00:00:08 <alise> i was only born 4 years after MI2 came out ;)
00:00:32 <alise> so half a decade after MI1 :P
00:00:44 <Vorpal> alise, lets see, when was that
00:00:48 <alise> but who noticed that sort of crap as a kid?
00:00:52 <alise> Vorpal: i was born in 1995.
00:00:55 <alise> MI1 came out in 1990; MI2, 1991.
00:02:36 <Vorpal> hm I can't say I played MI1 either, being slightly less than 1 year I couldn't really get enough pressure on the keyboard however much I tried. ;)
00:03:26 <alise> i had MI1/MI2/MI3 on three CDs (MI3 took up two CDs) and copied save games to a floppy disk :)
00:03:36 <alise> (separated parents, so i took it along... ...yeah, I was a nerd.)
00:03:49 <alise> and I had Universal Hint Service with the data files on another floppy
00:04:20 <Vorpal> alise, that's cheating
00:04:36 <alise> no, that's being young and having no fucking idea for a puzzle for something like two hours straight
00:04:41 <alise> and remember UHS gave very obscure hints first
00:04:43 <alise> then slightly less obscure
00:04:47 <alise> until it finally told you the literal solution
00:04:52 <alise> so often it served as a mere memory jog
00:05:14 <Vorpal> I never played MI, never got around to it
00:05:23 <Vorpal> so many other things to do.
00:06:23 <Vorpal> alise, and I'm more into RPGs than point and click anyway
00:07:18 <Vorpal> alise, anyway I currently use tmp as a real tmp in my current home dir
00:07:23 <Vorpal> and move stuff to videos or such
00:21:12 -!- FireFly has quit (Quit: swatted to death).
00:25:33 <alise> Wow, they've revamped tripod.com...
00:26:03 <alise> At least the frontpage design.
00:26:05 <alise> http://www.tripod.lycos.com/
00:26:14 <alise> "Tripod - Succeed Online"
00:26:20 <alise> Example site: "Ultra Mega Pizza".
00:26:39 <alise> I like the random person trapped in the webpage, trying to climb out.
00:27:57 -!- Wamanuz has quit (Remote host closed the connection).
00:30:53 <Gregor> alise: Web pages eat tens of people every year.
00:31:00 <Gregor> alise: It is a serious problem!
00:31:40 -!- Mathnerd314 has joined.
00:56:39 -!- Sgeo|web has joined.
01:01:09 -!- quintopia has joined.
01:08:27 <Gregor> quintopia: "Hi" is unlikely to elicit a reaction from most of the people in this channel.
01:09:03 <quintopia> i didn't have to elicit a reaction. people were already talking.
01:09:10 <oerjan> well, i _considered_ making the lo pun again
01:10:06 <quintopia> someone explain how fungot has a net connection? i wasn't aware funge98 had networking?
01:10:06 <fungot> quintopia: the interesting parts of the library and use a loop to write in the program we continue to), which is frowned upon.
01:10:15 <Gregor> oerjan: "Lo Pun" would be a good fake name for Chinese food :P
01:10:27 <oerjan> it uses a fingerprint for networking, i think
01:10:36 <Gregor> quintopia: I don't know about fungot in particular, but I imagine its stdout and stdin are just mapped to a socket connected to IRC.
01:10:36 <fungot> Gregor: i found a way to dynamically bind it.
01:10:40 <quintopia> oerjan: it's okay. i use lo as the default response too.
01:10:50 <oerjan> unless it just uses netcat and stdin ... right
01:11:09 <quintopia> that would, uh, be the smart thing
01:11:22 <Gregor> Works well for IRC anyway.
01:11:28 <Gregor> For other protocols it would ... not work so well :P
01:11:42 <oerjan> quintopia: well fungot already uses lots of _other_ fingerprints i believe, so it wouldn't be strange if it also used a networking one
01:11:42 <fungot> oerjan: the only funny knock knock jokes are the funniest... i'm into software!
01:12:07 <oerjan> i'm not entirely sure about this.
01:12:43 <oerjan> i recall ais523's thutubot used the stdin/stdout trick though, since thutu has only simple I/O
01:14:13 <oerjan> quintopia: but as a general rule, i think funge98 is one of the implemented esolangs with most available extra stuff. maybe modern INTERCAL can compete.
01:14:37 <oerjan> the majority of esolangs only do stdin/stdout at best
01:14:54 <Sgeo|web> Hey, add PSOX, and they can do more!
01:17:00 <fungot> oerjan: i don't think you need just some 50 more metres and i'm home. away, must to h to take care of family business first), what is
01:17:22 <oerjan> 's first responses looked eerily appropriate again
01:17:44 * quintopia notes the annoying placement of ' on most keyboards
01:18:00 <Sgeo|web> If I use Ubuntu, can I load it into memory and then take out the disk?
01:18:52 <quintopia> sgeo: if you use a stripped-down-enough build, possibly. most ubuntus are very fat tho
01:18:57 <alise> <Gregor> quintopia: I don't know about fungot in particular, but I imagine its stdout and stdin are just mapped to a socket connected to IRC. ;; wrong
01:18:58 <fungot> alise: reminds me of something my dad once ran a script on the website somewhere
01:19:00 <alise> it uses SOCK and SCKE
01:19:22 <alise> <oerjan> quintopia: but as a general rule, i think funge98 is one of the implemented esolangs with most available extra stuff. maybe modern INTERCAL can compete. ;; the only networking INTERCAL can do is its crazy INTERCAL-specific system of theft
01:19:59 -!- alise has quit (Quit: Leaving).
01:20:07 <oerjan> alise: doesn't INTERCAL have an ffi though... bah
01:20:25 * Sgeo|web wonders what alise's claimed advantages of Ubuntu over Knoppix are
01:20:30 <Sgeo|web> I don't need installability right now
01:21:17 <oerjan> no idea, it was just too obvious not to say
01:21:22 <Sgeo|web> Anyways, I have a disk with ddrescue
01:21:36 <Sgeo|web> And I can always burn Ubuntu later
01:25:44 <Sgeo|web> Also, I haven't used KDE in a very long time
01:25:57 <Sgeo|web> And Knoppix helped me get into Linux in the first place
01:29:17 <ais523> oerjan: C-INTERCAL has FFIs to C and Befunge-98
01:29:45 <ais523> also, CLC-INTERCAL supports good old-fashioned TCP as well, as an extension
01:32:52 -!- Sgeo|web has quit (Ping timeout: 252 seconds).
01:34:36 <quintopia> i don't really get why people put so much effort into INTERCAL
01:35:05 <ais523> to show serious language designers/ecosystems what they /should/ be doing
01:35:13 <quintopia> it doesn't really demonstrate any interesting possibilities in computing
01:35:39 <quintopia> it being so incredibly dreadfully slow at the most basic compute tasks
01:37:36 <ais523> INTERCAL is much faster at arithmetic
01:37:40 <ais523> because it operates in binary rather than unary
01:38:20 <ais523> adding two integers, for instance, is O(n) in BF, O(log n) in INTERCAL
01:38:28 <ais523> although admittedly, in a saner language it would be O(1)
01:39:51 <ais523> O(log n) for both for bignums
01:40:08 <ais523> but it's probably simpler in INTERCAL
01:40:41 <quintopia> perhaps it was the implementations then
01:40:43 <pikhq> ais523: Well, except for optimising implementations of BF, which make addition O(1).
01:40:47 <pikhq> But I'm not sure that counts.
01:42:37 <ais523> pikhq: except that optimising implementations of INTERCAL make (standard library) addition O(1) too?
01:42:47 <ais523> so it's not really a sensible way to compare
01:43:04 <ais523> (C-INTERCAL doesn't currently optimise the single-statement multithreaded version)
01:43:15 <quintopia> actually, order approximations are not a good way to compare in general
01:43:50 <quintopia> by order approximations, apples and oranges are equivalent
01:44:06 <ais523> well, if order approximations show things are different, they're /different/
01:44:20 <ais523> and everything else depends on things like the relative speeds of the computers you test on, what languages you compile via, etc
01:44:31 <pikhq> ais523: And of course, a sufficiently optimising compiler may turn your complex NP problem into a simple run on a quantum coprocessor. :P
01:44:55 <quintopia> unless BPP is a strict subset of NP
01:45:12 <ais523> quintopia: pikhq didn't say NP-complete
01:45:37 <pikhq> And it is known that there are at least some NP problems with efficient quantum implementations.
01:45:42 <ais523> not to mention, NP-complete problems tend to be solved quite quickly in practice in most cases
01:45:59 <ais523> obviously not in general, but you can come up with algorithms that work, say, 99% of the time
01:46:07 <quintopia> well, all problems of a fixed size are O(1)
01:47:08 <quintopia> ah, you mean like how the Simplex alg has exponential time worst cases for some convex bodies under ANY heuristic?
01:47:14 <quintopia> (isn't that the way that result works?)
01:48:24 <ais523> also, the way that you can efficiently come within quite a narrow margin of error, 5% or so, in the travelling salesman problem
01:48:53 <ais523> yes, that's a different point
01:48:54 <quintopia> directed TSP doesn't have anything better than a 2-approx at the moment, iirc
01:49:33 <quintopia> there was a neat algorithm involving simplifying spanning trees to make it a non-directed alg, it may have been slightly better than 2
01:49:43 <quintopia> but either way, there is certainly no PTAS at present
02:02:49 -!- augur has joined.
02:12:47 <quintopia> what would be the best sort of programmable chip to get on which to implement my own ISA?
02:13:17 <pikhq> quintopia: FPGA of course.
02:14:13 <quintopia> specific brand though, in terms of speed, price, ease of use?
02:14:20 <quintopia> is there one with a built-in clock?
02:18:18 <ais523> although you generally get them on evaluation boards which have clocks
02:18:31 <ais523> the manufacturers seem to go out of their way to make them hard to use, though
02:18:42 <ais523> at work, me and my boss spent over a month trying to set one up and failing
02:18:53 <ais523> in the end, they hired someone who'd already managed it at a different job to work on it fulltime
02:19:05 <quintopia> i want one i can create a circuit and sim it on my computer, and then just plug in the chip, press a button, and it'll program it
02:19:12 <ais523> I'm not sure how it went, but it's telling that the board that's in there /now/ is by a different manufacturer
02:22:12 -!- Mathnerd314 has quit (Ping timeout: 245 seconds).
02:35:09 -!- Sgeo|PM has joined.
02:35:37 <Sgeo|PM> According to wget, 250M has been downloaded
02:35:53 <Sgeo|PM> According to Task Manager, 159MB is in use
02:41:55 <quintopia> alright what's the best regular ole user-programmable chip out there in terms of price/speed/ISA feature set?
02:44:17 -!- cal153 has quit (Ping timeout: 276 seconds).
02:49:37 <Sgeo|PM> Dear task manager: There is no way in fscking hell that only 183MB are being used right now
02:50:26 <Sgeo|PM> What's the easiest way to find out where / is mounted?
02:52:32 <oerjan> (the ? was not part of the command btw)
02:53:08 <Sgeo|PM> How usable as a temporary OS is DSL?
02:53:26 <Sgeo|PM> More usable than, say, Parted Magic?
02:53:52 <quintopia> i suspect puppy would be the best bet here
02:54:24 * Sgeo|PM does not particularly feel like booting up his old comp
03:02:49 <Sgeo|PM> I can save to my writable DVD with Puppy?!?!?!?
03:03:48 <quintopia> that seems tricky. you may need to install some things.
03:05:37 <Sgeo|PM> Parted Magic thinks that Chromium should be the default filesystem browser
03:09:09 <Sgeo|PM> Let me write to the DVD, dang you!
03:14:42 -!- augur has quit (Ping timeout: 245 seconds).
03:18:00 <Sgeo|PM> DOES PARTED MAGIC NOT INCLUDE DRIVERS FOR WRITING TO DVDs?
03:18:08 * Sgeo|PM has old Knoppix lying around
03:18:20 <pikhq> Sgeo|PM: Odd; it should.
03:18:40 <pikhq> Protocol-wise, a DVD drive appears as a CD drive with a very large disc.
03:19:17 <Sgeo|PM> dd if=lupu-511.iso of=/dev/sr0
03:19:36 <Sgeo|PM> dd: opening `/dev/sr0': Read-only file system
03:19:57 <Sgeo|PM> Is there any way to confirm this GUI thingie's claim that the drive is /dev/sr0?
03:20:06 <Sgeo|PM> /dev/scd0 also gives the same thing
03:20:56 -!- Mathnerd314 has joined.
03:21:29 <Sgeo|PM> /dev/sr0 /media/cdrom1 auto defaults,ro 0 0 #cdrom
03:21:34 <pikhq> That's *not how you burn a DVD*.
03:21:47 <pikhq> Pull up the cdrecord man page.
03:23:45 <Sgeo|PM> -multi can't hurt, can it?
03:24:07 <pikhq> You probably want it for a Puppy disc.
03:25:10 <Sgeo|PM> cdrecord -multi lu...whatever the thing's name was
03:26:12 -!- Sgeo|PM has quit (Remote host closed the connection).
03:27:39 -!- cal153 has joined.
03:31:19 <quintopia> what would it take to convince skype to bring skype for linux up to version 5?
03:33:15 -!- bsmntbombdood has joined.
03:35:41 <pikhq> Eliminating all other OSes.
03:35:49 <pikhq> With sexual favors.
03:36:47 -!- Sgeo|PL has joined.
03:37:26 <oerjan> i guess sexual favors help with that, too.
03:37:33 <Sgeo|PL> This client feels like Pidgin :/
03:39:35 <Sgeo|PL> I think I'm in love with Puppy Linux
03:41:06 <pikhq> Sgeo|PL: Well, Puppy Linux is basically what happens when you start with Ubuntu and make it not suck.
03:41:35 <Gregor> Well hello there Windows. How about you ... cease existence and become a service that runs on top of Linux. Maybe if I ... /sweeten/ the deal?
03:42:05 <Sgeo|PL> PL's version of Firefox is old
03:42:22 <Sgeo|PL> version of Flash player, which it shouldn't have.
03:42:32 <Sgeo|PL> So I guess Firefox's start page is hallucinating
03:44:55 <Sgeo|PL> Can Flash be installed on Puppy Linux?
03:46:35 <pikhq> Should be in the package manager.
03:46:40 <pikhq> If not, apt-get install
03:47:09 <quintopia> well, ever since it came to pass that windows applications can hose your bootloader, there is no other safe way to run windows except in a VM on top of linux.
03:47:29 <Sgeo|PL> main, multiverse, universe?
03:49:00 <Sgeo|PL> Maybe I should attempt to get Knoppix to work :/
03:49:15 <Sgeo|PL> This is not particularly comfortable
03:53:03 <Sgeo|PL> Did Puppy Linux just come with Flash /preinstalled/?
03:53:09 <Sgeo|PL> I just downloaded the Flash .pet
03:53:19 <Sgeo|PL> Didn't confirm install yet
03:57:40 <Sgeo|PL> I really, really, like the lock thing
03:58:27 -!- Sgeo|PL has quit (Quit: Ayttm logging off).
04:06:53 -!- sshc_ has joined.
04:09:53 -!- sshc has quit (Ping timeout: 276 seconds).
04:16:13 -!- sshc has joined.
04:17:53 -!- sshc_ has quit (Ping timeout: 276 seconds).
04:18:28 * quintopia tries Illumination Software Creator
04:31:33 <quintopia> well, i'd like to try it anyway, but i, uh, can't figure out the name of the executable it installed
04:37:00 <pikhq> Erm. Insufficiently German.
04:40:37 -!- Sgeo|Puppy has joined.
04:41:38 <oerjan> You're cursed, i've said. You cannot be saved!
04:42:07 <pikhq> Didſt þou try archaic orþographic conventions?
04:43:11 <pikhq> oerjan: Hey, it's not *juſt* Japaneſe þat gets þe unuſual orþography treatment from me.
04:44:11 <pikhq> Pity I'm too lazy to uſe ð right.
04:44:56 <pikhq> It requires me to þink about þe phoneme þat way!
04:46:45 <Sgeo|Puppy> To find out more about the differences between misdemeanor and felony statutory rape, including the statutes of limitations on each, contact your state Attorney Generals Office.
04:47:04 <Sgeo|Puppy> Yeah, just go imply to the AG that you... wait, is there some sort of confidentiality thing there?
04:47:12 <Sgeo|Puppy> Surely it would make sense to ask a private lawyer for advice?
04:47:27 <Sgeo|Puppy> No, I have not committed statutory rape, in case anyone is wondering
04:50:37 * quintopia sucks the oil of oerjan's seeds to make paint
04:51:01 <quintopia> if at first you don't suck seed...
04:51:02 <pikhq> quintopia: Yeah, but they're all NORSE and stuff
04:51:40 <quintopia> THEY COME FROM THE LAND OF THE ICE AND SNOW WHERE THE MIDNIGHT SUN AND THE HOT SPRINGS BLOW
04:52:24 <oerjan> AND THE POLAR BEARS ROAM THE STREETS
04:55:45 <quintopia> "polar bear" and "jkullhlaupt" confused?
04:55:58 <quintopia> feel free to move that diaresis as needed
04:56:14 <Sgeo|Puppy> I need to configure touchpad behavior somehow
04:56:38 * oerjan adjusts quintopia's jokemeter
04:57:10 <Sgeo|Puppy> I tried some configuration thing (not gsynaptics, I think)
04:57:20 <Sgeo|Puppy> Complained about possibly a missing synaptics driver
04:57:35 <quintopia> you were making fun of possible uncommon misconceptions of iceland, whereas i was making fun of the fact that THE RING ROAD GETS WIPED OUT REGULARLY BY VOLCANOS
04:58:33 <oerjan> actually i wasn't particularly talking about iceland
04:58:50 <quintopia> but it's the last bastion of the norse!
04:59:15 <quintopia> and possibly the only source of norse epics!
05:01:37 <quintopia> what's the best personal programmable microchip in terms of balancing price with usability?
05:02:51 <quintopia> i don't need anything complicated. just good speed and a bunch of I/O pins.
05:02:58 <quintopia> and easy to use interface board/software
05:06:27 <Sgeo|Puppy> Into a witness-protection-program-like-thing
05:08:08 <quintopia> if you take care with hostmasks and routing and such, you could return to your old online identities being careful not to let them got back to your new real life identity
05:09:17 <Sgeo|Puppy> Was worried about the fact that I've revealed my RL identity, but that identity would be disappeared
05:15:25 * oerjan hopes this is very very hypothetical
05:16:30 * oerjan swats quintopia -----###
05:16:43 <Sgeo|Puppy> http://www.cbsnews.com/stories/2010/09/16/national/main6873992.shtml is what started me thinking about this
05:17:17 <Sgeo|Puppy> I promise it has no connection to my statute-of-limitations questions earlier
05:17:47 * oerjan doesn't feel like reading such stories anyhow
05:26:18 <pikhq> I do hereby declare US politics fundamentally broken. Nothing short of eugenics can fix it.
05:26:39 <Sgeo|Puppy> Did something happen in the last few minutes?
05:26:43 <quintopia> eugenics broke it in the first place
05:26:51 <pikhq> Sgeo|Puppy: Nothing politically notable, no.
05:27:12 <pikhq> Sgeo|Puppy: Just a long-delayed reaction to the sheer stupidity of EVERY SINGLE FUCKING THING.
05:27:49 <pikhq> Basically, US politics for the span of, oh, my *life* has been THE SINGLE MOST STUPID FUCKING THING.
05:28:13 <pikhq> "Hmm. The muslims hate us and want to kill us. What should we do?" "Make them hate us?" "BRILLIANT! BURN THE KORAN!"
05:28:21 <quintopia> your life is far too short of a span to measure the stupidity of american politics on
05:28:29 <quintopia> i suspect you were not alive for the red scare
05:28:45 <pikhq> quintopia: Ah, yes. It has been pretty stupid for the past century.
05:28:52 <pikhq> FFFFUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUu
05:29:15 <pikhq> Oh, wait, before that we had the argument that slavery was beneficial to black people.
05:29:20 <pikhq> FFFFFFFFFFFFFFFFFFFFUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU
05:29:28 <Sgeo|Puppy> At least most people consider the Red Scare an embarrasement
05:29:58 <Sgeo|Puppy> Except for some people on the board of education in Texas
05:30:49 <quintopia> pikhq: don't forget between that argument and the bomb, there was a legitimate govt-sanctioned eugenics program
05:32:05 <Sgeo|Puppy> "*install python first*" says a thing in my package manager
05:32:30 <quintopia> pikhq: perhaps you should take some time off and study the political histories of other countries, so that you can generalize "US politics are stupid" to "politics are stupid"
05:32:57 <pikhq> quintopia: I am somewhat familiar with other nations' politics. The US's is just DRAMATICALLY stupid.
05:33:18 <pikhq> Perhaps it's more just that the past's stupidity didn't get recorded as well as modern-day stupidity does.
05:33:27 <pikhq> But WHAT THE HELL PEOPLE.
05:33:40 <pikhq> "THEY HATE OUR RELIGIOUS FREEDOM NOW TAKE AWAY THEIR RELIGIOUS FREEDOM"
05:33:47 <pikhq> HOW DOES THIS MAKE ANY FUCKING SENSE YOU MORONS
05:34:18 <quintopia> at least i have the political freedom to choose which sort of stupidity i like best
05:34:43 <quintopia> it may be inefficient, but it's that or have some single person's stupidities imposed on me
05:35:02 <quintopia> i'd much rather be forced to go with the will of the many than the will of one
05:35:05 <pikhq> I welcome the robot uprising.
05:35:14 <pikhq> I fucking welcome the motherfucking robot uprising.
05:35:25 <quintopia> now that will be one thing the human race can be proud of
05:35:48 <ais523> UK politics is pretty stupid, but US politics is clearly even stupider
05:35:55 <quintopia> although that reminds of that early subnormality about the singularity...
05:36:01 <ais523> at least in the UK we get the occasional draw
05:36:05 <pikhq> ais523: And you guys have *actual, honest-to-god nazis*.
05:36:14 <pikhq> And somehow the US is stupider.
05:36:28 <ais523> pikhq: not actually in politics, not at the moment
05:36:50 <ais523> they got voted out again almost instantly when people noticed that they'd been voted in, everything before that was just a lack of paying attention
05:36:59 <pikhq> The BNP has two seats in the European Parliament still.
05:37:25 <ais523> the thing that riles me most about the BNP is that even if you completely ignore their policies on immigration, the rest of their policies are batshit insane too
05:37:27 <quintopia> and this one: http://www.viruscomix.com/page338.html
05:37:42 -!- Sgeo|Puppy has quit (Read error: Connection reset by peer).
05:38:03 <ais523> still, I'm glad that sort of party exists, just to help prevent the other parties collapsing in a mess in the middle US-style
05:38:06 -!- Sgeo|Puppy has joined.
05:38:22 <pikhq> ais523: They want to end the freaking EU. I cannot fathom the stupidity that could bring this about.
05:38:28 <ais523> the remaining parties have to be careful not to become /too/ identical in order to prevent the BNP (or at the other extreme the Communist Party) actually winning
05:38:39 <ais523> pikhq: that's actually quite a prevalent viewpoint in the UK
05:38:46 <ais523> even though I personally disagree with it
05:38:51 <pikhq> Oh, right, the UK has a bunch of morons.
05:39:13 <pikhq> That's been one of the few undeniably good things in international relations in the past, oh, century!
05:39:41 <ais523> the main argument against is that it's very expensive and doesn't really do a lot
05:40:12 <pikhq> Seems to me to have done rather a lot.
05:40:32 <pikhq> Schengen alone justifies it.
05:40:46 <pikhq> Except the UK isn't in the Schengen Zone.
05:41:06 <pikhq> Except the UK isn't on the Euro.
05:41:23 <ais523> I actually think the Euro is a bit of a dubious idea
05:41:25 <Sgeo|Puppy> Why can't the UK leave and the rest of the EU remain intact?
05:41:35 <ais523> in that it's nailed down a lot of the handles used to help manipulate the economy
05:41:49 <ais523> and the ECB are stuck trying to balance, say, Greece's and Germany's economy, and that doesn't really work
05:42:00 <pikhq> ais523: There's a few implementation flaws which have just now shown.
05:42:55 <pikhq> I'm going to guess, though, the ECB is a bit unused to handling a somewhat disparate economy.
05:43:09 <pikhq> Judging from the US, so's the Fed.
05:43:24 <ais523> part of the issue is that the economies in Europe simply don't make any sense with respect to each other
05:43:35 <ais523> in the UK, both prices and wages are higher than in most of mainland Europe, for instance
05:43:49 <ais523> this doesn't really make logical sense, as you'd assume people would just buy all their stuff abroad
05:43:57 <ais523> (some people actually do this, but most people can't be bothered)
05:43:58 <pikhq> Nor does the US. Keep in mind, it's not so much one nation as it is 50 nations in a federation.
05:44:25 <pikhq> Less so culturally, but very much so economically.
05:44:39 <ais523> yep, but I assume the economy's become at least slightly homogenous due to people being able to trade at will
05:44:47 <pikhq> Not even slightly.
05:45:14 <pikhq> In some areas, $30,000 is a living wage. In others, $100,000 will barely get you a roof over your head.
05:46:01 <quintopia> sorry this is late but i found it: http://www.viruscomix.com/page408.html
05:46:21 <quintopia> hopefully this is how the robotocalypse will go down
05:47:17 <ais523> pikhq: isn't that related to some extent to things like the cost of housing in the areas?
05:47:34 <pikhq> ais523: That is *one* of the major factors, yes.
05:47:35 <ais523> it makes sense for things you can't viably travel far from home for (such as food and housing)
05:47:47 <ais523> less sense for things you could just import
05:48:20 <pikhq> (keep in mind that in the US you need a car to go almost *anywhere*.)
05:48:36 <pikhq> Incidentally: FFFFFFFFFFFFFFFFFUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU
05:50:48 <ais523> from what I heard, the US mostly has no viable public transport (with a couple of exceptions in a couple of cities)
05:52:20 <pikhq> Nor are the cities at all walkable, except in the downtown areas of some of the older cities.
05:52:28 <pikhq> (looking at you, East Coast)
05:52:29 <ais523> meanwhile, I pay about £360 a year for unlimited bus travel with the major bus operator in Birmingham
05:52:45 <ais523> and rarely end up asking for lifts as a result
05:53:03 <ais523> (I never learnt to drive, and don't plan to; letting me at the controls of a car probably ought to be illegal)
05:53:12 <pikhq> I'd probably use the local bus system more if it were *at all useful*.
05:53:15 <ais523> well, it's a student discount, the typical value would be quite a bit higher
05:53:36 <pikhq> The bus route near here comes at 8 in the morning and 6 at night.
05:53:51 <pikhq> I have no idea how that's supposed to be usable.
05:54:17 -!- Sgeo|Puppy has quit (Ping timeout: 245 seconds).
05:54:33 <pikhq> Oh, and I could not walk anywhere aside from the grocery store. And even that's a couple miles.
05:55:04 <ais523> hmm, even the least frequent bus services here are once an hour
05:55:13 <ais523> apart from extra ones that follow unusual routes for the rush hour
05:55:43 <ais523> the most frequent come about 20 times an hour, in all sorts of slightly different variations
05:55:57 <ais523> (and would be much less annoying if it was "once every 3 minutes", but they tend to end up bunched)
06:00:41 * pikhq is still amazed how Bush got into office the first time 'round
06:01:31 <pikhq> If you think he actually won the election, you'd be wrong.
06:04:04 <ais523> well, he might have; I believe there isn't sufficient evidence to know whether he actually won or not
06:05:08 <pikhq> Formally speaking, he never won the election. The electoral college votes were never submitted.
06:05:19 <pikhq> Instead, the Supreme Court just declared that Bush won.
06:05:42 <pikhq> All Republicans in favor of the decision, all Democrats against.
06:05:47 <pikhq> I wish I were joking.
06:10:55 <coppro> in other news, whiteboards are awesome, and I <3 math
06:13:32 -!- ski has quit (Read error: Connection reset by peer).
06:13:46 -!- cheater99 has joined.
06:14:48 <coppro> first assignment in uni is far more interesting than anything in high school
06:17:43 <pikhq> Somehow, I believe in human rights and yet want to kill off a massive chunk of humanity right now because the STUPID IT FUCKING BURNS
06:29:49 -!- GreaseMonkey has joined.
06:55:36 -!- tombom has joined.
07:06:47 -!- oerjan has quit (Quit: Good night).
07:07:18 -!- Quadlex has joined.
07:16:14 -!- augur has joined.
07:48:57 -!- tombom has quit (Quit: Leaving).
07:59:26 -!- cheater99 has quit (Ping timeout: 252 seconds).
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:57:43 -!- atrapado has joined.
09:17:28 <Vorpal> <ais523> meanwhile, I pay about £360 a year for unlimited bus travel with the major bus operator in Birmingham <-- thats expensive
09:17:41 <ais523> not for the UK it isn't
09:17:47 <ais523> it's a lot cheaper than driving, at least
09:18:03 <ais523> but the buses are generally overpriced compared to the rest of Europe, I imagine
09:19:04 <Vorpal> <ais523> hmm, even the least frequent bus services here are once an hour <-- city or countryside?
09:19:11 -!- Gracenotes has quit (Remote host closed the connection).
09:19:25 <Vorpal> ais523, okay that is quite infrequent then
09:19:32 <ais523> yep, most are much more frequent
09:19:38 <Vorpal> I would say it was average for countryside
09:19:49 <ais523> the once-an-hour buses go quite a way out of the city, maybe even to neighbouring cities
09:20:06 <Vorpal> yeah, about the same around here
09:22:06 <Vorpal> ais523, or rather, about the same in the large city near here. I live in a small town and... apart from the "about hourly" buses to the city there is also one small half-length bus that goes around the town, about three times per day or such.
09:22:30 <ais523> in Birmingham, at least, buses tend to go high rather than long
09:22:40 <ais523> when they want to make them larger
09:22:49 <Vorpal> of course the city bus and the countryside bus never match up for me (I need to switch to get to the university)
09:22:59 <Vorpal> it is always about half an hour waiting between them
09:23:06 <Vorpal> either that or -5 minutes
09:24:47 <Vorpal> ais523, well, the small bus in this town is like a single-deck bus shortened to slightly larger than a van.
09:25:06 <Vorpal> ais523, but here buses go articulated rather than high
09:25:24 <fizzie> I just got a bus ticket for the Espoo region -- the whole Helsinki/Espoo/Vantaa/Kauniainen metropolitan area share bus ticketing systems, and you can buy either one-muncipality tickets or regional tickets -- from today to December 23rd, and that was pretty close to 130 eur.
09:25:30 <Vorpal> a double decked bus wouldn't fit in many places
09:26:13 <Vorpal> fizzie, wait, a ticket for that period?
09:26:35 <Vorpal> for such stuff they use plastic cards, rather than paper tickets around here
09:27:05 <fizzie> Well, I just used the abstract sense of "ticket"; it's a remote-readable card thing that I loaded the "ticket" onto.
09:27:27 <Vorpal> I don't go by bus often enough for it to be worth the cost, so I have one of those plastic cards that you can refill with money. But it costs half the price compared to buying tickets.
09:27:28 <fizzie> Would've been 269.90 EUR for the regional ticket for the whole four-city area, for 98 days.
09:28:33 <fizzie> A 366-day (the longest possible) period would be 982,80 EUR for the regional ticket, 473,10 for inside Espoo; so that's not so far from £360.
09:29:03 <fizzie> I miss the time I was a "real" student; those get a 50 % discount.
11:28:25 -!- GreaseMonkey has quit (Remote host closed the connection).
12:05:39 -!- ais523 has quit (Remote host closed the connection).
12:13:09 <fizzie> For future log-plotting, made a this sort of thing: http://p.zem.fi/esolog (up until now I've had all kinds of parameter-parsing and logfile/logdb-reading repeated in all of the scripts; also Python just for the change in flavor)
12:32:39 -!- FireFly has joined.
12:36:41 <Vorpal> fizzie, why not haskell or some such?
12:37:30 <fizzie> No real reason, except that I was more familiar with database and HTTP APIs and such in Python.
12:37:57 <Vorpal> fizzie, for example, if you do some intensive processing of the data to generate an extremely fancy graph you might want to take advantage of multiple cores to speed it up. Which sounds a bit painful in python, with it's global lock thing and so on
12:38:33 <fizzie> I'll probably delegate any intensive processing over to MATLAB anyway.
12:39:05 <fizzie> These are from clog logs this time (2004 and later), since my logs are a bit gappy.
12:39:23 <Vorpal> fizzie, clog also misses stuff sometimes
12:39:28 <Vorpal> so a merge might be useful
12:40:09 <Vorpal> fizzie, and if you and clog were on different sides in a netsplit
12:40:27 <Vorpal> hm how large are the clog logs in total?
12:40:42 <fizzie> In bytes, message counts, or what?
12:40:53 <Vorpal> fizzie, mb if I were to download all of them
12:41:09 <fizzie> In the raw clog format, around 117 MB.
12:42:32 <Vorpal> and db schema sounds simple: serial, timestamp (normalised to UTC, iirc clog switched timezones a few times?), sender, type, data. Where type is like message, notice, part, join
12:42:46 <Vorpal> and data would be message body, quit message or whatever
12:43:56 <Vorpal> for join it could be hostmask
12:44:07 <Vorpal> fizzie, how does your schema look?
12:44:30 <Vorpal> and hm, how would one normalise the daylight saving stuff and such
12:44:41 <Vorpal> would need a function to compute unix timestamp from date or such
12:45:16 <fizzie> The sqlite db I have for clog has two tables: one with logfiles and their line-counts (to find out which lines are new after fetching updated versions), and the other with columns tstamp/nick/uhost/type/body, where type is a small integer in 0..7 (msg, act, join, part, quit, nickchange, topicchange, modechange) and the rest are strings (since sqlite doesn't really do other types); timestamps are in UTC.
12:46:04 <Vorpal> fizzie, how did you transform the timestamps back to UTC?
12:46:37 <Vorpal> fizzie, 1) I'm pretty sure it uses daylight saving 2) I seem to remember the server in question moved to a different timezone some years ago
12:47:44 <fizzie> With a Pacific timezone → UTC conversion from pytz's timezone database; the only problem there is the one hour of ambiguous times when clocks are moved back, for those I just let the system guess, unless I actually see in the timestamps a backwards jump; in that case I'll use that to disambiguate.
12:47:53 <fizzie> It does go wrong if it has changed timezones, though.
12:48:37 <Vorpal> fizzie, you could compare it against your own logs. Like, match up messages, if they aren't within a few seconds of each other something is wrong.
12:49:13 <Vorpal> you don't even need to do it for every message
12:49:20 <fizzie> If there's only messages at 00:30, 01:30 and 02:30 on the day when the time jumps backwards from 02 to 01, you really can't tell from that data what the 01:30 time is.
12:49:55 <fizzie> I could try cross-correlating with my own logs, that's true. I think our timezone warp is at different day, too, so that'd work as a disambiguator.
12:50:26 <Vorpal> fizzie, UTC doesn't wrap so that should make it easy after you normalised the dates to UTC
12:51:08 <fizzie> The actual date of the timezone change has actually changed also in the US too between 2006/2007, but I assume that's taken into account in the timezone tables.
12:51:15 <Vorpal> and, if this would be intensive you don't need to check all, just a handful each day, randomly spread out. Should detect majorly wrong clock as well as moving timezones.
12:52:41 <fizzie> I'll look into that at some point when I have my own logs online; the computer's off right now.
12:53:06 <Vorpal> fizzie, anyway, since you need to keep two messages the same second in the correct order and thus need a counter as well as timestamp, you could use that to check for jump backwards
12:53:57 <fizzie> Oh, I don't bother keeping the message order consistent; it changes depending on which server you're looking from, anyway.
12:54:15 -!- Wamanuz has joined.
12:55:14 <Vorpal> fizzie, hm, less useful for checking back at old convos then. Still useful for stats though
12:55:39 <fizzie> I think we did have a talk about how there's actually no "proper" way of ordering IRC messages, when trying to decide who was first. (Unless you specify a reference server, of course.)
12:56:30 <Vorpal> fizzie, maybe we should use sequence numbers, like TCP does.
12:58:09 <fizzie> I *guess* the messages from one particular person will retain their internal ordering, even when looking at it from another server -- the IRC servers form a tree, after all, there's no alternative transportation paths that could reorder messages -- which is something my log-db doesn't necessarily, but I'm not sure I care so much.
12:58:37 <Vorpal> fizzie, you can order them if all clients are running on properly ntp synced computers and you can trust everyone to provide correct local timestamp with sub-second accuracy
12:58:48 <Vorpal> not that I even log sub-second in my own logs
12:59:01 <Vorpal> <fizzie> I *guess* the messages from one particular person will retain their internal ordering <-- yes
13:00:05 -!- Phantom_Hoover has joined.
13:00:14 <Vorpal> fizzie, there are also a few other things, like if you have three persons, A, B, C. and have: <A> 1 <B> I saw your 1. then C can not see them in the reverse order
13:01:17 <fizzie> Yes, but I don't really want to do semantic analysis on the message contents to determine whether B is referring to something A said before.
13:02:04 -!- atrapado has quit (Quit: Abandonando).
13:02:04 <Vorpal> I was just pointing out another property where it is known to have a specific order
13:02:56 <fizzie> http://p.zem.fi/on5o -- distribution of how many events there are for each unique "Y-m-d H:M:S" timestamp.
13:02:58 <Vorpal> meshed irc is tricky, there have been attempts
13:03:11 <Vorpal> but the "one-to-many" property of channels makes it tricky to pull off
13:03:47 <Vorpal> fizzie, how the heck did that happen
13:04:10 <fizzie> Or clog lag, or something.
13:04:24 <Vorpal> anything above 6 sounds unlikely without spam or lag indeed
13:04:30 <fizzie> sqlite> select count(*) as k, tstamp from logs group by tstamp having k > 50 order by k desc limit 5;
13:04:30 <fizzie> 86|2008-12-30 22:04:50
13:04:30 <fizzie> 60|2007-08-13 19:23:13
13:04:30 <fizzie> 60|2009-09-06 23:59:07
13:04:30 <fizzie> 59|2009-12-16 22:19:08
13:04:31 <Vorpal> and even 6 is quite unlikely
13:04:38 <fizzie> There's the top-5 seconds, according to clog.
13:04:54 <fizzie> Those have been UTCified, though.
13:05:00 <Vorpal> fizzie, could be bugs in that?
13:05:43 <Vorpal> fizzie, you said you had uhost? What is it?
13:05:51 <fizzie> Could be. Well, it's UTC-7 or UTC-8 in any case, so... http://tunes.org/~nef/logs/esoteric/08.12.30
13:05:54 <Vorpal> since clog logs doesn't include anything but nick
13:06:22 <fizzie> It's there for those messages where it is included. I just put in SQL NULLs for non-existing fields.
13:06:40 <fizzie> 14:04:50 seems to have a whole lot of comments.
13:07:25 <fizzie> Must be some sort of clog issue, then. For that particular second my message ordering's going to be a bit confusing, for reading purposes.
13:07:44 <Vorpal> fizzie, yeah, lag I presume
13:08:07 <fizzie> Heh, actually the discussion just above is in fact about clog.
13:08:16 <fizzie> 14:04:49 <oerjan> also, clog just stopped logging it seems
13:10:25 <Vorpal> fizzie, and yeah you could try to merge data from your own source with clog stuff. Useful for a multitude of reasons: 1) detect suspicious timestamps 2) correct for lag like the 08.12.30 case 3) fill in in case of netsplits to get both sides 4) fill in when clog was down
13:11:39 <fizzie> Might be a bit too much work, given that what I'd end up would still just be #esoteric logs.
13:11:51 <Vorpal> I might try to do some similar thing myself on clog + own logs, hm...
13:13:09 <Vorpal> fizzie, does clog not log notices to the channel?
13:13:19 <Vorpal> since you didn't list that as a type
13:13:52 -!- Phantom_Hoover has quit (Ping timeout: 245 seconds).
13:13:53 <Vorpal> the types I can think of are: msg, notice, act, join, part, quit, nick, topic, mode.
13:14:10 <Vorpal> oh well, for science I'm going to test this, I'm sure it will annoy some people
13:14:16 <fizzie> If it does, the formatting is something that got handled by the msg/act regexps.
13:14:51 <Vorpal> "05:14:35 -Vorpal(~AnMaster@unaffiliated/anmaster)- testing clog formatting"
13:15:08 <fizzie> Oh, it's that -nick(uhost)- mess; I only saw server-wide notices like that in the first logs, so I ignored it.
13:15:31 <Vorpal> hm what about other CTCPs....
13:15:44 <fizzie> Those are probably just with the raw ^As.
13:16:01 <Vorpal> I got this in my client from that: -clog- ERRMSG unknown CTCP: TESTINGCLOG
13:16:29 <fizzie> Yeah, it seems to ignore those then.
13:16:38 <Vorpal> what about ping and version though
13:17:08 <Vorpal> (sorry about that in case anyone got highlighted by it)
13:17:59 <Vorpal> anyway, server-wide notices would be from staffers presumably. Should be possible to find some way to detect that
13:18:34 <Vorpal> fizzie, got any sort of handy download script to share?
13:18:58 <Vorpal> fizzie, oh and did you import the esoteric logs from http://tunes.org/~nef/logs/old/ =
13:19:25 <fizzie> No, just the 2004 and later for now. I might add them later.
13:20:34 <fizzie> http://p.zem.fi/esolog-fetch.py -- that's what I used to fetch the files. It just downloads all the files to a "logs" subdirectory of current dir, and only fetches files that are missing or for which the server's modification timestamp is newer; then it sets mtime based on that after writing the file.
13:21:07 <Vorpal> wget http://p.zem.fi/esolog-fetch.py didn't do well
13:21:13 <Vorpal> fizzie, how do I get a raw download url for it
13:21:19 <Vorpal> instead of some html formatted thing
13:21:33 <fizzie> Use the download link on the page. Or just http://p.zem.fi/esolog-fetch
13:22:30 <fizzie> I guess something that does HEAD requests to old files to determine whether they need to be updated would be more "accurate", but at least that one has minimal server overhead when re-run, since it just fetches the file listing.
13:23:55 <Vorpal> fizzie, hm I looked at Gregor's hg repo for logs. why is there a fizzie directory in https://codu.org/projects/esotericlogs/hg/index.cgi/file/0ad952ce6894 ?
13:24:28 <fizzie> Those are probably some very old logs I had that clog didn't.
13:24:43 <fizzie> There's not much more than a few day's worth of stuff in there.
13:24:57 <fizzie> 1 These are raw logs from fizzie, starting in December 2002. These should fill in
13:24:58 <fizzie> 2 the short gap between 2002-12-14 and 2003-01-17, when clog's logs start.
13:25:28 <Vorpal> fizzie, why 2002-12-14 ?
13:25:53 <fizzie> That's when I joined. Or at least I joined the client that logs.
13:26:09 <fizzie> I think I was in the middle of messing around with the SparcStation at that time.
13:26:18 <Vorpal> well, I remember Gregor saying he didn't want someone to branch his repo due to bw load
13:27:06 <Vorpal> File "./esolog-fetch.py", line 53, in <module>
13:27:06 <Vorpal> f = open(logfile, 'w')
13:27:06 <Vorpal> IOError: [Errno 2] No such file or directory: 'logs/2004-01-01.log'
13:27:39 <Vorpal> oh maybe that simple *tries mkdir logs*
13:27:44 <fizzie> You should probably do that, yes.
13:27:58 <Vorpal> fizzie, you don't do persistent http?
13:28:31 <fizzie> Probably not, no. Unless urllib2 does it by trickery automatically.
13:28:51 <Vorpal> fizzie, it looked like you closed the connection yeah
13:29:38 <fizzie> I just took what I had in twitter-fungot for posting and adapted that.
13:29:39 <fungot> fizzie: if it's (a b c)
13:29:58 <Vorpal> hm it seems to be fetching over ipv6
13:30:02 <fizzie> Heh, the word length distribution looks a lot more uniform across people than the line lengths I plotted yesterday: http://www.cis.hut.fi/htkallas/wlen.png
13:30:07 <Vorpal> and yeah one connection per file
13:30:55 <Vorpal> fizzie, I suspect English imposes more constraints on what is possible for this
13:31:21 <Vorpal> and English is by far the most common language in here
13:31:32 <fizzie> In that plot "word" is defined by "things separated by whitespace", which probably explains the few >100-character words.
13:31:56 <Vorpal> fizzie, urls and such yeah
13:32:33 <Vorpal> fizzie, does it not go further than that? I'm pretty sure I pasted at least one url that turned out to be longer than the irc line length
13:32:47 <fizzie> It does, I just clipped it a bit to focus on the interesting part.
13:33:09 <fizzie> http://www.cis.hut.fi/htkallas/wlen2.png for the full view.
13:33:15 <Vorpal> fizzie, do you use gnuplot to do the plotting?
13:34:55 <Vorpal> fizzie, what is the y scale?
13:35:19 <fizzie> Log-frequency for one particular length.
13:37:51 <fizzie> You can compute the total number of lines for each person in the plot from the Y value of the "floor" where it cuts; for me it's somewhere close to -13, and that obviously corresponds to frequency 1/N where N is my total message count, so log(1/N) = -13 → -log(N) = -13 → N = e^13 =~ 442413.
13:39:52 <fizzie> sqlite> select count(*) from logs where nick like 'fizzie' and type = 0;
13:39:56 <fizzie> My logic seems to have failed.
13:41:59 <fizzie> Oh, I was looking at the word length graph. :p
13:42:08 <fizzie> So it's the total number of words, of course.
13:42:52 <fizzie> Yes, for len.png it's somewhere around -10.5, and e^10.5 =~ 36316.
14:13:11 * Vorpal ponders a db to use for log merging
14:13:21 <Vorpal> fizzie, how large is the sqlite file?
14:13:48 <Vorpal> fizzie, also that timezone normalising code, can you share that?
14:16:46 <fizzie> 244728832 bytes for the sqlite file.
14:17:23 <Vorpal> fizzie, about 233 MB then?
14:18:22 <fizzie> http://p.zem.fi/esologs-updatedb (with .py for HTML mess) for logs/*.log → logs.db sqlite3 update; it assumes the database has previously been created with e.g. the sqlite3 tool according to http://p.zem.fi/esologs-schema
14:19:18 <fizzie> It ignores those notices, and the "topic X, topic set by Y" messages clog presumably gets when it joins the channel.
14:20:44 <fizzie> There's also a single log file it ignores completely; it looks like the trailing half of some longer thing, it's missing timestamp and everything. Haven't checked which day's log its from.
14:23:42 <Vorpal> fizzie, yeah I will definitely adapt it
14:24:53 <Vorpal> fizzie, what are the numbers in the re_all array?
14:25:19 <fizzie> The type numbers it sticks in the "type" column.
14:25:37 <Vorpal> fizzie, and where is the pytz module from, I don't seem to have it
14:26:07 <Vorpal> oh found the package for it
14:26:14 <Vorpal> I typoed my first pacman search command
14:26:16 <fizzie> In Ubuntu it's called "python-tz" or something.
14:26:43 <Vorpal> python-pytz, but I managed to search for ptyz
14:26:46 <fizzie> (Which is a bit confusing, since a lot of stuff with "pyfoo" names is still called "python-pyfoo".)
14:29:24 <Vorpal> fizzie, what is "re_ig5 = re.compile(r'-\S*- ')" supposed to match?
14:30:49 <fizzie> -[anything]-, those notices.
14:32:07 <Vorpal> fizzie, so \S is like . ?
14:32:32 <fizzie> No, it's non-whitespace.
14:33:37 <Vorpal> fizzie, why the optional whitespace for act between the nick and the body
14:34:06 -!- augur has quit (Remote host closed the connection).
14:34:08 <Vorpal> and hm does "re_notice = re.compile(r'-(?P<nick>[^\( ]+)\((?P<uhost>.*)\)- (?P<body>.*)$')" look like a plausible regexp for notices? I'm more used to PCRE than python style regexp
14:34:19 <fizzie> Because there's one instance where there's just "* EgoBot" in the logfile.
14:35:48 <fizzie> It does look plausible, though I think from-server notices may have looked a bit different. Try it out and see; I wrote those regexps by running everything through (with db-writing disabled) and seeing which log lines that one "invalid body" assertion catches.
14:36:52 <Vorpal> fizzie, does it use the regexps in that array in order?
14:37:08 <Vorpal> since I intent to write ignores for the server ones
14:37:28 <fizzie> First match wins, yes.
14:38:14 <Vorpal> 2007-01-12.log:11:00:07 -ChanServ(ChanServ@services.)- andreou!i=ee4299@nemesis.ee.teiath.gr disabled notices of #esoteric access list changes
14:40:23 <fizzie> Here's another cryptic visualization: http://www.cis.hut.fi/htkallas/alise_vs_vorpal.png
14:40:50 <Vorpal> fizzie, pretty, what does it plot
14:41:01 <fizzie> But wouldn't that kill all the mystery in it?
14:41:23 <Vorpal> fizzie, hm does python anchor to start of line or something by default?
14:41:45 <fizzie> Yes, if you use re.match; re.search finds matches everywhere.
14:42:05 <fizzie> What is a bit confusing is that re.match won't anchor to end of line; usually it's either a totally free search, or anchored at both ends.
14:42:23 <fizzie> (Or actually start of string, not start of line; but these strings all are one line.)
14:42:28 <Vorpal> fizzie, which one do you use in that script
14:42:52 <fizzie> You can add a .* in the front if you want. Though it's anchored to the part that's after the timestamp.
14:43:23 <Vorpal> actually I wanted to anchor to start (and I did notice it cut away timestamp before)
14:44:29 <fizzie> The mystery plot's Y axis is time, divided into ten-minute bins (24*60/10 = 144); the X axis is message length in characters, with >100 clamped to 100. The color scale is log-frequency, where the frequencies have been normalized across the time bins, so each horizontal line sums to one.
14:45:13 <fizzie> The plot would show how your message length distributions differ during different parts of the day, but really it only shows that for the quiet parts of the day there's mostly noise, and for the rest of the time the distributions are pretty similar.
14:46:03 <Vorpal> fizzie, what about one for yourself
14:47:04 <fizzie> I have spoken significantly less here, I guess it might be even more noisy. Could try, of course.
14:48:43 <fizzie> http://www.cis.hut.fi/htkallas/avf.png -- it got a bit horizontally squeezed, can't quite recall how to change plot dimensions in MATLAB.
14:49:44 <Vorpal> fizzie, hm, they use the same value for the same colour?
14:50:20 <Vorpal> fizzie, why is your so much more blue basically
14:51:05 <fizzie> http://www.cis.hut.fi/htkallas/len.png -- for the same reason my peak is lower there.
14:51:39 <fizzie> Wider distribution, so the normalization makes the peak lower.
14:51:53 <Vorpal> fizzie, how long does esologs-updatedb.py take on average?
14:52:09 <Vorpal> as in, should it sit there quiet for minutes when turning off the insert into the db
14:53:21 <Vorpal> fizzie, hm now it failed:
14:53:23 <Vorpal> AssertionError: invalid log body: --- quit: woo11:09:38 <SimonRC> hi
14:53:47 <Vorpal> fizzie, what the heck is going on there, did you get that too?
14:53:50 <fizzie> Hm, that's strange; I didn't see that sort of line.
14:54:13 <Vorpal> fizzie, checksum for 2006-09-27.log ?
14:54:58 <fizzie> 1c8f11ac3f0df3b43de888ef793d35502b15986edbcb1b6c575a1f039b7086b271d3aa812f513178e59d5a9336e96e8fdb6730346dc7d48de299c0d91e895584 logs/2006-09-27.log
14:55:13 <fizzie> Yes, you seem to like it. :p
14:55:35 <fizzie> 10:21:32 --- quit: woo11:09:38 <SimonRC> hi
14:55:39 <fizzie> There is a line like that in there.
14:56:01 <Vorpal> fizzie, so something ignored it before my change?
14:56:03 <fizzie> Oh, of course: my notice-ignoring line actually ignores that one too.
14:56:18 <fizzie> "-\S*- " matches "--- " just fine.
14:56:50 <fizzie> It probably shouldn't, to catch lines like that; but on the other hand that's a clear mistake in the clog data.
14:57:04 <Vorpal> fizzie, you might miss other --- lines
14:57:17 <Vorpal> should I edit the file or ignore the line
14:57:42 <fizzie> I might; it sounds even likely that I do. Oh well, all my statistics so far have been about simple messages.
14:58:41 <fizzie> Oh, so that's how the aspect ratio is changed; updated avf.png. Now I'll have to leave, won't be back until... uh, now + 5 hours or so.
14:58:57 <Vorpal> fizzie, what you didn't make avf_b.png ? ;P
14:59:18 <fizzie> set(gcf, 'PaperSize', [16 6]);
14:59:18 <fizzie> set(gcf, 'PaperPosition', [0 0 16 6]);
14:59:23 <Vorpal> AssertionError: invalid log body: --- kick: SimonRC was kicked by lament (lament)
14:59:26 <fizzie> That's a really silly way to change plot sizes.
14:59:34 <fizzie> Hee, I'm missing all the kicks too.
14:59:51 <fizzie> Well, I'll rerun with fixed regexps later; now to the bus.
14:59:52 <Vorpal> fizzie, in the updated file your box looks wider
15:00:03 <Vorpal> fizzie, I will send you the updated regexps
15:03:46 <Vorpal> hm kick need two nicks
15:03:56 <Vorpal> I guess nick could be kicker and a new field target could be the kickee
15:04:11 <Vorpal> in case there are more such messages than just kicking
15:06:36 <fizzie> I have that sort of thing for nick-changes.
15:06:56 <fizzie> It puts old nick as "nick" and new nick as "body".
15:07:03 <Vorpal> fizzie, hm considering "Now I'll have to leave, won't be back until... uh, now + 5 hours or so." I assume you invented the time machine?
15:07:20 <fizzie> I'm in the bus, actually. :p
15:07:28 <Vorpal> fizzie, but kicks have a body too: the kick message
15:07:45 <Vorpal> fizzie, and using the uhost seems... weird
15:08:05 <Vorpal> fizzie, also what is "TEXT COLLATE NOCASE" exactly? Does it store it without case?
15:08:09 <Vorpal> or is it just for indexing?
15:08:39 <Vorpal> fizzie, the reason I wonder is that a nick change like Vorpal -> vorpal would be strange if it was stored without case
15:08:48 <Vorpal> ignoring broken log line: t okay i make this version of nopol simple and remove : and . later :)
15:08:50 <fizzie> The overcomplicated multi-table db schema I have for those postgresql logs (for multiple networks and channels) has a separate target field, I think.
15:08:51 <Vorpal> fizzie, any idea about that?
15:09:42 <fizzie> And "collate nocase" just makes the created index be usable for a case-insensitive "nick like 'foo%'" match.
15:10:07 <Vorpal> 10:17:15 <oklopol> but as that isn't all that easy
15:10:07 <Vorpal> t okay i make this version of nopol simple and remove : and . later :)
15:10:07 <Vorpal> 10:38:54 <oklopol> it's always hard for me not to have my langauge be tc right away
15:10:15 -!- cpressey has joined.
15:10:29 <fizzie> Looks like another clog thing; I just ignored that one.
15:10:42 <Vorpal> 2008-04-20.log is the log file
15:10:55 -!- BeholdMyGlory has joined.
15:10:56 <fizzie> It was the only line that didn't match the timestamp regex.
15:12:17 <Vorpal> hm for serial I guess I need to store that in the db
15:14:38 <fizzie> Most likely. The problematic case is when you run updatedb once with a log that ends with X:Y:Z timestamp, and then do fetch+updatedb at some later time, and it has had another X:Y:Z message just after your first fetch.
15:15:06 <Vorpal> ah I can get the serial for free with INTEGER PRIMARY KEY AUTOINCREMENT
15:15:25 <Vorpal> fizzie, hm? how do you mean?
15:16:58 <Vorpal> free as in "it will use the already existing internal row id column"
15:17:56 <Vorpal> fizzie, how do you avoid dupes/missing in that case currently?
15:18:08 <fizzie> If you wanted to have the "optimal" counter that has just values 0..N for each N+1 messages with a particular timestamp, you'd have to checl the earlier-added lines too. But the row ID is fine too.
15:18:47 <fizzie> There's the logfiles table that counts how many lines have been processed out of each log.
15:19:08 <Vorpal> I don't see the issue with another x:y:z timestamp really here
15:19:47 <fizzie> It's fine as long as you don't fetch in the middle of a line, but maybe those log writes are atomic enough.
15:21:58 <fizzie> The "another timestamp" is an issue only if you want messages 00:00:00-00:00:00-00:00:00-00:00:01 get counter numbers 0-1-2-0, but see only the two first messages during one updatedb.py run; you'd have to manually check the db for the latest message.
15:23:22 <fizzie> Small numbers use less storage, after all. (But the row number uses no storage, so it's even better from that viewpoint.)
15:23:26 <Vorpal> fizzie, what instead might be an issue is merging in other log data. Might be best to put merge resul into a new db anyway
15:24:03 <fizzie> Okay, now actualy really away.
15:34:35 -!- FireFly has quit (Quit: swatted to death).
15:34:36 -!- sftp has joined.
15:35:18 -!- FireFly has joined.
15:37:18 <Vorpal> what happens when you reach the highest inode on a disk? does it reuse ones old free ones (free due to file having been deleted or such)
15:37:42 -!- MigoMipo has joined.
15:42:07 -!- tombom has joined.
15:45:25 <pikhq> Vorpal: inodes are reference-counted, you know.
15:46:09 <Vorpal> pikhq, yes but if you touch two files on ext4 at least they seem to get inode numbers incremented by 1
15:46:21 <Vorpal> pikhq, so what I meant is, what happens when you reach MAX_INODE_NUM
15:47:28 <pikhq> It then has to resort to the pool of free inodes rather than doing naive allocation.
15:47:46 <pikhq> It's not like being on disk suddenly makes common memory allocation techniques stop working.
15:48:16 <Vorpal> pikhq, hm, so it uses a freelist for inode numbers or?
15:48:27 <pikhq> Pretty sure it does.
15:49:10 <Vorpal> pikhq, presumably not, since that would be inefficient if you had inode numbers being 32-bit and had like 10 files (all previous ones deleted) when you reached the max inode number
15:49:17 <Vorpal> then wouldn't the freelist take several GB?
15:49:38 <pikhq> I'm going with "clearly they're doing something".
15:49:53 <Vorpal> I go with "I sure hope that they do something..."
15:54:46 -!- tombom_ has joined.
15:56:50 -!- tombom has quit (Ping timeout: 264 seconds).
16:00:31 -!- jix has quit (Read error: Connection reset by peer).
16:03:27 <Vorpal> pikhq, some file systems seems to reuse inode numbers of deleted files right away, like ext4, while some, like ext3 only seem to do that sometimes.
16:03:39 <Vorpal> based on some simple tests
16:04:15 <Vorpal> hard to tell if there is a limit to this for ex4, could try to create lots of small files or such I guess
16:04:52 <Vorpal> hm it seems ext4 groups free inodes per directory!?
16:24:09 -!- webquint has joined.
16:25:52 <Vorpal> sqlite> select count(*) from logs where (nick like 'alise%' or nick like 'ehird%' or nick like 'tusho%') and nick not in ('alise', 'ehird', 'tusho'); gives 60455
16:26:30 <Vorpal> fizzie, did you match on ehird% for the graphs? After all ehird_ ehirdiphone and so on contribute a significant number of posts
16:27:13 <Vorpal> select count(*) from logs where (nick like 'vorpal%' or nick like 'anmaster%') and nick not in ('anmaster', 'vorpal'); gives just 2925
16:46:48 <Vorpal> fizzie, most common lines: http://sprunge.us/GEjK
16:47:42 <cpressey> it probably doesn't need to keep a freelist of inodes - it probably just picks the next number, sees if it already exists; if not, you get that one, but if so, increment and repeat
16:48:43 <cpressey> which means you could hit pathologically bad allocation behaviour in some circumstances that are probably quite rare in practice
16:49:54 <Vorpal> cpressey, hm, starting at 0 or starting at some higher base line?
16:50:14 <Vorpal> cpressey, and as I said it seems to keep inode numbers in groups for directories
16:50:16 <cpressey> Vorpal: probably starting [over] at 0
16:50:31 <Vorpal> so I don't think it does what you say
16:51:46 <cpressey> Vorpal: you have the source, so feel free to check instead of relying on the speculations of pikhq and myself.
16:52:00 <cpressey> If you do not have the source, to obtain a copy, please write to...
16:53:21 <Vorpal> sqlite> select count(*) from logs where body like '%☃%';
16:54:07 <Vorpal> sqlite> select count(*) from logs where type = 0 and body like '%☃%';
16:54:17 <Vorpal> type 0 is message in case anyone wonders
16:57:59 -!- webquint has quit (Quit: Page closed).
16:59:08 <Vorpal> sqlite> select count(*) from logs where type = 0;
16:59:13 <Vorpal> now waiting for "select count(*) from (select distinct body from logs where type = 0);"
16:59:34 <Vorpal> sqlite> select count(*) from (select distinct body from logs where type = 0);
17:04:59 -!- alise has joined.
17:06:08 <Vorpal> hm 2210 nicks that sent at least one PRIVMSG to the channel but 4913 mentioned in total in clog logs
17:06:30 <alise> it is cold in here
17:07:17 <Vorpal> select distinct nick from logs where nick not in (select distinct nick from logs where type = 0);
17:07:38 <Vorpal> alise, only 35 msg was sent prior to today including ☃
17:08:01 <Vorpal> alise, oh and the most common lines said on this channel were: http://sprunge.us/GEjK
17:08:43 <alise> what have you done?
17:09:00 <alise> well i don't remember any db :)
17:09:38 <Vorpal> alise, I fixed fizzie's script to import clog files into db (it ignored notices and as a side effect also kicks and some other stuff).
17:10:01 <alise> i can only top that with one thing
17:10:05 <Vorpal> alise, before you ask me to upload it:
17:10:22 <Vorpal> you can get the scripts though
17:10:28 <Vorpal> for downloading clog data and for importing it
17:10:39 <alise> i already have every log updated regularly.
17:10:48 <Vorpal> alise, I think you can't reuse previously downloaded clog files, it needs mtime to match that on server
17:10:54 <Vorpal> as far as I understood fizzie
17:10:55 <alise> + 2002-12 -- 2003-12 lgs
17:11:53 <Vorpal> alise, his download script: http://sprunge.us/TaPc
17:12:18 <Vorpal> and my fixed version of the db import script: http://sprunge.us/AigL
17:12:20 <alise> i already have a download script
17:12:25 <Vorpal> alise, "<Vorpal> alise, I think you can't reuse previously downloaded clog files, it needs mtime to match that on server"
17:12:50 <Vorpal> alise, which will in general not be the case if you don't do that manually
17:12:50 <alise> on the clog server? what?
17:12:58 <alise> that sounds like a rubbish script.
17:13:32 <Vorpal> alise, decide for yourself. I think that is how it can fetch partial info
17:13:46 <alise> as i said, there is a way i can top it
17:13:50 <Vorpal> alise, oh and you need the schema:
17:14:07 <Vorpal> http://sprunge.us/YdVT
17:14:39 <Vorpal> alise, also: http://www.cis.hut.fi/htkallas/avf.png
17:14:48 <Vorpal> alise, tell me if you can figure it out without reading log
17:15:14 <Vorpal> alise, nice try but no
17:15:30 <Vorpal> alise, it sounds like an interesting idea though
17:15:52 <Vorpal> alise, any other suggestion?
17:16:00 <alise> yeah, but i'll keep it >:)
17:16:07 <Vorpal> alise, why? I know what it is
17:16:32 <alise> 17:25:44 <Sgeo|web> Also, I haven't used KDE in a very long time
17:16:33 <alise> 17:25:57 <Sgeo|web> And Knoppix helped me get into Linux in the first place
17:16:33 <alise> 17:26:01 <Sgeo|web> I'm feeling nostalgic
17:16:33 <Vorpal> alise, I can tell you, if you don't want to keep guessing
17:16:37 <alise> this is why i told you to fucking use ubuntu
17:16:42 <alise> you're only doing it out of nostalgia
17:16:42 <Vorpal> but if you want to go on guessing that is fine too
17:16:46 <alise> Vorpal: sure, tell me :P
17:16:51 <alise> i thought you meant
17:16:54 <alise> suggestion for new graphs to do
17:17:06 <Vorpal> alise, well, y axis is time of day, divided in 10 minute bins
17:17:18 <alise> i was guessing line lengthg
17:17:30 <alise> this graphs us getting too tired to write lots of shit
17:17:58 <alise> well apart from me
17:17:58 <Vorpal> alise, no, fizzie claimed it was noise due to low activity
17:18:04 <alise> and don't get tired
17:18:55 <Vorpal> alise, well you still do according to your interpretation
17:19:11 <Vorpal> but really fizzie said it was due to lack of data due to low activity
17:19:14 <alise> well yeah but nothing like you do
17:19:19 <alise> i have yellow all the way down
17:19:22 <alise> you have big chasms
17:20:51 <Vorpal> alise, also I think we are a optimistic channel. the line "yes" is almost twice as common as the line "no"
17:21:13 <Vorpal> alise, and "yes" is the most common line
17:22:28 <Vorpal> alise, also how are you going to top this db thingy that both me and fizzie have now?
17:23:24 <alise> Vorpal: one freakin' word.
17:23:36 <Vorpal> what was that now again
17:24:26 <Vorpal> and I'm considering merging my personal logs to 1) detect/fix timestamp mismatches, there exist some due to clog lagging (180 messages in one second iirc) and DST changes 2) supply data for when clog was down 3) add missing data in case me and clog were on different sides of a netsplit
17:24:53 <alise> botte is an IRC bot. A ludicrous definition of course. it will do everything. including: hackego-style sandbox playgrounds... and a web server; it will have fizzie's old logs and all the clog logs, converted to UTC, with a nice web interface -- changeable timezone that takes this into account wrt the date, searches -- fulltext and regexp...
17:25:05 <alise> ...and log new things itself,
17:25:15 <alise> and now: the House of Pointless Statistics.
17:26:02 <alise> 18:35:37 <Sgeo|PM> According to wget, 250M has been downloaded
17:26:02 <alise> 18:35:53 <Sgeo|PM> According to Task Manager, 159MB is in use
17:26:11 <alise> Sgeo has never heard of a filesystem
17:28:23 <Vorpal> alise, how much code is written for it yet?
17:29:05 <Vorpal> alise, as for fizzie's old logs, I have them here, I haven't inserted them yet though... Different format to parse
17:29:38 <Vorpal> alise, I'm a bit surprised that wget uses 159 MB though
17:29:42 <Vorpal> if that is what he meant
17:29:49 <alise> enough swathes of code are right in my head
17:29:52 <alise> i just need the right language :)
17:29:59 <alise> nobody said there's a statue of limitations on oneupmanship
17:30:18 <Vorpal> alise, write it in an esolang. Otherwise fungot is by definition better than botte
17:30:18 <fungot> Vorpal: have you guys seen the 4chan quine? :)
17:30:30 <Vorpal> fungot, no I can't say I have.... that sounds nasty though
17:30:30 <fungot> Vorpal: it shouldn't be any pictures on your computer.)
17:30:37 <quintopia> I need to adjust my highlight list so quine doesn't ping me :/
17:30:46 <Vorpal> fungot, huh that was almost coherent
17:30:46 <fungot> Vorpal: the imperative " track" uses c. 0)) and not just insert a single value
17:30:53 <Vorpal> back to his normal self
17:31:29 <Vorpal> quintopia, idle less, then that highlighting you won't be an issue ;)
17:31:51 <quintopia> well, it's only when i'm not idling that highlights could bother me
17:32:02 <quintopia> i don't notice them when i'm afk...by definition
17:32:30 <alise> <Vorpal> alise, write it in an esolang. Otherwise fungot is by definition better than botte
17:32:30 <fungot> alise: the story tells better if i just spend all my time with schoolwork)))
17:32:32 <alise> fungot can't do this shit.
17:32:33 <fungot> alise: i'm really thinking of something that decides mutability for you. what will perfection in a scheme program and a network has servers is less important
17:32:50 <Vorpal> alise, sure but it is in an esolang!
17:33:44 <quintopia> even my awesome bot is not as awesome as a bot written in an esolang
17:33:47 -!- oerjan has joined.
17:34:25 <alise> Vorpal: it's such a stupid argument that not even you believe it, so i won't bother rebutting
17:34:36 <alise> (wrt anything written in a normal lang being inferior to an esolang)
17:34:43 <Vorpal> oerjan, can you figure out what http://www.cis.hut.fi/htkallas/avf.png is without reading about it in the logs
17:35:02 <quintopia> alise: who said anything about "anything"?
17:35:16 <Vorpal> alise, not inferior in general. Just less cool
17:35:20 <quintopia> the argument is that it takes a whole hell of a lot of coolness to trump "bot written in esolang"
17:35:33 <alise> <Vorpal> alise, write it in an esolang. Otherwise fungot is by definition better than botte
17:35:34 <fungot> alise: t'ain't working, mcgee
17:35:39 <alise> he knew nothing about botte
17:35:56 <alise> thus the conclusion, [unspecified thing] not in esolang < thing in esolang *by definition*
17:35:57 <Vorpal> alise, that was an exaggeration, you use them too you hypocrite
17:36:14 <alise> Vorpal: of course it is, but you kept defending it. also i said nothing personally insulting, so fuck you
17:36:18 <quintopia> and i will tell you if it has a chance of surmounting that mountain of Rule of Cool
17:36:48 <Vorpal> alise, I don't think I did either?
17:36:54 <alise> Vorpal: hypocrite is a personal insult
17:37:11 <Vorpal> alise, I would call it a factual description in this case.
17:37:11 <oerjan> Vorpal: exactly what, no, but it's obviously one of fizzie's log analyses
17:37:24 <Vorpal> oerjan, ah. Any guesses for what?
17:37:37 <alise> Vorpal: "motherfucker" could be factual too.
17:37:38 <oerjan> well there are three variables
17:38:00 <quintopia> however, "cunt" and "dick" are pretty much never factual
17:38:07 <Vorpal> quintopia, they could be
17:39:05 <Vorpal> oerjan, no further guesses? alise suggested it was a "swearogram" (if I remember the word he used correctly)
17:39:34 <alise> quintopia: all the features of fungot -- ... that is to say, babbling -- EgoBot's user-command definition + huge swathes of esolang interpreters (botte will have more) + HackEgo's arbitrary unix sandboxing stuff, applied throughout (thus many, many *non*-esolangs too) -- a lot of random commands such as pointless-karma-tracking, a quote database (dunno if i'll use hackegos), various random things, etc. -- and a full web interface to the logs from 2002-12 to
17:39:34 <alise> three seconds ago, rendered nicer than the tunes logs, with options to search (fulltext and regexp), user info pages with everything the bot knows about them, timezone adjusting including moving lines to other log dates if required... and a Statistics House -- oh yeah, and infobotty stuff too
17:39:34 <fungot> alise: the second rule of project douglasadams: do not say quine
17:39:46 <alise> fungot: my favourite project.
17:39:47 <oerjan> Vorpal: i really cannot be bothered to guess within a near-infinite space
17:40:12 <Vorpal> alise, why only 3 seconds ago?
17:40:34 <alise> Vorpal: non-literal. you use it to. "you hypocrite"
17:40:51 <Vorpal> oerjan, well y is time, divided in 10-minute buckets
17:41:02 <Vorpal> oerjan, x is line length, >100 are capped to 100
17:41:08 <Vorpal> oerjan, and colour is frequency
17:41:13 <quintopia> alise: and its webserver should serve its own source, so all can admire it (and also so that it is a quine too)
17:41:16 <Vorpal> oerjan, the time for y is time of day that is
17:41:41 <alise> quintopia: well, sure, if the source repository just isn't cool enough for you :P
17:41:58 <alise> oh yeah and downloadable logs too -- maybe in a repo
17:42:06 <alise> so they sync easily
17:42:17 <quintopia> alise: well, what if the repos is out-of-date ;)
17:42:17 <alise> with perhaps a more raw format than tunes, plus a script to render them more nicely
17:42:25 <alise> since tunes' is unreadable and not that easily parseable
17:42:30 <alise> quintopia: then it wouldn't be running on the server
17:43:00 <quintopia> alise: it would be cooler without the babbling >.>
17:43:12 <alise> quintopia: you don't like fungot's babbling?
17:43:24 <alise> it only happens when you mention its name! and even then it has a cutoff
17:43:32 <alise> probably i'll make it so that only "bot: ..." triggers a babble...
17:43:35 <oerjan> fungot still has one _major_ advantage - it's actually implemented *ducks*
17:43:36 <fungot> oerjan: it's obvious... the furs never reached istanbul... you were asking to!! all about macros. im confused by all the spaces.)
17:43:47 <alise> oerjan: as i said: no statue of limitations on oneupmanship
17:44:03 <quintopia> but it's kind of a useless feature...
17:44:07 <Vorpal> <alise> since tunes' is unreadable and not that easily parseable <-- true, the timezone stuff is annoying
17:44:08 <alise> iirc it's 5 messages before it ignores you
17:44:13 <alise> quintopia: but it /is/ a fun feature
17:44:14 <Vorpal> alise, other than that it is easy to parse tunes
17:44:22 <alise> Vorpal: not as easy as a nicer format
17:44:35 <alise> say, (msg unixts "name" "msg")
17:44:39 <alise> or even a binary format of some sort
17:44:40 <quintopia> alise: indeed, but fungot already does it, so why replicate it?
17:44:40 <fungot> quintopia: that sucked. rephrase: " fnord fnord
17:44:47 <Vorpal> alise, what about joins, parts, kicks, topic changes and so on?
17:44:48 <alise> i.e. binary-style but with only printable chars
17:44:54 <alise> Vorpal: they'd be different. obviously.
17:45:08 <alise> quintopia: mine would do it better >:| also, different corpus
17:45:15 <alise> and mine would probably try and use what you say to it to seed what it says
17:45:21 <Vorpal> alise, anyway, I suggest a simple line based format in that case
17:45:23 <alise> producing more pseudo-synchronicitious fun
17:45:25 <alise> Vorpal: maybe, yeah.
17:45:40 <alise> quintopia: a lovely "OS" project that has been very dead for a very long time
17:45:47 <Vorpal> alise, such as: unixts type sender message
17:45:55 <alise> "OS" is a misnomer, they basically reinvented computing fifty different times.
17:45:58 <alise> Vorpal: we'll see.
17:45:58 <Vorpal> for kick you would have target after sender
17:46:07 <Vorpal> but the basic idea is sound I think
17:47:01 <Vorpal> alise, type would be like "msg, act, notice, join, part, quit, kick, nick topic, mode"
17:47:12 <alise> Vorpal: yeah yeah yeah
17:47:19 <alise> quintopia: type /topic
17:47:37 <Vorpal> alise, those were the types needed for clog parsing btw
17:47:53 -!- Kiraz has joined.
17:48:46 -!- Kiraz has quit (Client Quit).
17:49:29 <alise> 20:18:28 * quintopia tries Illumination Software Creator
17:49:29 <alise> ehird@dinky:~$ sudo dpkg -i illumination64.deb
17:49:49 <alise> This trial version of Illumination Software Creator is fully functional but limited to 10 "Blocks" of functionality per project. Purchasing a license from our secure online store removes this limitation.
17:49:58 <alise> how can i revolugimise ajax2.0 now
17:50:08 <alise> "Build Project - Haiku - Desktop"
17:50:25 <Vorpal> isn't that some google thingy?
17:50:31 <quintopia> alise: you can get the full version for $5 if you liked it
17:50:34 <alise> haiku is the BeOS successor
17:50:39 <alise> quintopia: lol, or i could pirate it
17:50:58 <quintopia> alise: well, yes, but is it really worth pirating? It's kind of weak, IMO...
17:51:22 <alise> woo it quits when i hit the button
17:51:31 <alise> but, like Plain English, these things are hilarious
17:53:58 <Vorpal> (the current thing, not plain english)
17:54:12 <alise> http://www.radicalbreeze.com/illumination/ISCExampleTextEditor.png
17:54:47 <Vorpal> alise, so: extremely limited
17:55:28 <alise> 20:55:45 <quintopia> "polar bear" and "jökullhlaupt" confused?
17:55:28 <alise> 20:55:58 <quintopia> feel free to move that diaresis as needed
17:55:33 <alise> now i'm /certain/ it's not a diaeresis there
17:55:59 <Vorpal> alise, "jökullhlaupt"? Sounds like Icelandic
17:56:34 <Vorpal> but hm jökull is something related to volcanoes iirc?
17:56:55 <alise> quintopia: no, it's not a diaeresis
17:56:59 <alise> i have no idea as to the spelling
17:57:18 <quintopia> well, i could have sworn that was the english word for umlaut
17:57:28 <alise> but ö is a separate letter
17:57:33 <alise> at most i'd accept umlaut
17:57:39 <alise> no, diaeresis is the accent -- the same --
17:57:47 <alise> used to signify that two vowels do not form a dipthong
17:57:53 <alise> Zoë, coöperate, etc.
17:57:54 <Vorpal> oh yeah ö is a separate letter
17:58:03 <Vorpal> like in Swedish and other Nordic languages
17:58:08 <alise> the-letter-ö derives from German's o-with-umlaut
17:58:21 <Vorpal> alise, well okay, I'm just saying what it is today
17:58:21 <alise> so calling icelandic ö an o with an umlaut is acceptable, but it's definitely not a diaeresis
17:58:30 <alise> Vorpal: well you have to give it some name :-)
17:58:33 <quintopia> i should have spelled it joekull then it wouldn't have been a problem
17:58:40 <alise> what i'm being is a joekill
17:58:57 <quintopia> but i come to expect that from you
17:59:07 <alise> hey i'm funny upon occasion
17:59:14 <alise> just pedantic the rest of the time
17:59:27 <Vorpal> quintopia, åäö are separate letters in Swedish, not just aao with "decoration". Same goes for the other Nordic languages where relevant (not all have that set, Icelandic has some more stuff, iirc ð for example, Norwegian and Danish has ø instead of ö)
17:59:30 * quintopia donates a pile of dead horses to alise for SCIENCE
17:59:44 <alise> quintopia: you're looking for #beastiality
18:00:00 <quintopia> Vorpal: yes, it is a separate phoneme, but i need a name for JUST THE TWO DOTS
18:00:03 <alise> Vorpal: otoh you bastards won't give us /names/ for them that we can pronounce
18:00:06 <alise> quintopia: "umlaut"
18:00:11 <Vorpal> alise, dead horses he said. necrobeastiality?
18:00:17 <alise> or "diaeresis" depending on when it's used
18:00:26 <alise> there's no word for both of them :D
18:00:28 <Vorpal> quintopia, there isn't one for Swedish at least
18:00:35 <alise> An umlaut (pronounced /ˈʊmlaʊt/ OOM-lowt) is the orthographical representation of a type of sound shift in spoken language. A very similar diacritical mark (called diaeresis or "trema") is used to signify a linguistic hiatus. In modern computer systems (using Unicode), umlaut and diaeresis are represented identically: ä represents both a-umlaut and a-trema.
18:00:46 <quintopia> alise: what portion of your day do you spend rehashing day-old logs?
18:00:53 <alise> quintopia: very little of it
18:00:56 <alise> oerjan logreads too
18:01:00 <alise> you just haven't seen him do so yet
18:01:26 <quintopia> it just seems weird to be talking about conversations we had last night in such detail
18:01:34 <alise> this channel is a bit weird.
18:01:34 <Vorpal> <alise> Vorpal: otoh you bastards won't give us /names/ for them that we can pronounce <-- what do you mean?
18:01:46 <alise> Vorpal: for german we can say "u with umlaut"
18:01:52 <alise> we have no word for your ö without learning to pronounce it
18:01:59 <alise> even then, it will be a bit unclear what we mean to other english speakers
18:02:04 <alise> so name it; oobie or something
18:02:09 <Vorpal> I was about to say that we can't manage them either over here. Complain to the Icelandic ppl
18:02:13 <alise> otherwise we'll keep saying o-with-umlaut :)
18:03:20 <alise> is that ouml or oüml :D
18:03:28 <Vorpal> alise, what do you call д (a random Cyrillic symbol)
18:03:57 <Vorpal> alise, there are lots of letters that have no name in English.
18:04:07 <alise> Vorpal: giko neko's :O mouth
18:04:21 <alise> http://upload.wikimedia.org/wikipedia/commons/4/47/Gikoneko01.svg
18:04:25 <alise> no clue why that's svg but there you go
18:04:54 <Vorpal> alise, it's svg with embedded bitmap...
18:05:16 <alise> A, Be, Ve, Ge/He, Ge, De, Dje/Djerv...
18:05:18 <alise> I won't continue reciting.
18:05:25 <Vorpal> alise, that is transcribing
18:05:25 <quintopia> Vorpal: a jkullhlaupt is a mudslide resulting from a volcanic eruption under a glacier, not the volcano itself
18:05:30 <alise> Vorpal: good enough.
18:05:55 <Vorpal> alise, so transcribe åäö then
18:06:24 <Vorpal> alise, anyway you could just type åäö
18:06:26 <alise> Vorpal: is there a defined transcription?
18:06:32 <alise> also, yeah, try saying that out aloud
18:06:41 <alise> that's like saying
18:06:56 <alise> "the alphabet is ah buh cuh duh ef uh guh hu ey juh kuh..."
18:06:59 <Vorpal> that's the last three letters in the Swedish alphabet, in the right order
18:07:14 <alise> "the alphabet is ay, bee, cee, dee, ee, eff, gee, heigh, eye, jay, kay, ..."
18:07:16 <alise> the latter are the names
18:07:31 <alise> nobody would understand you if you referred to letters by their pronunciation as in the former
18:07:54 <alise> "Note: when replacing umlaut characters with plain ASCII, use ae, oe, etc. for German language, and the simple character replacements for all other languages."
18:08:01 <alise> an official WP recommendation to replace ö with o
18:08:09 <Vorpal> alise, hm yeah for English, as it happens "a" would coincide for them in Swedish. Same for a number of other vowels
18:08:52 <alise> you say "ah" for the letter a?
18:09:08 <Vorpal> alise, the letter a sounds like the same sound as in words like "al" alla" "arga" and so on
18:09:14 <Vorpal> in all the places for those
18:09:31 <Vorpal> well al more than alla, the first a in alla would be sort
18:09:36 <alise> 21:06:18 <Sgeo|Puppy> I could never go into hiding
18:09:36 <alise> 21:06:27 <Sgeo|Puppy> Into a witness-protection-program-like-thing
18:09:36 <alise> 21:06:36 <Sgeo|Puppy> I love my online identity too much
18:09:36 <alise> 21:07:07 <Sgeo|Puppy> I love the places too much
18:09:36 <alise> 21:07:20 <Sgeo|Puppy> Not the RL places. The online places
18:09:45 <alise> i'm glad those would take priority if your life was at risk.
18:10:05 <Vorpal> alise, a = long a, "al" = contains long a
18:10:16 <Vorpal> short a would not be the same though
18:10:45 <alise> quintopia: you mean /æ/ or [æ] then :p
18:10:51 <Vorpal> alise, hm? æ = Danish ä
18:11:00 <alise> i don't know that anyone uses /.../ for IPA
18:11:15 <alise> quintopia: (it seriously confused me; I still use æ occasionally.)
18:11:27 <quintopia> alise: where do i find Droid Sans and how do I install it?
18:11:37 <quintopia> i've never installed new fonts on here before...
18:11:43 <alise> 21:26:18 <pikhq> I do hereby declare US politics fundamentally broken. Nothing short of eugenics can fix it.
18:11:43 <alise> 21:26:43 <quintopia> eugenics broke it in the first place
18:11:54 <alise> i don't recall the US ever practising eugenics...
18:12:00 <alise> quintopia: sudo aptitude install ttf-droid
18:12:15 <oerjan> alise: i thought // was for phonemes and [] for actual sounds (phones?)
18:12:16 <alise> or [your favourite gui package manager] → "Droid"
18:12:18 <Vorpal> ah it is in the repos. Right
18:12:25 <alise> and "handheld device font with extensive style and language support"
18:12:33 <alise> oerjan: ah, perhaps
18:12:33 <Vorpal> oerjan, only in case of telemarketing
18:12:53 <alise> but not the marketing part
18:13:05 <Vorpal> alise, I had to come up with something before you did!
18:13:24 <Vorpal> alise, so I picked first word related to phones that came to my mind
18:13:32 <alise> you have a bad relationship with them huh
18:13:57 <Vorpal> alise, not any longer. I'm in the national "do not call for telemarketing" register :)
18:14:14 <alise> 21:35:48 <ais523> UK politics is pretty stupid, but US politics is clearly even stupider
18:14:28 <alise> Vorpal: we are too, but it's fun because credit card companies still ring you up to offer you a better deal
18:14:34 <alise> through some loophole; dunno which
18:14:51 <alise> pikhq: have you read The Metamorphosis of Prime Intellect?
18:15:17 <Vorpal> alise, or do you mean ones where you are an existing customer?
18:15:31 <Vorpal> hm then indeed not happening here
18:17:47 <alise> 21:38:22 <pikhq> ais523: They want to end the freaking EU. [...]
18:17:52 <alise> incorrect; they don't give a shit about what anyone else does
18:17:55 <alise> they just want to retreat from it
18:17:55 <cpressey> alise: oh yes, eugenics is but one small part of this nation's proud history
18:18:06 <alise> and spend the rest of their glorious lives living in a tiny bubble with the UK in it
18:18:22 <alise> cpressey: they actually did it awesome
18:18:36 <Vorpal> <alise> 21:38:22 <pikhq> ais523: They want to end the freaking EU. [...] <-- who?
18:18:45 <alise> Vorpal: UK right-wingers
18:18:47 <cpressey> "In 1907 Indiana became the first of more than thirty states to adopt legislation aimed at compulsory sterilization of certain individuals." etc etc --WP
18:18:59 <alise> they'd love for the rest of europe to burn to the ground
18:19:01 <alise> as long as we're not in the EU
18:19:11 <alise> cpressey: GO INDIANA
18:19:43 <Vorpal> alise, I presume you are joking right?
18:20:04 <quintopia> how long did it take kansas to get in on the act?
18:20:13 <alise> Vorpal: I NEVER JOKE.
18:20:32 <quintopia> (i think it took like 40 some odd years before the various state level anti-miscegenation laws were repealed.)
18:20:44 -!- oerjan has quit (Quit: Later).
18:21:07 -!- Sgeo|PartedMagic has joined.
18:21:20 <Sgeo|PartedMagic> lupu-511.iso 100% |******************************| 129M 00:00:00 ETAA
18:21:39 <alise> Sgeo|PartedMagic: ^C
18:21:51 <Vorpal> Sgeo|PartedMagic, "ETAA"?
18:22:25 <quintopia> they like to really hedge their estimates
18:22:54 <Sgeo|PartedMagic> Does anyone (besides me, soon) really use the save to DVD stuff?
18:23:05 <alise> most people save to usb
18:23:18 <alise> quintopia: in puppy.
18:23:37 <quintopia> alise: well from a netbook at least
18:25:34 <Vorpal> <Sgeo|PartedMagic> Does anyone (besides me, soon) really use the save to DVD stuff? <-- um I install on hdd and use that :P
18:25:49 <Sgeo|PartedMagic> There's some gui tool for settings, but it complains about a missing driver
18:25:53 <alise> yeah because Vorpal uses puppy
18:25:54 <Vorpal> I never saved state from a livecd to disk
18:26:01 <alise> puppy /is/ just a livecd
18:26:03 <Vorpal> alise, I used other livecds
18:26:10 <alise> Vorpal: puppy is nothing like them
18:26:20 <alise> ew they redesigned the puppy site
18:26:24 <alise> mistakes made in recent puppy history:
18:26:25 <Vorpal> alise, but I haven't actually tried to use it like a main system. I use them like I use systemrescuecd
18:26:26 <alise> - using ubuntu as a base
18:26:28 <alise> - oh god that website
18:26:41 <alise> Vorpal: puppy runs from ram, so the HD install is not very necessary
18:26:57 <Vorpal> alise, true, I know about knoppix and such
18:27:01 <Vorpal> I just don't see the point
18:27:10 <alise> it's nothing like knoppix, either
18:27:18 <alise> also, it's possible to install to hd
18:27:21 <alise> but it just loads it into ram immediately
18:27:29 <Vorpal> Sgeo|PartedMagic, that shouldn't be hard. Unless you live in the middle of nowhere
18:27:32 <alise> most people use a usb stick, since it's fast enough due to the ram-based operation, and leaves a portable system
18:27:58 <Vorpal> and even then there is these so called "web shops" that "ship" to you. Quite awesome!
18:28:16 <alise> so how's that data recovery going buddy
18:28:18 <Sgeo|PartedMagic> This way, I get to show off that my laptop's fully usable without a HD
18:28:34 <Vorpal> alise, if he doesn't have a hdd to recover it to, then what is he going to do
18:28:46 <Vorpal> <Sgeo|PartedMagic> This way, I get to show off that my laptop's fully usable without a HD <-- yes and?
18:28:49 <alise> Vorpal: use it as an opportunity to get attention from the fact that his laptop is running without an HD?
18:28:51 <Vorpal> how is that unexpected
18:29:03 <Sgeo|PartedMagic> Track 01: Total bytes read/written: 136114176/136114176 (66462 sectors).
18:29:07 <Vorpal> every computer I know that can boot from a cd can do so without a hd being in it
18:29:38 <Vorpal> alise, I can't see what's so unusual about that
18:29:47 <Vorpal> I mean, of course it is possible
18:29:48 <alise> Vorpal: well, that's never stopped 'im :)
18:29:58 <Vorpal> as long as you have something else to boot from
18:30:55 <quintopia> there was the fud. now for the nap?
18:32:10 <Vorpal> quintopia, do you mean FUD like "microsoft FUD"?
18:32:40 <Vorpal> google says fud stands for "fear, uncertainty and doubt" and I didn't see anything of that in here
18:34:42 <quintopia> which is a medium teal in RGB color
18:34:48 <Vorpal> $ echo -n fud | od -tx1
18:35:27 <Vorpal> quintopia, and F00D is just two components, F0, 0D. Which does not make up an RGB triplet
18:36:08 <quintopia> the initial 00 is implied by digit significance
18:36:14 -!- Sgeo|PartedMagic has quit (Ping timeout: 252 seconds).
18:37:03 <alise> i would say whoosh
18:37:04 <Vorpal> quintopia, still, fud is 66,75,64.
18:37:09 <alise> but i think Vorpal is actually so underground
18:37:16 <alise> that nothing can go level with or below his head
18:37:42 <Vorpal> alise, if you are saying that I don't keep up-to-date with popular culture? So what?
18:37:59 <alise> <quintopia> the hex value for fud is F00D
18:38:01 <alise> read this one more time
18:38:06 <alise> then call me an idiot
18:39:30 <Vorpal> well obviously "fud" is similar to "food". More so than most other three letter strings. That however does not explain "hex value"
18:40:05 <Vorpal> hm speaking of which, I wonder if there are any ascii -> hex quines.
18:40:26 <Vorpal> would have to be in the first 15 chars, all control codes.
18:41:07 <alise> Vorpal: you're stupid
18:42:25 <Vorpal> alise, then perhaps you could explain what the joke in that "the hex value for fud is F00D" statement is
18:42:32 <quintopia> when catherine forgot blair's name after just meeting him
18:42:42 <quintopia> so no one told her and made her try to guess
18:42:59 <quintopia> they said "hint: it's a kind of girl-sounding name, and it rhymes with 'blair'"
18:52:43 <alise> -rw-r--r-- 1 ehird ehird 11221 2010-09-16 20:57 by implanting explosive ammunition in VC ammo caches
18:52:43 <alise> -rw-r--r-- 1 ehird ehird 3142 2010-09-16 20:57 inessweek. More stories
18:59:10 <alise> anyone know how to do underload's a in dc?
18:59:14 <alise> that is, [foo] -> [[foo]]
19:05:58 <alise> popping is hard :D
19:09:36 <alise> "asX2 1>X" works but clobbers X
19:09:51 <alise> S/L don't work because the L pushes to the stack again
19:12:54 <alise> and also because there has to be something in X in the first place
19:18:24 <alise> okay i can translate everything but * and a
19:29:02 <fizzie> Vorpal: For today's graphs, they were just exact (well, case-insensitive) matches of "ehird", "tusho", "alise". Probably should've added the % in.
19:30:13 <fizzie> <Vorpal> 18:26:06> fizzie, did you match on ehird% for the graphs?
19:30:32 <alise> which graphs again? :D
19:31:02 <fizzie> www.cis.hut.fi/htkallas/len.png wlen.png alise_vs_vorpal.png avf.png if I recall all the names correctly.
19:31:29 <fizzie> Nothing very new there.
19:32:12 <Vorpal> fizzie, I mostly had the volume graphs graphs in mind
19:32:18 <alise> basically i am the best
19:32:20 <Vorpal> fizzie, that is the top 10 and such
19:32:24 <alise> fizzie: did you make the super-big all-time actichart?
19:32:47 <fizzie> alise: No, and now I don't feel like it since I have that new script framework. :p
19:33:07 <Vorpal> fizzie, so port the scripts over? :)
19:33:17 -!- Behold has joined.
19:33:20 <alise> fizzie: That was totally my idea. :|
19:33:28 -!- BeholdMyGlory has quit (Disconnected by services).
19:33:34 -!- Behold has changed nick to BeholdMyGlory.
19:33:50 <Vorpal> fizzie, anyway you may find the updated http://sprunge.us/HGcY useful
19:34:13 <alise> re_ig6 = re.compile(r'--- quit: woo11:09:38 <SimonRC> hi$')
19:34:15 <Vorpal> fizzie, since it handles notices and kicks and such. Mind the extra column for kick target
19:34:24 <Vorpal> alise, yeah there was one line that was broken
19:34:36 <fizzie> There's that other line that it already ignores that's broken, too.
19:34:52 <Vorpal> fizzie, but this one had a working timestamp
19:35:04 <Vorpal> fizzie, two ones in fact :P
19:35:36 <Vorpal> alise, anyway yes I know it is a hackish fix :P
19:35:52 <alise> Maybe I'll implement botte in SUPERLISP.
19:36:07 <Vorpal> alise, the proper fix would be to get the tunes admin to remove/correct broken lines in the logs on the server
19:36:17 <alise> the tunes admin /does not control/ clog
19:36:21 <alise> nobody controls clog
19:36:25 <Vorpal> alise, no, use plain english!
19:36:37 <alise> Faré won't do anything to them
19:36:45 <alise> only hcf would and he no longer has access to the server
19:36:54 <fizzie> Okay, I'll run the old script one last time to get a more hugey plot.
19:37:01 <Vorpal> alise, Halt and Catch Fire?
19:37:05 <alise> fizzie: Wider, and generally bigger.
19:37:16 <Vorpal> alise, I mean, did his nick stand for that
19:37:29 <alise> three-letter nicks used to be all the ræg.
19:37:33 <fizzie> Yeah, it should now be 3* the width and twice the height, and maybe with a bit better smoothing window too. Assuming it works.
19:37:45 <Vorpal> alise, sure but they are in short supply
19:37:52 <alise> which is why i said used to be.
19:37:57 <Vorpal> at least interesting ones
19:38:04 <fizzie> I used the nick "f" in IRCnet for a few weeks, but gave that up.
19:38:11 <Vorpal> alise, I know two letter nicks and one letter ones too
19:38:19 <fizzie> One-letter nicks look somehow wrongy.
19:38:24 -!- alise has changed nick to fcf.
19:38:30 <Vorpal> I used one letter ones on other networks, but mis-highlight quota is often high
19:38:31 <fcf> "Fornicate and Catch Fire"
19:38:39 <fcf> not registered!
19:39:05 -!- fcf has changed nick to qcf.
19:39:08 -!- wareya has quit (Read error: Connection reset by peer).
19:39:11 <qcf> "Quarks Catch Fire" - aw - registered.
19:39:12 -!- qcf has changed nick to fcf.
19:39:21 -!- fcf has changed nick to hqx.
19:39:23 <hqx> unregistered
19:39:50 <fizzie> hqx: http://zem.fi/~fis/biglog.png -- there, that's a bit bigger. (For any later updated versions, I'll try to get some sort of X scale going.)
19:40:09 -!- wareya has joined.
19:40:10 <hqx> no, this nick is awesome
19:40:26 <hqx> fizzie: Can you do the same for a relative plot? i.e. the sum is always 1
19:40:35 <Vorpal> fizzie, did that one include alise*, ehird* and so on?
19:40:53 <hqx> YEAH AND WHAT ABOUT MY NOVELTY NICKS
19:41:02 <hqx> ehird` is a bit important
19:41:06 <hqx> since i used it for a very long time
19:41:09 <hqx> after ehird got stolen
19:41:28 <Vorpal> hqx, you have like 20 000 messages with (ehird|alise|tusho).+
19:41:32 <Vorpal> notice the + at the end
19:41:33 <fizzie> Vorpal: No, it just contains three exact remappings. I can do that too.
19:41:47 <hqx> I like how I don't talk very much until AnMaster comes and then BOOM big spike.
19:41:48 <Vorpal> hqx, novelty ones are more work, provide us with a complete list!
19:42:02 <hqx> fizzie: ehird` -> ehird is quite relevant
19:42:11 <hqx> as it'll be why i'm not present leftwards
19:42:21 <hqx> well, -> alise
19:42:32 <Vorpal> fizzie, iirc for my anmaster* vorpal* cases it was more like 500 messages
19:42:37 <Vorpal> that were not on the base nicks
19:42:45 <hqx> Vorpal: Here, you have a DB; tell me what day I talked the most. :P
19:43:28 <fizzie> Computing the relative graph now with all nicks that start (case-insensitively) with "ehird", "alise" and "tusho"; that should catch some of them.
19:43:29 <hqx> That is, select top day sorted by count of messages by me that day.
19:43:32 <Vorpal> hqx, um, the date field is a timestamp, it is not split in separate date and time, otherwise it would be simple with some group by query
19:43:41 <Vorpal> hqx, now? a bit trickier
19:43:42 <hqx> Vorpal: I'm sure SQLite has datetime functions...
19:43:47 <fizzie> You can group by a substr expression.
19:43:55 <Vorpal> hqx, yeah but I would need to read the docs :P
19:44:11 <hqx> http://www.sqlite.org/lang_datefunc.html
19:44:29 <hqx> you might have to use strftime :P
19:44:32 <hqx> or something
19:44:41 <Vorpal> hqx, notice that anything not on indexes (and this won't be I bet) will probably take several minutes to run
19:44:48 <hqx> index EVERYTHING!
19:45:02 <Vorpal> hqx, nah, indexes on nick and timestamp atm iirc
19:45:28 <Vorpal> fizzie, why not a timestamp ?
19:45:39 <Vorpal> or does it not have a separate data type?
19:45:51 <hqx> probably not
19:45:55 <hqx> why not an integer, though?
19:45:56 <fizzie> sqlite only does integers, binary blobs and text strings.
19:46:28 <fizzie> And the date-and-time functions take timestamp strings, that's why not an integer.
19:46:35 <fizzie> sqlite> select substr(tstamp, 1, 10) as k, count(*) as num from logs where (nick like 'alise%' or nick like 'ehird%' or nick like 'tusho%') group by k order by num desc limit 5;
19:46:36 <Vorpal> hqx, I have strings like "2010-09-17 18:37:22" in the table. (blame fizzie!). You give me the "get date out of that" SQL code
19:46:47 <hqx> Thanks fizzie.
19:46:50 <hqx> I'll ask you next time :P
19:47:31 <hqx> Cool, it starts with me talking about trying to make a Windows editor that doesn't suck.
19:47:34 <Vorpal> fizzie, I like the limit 5 and you pasting 3 results :) (yeah yeah I know you didn't copy it all, but still)
19:47:35 <hqx> I don't recall why.
19:47:42 <fizzie> Hrrm, the relative graph got... strange.
19:47:44 <hqx> No, wait; a PDF reader.
19:47:46 <fizzie> http://zem.fi/~fis/biglogr.png
19:47:49 <hqx> fizzie: Strange is nice!
19:47:57 <hqx> That's not nice!
19:48:03 <Vorpal> hqx, it is however strange
19:48:03 <hqx> fizzie: That must be periods of no activity. (Wat?)
19:48:16 <hqx> But what we really learn is that I TOTALLY DOMINATE
19:48:32 <hqx> fizzie: Could you tell me which position is the start of 2010?
19:48:34 <fizzie> The periods are so long that I wonder why they didn't show up in the previous relative graph.
19:48:45 <hqx> There should be a complete flattening after that for a few weeks.
19:49:07 <Vorpal> hqx, according to that I spoke more during some periods than you. Assuming same height = equal number of messages at that period
19:49:17 <hqx> fizzie: It can't be right; the channel has never been dead when I've been around.
19:49:33 <hqx> Vorpal: Well, yeah, but I still dominate on the whole. Also, you probably weren't in a mental institution.
19:49:39 <fizzie> hqx: Yes, but I've been elsewhere; it's probably about that. These old scripts read just my own logs.
19:49:44 <Vorpal> hqx, hard to tell with no x scale
19:49:45 <hqx> I /think/ the bit where I follow the others line is the few-week period.
19:50:06 <hqx> Since before that it's very big, as I wasn't going to the unit.
19:50:13 <hqx> So there's some sense of scale, I guess.
19:50:16 <fizzie> Anyway, you can find the start of 2010 by counting three times the number of days in 2010 up until 2010-08-31 pixels.
19:50:22 <Vorpal> hqx, I was looking near the start after the last 0-place
19:50:48 <hqx> Vorpal: Consider that a lot of that red near the right was, apart from weekends, accumulated on a phone at night! Dedication :P
19:51:08 <fizzie> So 729 pixels from the end of the plot is where 2010 starts, approximately.
19:51:20 <Vorpal> fizzie, and is that before or after the last drop to zero?
19:51:24 <hqx> A day is three pixels? Surely not.
19:51:32 <fizzie> A day should be three pixels.
19:51:33 <hqx> It's not wide enough for that. Or is this just recent stuff?
19:51:53 <Vorpal> fizzie, then I doubt it is wide enough
19:52:15 <fizzie> It's 8443 pixels wide; that's 8443/3 = 2814 days; 2814/365 is about 7.7 years.
19:52:19 <hqx> 1095 horizontal pixels a year
19:52:19 <fizzie> Sounds approximately right to me.
19:52:43 <Vorpal> fizzie, anyway, those drops seem strange
19:52:54 <hqx> then i can only conclude that being in a mental institution monday to friday and never once connecting to the internet on the weekends must have had... no effect on my activity
19:53:00 <hqx> Vorpal: as fizzie said they were probably unlogged periods
19:53:05 <hqx> since he's using his logs
19:53:26 <Vorpal> hqx, yes but compare with http://zem.fi/~fis/biglog.png where it the last drop to zero doesn't show up
19:53:28 <fizzie> I'll try to add the X scale to that sooner or later. In fact I guess I could quickly hack in years.
19:53:56 <hqx> just less pronounced
19:54:03 <hqx> due to the non-relativity
19:54:17 <Vorpal> a log merging tool for these dbs would be useful
19:54:42 <Vorpal> anyway, maybe I should import this into postgre as well to compare speed for some queries (with the same indexes)
19:55:32 <hqx> maybe i'll use picoLisp for botte and put it in a picoLisp database!
19:55:52 <fizzie> A log merging tool would also be non-trivial, especially if you want to do it "right" (i.e. model the reliability of timestamps in different databases at different periods of time intelligently, then take some weighted-average timestamps to get most likely correct estimates, yet do it in a manner that doesn't mess message ordering, and so on; even the matching of messages needs some sliding-window stuff and maybe even approximate matching if some irreversible e
19:55:52 <fizzie> ncoding transformations have taken place).
19:56:24 <Vorpal> my 3 most active days btw:
19:57:04 <fizzie> 2009-09-06|373 -- aw, even my busiest day has been pretty unbusy by your standards.
19:57:14 <Vorpal> fizzie, I know my local computer has used ntp since before I started using irc.
19:57:28 <Vorpal> fizzie, so baring lag mine should be accurate.
19:57:42 <hqx> Vorpal: Now rewrite the query to sum up all message lengths.
19:57:50 <hqx> What's our top byte-emission days?
19:58:00 <hqx> I'm trying to be byte neutral. Save the planet and all.
19:58:05 <hqx> Only write as much as I read.
19:58:05 * Vorpal tries to remember the string length function of sql
19:58:19 <hqx> Try to edit existing lines with regular expressions instead of producing new ones. Recycling bytes.
19:58:20 <fizzie> It's just sum(length(body)) in place of count(*).
19:58:29 <fizzie> And you probably want a "and type = 0" in the where condition.
19:58:50 <Vorpal> you want that for the other ones too really
19:59:12 <fizzie> alise: 106190 characters in one day, not bad.
19:59:19 <Vorpal> sqlite> select substr(tstamp, 1, 10) as k, sum(length(body)) as num from logs where (nick like 'anmaster%' or nick like 'vorpal%') and type = 0 group by k order by num desc limit 5;
19:59:37 <Vorpal> the top two changed place for me
19:59:42 <fizzie> I should rebuild my logs.db with the fixed updatedb.pl.
20:00:04 <hqx> <fizzie> alise: 106190 characters in one day, not bad.
20:00:07 <hqx> that's .1 megabytes fuck yeah
20:00:11 <Vorpal> fizzie, if you use mine, note that type numbering changed, I inserted the new ones in the middle.
20:00:16 <hqx> hmm the second one is my top place in lines
20:00:22 <Vorpal> fizzie, rather than adding notice at the end
20:00:29 <hqx> first one i don't think is in the three you pasted
20:00:31 <Vorpal> it seemed nicer to put it together with msg and act
20:00:34 <hqx> indeed, it isn't
20:01:11 <Vorpal> hqx, actually the line count one includes part/joins
20:01:22 <Vorpal> sqlite> select substr(tstamp, 1, 10) as k, count(*) as num from logs where (nick like 'alise%' or nick like 'ehird%' or nick like 'tusho%') and type=0 group by k order by num desc limit 5;
20:01:25 <hqx> <fizzie> 2010-09-11|106190
20:01:26 <hqx> oh that's boring
20:01:31 <hqx> that's just when i pasted the entire Funge-98 spec in channel
20:01:37 <Vorpal> hqx, why did you do that?
20:01:59 <hqx> Vorpal: Quadr*scence and ch*ater99 (no ping) had been pissing everyone off for an hour or two
20:02:28 <hqx> then when aug*r decided to ask for pics and started commenting on Quadr*scence's, and I quote, "fuckab"ility...
20:03:16 <Vorpal> hqx, taking top line count for CTCP ACTION, generic PRIVMSG and NOTICE:
20:03:26 <Vorpal> sqlite> select substr(tstamp, 1, 10) as k, count(*) as num from logs where (nick like 'alise%' or nick like 'ehird%' or nick like 'tusho%') and type in (0,1,2) group by k order by num desc limit 5;
20:03:34 <Vorpal> sqlite> select substr(tstamp, 1, 10) as k, count(*) as num from logs where (nick like 'anmaster%' or nick like 'vorpal%') and type in (0,1,2) group by k order by num desc limit 5;
20:03:44 <hqx> Vorpal: can you list by "top non-PRIVMSGs"?
20:03:53 <hqx> to get join cycles, /mes, NOTICE, TOPIC, etc.
20:03:56 <Vorpal> hqx, in a sec, doing byte count now
20:04:02 <hqx> of everything? :D
20:04:16 <Vorpal> sqlite> select substr(tstamp, 1, 10) as k, sum(length(body)) as num from logs where (nick like 'anmaster%' or nick like 'vorpal%') and type in (0,1,2) group by k order by num desc limit 5;
20:04:19 <Vorpal> sqlite> select substr(tstamp, 1, 10) as k, sum(length(body)) as num from logs where (nick like 'alise%' or nick like 'ehird%' or nick like 'tusho%') and type in (0,1,2) group by k order by num desc limit 5;
20:04:22 -!- Sgeo|WebPuppy has joined.
20:04:29 <Vorpal> hqx, not of everything no
20:04:41 <Sgeo|WebPuppy> What do you mean, I can't use emacs with rxvt-unicode?
20:04:52 <Vorpal> hqx, so you want everything except /msg and /me?
20:04:59 <Vorpal> or did you want to include /me?
20:05:06 <hqx> Vorpal: exclude /me i guess
20:05:13 <hqx> Sgeo|WebPuppy: of course you can.
20:05:15 <fizzie> Sgeo|WebPuppy: That sounds really very strange, since I use emacs with rxvt-unicode.
20:05:38 <Vorpal> sqlite> select substr(tstamp, 1, 10) as k, count(*) as num from logs where (nick like 'alise%' or nick like 'ehird%' or nick like 'tusho%') and type not in (0,1) group by k order by num desc limit 5;
20:05:50 <Sgeo|WebPuppy> emacs: Terminal type rxvt-unicode is not defined. If that is not the actual type of terminal you have, use the Bourne shell command `TERM=... export TERM' (C-shell: `setenv TERM ...') to specify the correct type. It may be necessary to do `unset TERMINFO' (C-shell: `unsetenv TERMINFO') as well.
20:05:59 <hqx> Vorpal: That's so boringly few.
20:06:01 <fizzie> Your terminfo database sounds sucky, then.
20:06:05 <Vorpal> sqlite> select substr(tstamp, 1, 10) as k, count(*) as num from logs where (nick like 'anmaster%' or nick like 'vorpal%') and type not in (0,1) group by k order by num desc limit 5;
20:06:06 <hqx> Sgeo|WebPuppy: TERM=xterm emacs
20:06:13 <hqx> TERM=xterm-unicode emacs # might also work
20:06:16 <hqx> or whatever it is
20:07:32 <Sgeo|WebPuppy> Does Puppy Linux have a nice way to read .pptx files?
20:07:43 <hqx> i wish there was a variant of sql with symbols instead of those pointless words :)
20:07:47 <hqx> Sgeo|WebPuppy: probably not.
20:07:55 <hqx> and it would not be a feature of Puppy if there was.
20:08:57 <fizzie> hqx: Relational algebra has a well-defined set of funky symbols ready, even.
20:09:01 <hqx> fizzie: Indeed!
20:09:05 <hqx> Vorpal: new ppt
20:09:08 <hqx> ISO Standard powerpoint!
20:09:11 <hqx> *PowerPoint!
20:09:15 <hqx> was it ISO?
20:09:18 <hqx> i fogret. eh.
20:09:20 <hqx> yes, fogret
20:09:22 <hqx> fizzie: Indeed!
20:09:30 <hqx> Vorpal: Try pronouncing it!
20:09:37 <hqx> Chlschszzzxxxx.
20:09:46 <Vorpal> hqx, I was wondering if that was what it was
20:09:50 <hqx> fizzie: There's a fetish for that!
20:09:51 <Vorpal> since it would be so silly
20:09:59 <hqx> why does SQL even have "from TABLE", why isn't it a conditional?
20:10:00 <Sgeo|WebPuppy> Before the X config thing, I was able to click the touchpad but not scroll
20:10:11 <Sgeo|WebPuppy> Now that I did it, I can scroll, but it doesn't recognize taps as clicks
20:10:28 <fizzie> Vorpal: .xlsx is what it is, among others, like .xltx for templates and so on.
20:10:40 <cpressey> hqx: HI ARE YOU NEW HERE ALISE
20:10:48 <Vorpal> what I love (not) with arch: pacman -S postgresql *waits* hm no post install message, now what
20:11:05 <Sgeo|WebPuppy> Ok, how do I make rxvt or whatever not be bright pink?
20:11:11 <Vorpal> I mean gentoo told you in a post install message "go to /var/foo run postgresql-bar to set up the db stuff"
20:11:16 <Vorpal> arch doesn't tell you anything
20:11:39 <Vorpal> of course the arch wiki has info
20:11:55 <hqx> (k := substr(tstamp, 1, 10), num := count(*) | ∈ logs /\ (nick ~ 'alise%' \/ nick ~ 'ehird%' \/ nick ~ 'tusho%') /\ type ∉ {0,1}
20:11:58 <hqx> can't think or group/order syntax :P
20:12:03 <hqx> cpressey: HI.
20:12:18 <hqx> Vorpal: Funny that, with Debian it automatically starts it.
20:12:33 <Vorpal> hqx, indeed. Arch of course doesn't do that
20:12:40 <Vorpal> would be completely unlike arch
20:12:50 <hqx> It would be far too reasonable >:)
20:12:55 <fizzie> Sgeo|WebPuppy: You might have to tell us exactly what *is* bright pink there; it's not like it's a feature of the terminal emulator what color the things run inside it are.
20:12:59 <hqx> "Ooh, you went there."
20:12:59 <cpressey> hqx: btw, I implemented Pixley in C last night. Except... it doesn't have a garbage collector yet.
20:13:17 <Vorpal> hqx, actually, I had to stop it on ubuntu and redo it, since I needed some custom initdb options
20:13:26 <Sgeo|WebPuppy> fizzie, when I use the ssh link thingy, it starts the terminal as bright pink
20:13:27 <hqx> cpressey: Ooh, parallel! Mark and sweep! (Gotta rhyme, gotta rhyme) Gave birth to a garbage collector in the back of a Jeep!
20:13:31 <hqx> Best jingle ever.
20:14:23 <cpressey> Well, first I'm rewriting the interpreter in continuation-passing style to make GC easier
20:14:52 <hqx> cpressey: CHENEY ON THE MTA
20:14:54 <hqx> Do it do it do it
20:14:55 <Vorpal> hqx, hm, you forgot concurrent there
20:15:02 <fizzie> Sgeo|WebPuppy: Can't say I know what a "ssh link thingy" is, so... anyway, you can configure the default fg/bg/text colors with X resources. And I guess also the numbered colors, but those have pretty well-defined meanings already.
20:15:04 <cpressey> It makes me wonder whether choosing reference counting isn't sometimes influenced by "oh man my root set is on the C stack, that's a pain"
20:15:07 <hqx> Vorpal: That's rather more difficult.
20:15:16 <hqx> cpressey: http://home.pipeline.com/~hbaker1/CheneyMTA.html
20:15:19 <hqx> cpressey: DO IT DO IT DO IT
20:15:21 <Sgeo|WebPuppy> I'll just use gFTP and send the completed thing there
20:15:33 <hqx> Vorpal: mark and sweep? dunno about that
20:15:55 <Vorpal> hqx, with specialised hardware it could be done
20:17:43 <hqx> Vorpal: how?
20:17:47 <hqx> i don't think mark and sweep is...
20:17:49 <hqx> wait of course it's easy
20:17:59 <hqx> by parallel i mean at the same time as code continues to run
20:18:00 <hqx> which is easy
20:18:02 <hqx> concurrent is also easy
20:18:31 <Vorpal> hqx, wait, isn't the first property concurrent?
20:18:54 <hqx> i dunno i try and avoid using the terms because they suck :D
20:19:03 <Vorpal> hqx, so what do you mean by concurrent then
20:19:15 <hqx> i was totally wrong about everything, both are very easy
20:19:27 <Vorpal> hqx, "while the code continues running" is what I meant by concurrent
20:19:33 <Vorpal> but that is not the case you say?
20:19:33 <hqx> that's easy
20:20:00 <hqx> well not supre easy, but easy in a slightly inefficient manner and probably possible in a more efficient manner
20:20:18 <fizzie> It's nice how the categories go "easy" and then "probably possible".
20:20:45 <fizzie> It's like mathematicians, where the next step up from "trivial" is "open research question".
20:21:19 <fizzie> "Nope... No...This problem can't be done AT ALL. This one--maybe, but only with two yaks and a sherpa. ..."
20:22:01 <hqx> "...Possible, but it'll get you on the sex offenders registry in 32 states..."
20:22:10 <hqx> "...and besides, doing that to a horse is just /wrong/."
20:22:11 <Vorpal> `addquote <fizzie> It's like mathematicians, where the next step up from "trivial" is "open research question".
20:22:20 <hqx> it's slow.
20:22:22 <HackEgo> 225|<fizzie> It's like mathematicians, where the next step up from "trivial" is "open research question".
20:22:29 <hqx> can I add <fizzie> "Nope... No...This problem can't be done AT ALL. This one--maybe, but only with two yaks and a sherpa. ..." to that quote?
20:22:32 <hqx> it's funnier that way
20:22:49 <hqx> `addquote <fizzie> It's like mathematicians, where the next step up from "trivial" is "open research question". <fizzie> "Nope... No...This problem can't be done AT ALL. This one--maybe, but only with two yaks and a sherpa. ..." to that quote?
20:22:51 <HackEgo> 226|<fizzie> It's like mathematicians, where the next step up from "trivial" is "open research question". <fizzie> "Nope... No...This problem can't be done AT ALL. This one--maybe, but only with two yaks and a sherpa. ..." to that quote?
20:22:52 <hqx> Vorpal: we could add /two/ quotes :P
20:22:53 <fizzie> In interests of full disclosure, that one was from the Nukees comic.
20:23:03 <hqx> fizzie: what we don't know can't hurt us
20:23:11 <hqx> OH GOD CTHULHU HAS ARRIVED
20:23:16 <hqx> AAAAAAAAAAAAAAAAAAAAAAAAAAAA
20:23:20 <hqx> ...correction
20:23:24 <hqx> What we don't know can hurt us
20:23:29 <hqx> AAAAAAAAAAAAAAAAAAA MY LIMBS
20:24:05 <Vorpal> I do get the lovecraft reference but... why now?
20:24:09 <fizzie> Also, updatedb.py is curiously slow; it was a lot faster at work. (It seems to be mostly waiting for IO, though.)
20:24:27 <Vorpal> fizzie, it was the same speed both before and after the changes for me
20:24:29 <Sgeo|WebPuppy> So, even with the HD not in any use at all, not even inside the computer, it's degrading?
20:24:41 -!- poiuy_qwert has joined.
20:24:45 <Vorpal> fizzie, running on a sempron 3300+ with python -O
20:24:58 <fizzie> Vorpal: Yeah, I don't suppose the changes will affect IO times much.
20:25:11 <Vorpal> fizzie, and it is fast for me once I'm past the initial huge import
20:25:12 <hqx> inefficient way: thread { our_free_list := copy free list; deutschmarks := automatically_expanding_bit_list_of_size(heap_size); mark_into(deutschmarks); foreach bit <- bitlist { if !bit { if relevant_object(bit) free_according_to our_free_list { /* just allocated */ } else { free(relevant_object(bit)) } } }
20:25:19 <hqx> of a with-code mark and sweep
20:25:24 <Vorpal> fizzie, one thing that helped during initial import was to move db.commit() out to top level
20:25:32 <Vorpal> fizzie, rather than the intended level you had it at
20:25:49 <cpressey> hgx: you seem to have added "to that quote?" to that quote
20:26:06 <hqx> `addquote <fizzie> It's like mathematicians, where the next step up from "trivial" is "open research question". <fizzie> "Nope... No...This problem can't be done AT ALL. This one--maybe, but only with two yaks and a sherpa. ..."
20:26:11 <HackEgo> 227|<fizzie> It's like mathematicians, where the next step up from "trivial" is "open research question". <fizzie> "Nope... No...This problem can't be done AT ALL. This one--maybe, but only with two yaks and a sherpa. ..."
20:26:13 <Vorpal> hqx, now you did double revert
20:26:15 <fizzie> Oh well, it's already ~75% done, based on the size of the previous .db file, so I guess I'll just let it disk-thrash. But yes, I guess coimmit involves actual writes.
20:26:22 <Vorpal> hqx, what were you thinking!?
20:26:35 <hqx> i didn't see yours
20:26:42 <hqx> Is double revert like the opposite of double compile?
20:26:52 <hqx> You fix it, it's your fault for taking over my job. :--|
20:27:01 <Vorpal> hqx, no I don't. anyway look at the numbers
20:27:08 <Vorpal> I say all these are in
20:27:20 <hqx> Hell what?
20:27:32 <hqx> http://codu.org/projects/hackbot/fshg/index.cgi/rev/e12a6c39ae21
20:27:34 <hqx> "changeset 7"
20:27:35 <Vorpal> they got increasing ids
20:27:38 <hqx> that's not right
20:27:47 <hqx> `quote 225
20:27:50 <HackEgo> 225|<fizzie> It's like mathematicians, where the next step up from "trivial" is "open research question".
20:27:52 <hqx> `quote 226
20:27:54 <cpressey> and, writing the interpreter in CPS has made me wish I picked let and if instead of let* and cond (even though those lead to an obviously smaller implementation in Pixley)
20:28:00 <hqx> `quote 225
20:28:05 <HackEgo> 225|<fizzie> It's like mathematicians, where the next step up from "trivial" is "open research question".
20:28:14 <hqx> `quote 225
20:28:22 <Vorpal> hqx, maybe it uses a db outside the hg repo?
20:28:26 <hqx> was that the revision of--
20:28:28 <hqx> your adding the quote
20:28:30 <hqx> or the one before?
20:28:35 <Vorpal> http://codu.org/projects/hackbot/fshg/index.cgi/rev/e12a6c39ae21
20:28:36 <hqx> i think it /reverts/ the one you provide, thus leaving us one too far back
20:28:37 <hqx> not sure though
20:28:38 <hqx> `quote 224
20:28:40 <HackEgo> 224|<Gregor> It's impossible to add fake quotes.
20:28:44 <hqx> oh, never mind then
20:28:44 <hqx> `addquote <fizzie> It's like mathematicians, where the next step up from "trivial" is "open research question". <fizzie> "Nope... No...This problem can't be done AT ALL. This one--maybe, but only with two yaks and a sherpa. ..."
20:28:50 <HackEgo> 225|<fizzie> It's like mathematicians, where the next step up from "trivial" is "open research question". <fizzie> "Nope... No...This problem can't be done AT ALL. This one--maybe, but only with two yaks and a sherpa. ..."
20:28:50 <hqx> disaster perverted
20:29:09 <hqx> now /that/ would be fucked up
20:29:11 <Vorpal> hqx, yeah it reverts to the one you give
20:29:22 <Vorpal> hqx, well it could be doing cherry picking like darcs
20:29:27 <Vorpal> reverting exactly the one you gave
20:29:39 <Vorpal> rather than reverting to the one you gave
20:30:18 <Vorpal> hqx, wrt number 7: http://codu.org/projects/hackbot/fshg/index.cgi/graph/623d00b81b79
20:30:18 <hqx> i think reverts used to show in the commit tree
20:30:24 <hqx> and we were certainly at higher commit numbers
20:30:40 <fizzie> There's a "initial reimport" 10 days ago; seems like some maintenance.
20:30:46 <hqx> <ais523> addquote <cpressey> < ais523> then running repeatedly until you get the right sequence of random numbers < ais523> and just completely ignoring the input <-- some people live their entire lives this way, i reckon
20:30:49 <hqx> FFFFF IT HAS THE SPACE AFTER <
20:30:55 <hqx> `quote 223
20:30:56 <fizzie> Doing cherry-picking to a binary sqlite database would be quite impressive.
20:30:57 <HackEgo> 223|<alise> Why do you use random acronyms you know we don't know the expansions of? <pikhq> alise: TLAAW
20:30:59 <hqx> `quote 224
20:31:01 <HackEgo> 224|<Gregor> It's impossible to add fake quotes.
20:31:03 <hqx> `quote 222
20:31:04 <HackEgo> 222|< alise> Why do you use random acronyms you know we don't know the expansions of? < pikhq> alise: TLAAW
20:31:14 <hqx> It's all fucked! All fucked!
20:31:17 <hqx> Okay, I have a punning clan.
20:31:21 <hqx> `quote 219
20:31:23 <HackEgo> 219|<oklofok> colon is where your ass comes from right
20:31:33 <hqx> `revert d9be89875ec0
20:31:36 <hqx> `quote 220
20:31:38 <Vorpal> HackEgo, can you please
20:31:38 <HackEgo> 220|<cpressey> < ais523> then running repeatedly until you get the right sequence of random numbers < ais523> and just completely ignoring the input <-- some people live their entire lives this way, i reckon
20:31:42 <hqx> This will not take long.
20:31:59 <hqx> Vorpal: Umm, it's less spammy than a lot of fungot usage.
20:32:00 <fungot> hqx: ei oikeen tied ett fnord kokeilla fnord vai vihaako ne mua
20:32:02 <Vorpal> (like this intentionally is to show how it is to other people)
20:32:07 <fizzie> Ooh, Finnish from fungot!
20:32:08 <fungot> fizzie: and i was thinking of for example having an iterative fibonacci function wrapped inside one that only takes predicates and has an assumed decimal point for arithmetic operations).
20:32:17 <hqx> Now I'm going to continue repairing the timeline.
20:32:20 <Vorpal> fizzie, can you translate it?
20:32:28 <Vorpal> hqx, can you take it to /msg please
20:32:35 <hqx> Only if someone else complains.
20:32:45 <hqx> `quote 220
20:32:52 <Vorpal> hqx, what is wrong with quote 220?
20:32:52 <hqx> `quote 219
20:32:54 <HackEgo> 219|<oklofok> colon is where your ass comes from right
20:32:55 <fizzie> Vorpal: It's pretty ungrammatical, here's an approximation; "hqx: don't even know that fnord try fnord or whether they hate me".
20:33:09 <hqx> `addquote <cpressey> < ais523> then running repeatedly until you get the right sequence of random numbers < ais523> and just completely ignoring the input <-- some people live their entire lives this way, i reckon
20:33:11 <HackEgo> 220|<cpressey> < ais523> then running repeatedly until you get the right sequence of random numbers < ais523> and just completely ignoring the input <-- some people live their entire lives this way, i reckon
20:33:13 <hqx> Vorpal: the timeline was disturbed
20:33:15 <Vorpal> fizzie, not enough Finnish data to work well I guess
20:33:16 <hqx> I am re-adding quotes
20:33:19 <hqx> but eliding those meant to have been reverted
20:33:29 <hqx> `addquote <alise> "Europe is the national anthem of the Republic of Kosovo." <cpressey> alise: I <cpressey> I was going to say something then your last line floored me
20:33:31 <HackEgo> 221|<alise> "Europe is the national anthem of the Republic of Kosovo." <cpressey> alise: I <cpressey> I was going to say something then your last line floored me
20:33:36 <fizzie> Vorpal: There shouldn't be any, that was probably from #douglasadams.
20:33:37 <hqx> `addquote <alise> Why do you use random acronyms you know we don't know the expansions of? <pikhq> alise: TLAAW
20:33:40 <HackEgo> 222|<alise> Why do you use random acronyms you know we don't know the expansions of? <pikhq> alise: TLAAW
20:33:44 <hqx> `addquote <Gregor> It's impossible to add fake quotes.
20:33:49 <HackEgo> 223|<Gregor> It's impossible to add fake quotes.
20:33:51 <Vorpal> fizzie, you used some Finnish in here
20:33:52 <hqx> `addquote <fizzie> It's like mathematicians, where the next step up from "trivial" is "open research question". <fizzie> "Nope... No...This problem can't be done AT ALL. This one--maybe, but only with two yaks and a sherpa. ..."
20:33:54 <HackEgo> 224|<fizzie> It's like mathematicians, where the next step up from "trivial" is "open research question". <fizzie> "Nope... No...This problem can't be done AT ALL. This one--maybe, but only with two yaks and a sherpa. ..."
20:33:56 <Vorpal> fizzie, to oklopol and such
20:34:10 <hqx> I should have added more spaces in-between the messages while I had the chance. Oh well.
20:34:56 <Vorpal> hqx, which one did you remove?
20:35:41 <hqx> Just the one with <>s starting with a space that I already tried to revert, and failed without knowing so. I then readded it without the extra spaces, which I have restored in this timeline.
20:35:48 <hqx> I've just compensated for `revert being broken.
20:36:14 <Vorpal> <hqx> `addquote <cpressey> < ais523> then running repeatedly until you get the right sequence of random numbers < ais523> and just completely ignoring the input <-- some people live their entire lives this way, i reckon <-- it is right there?
20:36:21 <Vorpal> hqx, and don't touch it
20:36:24 <hqx> Vorpal: not that one
20:36:26 <Vorpal> the space should be there
20:36:26 <hqx> a different one
20:36:33 <hqx> find out your fucking self
20:36:40 <hqx> i preserved it in /that/ one because cpressey quoted it that way
20:36:43 <hqx> and it was his message
20:36:53 <fizzie> Heh, I was trying to take a sum(length(body)) over all messages in the db, and forgot the length() part; rather curious output:
20:36:54 <fizzie> sqlite> select sum(body) from logs where type = 0;
20:37:17 <Vorpal> fizzie, curious indeed
20:37:34 <fizzie> I guess it coerces everything to numerics, and there's something that starts with "Inf" that gets turned to +infinity.
20:37:58 <hqx> fizzie: Are you cool enough to upload the DB itself? :|
20:38:29 <hqx> uploading to zem.fi would just be a cp if i understand your home-network correctly :P
20:38:30 <Vorpal> fizzie, do you know any locale like C but with UTF-8?
20:38:33 <hqx> or an rsync i guess
20:38:42 <hqx> Vorpal: en-US' UTF-8 locale
20:38:52 <Vorpal> hqx, no it doesn't sort the right way
20:39:04 <Vorpal> since I need that property here
20:39:05 <hqx> Sgeo|WebPuppy: no
20:39:24 <Vorpal> hqx, does that exist? hm
20:39:39 <hqx> Vorpal: probably not
20:39:53 <Vorpal> it doesn't work, it defaults to C
20:39:57 <hqx> :( my touchpad isn't cool enough to do two-finger scrolling
20:39:59 <fizzie> hqx: Well, I guess I can put it webbable for a while, but thanks to my crummy upload pipe, it'll take a while to download.
20:40:13 <hqx> fizzie: how slow's the pipe?
20:40:21 <hqx> also, it'd take that long anyway
20:40:26 <hqx> since you'd have to upload it somewhere
20:40:27 <Vorpal> <hqx> fizzie: Are you cool enough to upload the DB itself? :| <-- just use the script itself
20:40:36 <Vorpal> since the db is quite large
20:40:40 <Vorpal> hqx, it is mostly automatic anyway
20:40:42 <hqx> no, i haven't got all the logs in the format it wants
20:40:46 <hqx> and that would take longer
20:40:53 <Vorpal> hqx, no, tunes.org is fast
20:41:02 <hqx> i don't want two copies of the text logs.
20:41:21 <Vorpal> hqx, you can't update the db then
20:41:28 <Vorpal> hqx, and what format do you have them in?
20:41:36 <fizzie> Vorpal: Can't you just set the LC_CTYPE category to en_US.UTF-8 but LC_COLLATE to C (or POSIX, I guess it's identical).
20:42:04 <fizzie> hqx: But then if I uploaded to some place with a fast pipe, all the interested parties (a total of you) could get it.
20:42:11 <Vorpal> fizzie, hm... doesn't seem to work for the initdb stuff of postgres
20:42:29 <Vorpal> oh wait, I think I found a way around
20:42:30 <hqx> Vorpal: tunes format
20:42:46 <Vorpal> hqx, he has it in that too? Just with mtime being set as on server
20:42:56 <fizzie> Vorpal: "The category names translate into names of initdb options to override the locale choice for a specific category. For instance, to set the locale to French Canadian, but use U.S. rules for formatting currency, use initdb --locale=fr_CA --lc-monetary=en_US."; something like that.
20:43:14 <hqx> Vorpal: precisely
20:43:21 <Vorpal> fizzie, yeah I need to modify which flags /etc/rc.d/posgresql passes to initdb
20:43:24 <hqx> anyway grr databases suck (botte needs to have this crap in a db too)
20:43:27 <Vorpal> so I can use it by hand
20:43:39 <hqx> there needs to be a ... LISPABASE!
20:43:50 -!- Sgeo|WebPuppy has quit (Ping timeout: 252 seconds).
20:45:59 <Vorpal> hqx, and yes databases suck
20:46:15 <Vorpal> hqx, but give be a better alternative for this. object db doesn't seem very useful here
20:46:25 <hqx> as I said: Lispabase!
20:46:42 <hqx> Do not ask what it is! I am formulating it right now!
20:47:30 <Vorpal> hqx, so... what is it? :D
20:47:38 <fizzie> hqx: Okay, it's at http://zem.fi/~fis/logs.db -- might spew it out as text/plain, do use wget or curl -- but if you want to keep the .db file up-to-date, you'll need the text logs. (And the mtime trick is only used by the fetch script, the updatedb.py one should run just fine if you just have the text files there; though it might need a two-line tweak if you have the original file names and not my YYYY-mm-dd.log ones.)
20:48:04 <hqx> Does anyone know of a nicer way to write the prolog bagof(X,p(X),X)?
20:48:20 <hqx> fizzie: 403 Joe for-Biden.
20:48:52 <hqx> fizzie: Do they work with your log format in the olde logs?
20:49:01 <hqx> Vorpal: If you start wgetting that too I'll murder you.
20:49:09 <Vorpal> hqx, why would I wget it
20:49:11 <hqx> fizzie: Guh, your upload really is the terrible. Oh well; I can wait 41 minutes.
20:49:13 <hqx> Vorpal: To slow mine down.
20:49:29 <fizzie> hqx: No, updatedb.py just handles the tunes.org text. And the upload's 1 Mbps, approximately.
20:51:39 <cpressey> I would have said Sexpbase, but I admit that does not roll off the tongue as easy
20:52:14 <hqx> The problem with a Lispabase is, like all Lisp database people have realised, the perfect database is basically sub-TC Prolog. So they end up just implementing that. And admittedly, Allegro Cache is probably amazing.
20:52:51 <cpressey> for retrieval expressivity, yes, sub-TC prolog
20:54:10 <cpressey> also, most web services end up needing a query language like that too
20:54:21 <hqx> the problem is making it efficient, ofc
20:54:37 <cpressey> and... undoability is... a parallel problem i guess
20:54:50 <hqx> undoability? like atomic transactions or what?
20:55:01 <hqx> :- dynamic(stored/1).
20:55:01 <hqx> memo(Goal) :- ( stored(Goal) -> true ; Goal, assertz(stored(Goal)) ).
20:55:03 <hqx> oh my god that is beautiful
20:56:08 <cpressey> atomicity is... another parallel problem... maybe if you have it versioned like a VCS... well anyway
20:56:15 <fizzie> Heh, you *can* get gnuplot to produce histograms: just use something like "plot 'data' using (0.2*int($1/0.2)):2 smooth frequency" -- the 0.2*int(x/0.2) bit collapses bins into one X value, and "smooth frequency" makes it sum together such values.
20:56:44 <hqx> hmm, in prolog dbs how do they do multiple columns?
20:56:45 <hqx> i'd guess like
20:56:52 <hqx> col1(ID, Col1)
20:56:55 <hqx> col2(ID, Col2)
20:57:01 <hqx> but what if you don't want an ID?
20:57:06 <hqx> just table(Cols) where Cols is a list?
20:57:22 <cpressey> hell, just use full-TC prolog. you're going to have queries that time out *anyway*
20:57:48 <hqx> <cpressey> hell, just use full-TC prolog. you're going to have queries that time out *anyway*
20:57:55 <hqx> full-TC prolog is even more inefficient, i think
20:59:56 <cpressey> sufficiently clever compiler ftw
21:01:03 <cpressey> (not serious. my ftws are generally at least 50% facetious)
21:01:44 <hqx> prolog really needs that wonderful multiplexer thing. sigh.
21:02:34 <fizzie> Like a man once said, sufficiently clever compilers are indistinguishable from magic.
21:07:53 <hqx> 08:45:54 <ais523> humans are very good at pretending to be IRC bots, I find
21:08:05 <hqx> the great thing about this log is that i've forgotten whether it was really even a bot or not
21:08:18 <hqx> <hqx> prolog really needs that wonderful multiplexer thing. sigh. ;; er, prolog-DB-language :P
21:08:51 <hqx> ahh optbot
21:08:53 <hqx> the mem'rys
21:14:15 <Vorpal> why does postgresql have that elephant logo?
21:14:25 <hqx> why is linux a penguin?
21:14:32 <hqx> also, elephants never phorget.
21:14:47 <fizzie> I would have guessed the "elephants equals memory" thing too.
21:15:23 -!- Gracenotes has joined.
21:15:43 <Vorpal> <hqx> why is linux a penguin? <-- I heard the story behind that
21:15:46 <hqx> what's the less(1) for "stfu and display the ansi codes"?
21:15:52 <hqx> Vorpal: that story was one after the fact
21:15:54 <hqx> not before it was drawn
21:15:58 <hqx> unless it's a different one
21:16:01 <Vorpal> as the cmd line switch
21:16:13 <hqx> yeah, -r or -R
21:16:16 <Vorpal> if you mean it mangles terminal control codes
21:16:22 <Vorpal> hqx, one of those yeah
21:16:29 <hqx> $ grep --colour -r optbot . | less -R
21:16:33 <hqx> i used to know how to make this work...
21:16:37 <fizzie> The internet claims this was the source: http://www.pgsql.ru/db/mw/msg.html?mid=1238939
21:16:38 <hqx> still not working though
21:16:43 <Vorpal> hqx, the one about Torvalds and penguins?
21:16:54 <hqx> obscure: a revolver/hit man (Grosse Pt is an anagram of Postgres, and an
21:16:54 <hqx> abbreviation of the title of the new John Cusack movie)
21:17:13 <hqx> Vorpal: well, i remember Torvalds' description of it as just having eaten and then copulated.
21:17:20 <hqx> so, uh, there is that
21:17:32 <Vorpal> hqx, didn't it try to bite him?
21:17:45 -!- poiuy_qwert has quit (Quit: Leaving).
21:17:47 <Vorpal> must have misremembered
21:17:55 <hqx> i don't think there was a real penguin.
21:18:09 <Vorpal> hqx, I think there was one when he was down at some zoo or something
21:18:27 <hqx> i'm not about to grab my copy of Just For Fun and look it up :)
21:18:49 -!- Sgeo|Puppy has joined.
21:19:24 <fizzie> I do remember some sort of penguin assault story, too.
21:19:34 <hqx> so what's the grep | less thing to get the lovely colours?
21:19:54 <hqx> --colour=force
21:20:02 <hqx> $ grep --colour=force -r optbot . | less -R
21:20:06 <hqx> Sgeo|Puppy: what?
21:20:28 <hqx> ./08.08.07:13:55:01 --- nick: OerjansTerribleP -> otpbot
21:20:40 <Sgeo|Puppy> On my HD, there was a NTFS partition and a partition of type 88
21:20:56 <Sgeo|Puppy> Which according to my friend's Knoppix's fdisk -l, is Linux Plaintext
21:21:06 <cpressey> plaintext partition. sounds yummy
21:21:20 <hqx> might merely be corrupted data
21:21:22 <hqx> which you know you have
21:22:01 <hqx> 02:49:42 * tusho writes a parser for a restricted subset of querying english
21:22:01 <hqx> 02:49:53 <tusho> rabbit's tails' lengths -> rabbit.tails.map(length)
21:22:01 <hqx> 02:50:08 <tusho> and then I can parse "x is y", "x has y" and build up a silly little knowledge databse
21:22:01 <hqx> 02:50:11 <tusho> *database
21:22:01 <hqx> 02:50:14 <tusho> via irc!
21:22:06 <hqx> I sure wish you finished writing that, tusho!
21:22:15 <hqx> 03:01:26 <tusho> "the length of the rabbit's tail is 5cm" -> (is (rabbit tail length) (5cm)) is kind of non-trivial...
21:22:19 <hqx> Eff you! Keep working!
21:22:41 <fizzie> Usually it should be 83 or 8e or fd.
21:23:00 <fizzie> "What to do afterwards? Last year I made a hack, reserving type 88 hex for a Linux plaintext partition table. You must be able to find the kernel patch somewhere on Google, otherwise ask. No fdisk required, the partition table is just plaintext that you edit using emacs or vi."
21:23:03 <hqx> fizzie: Here, you write that parser.
21:23:04 <fizzie> Doesn't sound very widely used.
21:23:21 <fizzie> I'll delegate to our natural-language folks.
21:23:27 <hqx> AUUUGUUUUR
21:23:34 <cpressey> regarding linux plaintext partitions
21:23:40 <hqx> i doubt it's really that
21:23:43 <hqx> since he's lost huge swathes of data
21:23:45 <cpressey> nor is gogol, but he has an excuse
21:23:51 <hqx> some stuff is bound to just be corrupted
21:24:23 <cpressey> i like how the list of partition types has come down through the ages like a mythological narrative
21:24:41 <fizzie> Well, it's helpful if you want to know the reason, but I can't seem to be able to find the patch in question.
21:24:50 <cpressey> 88? oh, that's a Linux Plaintext Partition! What's that? Oh, it's...
21:24:58 <Vorpal> <hqx> 02:49:42 * tusho writes a parser for a restricted subset of querying english <-- W|A?
21:25:20 <hqx> Vorpal: no :P
21:25:21 <fizzie> cpressey: It's actually "Linux Plaintext Partition Table", though; seems that some places have cropped the name.
21:25:25 <hqx> all it would do is
21:25:45 <hqx> "x's y" -> x.y; "xs' y" -> xs.map(y)
21:25:49 <hqx> "x of y" -> y.x
21:25:57 <hqx> "the x" -> x, more or less, or perhaps member-of(xs)
21:26:07 <hqx> and infix stuff like "X is Y" -> is(X,Y)
21:26:07 <cpressey> fizzie: that makes more sense. but also provides even more support for my "oral tradition" analogy.
21:26:19 <hqx> plus some stuff for handling plurals like person/people
21:26:40 <cpressey> but, still, writing a partition table in plaintext? that's pretty weird
21:26:55 <Vorpal> cpressey, yeah apart from a handful of well known ones no one really knows
21:27:12 <Vorpal> cpressey, personally I only have types fd and 82
21:27:21 <fizzie> There are a few lists of the types, with descriptions.
21:27:36 <Vorpal> cpressey, yeah you should use XML
21:28:32 <fizzie> Maybe they could invent a XML-based GPTv2 for EFI next.
21:28:44 <fizzie> Then everyone would have to have a XML parser in BIOS.
21:28:49 <fizzie> Now that'd be the awesome.
21:29:28 <Vorpal> fizzie, anyway 83 is the normal linux one. fd is raid with kernel auto config, 8e is linux lvm.
21:29:37 <Vorpal> just to help with any confusion
21:29:56 <fizzie> I know those few; that's why I listed them in particular.
21:31:00 <Vorpal> fdisk lists "a0 IBM Thinkpad-vi" here
21:31:12 <Vorpal> and what is it used for
21:31:25 <fizzie> "Reported for various laptops like IBM Thinkpad, Phoenix NoteBIOS, Toshiba under names like zero-volt suspend partition, suspend-to-disk partition, save-to-disk partition, power-management partition, hibernation partition. Usually at the start or end of the disk area."
21:31:33 <fizzie> Says my handy list at http://www.win.tue.nl/~aeb/partitions/partition_types-1.html
21:31:43 <fizzie> (The list also contains some links to other lists.)
21:31:46 -!- Sgeo|Puppy has quit (Ping timeout: 245 seconds).
21:32:02 <fizzie> Don't know why "-vi" in particular, though.
21:32:14 -!- Slereah has quit (Ping timeout: 264 seconds).
21:32:54 <fizzie> My fdisk lists a0 as "IBM Thinkpad hi", and it probably should continue "...ernation" or something.
21:34:17 <Vorpal> fizzie, seems to have LANG set
21:35:28 <fizzie> Heh. According to strings (and probably the corresponding translation files) they have longer names in there, I just don't know how to make fdisk display them.
21:37:18 -!- BeholdMyGlory has quit (Remote host closed the connection).
21:37:36 -!- BeholdMyGlory has joined.
21:37:46 -!- Slereah has joined.
21:48:56 -!- Sgeo|Puppy has joined.
21:55:27 -!- prettyrobots has joined.
21:55:48 -!- prettyrobots has left (?).
22:03:05 -!- hqx has changed nick to alise.
22:05:47 <Sgeo|Puppy> Puppy Linux, there's all these little things getting on my nerves
22:07:59 <fizzie> http://zem.fi/~fis/act.png -- first gnuplot-based log-activity plot; doesn't do cumulative yet, does only rectangular-window smoothing, only 2010-01 -- 2010-09 (but with labels on the axis), default color specification isn't so hot.
22:09:25 <fizzie> Also, Y-scale lies a lot. Well, for some values of lying.
22:10:36 <alise> See that drop to the buttom?
22:11:46 <alise> FUCK NSPLUGINWRAPPER
22:21:47 <Vorpal> fizzie, what package do you suggest for interfacing with postgres in python?
22:22:57 <Vorpal> <alise> FUCK NSPLUGINWRAPPER
22:24:12 <Vorpal> alise, now I have that star trek parody song you linked a few days ago (or was it yesterday?) stuck on my head
22:24:32 <Vorpal> fizzie, cumulative is easier to read
22:25:25 <Vorpal> fizzie, but as you can see you dropped below my level several times. But at other times you stood for more than those not on the 10-top list
22:26:24 <fizzie> Vorpal: I don't remember which Postgres/Python thing I've used; something that does the usual dbapi2 or whatever they called that.
22:26:49 <Vorpal> fizzie, both things I can find in arch repos does the dbapi2 stuff
22:26:57 <Vorpal> fizzie, I want to know which one though :/
22:26:58 <fizzie> Psycopg sounds familiar.
22:27:22 <Vorpal> fizzie, does "pypgsql" sound familiar as well?
22:27:32 <fizzie> http://wiki.postgresql.org/wiki/Python -- there certainly seems to be many of them.
22:28:26 <Vorpal> the pypsql is dead anyway it seems
22:30:04 <fizzie> http://zem.fi/~fis/act.png -- updated for the full clog range, but not yet cumulative/relative. To get the cumulative sums is easy, but I'm not yet sure how to do filled areas with gnuplot. (That one has a time-step of one day per window, and a window length of one full month, so it averages more.)
22:31:38 <olsner> Nants ingonyama bagithi Baba!
22:32:30 <fizzie> I think I can write cumulative with just a suitable "using" spec, but it's going to be ugley; maybe I should have Python sum them up.
22:33:42 <olsner> hmm, I don't think it's necessarily wise to let python do anything
22:34:31 <fizzie> Python's ternary operator is just freaky, though.
22:34:58 <fizzie> Sure; it's written "x if c else y", and it does "c ? x : y".
22:35:21 <Vorpal> fizzie, it reads like English though?
22:35:22 <fizzie> But it's a real expression, so you can stick it wherever.
22:35:52 <fizzie> It reads a bit like english, but not so much.
22:35:54 <Vorpal> fizzie, I remember seeing some rationale for it
22:36:19 <fizzie> They do have a rationale about using it in cases where x is the "usual" case, and y the "exceptional" value.
22:36:46 <olsner> rationale? if it's more articulate than "ook? ook!" I'll be officially surprised
22:37:08 <Vorpal> olsner, fizzie posted it
22:37:39 <olsner> Vorpal: oh, is that what he meant by "ook? ook!"? I see :)
22:39:13 <Vorpal> fizzie, hm python docs seem to suggest using a cursor and execute on that
22:39:22 <Vorpal> fizzie, rather than doing it directly on the connection
22:39:46 <fizzie> I don't know either; doing it directly on the connection creates a cursor and runs execute on the new one.
22:39:58 <olsner> hmm, perhaps they stole it from a VB style-guide or something?
22:41:49 <fizzie> Hm, well, a cumulative plot: http://zem.fi/~fis/actf.png
22:42:49 <Vorpal> AttributeError: 'psycopg2._psycopg.connection' object has no attribute 'execute'
22:42:59 <Vorpal> fizzie, not for psycopg2 it seems
22:43:04 <fizzie> Heh, maybe they haven't defined the shortcuts, then.
22:50:51 <fizzie> Well, to complete the trifecta, relative-activity plot from gnuplot: http://zem.fi/~fis/actr.png -- I especially like it how it put the nicknames behind the filled plot, not in front.
22:53:09 <fizzie> Incidentally, according to this data there has been one 31-day period during which alise has been responsible for 46.1 % of all the lines said on the channel. It's not quite half, but impressively close. (It probably would go well over one half if I shortened the window from month to, say, wekk.)
22:55:03 <fizzie> Oh, actually that's just an artefact; the 46.1 % peak is the window at 2010-08-29, which isn't a full month long, only half. But still, there's something at least very close to 45 % in autumn 2009.
22:57:59 <fizzie> Also funny how pre-2006 summer there's been a completely different set of people talking.
22:58:37 -!- MigoMipo has quit (Read error: Connection reset by peer).
23:00:18 <fizzie> (Cf. http://zem.fi/~fis/act-old.png )
23:01:02 -!- BeholdMyGlory has quit (Read error: Connection reset by peer).
23:02:01 <fizzie> (Er, except actually I; I had managed to hide my own stuff from the actr.png I had up earlier, fixed now.)
23:02:58 <Vorpal> fizzie, hrrm: http://wiki.postgresql.org/wiki/Todo:Collate
23:03:53 <fizzie> Vorpal: Heh, sounds messy.
23:04:56 <Vorpal> fizzie, can you fix the nick labels on http://zem.fi/~fis/actr.png ?
23:05:05 <fizzie> Vorpal: From the section on index use on LIKE expressions, which is basically where the collation makes a difference: "The optimizer can also use a B-tree index for queries involving the pattern matching operators LIKE and ~ if the pattern is a constant and is anchored to the beginning of the string — for example, col LIKE 'foo%' or col ~ '^foo', but not col LIKE '%bar'. However, if your server does not use the C locale you will need to create the index with
23:05:06 <fizzie> a special operator class to support indexing of pattern-matching queries. See Section 11.8 below."
23:05:28 <fizzie> I can make the graph endpoint to something like middle of 2011, that should unhide the nicks.
23:05:33 <fizzie> I don't quite know how to get them on top.
23:05:47 <Vorpal> fizzie, couldn't you put them on the side
23:05:56 <fizzie> I guess I can do that, too.
23:06:08 <fizzie> It makes the graph area smaller, but then again so does moving the endpoint.
23:06:22 <Vorpal> fizzie, just made the total drawing wider!
23:07:26 <fizzie> I'm not quite sure what the optimal width is.
23:07:30 <Vorpal> fizzie, being able to do:
23:07:33 <Vorpal> AS ENUM ('msg', 'act', 'notice', 'join', 'part',
23:07:33 <Vorpal> 'quit', 'kick', 'nick', 'topic', 'mode');
23:07:48 <fizzie> Anyway, actr.png quick-fixed with "set key outside right center".
23:08:56 <fizzie> Also, I'unno; I'd rather do the types in the client app, just in case the SQL API can't handle customizations like that. The esolog.py that does the esolog.logrange() generator on-the-fly maps the numbers to strings like that and back.
23:09:18 <Vorpal> Sgeo|Puppy, you are, in *
23:09:25 <Vorpal> Sgeo|Puppy, you simply don't talk enough
23:09:36 <fizzie> Either that, or you use non-fixed nicks too much.
23:09:37 <Vorpal> fizzie, you could try to merge Sgeo* to see if it matters
23:09:57 <Vorpal> fizzie, like you should merge fizzie fizzien900
23:10:07 <fizzie> I could, yes. I did get Sgeo in the 2010-only plot.
23:10:17 <Vorpal> fizzie, I always liked to read the latter fizzien 900
23:10:22 <Vorpal> rather than fizzie n900
23:10:32 <Sgeo|Puppy> Also, I want to know the date I first appeared here
23:10:44 <Vorpal> Sgeo|Puppy, just import the thing into your db
23:11:00 <Vorpal> Sgeo|Puppy, complete clog logs
23:11:25 <Vorpal> I'll check for first date for you
23:11:28 <fizzie> Sgeo|Puppy: 2005-11-10 21:48:04, according to clog.
23:11:43 <fizzie> (That's a nick-change from Sgeo to Sgep.)
23:11:50 <Sgeo|Puppy> Didn't realize there was a full log thing
23:12:30 <Vorpal> 91399|2005-11-10 21:48:04|Sgeo|||7|Sgep
23:12:39 <Vorpal> I think it is a nick change
23:12:50 <fizzie> Vorpal: <fizzie> Sgeo|Puppy: 2005-11-10 21:48:04, according to clog.
23:12:55 <fizzie> Vorpal: <fizzie> (That's a nick-change from Sgeo to Sgep.)
23:13:00 <fizzie> "Is there an echo in here?"
23:13:13 <fizzie> Aaanyway, link to logs is in the topic.
23:13:18 <fizzie> It's always been in the topic.
23:13:21 <Vorpal> fizzie, first line you said was in: 107493|2006-01-23 02:05:47|Sgeo|||0|Bye all
23:13:25 <Vorpal> Sgeo|Puppy, ... which bot?
23:13:49 <Vorpal> Sgeo|Puppy, what do you think we did for these stats?
23:14:01 <fizzie> Sgeo|Puppy: There scripts we've been discussing have been posted, though.
23:14:01 <Vorpal> Sgeo|Puppy, we imported them into a db from those files!
23:14:30 <Vorpal> just grep the last log for sprunge
23:14:42 <Vorpal> fizzie, unless you made any further updates?
23:14:46 <fizzie> Sgeo|Puppy: Or: http://p.zem.fi/esolog-fetch.py http://p.zem.fi/esolog-updatedb.py
23:14:56 <fizzie> That's not "further updates", but it does include your fixed regexps.
23:15:05 -!- Sgeo|Puppy has quit (Read error: Connection reset by peer).
23:15:09 <Vorpal> fizzie, what you update in place?
23:15:19 <fizzie> Vorpal: Sure, why not?-)
23:15:37 <Vorpal> fizzie, <insert something about keeping urls unchanging once published>
23:16:11 <fizzie> There's no version number, so any changes are fair game.
23:16:22 -!- Sgeo|FsckThatCli has joined.
23:16:32 <Vorpal> fizzie, it is strange that the first sgeo message with type 0 is "107493|2006-01-23 02:05:47|Sgeo|||0|Bye all"
23:16:42 <fizzie> If it makes you feel better, consider those URLs to unchangingly point to the latest version.
23:16:54 <Vorpal> Sgeo|FsckThatCli, 170 MB or such, >200 MB when importing it into the db
23:16:57 <Vorpal> for the db file itself
23:17:37 <Vorpal> the latter contain all *.log
23:17:55 <fizzie> 66574|2005-09-17 21:16:51|Sgep|||0|Hi all
23:17:59 <fizzie> That's the first Sgep message.
23:18:42 <alise> <fizzie> Incidentally, according to this data there has been one 31-day period during which alise has been responsible for 46.1 % of all the lines said on the channel. It's not quite half, but impressively close. (It probably would go well over one half if I shortened the window from month to, say, wekk.)
23:18:42 <alise> <fizzie> Oh, actually that's just an artefact; the 46.1 % peak is the window at 2010-08-29, which isn't a full month long, only half. But still, there's something at least very close to 45 % in autumn 2009.
23:18:44 <alise> i am fucking awesome
23:19:08 <Vorpal> <fizzie> That's the first Sgep message. <-- ah yes
23:19:10 <fizzie> alise: I'm sure I could think of alternative adjectives, too.
23:19:32 <fizzie> Vorpal: Maybe a off-by-one error incremented his nick accidentally.
23:19:46 <alise> <Sgeo|Puppy> Yes, I know where the 1 day logs are
23:19:46 <alise> <Sgeo|Puppy> Do I have to combine them myself?
23:20:51 <Vorpal> 66580|2005-09-17 21:16:51|Sgep|||0|Hi all
23:20:52 <Vorpal> 66649|2005-09-17 22:39:11|Sgep|||1|is using konq, and can't figure out Java, so I can't really see EsoShell :-(
23:20:56 <Vorpal> Sgeo|FsckThatCli, what was esoshell?
23:21:22 <Sgeo|FsckThatCli> 15:37:23 <calamari> ihope: http://esoshell.kidsquid.com/ 15:37:45 <calamari> ihope: it allows you to use certain esoteric languages straight from your web browser
23:21:49 <fizzie> Heh, last first-page google-result for esoshell: "Search Kazaa for esoshell brainfuck - esolang albums. Page 1 of 0 ..." Yeah, I'm sure there's a lot of albums like that.
23:22:23 <Vorpal> http://esoshell.kidsquid.com/ = 403
23:22:37 <Vorpal> anyone seen calamari recently?
23:23:04 <Vorpal> 1804491|2010-09-15 07:16:24|calamari|||5|Quit: Leaving
23:23:41 <Vorpal> Sgeo|FsckThatCli, what was PESOIX?
23:24:26 <Vorpal> sqlite> select count(*) from logs where type = 0 and body like '%?';
23:24:31 <Vorpal> that's a lot of questions
23:24:45 <Sgeo|FsckThatCli> Hear complains about the name being too similar, so go with phonetics: P-ESO-X = P S O X
23:25:13 <fizzie> sqlite> select count(*) from logs where type = 0 and body like '%? ';
23:25:16 <fizzie> That's probably mostly me.
23:25:53 <fizzie> sqlite> select nick, count(*) as k from logs where type = 0 and body like '%? ' group by nick order by k desc;
23:25:58 <fizzie> Well, it's not *only* me.
23:26:13 <Vorpal> fizzie, why is it mostly you though
23:26:26 <Vorpal> that is the number of lines ending in ?
23:26:37 <fizzie> I'm sure I've told the story before; we had a babblebot that tried to answer anything that ended with "?" on another channel, so I got into a habit of adding a space after a trailing ? to keep it quiet.
23:26:47 <Vorpal> select nick, count(*) as cnt from logs where type = 0 and body like '%?' group by nick order by cnt desc LIMIT 20;
23:27:10 <quintopia> fizzie: that sounds like the worst bot ever!
23:27:13 <alise> <fizzie> alise: I'm sure I could think of alternative adjectives, too. ;; awesomely fucking?
23:27:30 <Vorpal> alise, when did fizzie say that?
23:28:04 <Vorpal> fizzie, how do you merge ehird* into one post for the count here
23:30:20 <Vorpal> alise, hah I win over you. I asked more questions than you. Not a large difference (24763 vs 32918).
23:31:05 <Vorpal> sqlite> select count(*) as cnt from logs where (nick like 'Sgeo%') and type = 0 and body like '%?' LIMIT 20;
23:31:16 <Vorpal> Sgeo|FsckThatCli, you asked quite a few. Hm should do questions / total lines
23:31:19 <Vorpal> that would be interesting
23:32:18 <alise> <Vorpal> alise, hah I win over you. I asked more questions than you. Not a large difference (24763 vs 32918). ;; so i'm wiser.
23:32:20 <fizzie> Vorpal: You have more comments in general, though; you should compare the question fractions instead.
23:32:31 <fizzie> sqlite> select qn, (kq+0.0)/k from (select nick as qn, count(*) as kq from logs where type = 0 and body like '%?' group by nick) join (select nick as tn, count(*) as k from logs where type = 0 group by nick) on qn = tn where k > 100 order by (kq+0.0)/k desc limit 10;
23:32:31 <fizzie> ELIZA|0.675496688741722
23:32:31 <fizzie> amca|0.461077844311377
23:32:31 <fizzie> KajirBot|0.458536585365854
23:32:45 <fizzie> That's "fraction of questions" for people with more than 100 comments in total.
23:33:01 <fizzie> (There are a few nicks with only questions, which is pretty boring.)
23:33:09 <Vorpal> <fizzie> Vorpal: You have more comments in general, though; you should compare the question fractions instead. <-- hm what do you mean
23:33:18 <Vorpal> I thought that was what I said
23:33:42 <fizzie> Vorpal: Yes, it was a reply to the earlier comment on you winning; I was busy writing that statement.
23:33:44 <alise> ./07.11.24:16:50:37 --- quit: KajirBot (Remote closed the connection)
23:33:44 <alise> ./09.01.19:10:12:16 <ehird_> nope, KajirBot and seabot
23:34:21 <fizzie> Filtering on >1000 comments gives real people: http://p.zem.fi/zpw4 for top-10.
23:34:36 <fizzie> (Also, Sgeo wins there.)
23:35:06 -!- cpressey has quit (Quit: leaving).
23:36:13 <Vorpal> Vorpal: 32918 / 223498 = 0.147285
23:36:14 <Vorpal> alise: 24763 / 392906 = 0.063025
23:36:14 <Vorpal> Sgeo: 5615 / 25044 = 0.224205
23:36:26 <Vorpal> fizzie, that is with nick aggregation
23:36:35 <Vorpal> alise, no it means that you are not curious
23:37:12 <alise> i just work things out myself instead :)
23:37:31 <Vorpal> alise, no I much prefer my interpretation here!
23:37:43 <alise> Sgeo|FsckThatCli: no: we're the ones suffering. trust me.
23:37:54 <alise> ./09.01.19:10:13:23 <KajirBot> sorry, i haven't eaten something before. have you got ten black holes?
23:37:59 <alise> N... not really, no.
23:38:22 <alise> 08:10:36 <Sgeo> Why is it that it seems no one cares?
23:38:22 <alise> 08:12:01 * pikhq wakes
23:38:22 <alise> 08:12:15 <Sgeo> hi pikhq
23:38:22 <alise> 08:12:22 <ehird`_> Sgeo: because people do other things than talk about psox
23:38:22 <alise> 08:12:56 * Sgeo goes to #PSOX
23:38:23 <Vorpal> alise, markov chain I presume? Or something else
23:38:33 <alise> Vorpal: no, that was when you tried to eat it anything other than a ridiculous item
23:39:05 <alise> pikhq: Have you read The Metamorphosis of Prime Intellect?
23:39:07 <Vorpal> that makes a lot more sense
23:39:44 <Vorpal> no, don't pollute the sun!
23:40:34 <Sgeo|FsckThatCli> It suddenly occurs to me that that will not result in a black hoe
23:41:06 -!- oerjan has joined.
23:42:53 <Vorpal> Sgeo|FsckThatCli, merging sgep* into your stats too your question ratio goes down to 0.223952
23:43:15 <Vorpal> this would indicate you have become more questioning over time perhaps
23:43:21 -!- tombom_ has quit (Quit: Leaving).
23:43:27 <Vorpal> probably within the statistical error margin though
23:44:14 <Vorpal> 0.2 exactly. but 53/265 shows how small the data set is
23:44:18 <Sgeo|FsckThatCli> We should look for days where for some reason, questioning rations for people, at least one person, are particularly high
23:44:21 <pikhq> alise: Yes, I have.
23:44:26 <Vorpal> so unlikely to be statistically significant
23:44:39 <alise> pikhq: I'm reading it now.
23:44:57 <Vorpal> Sgeo|FsckThatCli, I'm matching on Sgeo% not Sgeo
23:45:24 <Vorpal> <Sgeo|FsckThatCli> We should look for days where for some reason, questioning rations for people, at least one person, are particularly high <-- you do that, not me
23:46:35 <Vorpal> alise, anyway I suspect if we cut away your monologues you will have a lot higher question ratio
23:46:49 <alise> 11:11:35 * Sgeo is here only for a few min right now
23:46:50 <alise> 11:11:42 <Sgeo> I'll be back laterish maybe
23:46:50 <alise> 11:12:31 <Sgeo> um, actually, I can't risk the computer beeping
23:46:50 <alise> 11:12:31 <Sgeo> Bye
23:46:55 <alise> Vorpal: my monologues usually have a lot of questions...
23:47:01 <alise> I prefer "soliloquies"
23:47:04 <fizzie> Vorpal: Group over year to find change-over-time, something like this:
23:47:08 <fizzie> sqlite> select qyr, qn, (kq+0.0)/k from (select substr(tstamp,1,4) as qyr, nick as qn, count(*) as kq from logs where type = 0 and body like '%?' group by qyr, qn) join (select substr(tstamp,1,4) as tyr, nick as tn, count(*) as k from logs where type = 0 group by tyr, tn) on qyr = tyr and qn = tn where k > 1000 order by (kq+0.0)/k desc limit 10;
23:47:14 <fizzie> 2007|Sgeo|0.278288868445262
23:47:18 <fizzie> 2008|Sgeo|0.264714946070878
23:47:21 <fizzie> 2010|Phantom_Hoover|0.225965296232609
23:47:25 <fizzie> 2007|ihope|0.224911749873928
23:47:28 <fizzie> 2009|Sgeo|0.221869590305828
23:47:29 <fizzie> (Apologies for the trailing spaces there.)
23:47:44 <Vorpal> alise, well yes but this only matches end of line
23:48:01 <Vorpal> <fizzie> (Apologies for the trailing spaces there.) <-- where
23:48:30 <Vorpal> fizzie, anyway this doesn't do the all-important nick aggregation
23:48:36 <fizzie> Those lines have trailing spaces; probably not noticeable on some clients, or screens wider than my phone.
23:48:38 <Vorpal> without which I consider the results useless
23:48:57 <Vorpal> fizzie, they were stripped here
23:49:01 <fizzie> It's a hack, not something "all-important".
23:49:19 <Vorpal> fizzie, it's all important, it lowers my question ratio a bit
23:49:50 <alise> Sgeo|FsckThatCli: Uhh?
23:49:52 <alise> What makes you think that?
23:49:55 <fizzie> I consider any results that just match for "?" at eol "useless", so there.
23:50:19 <alise> they act nothing alike.
23:50:30 <alise> beep may have been in response to
23:50:31 <alise> <alise> 11:11:35 * Sgeo is here only for a few min right now
23:50:31 <alise> <alise> 11:11:42 <Sgeo> I'll be back laterish maybe
23:50:31 <alise> <alise> 11:12:31 <Sgeo> um, actually, I can't risk the computer beeping
23:50:31 <alise> <alise> 11:12:31 <Sgeo> Bye
23:50:32 <quintopia> perhaps i am and have changed a lot?
23:50:38 <alise> quintopia: are you?
23:50:45 <alise> you still play agora
23:50:48 <alise> and act exactly like you did
23:50:49 <alise> so i doubt that :p
23:51:10 <alise> Sgeo|FsckThatCli: irrelevant
23:51:20 <quintopia> the beep was in response to a wall of text i didn't feel like reading
23:52:02 <quintopia> i haven't finished reading subnormality yet. i'll get to it
23:52:04 <Vorpal> #feather-lang is lonely, while waiting for ais
23:52:28 <fizzie> Okay, here's something that doesn't need nick-aggregation: the most confused month of #esoteric:
23:52:31 <fizzie> sqlite> select qyr, (kq+0.0)/k from (select substr(tstamp,1,7) as qyr, count(*) as kq from logs where type = 0 and body like '%?' group by qyr) join (select substr(tstamp,1,7) as tyr, count(*) as k from logs where type = 0 group by tyr) on qyr = tyr order by (kq+0.0)/k desc limit 10;
23:52:36 <fizzie> 2004-03|0.153846153846154
23:52:40 <quintopia> it does indeed have a lot of text, but it isn't frequently published, so it's easy to stay afloat
23:52:40 <alise> Vorpal: you realise he hates being pestered about Feather?
23:52:57 <alise> * Sgeo|FsckThatCli pushes Virus Comix onto quintopia ;; he linked two yesterday :P
23:53:06 <Vorpal> alise, actually he bought it up a few weeks ago, saying he had written a bit of code on it
23:53:11 <Vorpal> alise, without me mentioning it
23:53:20 <fizzie> 15.4 % questions. (The second place goes to 2007-04 with 13.0 %.)
23:53:22 -!- zzo38 has joined.
23:53:31 <Vorpal> fizzie, what about last full month
23:53:45 <alise> Sgeo|FsckThatCli: ?
23:53:51 <alise> Vorpal: *brought, and yes
23:53:56 <alise> Vorpal: but he wasn't pestered about it
23:54:06 <Vorpal> fizzie, still quite confused
23:54:08 <alise> fizzie: fff no space between number and %
23:54:19 <alise> 15:37:21 <Sgeo> erm, I don't know C
23:54:19 <alise> 15:37:27 <Sgeo> What's write(), and how does it work?
23:54:47 <Vorpal> alise, i should start writing like augur, just to annoy you
23:55:21 <Vorpal> write() is not C, it's POSIX
23:55:44 <fizzie> alise: Finnish grammar rules put a space there, blame them for forcing the habit on me.
23:55:56 <quintopia> poll: what's you favorite major syntax style? C, Scheme/LISP, Smalltalk, Forth, Prolog, Ada, assembly?
23:56:04 <alise> quintopia: AliseLang
23:56:07 <Vorpal> I think Swedish put a space in front of % too
23:56:15 <alise> forth has no syntax
23:56:21 <quintopia> alise: since when is that "major"?
23:56:30 <Vorpal> quintopia, something between lisp and perl is my answer
23:56:35 <zzo38> alise: Yes it has no syntax that is why sometimes you make syntax for some commands if they are needed
23:56:36 <quintopia> it does. it consists of sentences separated by spaces
23:56:37 <Sgeo|FsckThatCli> US English grammar rules require punctuation in ", and I can't stand it
23:56:41 <alise> quintopia: i refuse to answer populist questions
23:56:49 <alise> because you can write a word that parses it as anything
23:56:56 <alise> technically, the default compiler word separates by space, yes
23:56:59 <alise> but that's just one of many
23:57:08 <alise> and there's nothing stopping you writing a word "c" that interprets everything after it as regular C code
23:57:11 <alise> so no: Forth has no syntax.
23:57:23 <alise> <Vorpal> quintopia, something between lisp and perl is my answer
23:57:27 <alise> what's your favourite letter of the alphabet?
23:57:31 <alise> "Something between A and Z."
23:57:42 <Vorpal> alise, yes, closed interval!
23:57:54 <alise> i refuse to answer :)
23:58:13 <quintopia> i've got a new lang in mind, and i want it to be very comfortable to write
23:58:15 <fizzie> The question stands on its head and wags its tail.
23:58:22 <Vorpal> quintopia, I would say Erlang or Haskell then. They are not in the list but the list doesn't contain any nice syntaxes.
23:58:24 <quintopia> so i just wanted to know what makes people feel msot at home
23:58:31 <alise> erlang's syntax is gross
23:58:38 <alise> it looks like prolog while acting nothing like prolog
23:58:43 <alise> and prolog doesn't have the prettiest syntax in the first place
23:58:50 <alise> quintopia: fuck most people
23:58:52 <alise> write what you like
23:58:59 <quintopia> i'm going to lump them all together haskell/prolog/erlang
23:59:06 <alise> haskell is nothing like the others
23:59:08 <Vorpal> quintopia, they are not like each other
23:59:14 <zzo38> alise: For favorite letter of alphabet ,but what if you want alphabet other than English alphabet, then it won't be A to Z?
23:59:15 <alise> anyway i refuse to answer because they'll dilute your language into something everyone else likes
23:59:19 <quintopia> yes, well, i just did it! mwahahaaha
23:59:20 <alise> instead of something you like, which is where innovation comes in
23:59:29 <Sgeo|FsckThatCli> quintopia, it is obvious that you know no Haskell, Prolog, or Erlang
23:59:36 <alise> zzo38: en:alphabet == latin alphabet almost always.
23:59:38 <quintopia> alise: i resent the implication that i can't stand on my own and be innovative