pyroelectricity in Polymers
(2008-03-23 21:34:57) kanzure: First legally accessed, paywalled scientific paper
(2008-03-23 21:35:10) kanzure: this should go in my baby book or something
(2008-03-23 21:35:32) fenn: today is something special.. i forget
(2008-03-23 21:35:56) kanzure: there was chocolate today
(2008-03-23 21:35:59) kanzure: so I'd venture a guess of easter
(2008-03-23 21:36:14) fenn: aha
(2008-03-23 21:36:34) fenn: right, we had jehova's witnesses
(2008-03-23 21:36:42) fenn: (not for dinner)
(2008-03-23 21:37:19) fenn: international bunny day - day of the replicator!
(2008-03-23 21:37:57) fenn has changed the topic to: happy bunny day - day of the replicator!
(2008-03-23 21:40:14) kanzure: epitron: I'll be adding a 50 MB zip file to the server later tonight.
(2008-03-23 21:41:45) kanzure: woot, nature!
(2008-03-23 21:43:01) kanzure: American Physical Society !
(2008-03-23 21:58:13) 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.
(2008-03-23 21:58:34) 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.
(2008-03-23 21:58:53) 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"
(2008-03-23 22:03:01) fenn: if (grep '<a href=".*\.pdf">'): wget $1; else ... ?
(2008-03-23 22:03:20) fenn: note that was pseudo-code
(2008-03-23 22:05:46) kanzure: nah,
(2008-03-23 22:05:53) kanzure: the url doesn't necessarily have .PDF in it
(2008-03-23 22:05:58) kanzure: but you can scan for PDF next to the link
(2008-03-23 22:06:11) kanzure: there's a way to process Google pages, they have a nice systematic format with div tags and so on
(2008-03-23 22:06:22) 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
(2008-03-23 22:06:59) kanzure: WWW:Mechanizer might be what I want.
(2008-03-23 22:07:17) fenn: you need to get those cookies though, which might not be straightforward with a non-browser interface
(2008-03-23 22:07:54) kanzure: ooh
(2008-03-23 22:08:02) kanzure: WWW::Mechanize does something interesting though
(2008-03-23 22:08:03) fenn: especially since scientific publishers arent known for their browser compatibility
(2008-03-23 22:08:08) kanzure: use WWW::Mechanize;
(2008-03-23 22:08:08) kanzure: use HTML::TokeParser;
(2008-03-23 22:08:11) kanzure: my $agent = WWW::Mechanize->new();
(2008-03-23 22:08:11) kanzure: $agent->get("http://www.radiotimes.beeb.com/");
(2008-03-23 22:08:11) kanzure: $agent->follow("My Diary");
(2008-03-23 22:08:15) kanzure: $agent->field("email", $email);
(2008-03-23 22:08:15) kanzure: $agent->click();
(2008-03-23 22:08:24) kanzure: so in other words it's like doing the physical actions
(2008-03-23 22:08:29) kanzure: except there's no browser UI
(2008-03-23 22:13:47) kanzure: Okay, first part done. Neat.
(2008-03-23 22:23:03) kanzure: Hm, the rest is testing my knowledge of regexps. I want to extract the first link after <b>[PDF]</b>.
(2008-03-23 22:24:06) fenn: $agent->follow("[PDF]") right?
(2008-03-23 22:24:46) kanzure: if the name of the link was [PDF], yes
(2008-03-23 22:24:49) kanzure: but PDF is not part of the link name
(2008-03-23 22:24:51) kanzure: it's just 'near' it
(2008-03-23 22:25:02) fenn: unless you want to do something like ~= <b>\[PDF\]</b>.*<a href="(.*)">
(2008-03-23 22:25:07) kanzure: so I'm thinking of splitting up the string
(2008-03-23 22:25:12) kanzure: exploding it just before [PDF]
(2008-03-23 22:25:16) kanzure: and then after the next </a>
(2008-03-23 22:25:18) fenn: oh, 'follow' means click on the link
(2008-03-23 22:25:23) kanzure: right
(2008-03-23 22:25:57) kanzure: will your regexp work?
(2008-03-23 22:26:12) kanzure: $1 would be my url, right?
(2008-03-23 22:26:22) fenn: ~= '<b>\[PDF\]</b>.*<a href="(.*)">(.*)</a>' $1 is the url $2 is the link text
(2008-03-23 22:27:19) kanzure: <td valign=top><p class=g><span class="w"><font size=-2><b>[PDF]</b></font>&nbsp;<a href="http://www.stat.yale.edu/~jtc5/GeneticNetworksGroup/GibsonAndBruck2000.pdf" onmousedown="new Image().src='/scholar_url?sa=T&url=http://www.stat.yale.edu/~jtc5/GeneticNetworksGroup/GibsonAndBruck2000.pdf';"><b>Efficient exact stochastic simulation </b>of <b>chemical systems </b>with <b>many species </b>and <b>many channels</b></a></span> - <a class=fl href="/scholar?hl=en&lr=&cluster=216899736223511881">all 10 versions &raquo;</a><font size=-1><br><span class="a">MA Gibson, J Bruck - J. Phys. Chem. A, 2000 - stat.yale.edu</span><br>Page 1. <b>Efficient</b> <b>Exact</b> <b>Stochastic</b> <b>Simulation</b> of <b>Chemical</b> <b>Systems</b> with <b>Many</b><br>
(2008-03-23 22:27:19) kanzure: <b>Species</b> and <b>Many</b> <b>Channels</b> Michael A. Gibson* and Jehoshua <b>...</b>
(2008-03-23 22:27:21) kanzure: anyway, it gets nasty
(2008-03-23 22:27:29) kanzure: so you can't assume it's <a href=""> precisely
(2008-03-23 22:27:53) kanzure: hm, it's kind of smart with the "on-mouse-down" function
(2008-03-23 22:28:05) kanzure: but that's no problem - that's what the follow() idea is, no? :)
(2008-03-23 22:28:07) kanzure: or the "click" hehe
(2008-03-23 22:28:59) fenn: it gives the url in the <a href="
(2008-03-23 22:29:01) kanzure: ooh
(2008-03-23 22:29:04) kanzure: yes, but
(2008-03-23 22:29:07) kanzure: look at what they do, they have this:
(2008-03-23 22:29:15) kanzure: <a href="http://www.stat.yale.edu/~jtc5/GeneticNetworksGroup/GibsonAndBruck2000.pdf" onmousedown="new Image().src='/scholar_url?sa=T&url=http://www.stat.yale.edu/~jtc5/GeneticNetworksGroup/GibsonAndBruck2000.pdf';">
(2008-03-23 22:29:30) fenn: you want to ignore that? or what?
(2008-03-23 22:29:32) kanzure: <a href="xxx" onmousedown="xxx">
(2008-03-23 22:29:37) kanzure: erm
(2008-03-23 22:29:40) kanzure: <a href="xxx" onmousedown="yyy">
(2008-03-23 22:29:51) kanzure: where we want the xxx, and then after the yyy"> we can get the title
(2008-03-23 22:30:20) fenn: ~= '<b>\[PDF\]</b>.*<a href="(.*)" onmousedown="(.*)"' $1 is the url $2 is the onmousedown
(2008-03-23 22:30:39) kanzure: do I have to say $var ~= ... ?
(2008-03-23 22:30:58) Enki-2 left the room (quit: Read error: 104 (Connection reset by peer)).
(2008-03-23 22:31:02) fenn: $var should be the whole document i think
(2008-03-23 22:31:09) kanzure: ah, right
(2008-03-23 22:31:09) kanzure: okay
(2008-03-23 22:31:13) fenn: ~= loads all that crap into special variables
(2008-03-23 22:31:30) fenn: did i mention i hate perl? :)
(2008-03-23 22:31:58) kanzure: hehe
(2008-03-23 22:33:18) fenn: oh joy they use single quotes in the url
(2008-03-23 22:34:32) kanzure: hm?
(2008-03-23 22:37:41) Enki-2 [n=weechat@c-71-234-190-248.hsd1.ct.comcast.net] entered the room.
(2008-03-23 22:39:20) kanzure: $searchresults ~= '<b>\[PDF\].*<a href="(.*)" onmousedown="(.*)">(.*)</a>';
(2008-03-23 22:39:23) kanzure: parse error
(2008-03-23 22:39:42) kanzure: hm
(2008-03-23 22:44:39) fenn: ah i knew it would be something stupid
(2008-03-23 22:44:44) fenn: =~ not ~=
(2008-03-23 22:47:17) fenn: http://fennetic.net/pub/irc/test.pl
(2008-03-23 22:53:03) kanzure: hahaha
(2008-03-23 22:53:06) kanzure: this is awesome
(2008-03-23 22:53:58) fenn: i wish they would have used a different operator
(2008-03-23 22:55:01) kanzure: epitron, fenn -- Cosmic coincidences ;) http://heybryan.org/shots/2008-03-23_autoscholar.png
(2008-03-23 22:55:46) fenn: who is eugen?
(2008-03-23 22:55:51) kanzure: I've never talked about him?
(2008-03-23 22:55:56) kanzure: http://eugen.leitl.org/ see for yourself
(2008-03-23 22:56:15) kanzure: he's also mentioned here - http://heybryan.org/stats/2008-03-23-extropy-chat.html
(2008-03-23 22:56:31) kanzure: he's also mentioned on http://heybryan.org/mediawiki/index.php/Roadmap
(2008-03-23 22:56:37) kanzure: http://heybryan.org/mediawiki/index.php/Getting_up_to_speed too (I think)
(2008-03-23 22:57:16) fenn: more like synthetic serendipity ;)
(2008-03-23 22:57:36) kanzure: Eugen Leitl is sort of a polymath.
(2008-03-23 22:57:41) kanzure: Good guy to know.
(2008-03-23 22:58:10) fenn: many a time i've come across chat logs of me asking the same question i was googling for
(2008-03-23 22:58:25) 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.
(2008-03-23 22:58:30) kanzure: same here
(2008-03-23 22:58:55) fenn: well, what do you do about that?
(2008-03-23 22:59:03) kanzure: hide
(2008-03-23 22:59:08) kanzure: but I have been a moron and am already on the net
(2008-03-23 22:59:18) kanzure: in fact, there's a very significant portion of content generated by me on the net
(2008-03-23 22:59:18) kanzure: heh
(2008-03-23 22:59:20) fenn: hm.. secrecy doesn't work
(2008-03-23 22:59:23) kanzure: right
(2008-03-23 22:59:33) kanzure: it seems that the best way to go about it is to just be open as much as possible
(2008-03-23 22:59:35) fenn: better to make yourself indispensable
(2008-03-23 23:00:01) fenn: and dont get involved in shady criminal activities as a hobby
(2008-03-23 23:00:17) fenn: because then there's a reasonable excuse for your mysterious disappearance
(2008-03-23 23:00:23) kanzure: secretly, I was a script kiddie when I was 11
(2008-03-23 23:00:47) fenn: there werent any scripts when i was 11 :(
(2008-03-23 23:00:51) kanzure: old fart
(2008-03-23 23:00:56) fenn: i'm not that old (25)
(2008-03-23 23:01:19) kanzure: okay, so the script is mostly working
(2008-03-23 23:01:25) kanzure: it gets the wrong URL though
(2008-03-23 23:01:53) kanzure: it gets the stuff after a href=" but then doesn't stop at the first "
(2008-03-23 23:02:06) kanzure: $searchresults =~ '<b>\[PDF\]</b>.*<a href="(.*)" onmousedown="(.*)">(.*)</a>';
(2008-03-23 23:02:28) kanzure: so I need those quotation marks to match exactly
(2008-03-23 23:02:30) fenn: oh, probably greedy vs non-greedy matching
(2008-03-23 23:02:36) kanzure: should I add /g to the end?
(2008-03-23 23:02:43) kanzure: or /ig ?
(2008-03-23 23:02:51) fenn: g is global?
(2008-03-23 23:02:53) kanzure: oh
(2008-03-23 23:02:54) kanzure: hrm
(2008-03-23 23:03:17) fenn: (.*?) i think
(2008-03-23 23:03:21) kanzure: oh
(2008-03-23 23:03:22) kanzure: no,
(2008-03-23 23:03:27) kanzure: =~ m/stuff/;
(2008-03-23 23:04:05) fenn: not following you
(2008-03-23 23:04:25) kanzure: nope, that didn't work
(2008-03-23 23:04:44) kanzure: you're right
(2008-03-23 23:07:28) kanzure: wget --output-file is not useful
(2008-03-23 23:07:34) kanzure: it just routes the wget output to that file
(2008-03-23 23:07:36) kanzure: not the downloaded file
(2008-03-23 23:08:50) fenn: --output-document
(2008-03-23 23:09:47) kanzure: Alright, neat. Now I need to figure out a way to fetch a document if there is no PDF available.
(2008-03-23 23:12:07) 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 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.
(2008-03-23 23:12:55) fenn: But (EOL)
(2008-03-23 23:14:47) kanzure: eh
(2008-03-23 23:14:55) 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.
(2008-03-23 23:15:14) kanzure: this seems pretty simple in comparison
(2008-03-23 23:15:35) kanzure: if javascript execution works
(2008-03-23 23:16:09) fenn: does WWW::Mechanise run javascript?
(2008-03-23 23:16:23) kanzure: asking #perl atm
(2008-03-23 23:16:46) kanzure: ah, shit, no
(2008-03-23 23:16:50) kanzure: http://search.cpan.org/dist/WWW-Mechanize/lib/WWW/Mechanize/FAQ.pod
(2008-03-23 23:18:08) fenn: I don't use JavaScript myself, so I don't have an itch to scratch. <- wtf does he use WWW::Mechanize for then?
(2008-03-23 23:20:08) kanzure: hahah
(2008-03-23 23:20:24) fenn: read 'So what can I do?' - is the url in the html source somewhere? (even if its not standard)
(2008-03-23 23:20:38) kanzure: yes, I figured I'd have to do that anyway
(2008-03-23 23:20:41) kanzure: and it turns out it's pretty simple in this case
(2008-03-23 23:20:44) kanzure: it's just a form call hehe
(2008-03-23 23:20:54) fenn: is it the same from publisher to publisher?
(2008-03-23 23:22:27) 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)
(2008-03-23 23:26:43) fenn: i think 'socialized eng kdb' should be changed, because 'socialized' actually means something rather different
(2008-03-23 23:27:22) fenn: socio-economic system in which property and the distribution of wealth are subject to control by the community
(2008-03-23 23:27:58) fenn: but you mean, distributed development and distributed knowledge right?
(2008-03-23 23:28:02) kanzure: right
(2008-03-23 23:28:10) kanzure: socialized knowledge as in, "social knowledge"
(2008-03-23 23:28:15) kanzure: social as in "community-known"
(2008-03-23 23:28:25) kanzure: but I'm good for a change
(2008-03-23 23:30:56) fenn: societal engineering knowledge db?
(2008-03-23 23:31:21) kanzure: sure
(2008-03-23 23:33:07) 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
(2008-03-23 23:34:57) kanzure: my local "should be a programmer but isn't yet" buddy just sent me this: http://www.osnews.com/images/comics/wtfm.jpg
(2008-03-23 23:37:01) kanzure: Bill: so like apt-get but for science journals
(2008-03-23 23:37:04) kanzure: smart guy too
(2008-03-23 23:38:13) fenn: heh. good for ui testing at least
(2008-03-23 23:38:32) kanzure: Java.swing :(
(2008-03-23 23:38:37) kanzure: I will never again use java swing
(2008-03-23 23:38:38) kanzure: ever
(2008-03-23 23:39:38) fenn: but what about the cool sketch-on-a-napkin skin?
(2008-03-23 23:39:44) kanzure: hm?
(2008-03-23 23:39:47) kanzure: xkcd ref?
(2008-03-23 23:40:01) fenn: http://napkinlaf.sourceforge.net/
(2008-03-23 23:41:20) fenn: good for a laf at least :)
(2008-03-23 23:41:46) kanzure: har har
(2008-03-24 00:03:32) kanzure: fenn: does .* have to match anything?
(2008-03-24 00:03:50) fenn: please restate the question in the form of an answer
(2008-03-24 00:03:53) kanzure: hm
(2008-03-24 00:04:23) fenn: .* means any character, any number of times
(2008-03-24 00:04:43) kanzure: see, I want to now match for <a href="(.*?)"(.*?)>.*PDF.*</a>
(2008-03-24 00:04:57) kanzure: so that I get the URL to links that have 'PDF' in their name
(2008-03-24 00:05:13) fenn: </a> might belong to a url that is later in the file
(2008-03-24 00:05:28) kanzure: hrm
(2008-03-24 00:05:29) kanzure: okay
(2008-03-24 00:05:34) kanzure: so I guess I can do (.*?) still
(2008-03-24 00:05:38) fenn: so i think the second .* should be .*?
(2008-03-24 00:05:43) kanzure: <a href="(.*?)"(.*?)>.*PDF.*?</a>
(2008-03-24 00:05:47) fenn: yes
(2008-03-24 00:05:51) kanzure: I don't really need the second (.*?)
(2008-03-24 00:05:57) kanzure: <a href="(.*?)".*?>.*PDF.*?</a>
(2008-03-24 00:06:11) fenn: () just changes what gets stored in special variables
(2008-03-24 00:06:15) kanzure: right
(2008-03-24 00:06:19) kanzure: so I just need the href variable
(2008-03-24 00:06:35) fenn: (in this example)
(2008-03-24 00:07:21) fenn: uh, it might be a bit late to mention that there are probably lots of xml or html parsing tools
(2008-03-24 00:07:30) kanzure: I was looking into them
(2008-03-24 00:07:41) kanzure: the perl community is obsessed with LinkExtor
(2008-03-24 00:07:48) kanzure: but it does not return the link name
(2008-03-24 00:07:55) kanzure: and for some reason I didn't think to go more generalized with XML
(2008-03-24 00:08:48) fenn: i think you need ? on all the .*'s
(2008-03-24 00:12:19) kanzure: if () { /* stuff * / } else if ($searchresults =~ /Get this article/) { } <--- parse error at } else if () { ........ wtf?
(2008-03-24 00:13:57) fenn: / get this article/ probably has a / in it somewhere
(2008-03-24 00:14:08) fenn: you can use other characters to bound the regexp
(2008-03-24 00:14:45) kanzure: ah, it was something else entirely
(2008-03-24 00:17:41) kanzure: how ridiculous
(2008-03-24 00:17:49) fenn: how Perl
(2008-03-24 00:17:51) kanzure: my password, I have to pass it to wget with --proxy-pass
(2008-03-24 00:17:55) kanzure: and it has a bang in it
(2008-03-24 00:17:57) kanzure: for some reason
(2008-03-24 00:18:00) kanzure: and this messes up bash
(2008-03-24 00:18:03) kanzure: apparently I can't escape it
(2008-03-24 00:18:06) kanzure: with \!
(2008-03-24 00:18:17) fenn: !! i think
(2008-03-24 00:18:21) kanzure: ugh
(2008-03-24 00:18:35) fenn: no that's no it
(2008-03-24 00:18:40) kanzure: hm
(2008-03-24 00:18:55) kanzure: echo ! works though
(2008-03-24 00:19:03) kanzure: oh
(2008-03-24 00:19:12) fenn: are you doing double quotes? try single
(2008-03-24 00:19:52) kanzure: ah
(2008-03-24 00:20:55) kanzure: wget still says it is ambiguous though
(2008-03-24 00:21:03) kanzure: wget --proxy-pass=hi!
(2008-03-24 00:21:18) fenn: wget --proxy-pass='hi!'
(2008-03-24 00:21:21) kanzure: right
(2008-03-24 00:21:22) kanzure: even that
(2008-03-24 00:21:59) fenn: it's --proxy-password
(2008-03-24 00:22:09) kanzure: bleh
(2008-03-24 00:22:50) kanzure: oh
(2008-03-24 00:22:53) kanzure: I just realized something
(2008-03-24 00:23:00) kanzure: Google pulls up numerous documents that are related to the title
(2008-03-24 00:23:05) kanzure: so you have to do *exact* title searches
(2008-03-24 00:23:09) kanzure: as in, intitle:"name"
(2008-03-24 00:23:14) kanzure: where "name" is the name of the article
(2008-03-24 00:23:17) kanzure: else you will get PDFs from other things
(2008-03-24 00:45:22) kanzure: fenn:
(2008-03-24 00:45:22) kanzure: http://search.cpan.org/dist/WWW-Mechanize/lib/WWW/Mechanize.pm#$mech->redirect_ok()
(2008-03-24 00:45:32) kanzure: you think there's a way to figure out the 'current URL' that it is sitting at ?
(2008-03-24 00:46:16) kanzure: ah, nevermind
(2008-03-24 01:01:16) kanzure: fenn: Here's my regexp, $r3 =~ '<a.*?href="(.*?)">.*?PDF.*?</a>';
(2008-03-24 01:01:16) kanzure: and here's what I need to match: <a title="Download PDF" class="articlenav2" target="_top" href="/nature/journal/v347/n6293/pdf/347539a0.pdf">Download PDF</a>
(2008-03-24 01:01:23) kanzure: for some reason, I am only getting back "/nature/" ...
(2008-03-24 01:20:18) fenn: dunno
(2008-03-24 01:20:41) fenn: $1 = "/nature/" ?
(2008-03-24 01:22:45) fenn: also print $r3 and make sure it's what you expect
(2008-03-24 01:23:51) kanzure: of course
(2008-03-24 01:23:56) kanzure: #perl has convinced me to switch to LinkExtractor
(2008-03-24 01:24:01) kanzure: http://search.cpan.org/~podmaster/HTML-LinkExtractor-0.13/LinkExtractor.pm
(2008-03-24 01:24:09) kanzure: now I'm figuring out how to access the _TEXT variable
(2008-03-24 01:24:43) kanzure: foreach my $x ($LX->links) { $x->_TEXT; } or something ...
(2008-03-24 01:26:31) fenn: huh. linkextractor just dumps you back in the same situation
(2008-03-24 01:26:41) fenn: there's no variable containing 'I am a LINK!!!'
(2008-03-24 01:27:16) kanzure: _TEXT is it.
(2008-03-24 01:27:21) kanzure: you just have to strip out the HTML.
(2008-03-24 01:27:41) fenn: then why bother with linkextractor :)
(2008-03-24 01:27:56) kanzure: blah, #perl is telling me to go back to the newbie tutorials
(2008-03-24 01:28:12) kanzure: I just want to get the _TEXT variable :(
(2008-03-24 01:28:19) kanzure: I thought it might be $x{_TEXT} but that wasn't it either
(2008-03-24 01:28:37) kanzure: it's a hashref, I know that much
(2008-03-24 01:29:33) fenn: $LX->links{_TEXT}?
(2008-03-24 01:30:02) fenn: for my $Link( @{ $LX->links } ) {
(2008-03-24 01:30:02) fenn: ## new modules are linked by /author/NAME/Dist
(2008-03-24 01:30:02) fenn: if( $$Link{href}=~ m{^\/author\/\w+} ) {
(2008-03-24 01:30:03) fenn: print $$Link{_TEXT}."\n";
(2008-03-24 01:30:20) fenn: change author to PDF
(2008-03-24 01:30:45) kanzure: bleh
(2008-03-24 01:30:49) kanzure: well, not quite
(2008-03-24 01:31:16) kanzure: $$Link{href} is assuming there's a pdf in the URL
(2008-03-24 01:31:17) fenn: i forget how to get an element from an array
(2008-03-24 01:31:58) fenn: oh, change href to _TEXT then
(2008-03-24 01:32:29) fenn: and then print the href :)
(2008-03-24 01:32:37) kanzure: yep
(2008-03-24 01:32:45) kanzure: hm, there's a few other parts of the code I have to fix
(2008-03-24 01:32:51) kanzure: because nature.com is a bitch and provides relative links
(2008-03-24 01:33:02) kanzure: but that's simple enough
(2008-03-24 01:55:35) kanzure: fenn: It works.
(2008-03-24 01:56:13) 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.
(2008-03-24 01:56:33) kanzure: (2) Doesn't yet do a list of inputs, but this is easy.
(2008-03-24 02:05:35) fenn: know anything about webs of trust?
(2008-03-24 02:05:58) fenn: is it possible to 'import' a web that's already existing? for example robots.net or advogato
(2008-03-24 02:07:44) kanzure: don't know what you are talking about
(2008-03-24 02:08:04) fenn: if software doesn't work it's no big deal
(2008-03-24 02:08:26) fenn: but say you have a big dependency tree leading up to some hardware project, that costs money and time and effort
(2008-03-24 02:08:47) fenn: so you'd like to know in advance the trustworthiness of the author (in the field of the invention)
(2008-03-24 02:09:44) fenn: there exist already, open systems to gauge the amount of trust an individual has acquired
(2008-03-24 02:09:45) kanzure: http://heybryan.org/projects/autoscholar/
(2008-03-24 02:10:48) 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
(2008-03-24 02:11:07) fenn: and they wont necessarily contribute to our project and become center nodes
(2008-03-24 02:11:54) fenn: so it'd be better if we could use the whole web of trust from elsewhere that already exists
(2008-03-24 02:12:12) fenn: rather than trying to build community and trust all over again for every web site
(2008-03-24 02:12:45) fenn: i get annoyed about having to make new logins all the time
(2008-03-24 02:13:30) kanzure: have you heard of OpenID?
(2008-03-24 02:14:04) fenn: nope
(2008-03-24 02:15:11) ***fenn looks
(2008-03-24 02:18:09) kanzure: oh shit
(2008-03-24 02:18:13) kanzure: yeah, that's a good thing to know about
(2008-03-24 02:18:17) kanzure: http://claimid.com/ is what I use
(2008-03-24 02:18:20) kanzure: but there are others out there
(2008-03-24 02:18:22) kanzure: probably a few that are more legit
(2008-03-24 02:35:36) fenn_ [n=pz@adsl-75-62-114-103.dsl.bltnin.sbcglobal.net] entered the room.
(2008-03-24 02:35:54) fenn left the room (quit: Nick collision from services.).
(2008-03-24 02:36:06) fenn_ is now known as fenn
(2008-03-24 02:44:02) fenn_ [n=pz@adsl-75-60-174-103.dsl.bltnin.sbcglobal.net] entered the room.
(2008-03-24 02:49:18) fenn__ [n=pz@adsl-75-62-114-201.dsl.bltnin.sbcglobal.net] entered the room.
(2008-03-24 02:57:01) fenn left the room (quit: Read error: 110 (Connection timed out)).
(2008-03-24 03:04:28) fenn_ left the room (quit: Read error: 110 (Connection timed out)).
(2008-03-24 03:19:38) 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
(2008-03-24 04:13:38) fenn [n=pz@adsl-75-60-172-195.dsl.bltnin.sbcglobal.net] entered the room.
(2008-03-24 04:25:09) fenn_ [n=pz@adsl-76-248-68-214.dsl.bltnin.sbcglobal.net] entered the room.
(2008-03-24 04:29:41) fenn__ left the room (quit: Read error: 110 (Connection timed out)).
(2008-03-24 04:41:36) fenn left the room (quit: Read error: 110 (Connection timed out)).
(2008-03-24 04:52:28) fenn [n=pz@adsl-75-60-175-237.dsl.bltnin.sbcglobal.net] entered the room.
(2008-03-24 05:13:49) fenn left the room (quit: Read error: 104 (Connection reset by peer)).
(2008-03-24 05:15:08) fenn [n=pz@adsl-76-251-84-195.dsl.bltnin.sbcglobal.net] entered the room.
(2008-03-24 05:16:38) fenn_ left the room (quit: Read error: 110 (Connection timed out)).
(2008-03-24 05:43:20) fenn left the room (quit: Read error: 104 (Connection reset by peer)).
(2008-03-24 05:43:58) fenn [n=pz@adsl-76-251-86-219.dsl.bltnin.sbcglobal.net] entered the room.
(2008-03-24 05:55:16) fenn_ [n=pz@adsl-75-62-113-247.dsl.bltnin.sbcglobal.net] entered the room.
(2008-03-24 06:11:09) fenn left the room (quit: Read error: 110 (Connection timed out)).
(2008-03-24 06:18:09) kanzure: probably
(2008-03-24 06:18:32) fenn_ left the room (quit: Read error: 110 (Connection timed out)).
(2008-03-24 06:31:18) Enki-2 left the room (quit: Read error: 104 (Connection reset by peer)).
(2008-03-24 06:37:31) Enki-2 [n=weechat@c-71-234-190-248.hsd1.ct.comcast.net] entered the room.