00:01:14 -!- Rugxulo` has joined.
00:03:01 -!- kar8nga has quit (Remote closed the connection).
00:10:07 -!- Rugxulo has quit (Read error: 148 (No route to host)).
00:44:30 -!- Halph has joined.
00:51:24 -!- Halph has changed nick to coppro.
01:07:48 -!- coppro has quit ("I am leaving. You are about to explode.").
01:11:30 -!- coppro has joined.
01:52:59 -!- Gregor has joined.
02:00:08 -!- HackEgo has joined.
02:06:28 <ehird> http://xkcd.com/156/ hey, I remember when xkcd was still funny.
02:06:31 <ehird> i thought maybe i was imagining it
02:27:47 -!- Rugxulo` has left (?).
04:18:08 -!- comex has quit (Read error: 104 (Connection reset by peer)).
04:45:21 <ehird> FIRST THING SAID FOR TWO HOURS THREE MINUTES
04:45:33 <ehird> *AND THREE MINUTES FOR THAT MATTER OH YES OH YES OH YES INDEED
04:47:13 * oerjan calculates differently
05:26:02 <oerjan> xkcd reaches still new levels of absurdity
05:26:20 <oerjan> though not, i think, of humor
05:30:25 <ehird> hey, xkcd isn't even making a reference
05:30:31 <ehird> it's just a bunch of lines and some text
05:30:43 <ehird> except one reference in the alt text
05:30:49 <ehird> oregon trail! i bet he's been playing that game for months
05:30:54 <ehird> what with the two comics about it
05:31:18 <ehird> PREDICTION OF FORUM THREAD CONTENTS: "GET OUT OF MY HEAD RANDALL" "hahahaha best xkcd ever" "Can I fuck you Randall" "I WISH TO HAVE YOUR BABIES YOU ARE GENIUS"
05:31:36 * ehird wonders what word highlighted pikhq there
05:32:39 <ehird> http://echochamber.me/viewtopic.php?f=7&t=46236
05:32:41 <ehird> they're mostly critical
05:32:51 <ehird> [02:32] <pikhq> :)
05:32:55 <ehird> last thing said before me
05:41:32 -!- ehird has quit ("Ex-Chat").
05:43:34 <oerjan> which reminds me, is there supposed to be a pun in "Ex-Chat"?
05:50:12 -!- Pthing has quit (Remote closed the connection).
05:57:00 <oklopol> can i get EgoBot to http pages for me, kinda annoying to go all the way to the other room to reset the modem thingie
05:58:34 <EgoBot> Installed user interpreters: aol austro b1ff bc bct bfbignum brit brooklyn bypass_ignore chef chiqrsx9p choo cockney ctcp dc drawl dubya echo eehird ehird fudd funetak google graph gregor hello id jethro kraut num ook pansy pirate plot postmodern postmodern_aoler redneck reverse rot13 sadbf sfedeesh sffedeesh sffffedeesh sffffffffedeesh slashes svedeesh swedish valspeak warez yodawg
05:59:08 <oerjan> !echo http://oerjan.nvg.org/esoteric
05:59:45 <oerjan> oklopol: i conclude that as "yes, but"
06:02:20 <Gregor> You could hack up a script to do it more properly.
06:05:07 -!- acloglio has joined.
06:05:19 <acloglio> oerjan: artmetic is where you aren't bound by the usual axioms, but make your own rules
06:05:48 <acloglio> also the new xkcd is not absurd, and is definitely a joke. it's just a particularly bad graph joke
06:05:55 -!- acloglio has changed nick to oklofok.
06:06:49 <oklofok> also kinda annoying how mirc chooses completely randomly what nicks to keep as my alternatives
06:07:02 <oklofok> because it always chooses the ones i definitely never want to use again
06:07:13 <oklofok> almost wrote "want" as "onet"
06:07:19 <oerjan> in any case, the sign of a burnt-out man
06:09:01 <oerjan> oklofok: it doesn't just choose the last one that is available?
06:09:14 -!- FireFly has joined.
06:09:21 -!- oklofok has changed nick to oklokok.
06:09:31 <oklokok> it has two slots for nicks
06:09:52 <oklokok> let's see if i'm still acloglio when i jump next
06:11:00 <oklokok> i recall once changing my nick like 10 times to make sure some nick isn't reused
06:11:02 <oerjan> it _could_ be complicated by when it saves state
06:11:42 <oklokok> maybe, anyway it seems i should be at the uni, time just flies by when you're crying over bad software
06:12:03 <oerjan> time flies like an arrow
06:14:09 <oerjan> !sfedeesh Recursion is fun. Or is it?
06:14:09 <EgoBot> Recoorseeun is foon. Oor is it? Bork Bork Bork!
06:14:28 <oerjan> !swedish Recursion is fun. Or is it?
06:14:29 <EgoBot> Recoorseeun is foon. Oor is it? Bork Bork Bork!
06:18:53 <oerjan> !swedish Sju skönsjungande sjuksköterskor skötte sjuttiosju sjösjuka sjömän på skeppet Shanghai.
06:18:54 <EgoBot> Sjoo skönsjoongunde-a sjooksköterskur skötte-a sjootteeusjoo sjösjooka sjömän på skeppet Shungheee. Bork Bork Bork!
06:21:46 -!- oklopol has quit (Read error: 110 (Connection timed out)).
06:28:56 -!- ehird has joined.
06:29:21 <ehird> 21:43:34 <oerjan> which reminds me, is there supposed to be a pun in "Ex-Chat"?
06:29:26 <ehird> default quit message: Ex-Chat
06:32:32 <ehird> anyway I'm back on os x. not sure why
06:32:39 <ehird> font rendering sure is better thouh
06:33:05 <oerjan> reality distortion, probably
06:34:02 <oerjan> secretly reality is distorted every few minutes, but only the schizophrenics are able to notice it
06:34:16 <ehird> mac os x is too darn pretty
06:35:26 <ehird> I'm missing the hold-modifier-and-drag-to-move-window from metacity
06:36:29 * oerjan notices an anvil above, and says nothing
06:37:24 <ehird> i cannot feel a thi-
06:37:25 -!- ehird has quit.
06:37:41 -!- ehird has joined.
06:37:55 <ehird> sorry, connection dropped
06:38:07 <ehird> neither words in that sentence are!
06:39:01 <ehird> also, os x has still cracked mouse acceleration way better than other OSs
06:39:20 -!- oerjan has quit ("Cracking mice is hard to do cleanly").
06:41:55 <Asztal> Firefox 3.6/3.7 did its own mouse acceleration on windows at one point, then they scrapped it because there might already be mouse drivers doing it.
06:42:06 <Asztal> It makes far more sense for the system to do it :(
06:42:26 <ehird> Firefox... did its own... mouse accel- sorry, what?
06:42:34 <ehird> Windows really is on crack, isn't it?
06:43:29 <Asztal> For big scrolls I use the middle-mouse autoscroll thing, anyway. I kind of like how Firefox does that one, and none of the other browsers do it the same :(
06:43:47 <ehird> That should be done by the OS too!
06:44:06 <ehird> Asztal: So, question. The mouse driver did the delay before a menu opens in NT 4.0, at least.
06:44:08 <Asztal> https://bugzilla.mozilla.org/show_bug.cgi?id=462809 https://bugzilla.mozilla.org/show_bug.cgi?id=509651 https://bugzilla.mozilla.org/show_bug.cgi?id=513817
06:44:09 <ehird> Does it still handle that?
06:44:42 <Asztal> ehird: It's in a registry entry now, TweakUI (e.g.) has an option to change it.
06:44:57 <ehird> Well, at least that's saner
06:45:02 <Asztal> Most other UI delays are based on that delay time, too
06:45:08 <ehird> But WHY does the mouse driver handle mouse acceleration?
06:46:04 <Asztal> It doesn't *have* to. Some do, apparently. Because Windows doesn't do any of its own.
06:46:27 <ehird> Sure it does... it used to, at least? Right?
06:46:41 <ehird> The MS people as new as 7 can't seriously have never ever studied any mouse ergonomics.
06:47:12 <ehird> Also, gah why don't other OSs do OS X's scroll acceleration?
06:47:37 <ehird> You don't need any headache-inducing smooth scrolling because small movements just go a little bit, but if you flick it you can easily move across the whole document.
06:48:26 <ehird> I'm glad I don't use Firefox on any OS, though... it's clear they want to be an island and don't care about the platform.
06:50:26 * ehird wonders how to make Colloquy inform on every new message.
06:51:09 <ehird> Asztal: say something plz
06:51:51 <Asztal> Mr. Flibble is very cross.
06:52:02 <ehird> Mmph, uses Growl, not the dock icon.
06:52:18 <ehird> But I don't want to bounce it in the dock.
06:53:23 <ehird> I demand you speak!
06:53:47 * Asztal mumbles incoherently instead
06:54:12 <ehird> Ehh... I want a persistent, non-annoying blob so that I know there's messages to read, not a single bounce. :(
06:54:29 <Asztal> "So if we back out the acceleration model, we are left with the problem that Chrome is perceptually twice as fast as us at scrolling"
06:54:50 <ehird> I am so, so glad I do not use Firefox.
06:56:46 <Asztal> Firefox 3.6 includes new features such as not enumerating every single system font before even showing a window.
06:57:05 <Asztal> I don't know how you could stay away, really.
06:57:42 <ehird> Asztal: I do have to wonder why you torture yourself with bad programs on a bad OS. :P
06:59:18 <Asztal> Stupid windows-only games. Bad habit. :P
07:00:24 <ehird> Asztal: I've never understood how non-constant gaming can justify such horror...
07:00:28 -!- BeholdMyGlory has joined.
07:04:12 <Asztal> I've adapted to using bad software. I can even use Windows Media Player (!)
07:04:27 <ehird> You know, you've done serious psychological damage to yourself.
07:04:37 <ehird> You might need years of therapy.
07:05:20 <ehird> WTF, with MondoMouse I can't enable Cmd-drag to resize, only Opt-Cmd and move mouse.
07:05:31 <ehird> (or move, which is what I primarily want)
07:08:49 <ehird> http://db.tidbits.com/article/10624 Uhh, wow.
07:09:11 <ehird> Oh, it doesn't do the fun minimise effect promised.
07:09:13 <ehird> How disappointing!
07:45:52 -!- BeholdMyGlory has quit (Remote closed the connection).
07:58:30 -!- FireFly has quit ("Later").
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:08:48 -!- kar8nga has joined.
08:15:23 -!- Rugxulo has joined.
08:19:27 <Rugxulo> someone in here wrote a Befunge interpreter in sed, right?
08:19:36 * Rugxulo is vaguely curious how that would work, even if input is an issue
08:20:20 <fizzie> I wrote an incomplete one; I'm not sure if someone else wrote a more finished one.
08:22:50 <fizzie> I don't really remember the details, and my home desktop is off now, though. I don't think it did input, but it should be possible: just save all the interpreter "state" in the hold space when encountering ~ or &, and then have the main rules use the hold space to figure out where to continue, and the input for, well, the input.
08:23:39 <fizzie> I seem to think I had a reasonably clever way of manipulating the playfield, or at least to do get(x, y) on it, but I don't recall what that might've been.
08:23:43 <Rugxulo> has it been tested with non-GNU seds? (e.g. *BSD)
08:24:11 <Rugxulo> although in fairness, those are wimpy anyways, so you'd be lucky if it worked
08:25:15 <fizzie> No (as in, "no, it hasn't been tested"). And really, it was *very* incomplete; it was more of a collection of bits of groundwork for a Befunge interpreter.
08:25:50 <fizzie> I think there was a reasonable Brainfuck-in-sed though? Or maybe I'm mixing up with that brainfuck-to-ELF compiler which was written in sed.
08:26:15 <Rugxulo> Brainf***-to-ELF I found, haven't seen an interpreter for that
08:26:24 <Rugxulo> there was an Unlambda interpreter though
08:26:41 <Rugxulo> heh, well sed is kinda the wrong tool for the job (but fun) ;-)
08:27:07 <Rugxulo> Perl is said to be a replacement for AWK and sed, but it goes way way way beyond that
08:27:23 <fizzie> I did some of our "introduction to imperative programming" course homework assignments with sed, for the fun.
08:27:25 <Rugxulo> if mtve was ever here, he could probably elaborate
08:28:26 <Rugxulo> but heck, if even AWK was only designed for one-liners mostly, then you know sed wasn't exactly meant for stuff like this ;-)
08:28:43 <fizzie> Arithmetics are a bit painful, yes. (Except with unary math, but that's a bit... memory-intensive.)
08:30:14 <fizzie> Also function calls. I think in one of the programs I did a rudimentary stack, where you could push return labels on to, with a huge "s///; t foo; s///; t bar;" dispatcher to return.
08:31:24 <fizzie> I had some sort of a hundred-line maybe-not-so-optimal decimal adder in one sed program. It's easier with binary, though; but then you have to either convert that for output, or just have the user read long bitstrings.
08:31:50 <Rugxulo> it's a bit complicated, I know
08:32:05 <Rugxulo> and 'D' never did what I wanted
08:32:57 <fizzie> There's also a lot of "t dummy; : dummy; s///; t foo" style code for branching, since t looks for any successful s instructions since the last t.
08:33:48 <fizzie> Early lunchtime now. →
08:35:02 -!- jix has joined.
08:35:28 <Rugxulo> some seds (GNU? I forget ...) support 'T'
08:37:15 <Rugxulo> can be vaguely useful sometimes
08:40:55 <ehird> should i sleep or internet
08:41:46 <Rugxulo> if you're too tired, just sleep
08:41:52 <Rugxulo> (hard to think when tired anyways)
08:41:54 <ehird> however: internets
08:41:59 <ehird> they require no thought :P
08:42:06 <ehird> (but sleep! but internets! but zzzzzzzzzzzzzzzzzzzzz huh what)
08:53:41 <fizzie> But "T foo" is simply a "t skip; b foo; : skip".
08:54:03 <fizzie> Given the amount of unnecessary verbosity sed imposes by default, that's not much of a hassle to write.
08:55:00 <fizzie> These 03am-written presentation slides make no sense to me, and I'll have to give the presentation in an hour. Well, maybe everyone else will be equally tired.
08:56:41 -!- MigoMipo has joined.
08:58:42 <Rugxulo> drink a soda to pep you up
09:00:08 <ehird> i hink that would surpass my meagre abilities
09:00:17 <ehird> sleeping would be smart; i'ma trial it out
09:00:25 <ehird> TRY BEFORE YOU DIIEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE
09:00:27 -!- ehird has quit.
09:28:28 -!- jix has quit (Read error: 60 (Operation timed out)).
09:40:27 -!- Rugxulo has left (?).
10:02:55 -!- MigoMipo has quit ("Page closed").
10:25:38 <oklokok> when i was in the 6th grade or so, i had this habit of saying "in your dreams, motherfucker" to everyone, all the time.
10:45:30 -!- KingOfKarlsruhe has joined.
11:04:26 -!- kar8nga has quit (Remote closed the connection).
11:23:30 -!- Slereah has quit (Read error: 110 (Connection timed out)).
11:37:08 -!- MigoMipo has joined.
11:46:14 -!- dbc has quit (robinson.freenode.net irc.freenode.net).
11:46:14 -!- Deewiant has quit (robinson.freenode.net irc.freenode.net).
11:46:36 -!- dbc has joined.
11:46:36 -!- Deewiant has joined.
12:03:36 <fizzie> Wow, that was one popular Master's thesis presentation.
12:03:50 <fizzie> There were something like 30 people there.
13:01:34 -!- MigoMipo has quit ("Page closed").
13:13:45 -!- kar8nga has joined.
13:24:33 <oklokok> wait didn't you present yours like ages ago
13:25:49 <Deewiant> His was not the only Master's thesis presentation
13:26:43 <oklokok> well he gave *some* presentation
13:27:01 <oklokok> or do you mean the earlier one wasn't his
13:34:31 <oklokok> right k, then i guess it was a coincidence
13:47:40 <fizzie> This one wasn't mine, yes. I just was there to listen, since it was advertised so well.
13:48:09 <fizzie> I just wanted to add that I will present lots of pretty pictures, and
13:48:09 <fizzie> no formulas or boring technical details, so attending my presentation
13:48:09 <fizzie> today is an excellent alternative to actual research work,
13:48:09 <fizzie> administration and most other activities you normally get paid for
13:48:09 <fizzie> around here! (I mean the things you get paid for, not what you
13:48:09 <fizzie> actually do. I don't know if I can compete with the best flash games,
14:06:37 <oklokok> almost makes me book the train tickets even though it's already over
14:15:22 -!- kar8nga has quit (robinson.freenode.net irc.freenode.net).
14:15:22 -!- KingOfKarlsruhe has quit (robinson.freenode.net irc.freenode.net).
14:15:22 -!- Cerise has quit (robinson.freenode.net irc.freenode.net).
14:19:31 -!- kar8nga has joined.
14:28:25 -!- augur has quit (Read error: 113 (No route to host)).
14:38:17 <Deewiant> (1(00)*1|((0|1(00)*01)(11|10(00)*01)*(0|10(00)*1)))*
14:42:23 -!- ais523 has joined.
14:52:01 -!- oerjan has joined.
14:55:06 <ais523> my Evolution calendar has reappeared
14:59:32 -!- KingOfKarlsruhe has joined.
15:02:11 -!- Pthing has joined.
15:32:57 -!- coppro has quit ("I am leaving. You are about to explode.").
15:36:46 -!- FireFly has joined.
15:44:30 <AnMaster> ais523, started coding on the feather implementation?
15:45:12 <ais523> I'm planning to but haven't got around to it
15:48:38 -!- Slereah has joined.
15:50:42 <ais523> hmm... Feather's the only language I know in which you have to worry about portability within a program
15:53:34 -!- BeholdMyGlory has joined.
15:54:40 <ais523> because the interpreter might change beneath you
15:56:31 -!- oerjan has quit ("leaving").
16:06:04 -!- MigoMipo has joined.
16:13:19 -!- KingOfKarlsruhe has quit (Remote closed the connection).
16:19:57 -!- Slereah has quit (Read error: 104 (Connection reset by peer)).
16:20:10 -!- Slereah has joined.
16:32:05 -!- ais523 has quit (Read error: 110 (Connection timed out)).
16:37:36 -!- jix has joined.
16:43:51 -!- ais523 has joined.
16:52:24 -!- Jerry_ has joined.
16:52:24 -!- Jerry_ has changed nick to Cerise.
16:52:54 -!- Cerise has changed nick to Guest5773.
16:53:32 -!- Guest5773 has changed nick to Cerise.
17:22:02 -!- Rugxulo has joined.
17:24:14 -!- Zombie_Will has quit ("Do you like zombies? How about zombie games? Come join ##zombie and visit www.zombieexperiment.com to help us design a zombie).
17:26:01 -!- Rugxulo` has joined.
17:28:24 -!- Rugxulo` has quit (Client Quit).
17:28:34 -!- Rugxulo` has joined.
17:35:16 -!- kar8nga has quit (Remote closed the connection).
17:42:53 -!- Rugxulo has quit (Read error: 113 (No route to host)).
17:48:45 -!- Rugxulo` has changed nick to Rugxulo.
17:57:36 -!- augur has joined.
18:06:56 <AnMaster> to ehird when he joins: the power supply on my thinkpad turned out to be too weak to both charge the battery and sustain a heavy load (both cores at 95% or more, 3D graphics). It could sustain the load itself (so the battery didn't discharge, but nor did it charge)
18:07:09 <AnMaster> since he was considering a thinkpad he may be interested in this
18:07:34 <AnMaster> Rugxulo, computer bought this year, augusyt
18:07:41 <Rugxulo> or you just mean it can't run and charge at the same time
18:07:54 <AnMaster> Rugxulo, can't run heavy load and charge it at same time indeed
18:08:14 <Rugxulo> my laptop (6 cell battery) gets max (!) 2 hours, and only then on "Power Saver" :-/
18:08:36 <AnMaster> oh and for reference wlan was turned off while doing this. Was using ethernet
18:09:03 <AnMaster> Rugxulo, mine last 2 hours 50 minutes with light load
18:09:08 <Rugxulo> wifi, usb, etc. all use power, turn off all unnecessary stuff
18:09:22 <AnMaster> bluetooth is almost always off for me
18:09:41 <AnMaster> wlan however I do use quite a bit
18:09:53 <AnMaster> Rugxulo, and turning off usb? huh?
18:10:08 <AnMaster> well I can't realistically, since some internal devices show up on the usb bus
18:10:09 <Rugxulo> I mean don't leave unnecessary USB devices plugged in
18:10:15 <AnMaster> for example the bluetooth thingy
18:10:30 <AnMaster> Rugxulo, oh right, I almost never use any usb devices with it
18:10:42 <AnMaster> exception is sometimes at home when I plug in the printer
18:10:52 <AnMaster> for mouse at home I use desktop + synergy
18:12:17 <Rugxulo> what OS do you use? some are better at power than others
18:13:53 <AnMaster> Rugxulo, on the laptop? Ubuntu x86_64
18:14:04 <AnMaster> because arch didn't work very well
18:14:12 <AnMaster> and I had no time to dig into those issues
18:14:58 <Rugxulo> I tried XUbuntu 8.04.1 the other day, mainly for DOSEMU testing, seemed okay
18:15:09 <Rugxulo> 9.10 will be out soon (this month)
18:16:49 <AnMaster> Rugxulo, how long after that will 9.04 be supported?
18:19:28 * AnMaster is annoyed he has to choose whenever to run that heavy load or get the computer charged for tomorrow (will need it then...)
18:19:44 <Rugxulo> 8.04 is LTS, I think, hence it's still supported
18:19:45 <AnMaster> and the desktop is just a bit too old when it comes to CPU
18:21:48 <Rugxulo> 9.04 should be supported for a while, though, I'd imagine
18:21:59 <Rugxulo> even Fedora is supported until two versions later
18:25:26 <oklokok> AnMaster: that's why nights were invented
18:27:09 <AnMaster> oklokok, except when I get up tomorrow I need to catch the bus very early on. thus having no time to pack my backpack
18:27:26 <Rugxulo> can't bring the plug with you?
18:28:16 <oklokok> you wake up, and start running? i mean if you even put clothes on, the time to throw the comp in the bag will not dominate.
18:28:50 <AnMaster> <Rugxulo> can't bring the plug with you? <-- sure, but not the wall socket
18:29:16 <AnMaster> yeah, the house was built *before* laptops were invented and/or common
18:29:21 <oklokok> (technically you probably could take the socket!)
18:29:23 <AnMaster> thus not having a lot of outlets
18:29:26 -!- kar8nga has joined.
18:29:28 <ais523> oklokok: but it wouldn't provide electricity
18:29:45 <AnMaster> ais523, it would if I had a 20 km extension cable!
18:30:15 <AnMaster> oklokok, calculating limits is even harder.... *sigh*
18:31:17 <oklokok> anyway that's noobie stuff
18:31:18 <AnMaster> oklokok, the teacher said that was outside the scope of this module (iirc that is the English term)..
18:31:35 <AnMaster> so yeah from standard values and figuring it out.
18:31:54 <AnMaster> oklokok, as in "known basic ones"
18:32:22 <AnMaster> like you know what the trig functions are for pi/2, pi/4 and pi/6 is
18:32:30 <AnMaster> (well the three basic ones that is)
18:33:14 <oklokok> lim_{x->0} (x/x) = lim_{x->0} 1 HEY NOW IT DOESN'T DIVIDE BY X YAY I HAS LIMITS
18:34:04 <oklokok> if you don't do epsilons, you probably don't have a definition for limits, basically you have no idea what you're doing.
18:34:44 <AnMaster> \[\lim_{x\rightarrow\inf}\frac{x-x^{2}}{x^{2}+1}\]
18:35:10 <AnMaster> <oklokok> lim_{x->0} (x/x) = lim_{x->0} 1 HEY NOW IT DOESN'T DIVIDE BY X YAY I HAS LIMITS <-- sounds similar
18:35:59 <oklokok> for (x-x^2)/(1+x^2), without epsilons, you should go HEY THOSE X AND 1 LOOK REALLY SMALL THEY PROBABLY DON'T AFFECT ONCE YOU PUT THE NUMBER INFINITE IN PLACE
18:36:02 <AnMaster> oklokok, also he defined it with epsilon... just said that calculating with that was outside the scope of this module. Would be more of that during this spring iirc.
18:36:40 <AnMaster> oklokok, actually I think you are supposed to somehow get rid of x^2 first
18:36:56 <oklokok> then divide both the denominator and the numerator by something
18:37:01 <AnMaster> by factorising and getting x^(-2) and such
18:37:18 <AnMaster> seems to get the right answer more often too
18:38:56 <AnMaster> though it seems rewriting it as \[\lim_{x\rightarrow\inf}1-\frac{x}{x^{2}+1}\] is simpler. Maybe
18:39:17 -!- MigoMipo has quit ("Page closed").
18:39:27 <AnMaster> that needs some more parentheses
18:39:44 <AnMaster> \[\lim_{x\rightarrow\inf}\left(1-\frac{x}{x^{2}+1}\right)\]
18:41:29 <oklokok> one way: (x-x^2)/(1+x^2)-(-1) = (x-x^2+1+x^2)/(1+x^2) = (x+1)/(1+x^2) and notice the bottom is of a greater degree
18:41:48 <oklokok> other way: divide both den and num by x^2, to get stuff like 1+(1/x)
18:42:07 <oklokok> all 1/P(x) clearly go to 0 when degree of P > 0
18:42:59 <AnMaster> oklokok, the issue is, you should get the same answer through both methods
18:43:04 <AnMaster> when you don't, something went wrong
18:43:25 <oklokok> yeah, limits are unique in R
18:43:41 -!- oklokok has changed nick to oklopol.
18:45:03 <oklopol> also what kind of a uni does limits informally
18:45:21 <oklopol> although maybe our dumbed down math does that too, dunno
18:45:32 <oklopol> we have separate courses for the less fortunate
18:45:33 <AnMaster> oklopol, depends on what you are studying
18:45:46 <AnMaster> I mean, math or something else that *uses* math
18:47:06 <oklopol> i understand they might want to skip proofs, but skipping definitions is just plain cheating
18:48:58 <AnMaster> oklopol, also the answer should be that the limit above goes to either + or - infinity, otherwise the answer in the book (which is the oblique asymptote (english term taken from wikipedia, lets hope it is the correct translation) when x→inf wouldn't exist, at least as far as we were told)
18:49:28 <oklopol> the limit is -1, as i said
18:49:57 <oklopol> think about lim_{x->inf} 1, clearly it isn't + or - inf
18:51:12 <oklopol> for each e > 0 you need to find some point in the series after which no deviation > e happens from the limit; the limit is 1, because for any e > 0, the whole sequence stays within (1-e, 1+e)
18:52:41 <oklopol> anyway, believe it or not, i consider game theory harder than your homework problems
19:00:30 <oklopol> hmm, okay pasting from a pdf to pb.vjn.fi doesn't really work that well, but actually this looks roughly as scrambled in the book
19:00:30 <oklopol> http://www.vjn.fi/pb/p315154346.txt
19:00:59 <AnMaster> oklopol, ok that is plain unreadable
19:01:36 <AnMaster> oklopol, I thought game theory involved mostly combinatorics and probability and such?
19:01:44 <AnMaster> I can't tell from that pdf though
19:02:11 <oklopol> most of game theory involves games ;)
19:03:02 <oklopol> the book is mostly about different kinds of solutions
19:03:15 <oklopol> solution/equilibrium concepts
19:03:49 -!- kar8nga has quit (Remote closed the connection).
19:03:59 <AnMaster> <oklopol> most of game theory involves games ;) <-- well yeah, but the tools include combinatorics and probability right?
19:04:10 <oklopol> rationalizability is one of the weaker ones, where a move/strategy is created from scratch based on the assumption that all other players are infinitely intelligent, and knowledge about the game
19:05:10 <oklopol> the "infinitely intelligent" part means infinite sequences of sets, the notation gets so goddamn hairy
19:05:10 <AnMaster> oklopol, interpreting that paste may require infinitely intelligence too... So I guess it is intended for those who move in those circles ;P
19:30:40 <fizzie> Hee, I implemented a simplified DEFLATE variant in that rfk86 port. (It does the normal LZ77 encoding, but doesn't use the fancy code-length dual-wielded-Huffman thing to store the trees; and it compresses each -- very short -- message separately, so the compression ratios are a lot less, but that's the price you pay for being able to uncompress each separately without uncompressing the whole data stream. The TI-86 is not an especially fast machine.)
19:31:34 <Deewiant> You should do it in 20 blocks or something
19:31:50 <fizzie> Probably, but that's such a hassle.
19:32:05 <Deewiant> How much space does the current scheme save?
19:33:08 <fizzie> The 86p file (which has few bytes of headers) went from 23685 bytes to 16766 bytes, which is at least something. (Plus that's with an animated exit splash screen added.)
19:33:33 <fizzie> AnMaster: Like I said to Deewiant, you don't get a fancy grayscale introduction splash screen with TI-Basic. :p
19:33:46 <AnMaster> fizzie, didn't see you mention that
19:34:06 -!- jix has quit (Read error: 110 (Connection timed out)).
19:34:13 <AnMaster> fizzie, also what does the program do?
19:34:25 <fizzie> It's a robotfindskitten port.
19:34:34 <fizzie> It.. you know, finds kitten.
19:34:46 <fizzie> Deewiant: Yes, but it doesn't compress *that* much with "real" DEFLATE either.
19:34:47 <AnMaster> fizzie, haven't played that game
19:34:55 <fizzie> fis@eris:~/src/rfk86$ ./messages.pl
19:34:55 <fizzie> Uncompressed message data: 19454 bytes
19:34:55 <fizzie> LZ77 with trivial encoding: 18708 bytes
19:34:55 <fizzie> LZ77 with Huffman (w/o trees): 11685 bytes
19:34:55 <fizzie> LZ77 with Huffman (raw trees): 12236 bytes
19:34:56 <fizzie> LZ77 with Huffman (no dist tree): 12209 bytes
19:35:58 <AnMaster> fizzie, isn't deflate basically gzip?
19:36:25 <fizzie> Deewiant: The new version is not in the URL I used.
19:36:32 <AnMaster> fizzie, also how would you do grayscale?
19:36:44 <Deewiant> fizzie: But I presume messages.txt is unchanged? That's all I want
19:37:06 <fizzie> Deewiant: Oh, right. http://www.cis.hut.fi/htkallas/rfk86/ if I remember right.
19:37:12 <fizzie> I'll put up a "project page" for it at zem.fi when I get that power supply.
19:37:41 <fizzie> 7z a -tgzip -mx=9 does 18866 -> 8963 bytes here to messages.txt.
19:37:46 <AnMaster> fizzie, how do you do that grayscal!?
19:38:41 <fizzie> AnMaster: It's done by flickering two frames; the screen response times are so slow it doesn't really look flickery at all.
19:39:48 <fizzie> Typically you stick in an interrupt handler which flips screens so that frame 1 is shown for two timer interrupts, and frame 2 for one; that way you get a nice 4-level greyscale.
19:39:49 <ais523> my first programmable calculator had something like 106 bytes of memory
19:40:12 <fizzie> ais523: The TI-86 has 128K of RAM and a 256K ROM chip. Or thereabouts.
19:40:50 <AnMaster> ais523, 106 bytes for the program to use for data? or 106 bytes for the entire program code?
19:40:53 <fizzie> Though 96K of that RAM is technically speaking used for variables (program- and otherwise), and it would be pretty impolite for the running program to mess that.
19:41:33 <ais523> AnMaster: the data was 27 double-precision floats
19:41:38 <oklopol> in RFK, is the game basically just about solving TSP
19:41:44 <ais523> actually, probably decifloats as it was a calculator
19:41:55 <Rugxulo> 7-Zip uses a custom Deflate, BTW, better than "normal"
19:42:16 <fizzie> Well, it's just a custom deflate *encoder*; it's still the standard stream format.
19:42:38 <AnMaster> fizzie, it makes quite a bit of difference though
19:42:49 <AnMaster> oklopol, no you don't know when you will find the kitten
19:43:34 <fizzie> Yes, and since you don't know that, the shortest path to visit all items is the optimal you can choose.
19:44:20 <fizzie> Anyway, a full DEFLATE decoder theoretically speaking needs to keep up to 32K of the decoded output stream in memory (for backreferences), I don't really have space for that on the TI.
19:44:35 <oklopol> right. i think it is, assuming you want to minimize expected time to find the kitten. not for instance going for a world record.
19:44:35 <AnMaster> fizzie, if deflate == gzip I thought it was seekable?
19:44:57 -!- MigoMipo has joined.
19:45:12 <oklopol> what if you have, at distance 10, to the right, one item, and to the left, at distance 20, a hundred items
19:45:15 <fizzie> It's not really seekable, no. Though you can flush the encoder state to generate gzip files that you can seek in, in a limited fashion.
19:45:46 <fizzie> I guess it's not quite TSP, right.
19:45:59 <oklopol> i guess that too, although that probably isn't a counterexample
19:46:15 <AnMaster> fizzie, hm... pretty sure *.tgz are seekable
19:46:17 <oklopol> because in that case it doesn't matter which way you go first
19:46:31 <oklopol> if tsp gives either way, then it doesn't solve optimally
19:46:49 <fizzie> AnMaster: I don't think you can seek an arbitrary DEFLATE stream.
19:47:09 <AnMaster> oklopol, why would "either way" be invalid? Lets say you have two points + your starting point
19:47:18 <oklopol> i'll ask the algo profs, they usually know this stuff
19:47:23 <AnMaster> they are placed as an equilateral triangle
19:47:35 <oklopol> AnMaster: going right is clearly stupid, because the kitten is probably to the left
19:47:36 <AnMaster> then it shouldn't matter which of those two points you visit first
19:47:55 <oklopol> just one counterexample proves TSP isn't optimal.
19:48:16 <AnMaster> <fizzie> AnMaster: I don't think you can seek an arbitrary DEFLATE stream. <-- hm ok, maybe it was block based or something
19:48:18 <oklopol> well, that solving TSP doesn't necessarily find an optimal solution to the problem
19:48:36 <Deewiant> nanozip brings messages.txt down to 7281, but I guess that's tricky to implement
19:49:49 <oklopol> a counterexample in which you actually have to go against tsp is when you have this large circle on the border of which you start, and on the opposite side there is a large cluster, tsp is solved by going round, RFK by moving right to the cluster, assuming |cluster| > big
19:50:45 <fizzie> Deewiant: Anything that involves keeping the whole uncompressed messages.txt in memory (which is what happens for deflate with the 32k sliding window) will be tricky to implement, since there are no contiguous memory ranges big enough.
19:51:49 <fizzie> Deewiant: I already managed to waste 150 bytes of code in the simplified-deflate decoder.
19:54:27 <fizzie> No, but it doesn't sound any simpler either. I seems vaguely bzip2-related; at least BWT is mentioned there.
19:57:01 -!- jix has joined.
19:58:36 <Rugxulo> no, NanoZip is something complex, maybe even context mixing
19:58:42 <Rugxulo> see Encode.ru/forum/ (I don't remember the details)
20:07:54 <Deewiant> I'm annoyed that I got the best result on pretty much my first try (nanozip)
20:08:17 <Deewiant> 15 programs later and this is the only one that even gets close :-P
20:10:03 <Rugxulo> context mixing is pretty powerful
20:14:16 <Rugxulo> and BTW, PAQ8L by default uses -5 (lots of RAM, slow)
20:14:56 <Deewiant> Or that's what it claimed, I didn't check.
20:16:34 -!- ehird has joined.
20:17:50 <Rugxulo> latest fork of PAQ8 is PAQ8px
20:18:11 <Rugxulo> original author (Matt Mahoney) is more focused on ZPAQv1, though
20:18:23 <Rugxulo> (which ain't as good but less likely to break upon updates)
20:19:53 <Deewiant> paq8px -6 and up does 7264! Yay!
20:21:05 <Rugxulo> can't remember, it might also support -mx6 (or -m6)
20:21:23 <Rugxulo> not sure if that is default or not with -6, but I know it isn't active in lower (I think??)
20:21:33 <Deewiant> Oh, -9 isn't actually a supported flag, that'd explain it
20:23:29 <Rugxulo> maybe I'm thinking of paq8q (which isn't as good I think, mostly an uber-experimental port, testbed for better UTF-8 handling)
20:23:45 -!- jix has quit (Read error: 110 (Connection timed out)).
20:24:02 <Deewiant> CCMX doesn't even come close at 8708
20:24:57 <Rugxulo> see http://www.mattmahoney.net/dc/text.html
20:26:17 <Rugxulo> that should keep you busy :-)
20:27:26 <Deewiant> I was seeing http://www.maximumcompression.com/data/text.php
20:29:48 <Deewiant> Although that one pointed out durilca's -t switch
20:30:38 <Rugxulo> www.uclc.info might also be useful
20:31:00 <Deewiant> Less than a third of the original
20:45:13 <fizzie> I don't suppose any of those methods would create something I could easily extract specific bytes from, with the added restrictions that (1) they shouldn't really use more than a kilobyte or two of working memory, (2) the code shouldn't take much more space than my 150-byte deflate decoder, and (3) it would be something I could cobble together with my very shaky z80 asm skills in a day or two?
20:45:49 <Deewiant> For (1), I think these all use at least 30 megs or so
20:46:43 <Deewiant> bzip2 -1 manages 8467 which might be doable
20:48:18 <ehird> Hi there from scissor-switch keyboardism.
20:49:57 <ehird> It's a Cherry. I gather it should be good, then.
20:50:12 <ehird> Admittedly it feels quite satisfying to press a key, but I seem to be making errors.
20:51:05 <Rugxulo> I think bzip2 -s is "small" (-2 slowly)
20:51:15 <ehird> Same thing, surely?
20:51:33 <Rugxulo> paq8f -1 uses 21 MB (except on .JPG), I think, that's the least I know of offhand
20:51:46 <Rugxulo> lpaq5 (?) and less can use 6 MB minimum
20:51:53 <Deewiant> I just said -1 since -1 and -9 give the same result, and -1 presumably uses the least memory
20:52:18 <Rugxulo> for files < 900k it's the same, the author refuses to check filesize, assumes -9 by default
20:52:41 <Deewiant> Oh right, it's just the block size
20:53:18 <Rugxulo> -s "uses less memory (at most 2500k)"
20:53:57 <Deewiant> DURILCA completes too fast to be able to measure its memory usage
20:54:00 <ehird> I like how the modifier keys being too big is causing me troubles.
20:54:13 <ais523> ehird: you're trying out a new keyboard?
20:54:13 <ehird> Ooh, that capslock LED is fancy.
20:54:15 <ais523> or actually bought one?
20:54:28 <Deewiant> It does have a "use N MB memory" setting but it doesn't seem to do much
20:54:40 <ehird> ais523: This is in the possession of me, yes; I couldn't bear that Saitek one another second.
20:54:47 <Deewiant> Well, with -m2048 it says "out of memory", which I doubt is correct
20:54:58 <ehird> It's Cherry, which make all those mechanical keyswitches, so I guuess it's high quality.
20:54:58 <ais523> ehird: did you have it before, or have you just obtained it?
20:55:07 <ehird> Although that non-typed duplicated u makes me suspicious.
20:55:15 <ehird> It's scissor-switch; what they put in laptop keyboards.
20:55:21 <ehird> Deewiant: It's not a mechanical one, alas.
20:55:35 <Deewiant> Yes, that's the third time you've stated that
20:55:41 <Deewiant> Which keyboard model, not switch model
20:55:48 <ehird> http://www.cherrycorp.com/english/keyboards/Desktop/G85-23100/index.htm
20:56:19 <ehird> I would love to know the difference.
20:56:23 <Deewiant> Maybe I misremembered what ML are.
20:56:55 <ehird> ehird: It's Cherry, which make all those mechanical keyswitches, so I guuess it's high quality.
20:57:17 <ehird> I must admit, this thing is uber-sleek.
20:57:23 <Rugxulo> hmmm, pic is of a German one??
20:57:32 <ehird> Like a MacBook Air, but heavier and it only does typing!
20:57:39 <ehird> Deewiant: Six discreet ones at thhe very top.
20:57:52 <ehird> Hey, I like changing my volume with a key.
20:58:00 <ais523> why do people advertise plug & play on keyboards
20:58:03 <ais523> are there any that /don't/ do that?
20:58:05 <ehird> The media keys fail to be scissor-switch.
20:58:16 <ehird> I think this duplication might be that my OS is set to repeat too quickly for this kb.
20:58:34 <ehird> I think very fast typing is lagging it, but is that just the placebo effect? I think it is.
20:58:47 <ais523> also, I like media keys
20:58:48 <Deewiant> Or the keys are just chattering, in which case you should get it repaired or replaced.
20:58:51 <ais523> I use them to control media players
20:58:52 <ehird> Fun fact: If you double-click a word to highlight it in OS X, then hit delete, it deletes one of the adjacent spaces too.
20:59:02 <ehird> Deewiant: I think it seems fine.
20:59:14 <ais523> (it could make a difference in a word processed or rich-text app)
20:59:26 <ais523> (or if one of them was an unusual width)
20:59:34 <ehird> No, it only does " " I think.
20:59:37 <ehird> <Rugxulo> hmmm, pic is of a German one??
20:59:45 <ehird> Cherry are a German company; it was like that on the box, too, which worried me a bit.
21:00:09 <ehird> Well, I seem to be typing quickly with this, so as soon as I get used to the sizes and positions of the relevant keys I should be happy.
21:00:16 <Deewiant> Isn't the German layout keywise identical to UK?
21:00:27 <ais523> why are Z and Y the other way round on German keyboards?
21:00:31 <ehird> Also, control, insert, home, etc all have different labels, which would be annoying.
21:00:33 <ais523> it seems kind-of weird to swap just two keys
21:00:45 <ehird> But I like my labels to be correct if I have any labels.
21:00:47 <oklopol> ais523: perhaps because they use z, but not y?
21:00:52 <ehird> ais523: Supposedly it makes typing faster, but it doesn't really.
21:00:53 <Rugxulo> ehird, yeah, they don't typically speak English ;-)
21:00:59 <ehird> And making QWERTY faster: laughable!
21:00:59 <Rugxulo> "strg" -> "ctrl" (I think)
21:01:02 <ais523> oklopol: but QWERTY was originally designed to be slow...
21:01:10 <Deewiant> ais523: Because Z is more common than Y, especially next to T.
21:01:12 <ehird> It's like... strong C.
21:01:15 <ehird> I am commanding you to exit.
21:01:17 <ais523> I'm not sure it actually is slow, though
21:01:19 <oklopol> ais523: so you should pessimize it further?
21:01:28 <ais523> oklopol: there's no reason to pessimize it nowadays
21:01:29 <ehird> Deewiant: But the position where Y is in QWERTY isn't convenient to press at all.
21:01:30 <oklopol> i'm not sure i follow your logic
21:01:31 <Rugxulo> meant to be slow so the typewriters wouldn't jam
21:01:39 <ehird> Optimising QWERTY isn't that simple... you have to basically rewrite it.
21:01:39 <Rugxulo> Dvorak is faster (allegedly)
21:01:41 <ais523> although, I suppose german typewriters could have been pessimized
21:01:56 <Deewiant> ehird: I agree, but that's the argument.
21:01:56 <ehird> Rugxulo: Dvorak is faster, slightly. More importantly, it causes a lot less hand strain.
21:02:10 <fizzie> If that explains QWERTZ, what's the reason for AZERTY?
21:02:19 <Rugxulo> but unless you never have to use any other keyboard (as if), it kills your QWERTY skills
21:02:23 <oklopol> who cares where each letter is, when the whole keys are scattered completely randomly anyway
21:02:27 <ehird> Rugxulo: Totally wrong.
21:02:38 <Rugxulo> although the fastest typer in the world (recently deceased woman) used Dvorak to get like 220 wpm !!
21:02:43 <ehird> Rugxulo: That's FUD, etc.
21:02:56 <Rugxulo> FUD! TROLL! STRAWMAN! EEE!
21:02:57 <ehird> Learning Dvorak doesn't kill your QWERTY touchtyping; not typing in QWERTY kills your touchtyping.
21:03:10 <ehird> So, yeah, if you switch totally to Dvorak and only use it for months ou
21:03:23 <fizzie> AZERTY's the French-speaking layout, it swaps Q/A and Z/W w.r.t. QWERTY.
21:03:38 <Rugxulo> it's not that it's that big a deal, just you have to carry around your own keyboard or hope they support Dvorak layout in the OS (or whatever)
21:03:39 <ehird> And if QWERTY touchtyping matters, i.e. you have to use keyboards that aren't yours regularly, then it won't happen.
21:03:54 <ehird> So if such a situation is common like carrying around your keyboard would seem to imply...
21:03:54 <Deewiant> fizzie: I meant, I don't know the reasoning behind AZERTY.
21:03:59 <ehird> Then it won't actually be needed.
21:04:13 <Rugxulo> oh, and BTW, I think certain apps (e.g. vi) are impossible to use under Dvorak
21:04:53 <ehird> Vi isn't exactly ergonomic with its key placement, as far as I know.
21:04:56 <ais523> Rugxulo: some people play NetHack with the vi layour under Dvorak
21:05:02 <ehird> I don't know that it prefers the home row for common operations. Take a look at where i is.
21:05:13 <ais523> ehird: hjkl are pretty ergonomic for moving around
21:05:14 <ehird> Admittedly, hjkl suffer, but those are totally unintuitive to start with.
21:05:19 <Deewiant> i isn't a common operation :-P
21:05:25 <ais523> I suppose you could say that less is hurt under Dvorak
21:05:34 <ehird> Deewiant: Err, sure it is?
21:05:37 <Deewiant> But on Colemak it's on the home row anyway, so shrug.
21:05:52 -!- puzzlet_ has joined.
21:06:16 <Deewiant> ehird: Not very common compared to some others IMO.
21:06:33 <ehird> But what do f and j do?
21:06:36 <Deewiant> I, at least, often use c/s/a/i/o to get into insert mode.
21:06:38 <ehird> Well, j is a moving operation.
21:06:45 <ehird> That prime key is totally wasted.
21:06:54 <ehird> I don't even know what it does, but it probably isn't common.
21:06:54 <Rugxulo> f goes to the first occurance of letter specified
21:06:55 <Deewiant> fX -> find next X on the current line
21:07:12 <ehird> Yeah; why is that on one of the two easiest QWERTY keys to press?
21:07:14 <Deewiant> t is the same thing but it goes to just before that letter.
21:07:23 <ehird> In conclusion, apart from hjkl, using Dvorak for vi won't change a thing.
21:08:00 <Rugxulo> I can't imagine typing 200 wpm, even in Dvorak
21:08:14 <ais523> now I want to change vi's UI to be more like NetHack
21:08:15 <Deewiant> I use Colemak and move 8 keys around, I think
21:08:24 <ais523> you could have tX to do just before that lteter
21:08:35 <ehird> Rugxulo: I can type 100 wpm in QWERTY with a decent rubber dome keyboard.
21:08:36 <ais523> and f to go to just before whatever letter you had quivered
21:08:56 <ehird> I can imagine typing 200 wpm, I just don't care to. Diminishing returns. I don't think at 100 wpm, so I usually end up typing slower than that.
21:09:15 <ehird> I want to switch to Dvorak for the easier and less straining typing, though.
21:09:27 <ais523> ehird: what if you were retyping a printed document, for some reason?
21:09:28 <Rugxulo> well, the lady I mentioned was a former secretary, so I guess it was more important for her job
21:09:32 <Deewiant> Essentially because I use JKDF for navigation instead of HJKL and the rest is shuffling around to get back NEFT which were overwritten
21:09:52 <ehird> ais523: I can't hold a whole document in memory; not even a whole paragraph.
21:10:05 <ais523> ehird: you're meant to read the document and type as you read
21:10:07 <ehird> The latency introduced by having to look at a passage means that any typing over, oh, 50 wpm is fine.
21:10:10 <ais523> rather than looking at the screen or keyboard
21:10:19 <ehird> ais523: Well, yes. I mean that,
21:10:29 <ais523> ah, you don't read at 200wpm
21:10:30 <ehird> I can't stream out typing as fast as I look at the letters.
21:10:45 <ehird> I have to recognise a letter, which I can do instantly, but not while concentrating on typing another leetter.
21:10:48 <ehird> *letter (uhhhh ohhhh)
21:11:20 <ais523> you have to concentrate to type letters?
21:11:23 <ehird> Admittedly, I can retype at about the speed I type out normal sentences, so it doesn't matter.
21:11:27 <ehird> ais523: Subconsciously, sure.
21:12:03 -!- Ilari has quit (Read error: 104 (Connection reset by peer)).
21:12:56 -!- Ilari has joined.
21:13:11 <Rugxulo> 104 keys, aren't most Win keyboards 105? what does it miss?
21:13:36 <ehird> ESC, F1-F12, prt sc/sys rq, scroll, pause/break.
21:13:47 <ehird> `, 1-0, -, =, backspace
21:13:57 <ehird> tab, (letters), [, ], enter
21:14:07 <ehird> caps, letters, ; ' \
21:14:11 <ehird> caps, letters, ;, ', \
21:14:26 <ehird> shift, §, leltters, ,, ., /, shift
21:14:34 <ehird> control, windows, alt, space, alt gr, windows, menu, control
21:14:39 <ehird> insert, home, pageup, delete, end, pagedown
21:14:41 <ehird> up, down, left, right
21:14:51 <ehird> num, /, asterisk, -, 7 8 9 +
21:15:02 <Deewiant> 104 is the standard US layout with Windows keys, 105 the standard European, 109 the standard Asian
21:15:13 <ehird> This is the US layout; $.
21:15:23 <ehird> What does the European layout add?
21:15:43 <ehird> This is the European layout.
21:15:49 <ehird> I have a pound key on my 3 key, not a #.
21:15:51 <Deewiant> A non-letter next to left shift
21:15:58 <Ilari> AFAIK, the same thing as what does 102-key keyboard add to 101-key keyboard.
21:16:04 <ehird> Um, yes, I have that.
21:16:19 <ehird> It's marked broken-| on top, \ on bottom.
21:16:21 <Deewiant> So it's not the US layout, and probably 105-key.
21:16:35 <ehird> I much, much, much prefer the US layout where the enter key is as big as backspace and only takes up one row.
21:16:39 <ais523> ehird: where are " and @?
21:16:42 <ehird> And the | \ key is above enter.
21:17:01 <ehird> ais523: In my OS? shift-2 and shift-second key after l.
21:17:17 <Deewiant> I much prefer the inverted-L enter key
21:17:23 <ehird> ais523: In my OS? shift-second key after l and shift-2
21:17:26 -!- puzzlet_ has quit (Read error: 60 (Operation timed out)).
21:17:30 -!- puzzlet_ has joined.
21:17:35 <ehird> On my keyboard? Shift-2 and shift-second key after l
21:17:40 <ehird> Deewiant: You are of the devil.
21:17:46 -!- puzzlet has quit (Connection timed out).
21:17:48 <Rugxulo> so shift-3 is pound but shift-4 is what? '$' ?
21:17:56 <Deewiant> I always hit the \ key on US layouts when I want to press enter
21:17:58 <ehird> Shift-3 is # for me.
21:17:59 <fizzie> I prefer the inverted-L enter too; we are both of the devil, then.
21:18:02 <ais523> ehird: sounds like a mix of US and UK, then
21:18:03 <ehird> On the keyboard it's pound.
21:18:18 <ehird> Option-3 is £ in OS X.
21:18:22 <ais523> hmm... shift-2 is @ in your OS?
21:18:24 * Rugxulo doesn't remember inverted L being the mark of the beast
21:18:29 <ehird> ais523: No, it'ss the US layout on my OS
21:18:32 <ehird> FUCK THESE REPEATING KEYS
21:18:39 <ais523> you seem to have a UK keyboard layout
21:18:49 <ehird> I'm turning off key repeat and seeing if it sstill happens.
21:18:53 <ehird> If it does, fuck this keyboard with a rake.
21:19:08 <Deewiant> ehird: Seriously, see if you can repeat it by pressing just one button slowly
21:19:15 <fizzie> The iBook OS X Finnish layout is quite a lot different than the Windows/Linux Finnish one, when it comes to things like \ and |.
21:19:15 <ehird> Hello, I am typing and hoping this shit doesn't happen because if it does awesomeness is great and nice.
21:19:25 <ehird> I have no key repeating in my OS, and I am going to prove that it is my OS's settings.
21:19:37 <ehird> Tra la la shit shit shit shit shit shit shit shit shit bugger fuck maybe profaanities will--
21:19:47 <ehird> Anyone know how to return a keyboard to cherry?
21:20:00 <ehird> Rugxulo: Yeeessssssssss..................
21:20:07 <Deewiant> Contact a customer service rep
21:20:09 <fizzie> Deewiant: Profane nanites.
21:20:15 <ehird> Deewiant: Interact?! With... HUMANS?!
21:20:26 <Rugxulo> press Strg-Alt-SysReq R E I S U B
21:20:31 <Deewiant> Well, unless they have a standard return form or whatever
21:20:35 <Deewiant> Did you order it directly from them?
21:20:45 <ehird> ais523: German ctrl.
21:20:50 <Rugxulo> then Turbo + Macro + grey plus + altNumLock
21:20:52 <ehird> On the box it says strg because cherry are german,.
21:21:03 <ehird> I wonder if I could make some sort of evil keyboard driver that cancels out two identical keys in quick succession.
21:21:18 <ais523> ehird: Windows comes with an accessibility tool to do that, I think
21:21:21 <ehird> That would be the Trrue Hackeerly Solution(TM).
21:21:22 <Rugxulo> it's probably your OS, then, maybe you type too fast
21:21:24 <ehird> Wow, twice in one message.
21:21:27 <ehird> Rugxulo: No, it isn't.
21:21:30 <ehird> I disabled key repeating.
21:21:41 <Deewiant> Amazon might have something for returns
21:22:15 <Rugxulo> maybe you're just pressing too hard, not used to such sensitivity yet
21:22:41 <ais523> ehird: in Ubuntu, it's Preferences | Assistive Technologies | Keyboard Accessibility | Ignore fast duplicate keypresses
21:22:44 <Deewiant> If it happens with /every/ key then probably that
21:23:03 <ais523> although annoyingly, if you type, say, "exe" in it, it'll disregard the second e if it's too soon after the first e
21:23:11 <ais523> presumably you'd need to set it to a very short timeout, the way you type
21:23:13 <ehird> Rugxulo: I DISABLED KEY REPEAT IN MY OS
21:23:15 <ehird> and it still happened
21:23:19 <ehird> No key repeat = hold key, only types once.
21:23:24 -!- MigoMipo has quit ("Page closed").
21:23:36 <ehird> I know for a fact I'm not tapping, releasing and tapping again because it makes a click and requires quite a bit of pressure to do that.
21:23:38 <ehird> It is the keyboard.
21:24:12 <ehird> I''m glad yoou're ssory.
21:24:47 <Deewiant> Actuation is after the tactile point, then?
21:24:51 <ehird> http://www.google.com/search?hl=en&safe=off&client=safari&rls=en-gb&q=ignore+repeated+duplicate+key+presses+os+x&aq=f&oq=&aqi= ;; hmph
21:25:10 <ehird> As soon as the depress and tick happens, the key sends.
21:25:24 <Rugxulo> wow, two pounds thick, that's heavy
21:25:25 <Deewiant> I.e. you're sure it doesn't send the key before the click
21:25:34 <ehird> Well, it doesn't click; the clicking is it batting the frame.
21:25:43 <ehird> Rugxulo: It is indeed heavy.
21:25:47 <ehird> However, it's very thin.
21:25:51 <ehird> Even with the keys.
21:25:54 <Rugxulo> some netbooks are barely that heavy
21:26:01 <ehird> Most netbooks are about 1 pound.
21:26:14 <ehird> (I don't count the low-voltage proper-CPU craptops as neetbboos.)
21:26:20 <ehird> (^ legit errors, btw)
21:26:35 <ehird> Rugxulo: The MacBook Air, which is a full notebook albeit a little underpowererd, weighs 3 pounds.
21:26:52 <Rugxulo> no optical drive though, right?
21:26:53 <ehird> I'm preparing to throw this thing out of the window.
21:26:58 <ehird> I never use the optical drive.
21:27:05 <Rugxulo> just saying, that's part of the reason
21:27:16 <ehird> Rugxulo: Actually, the optical drive was omitted for thinness reasons.
21:27:20 <ehird> They only weigh, like, 200-400 grams.
21:27:27 <Rugxulo> your keyboard has a 3 yr warranty, so don't throw it out just yet
21:27:35 -!- kar8nga has joined.
21:27:50 <Deewiant> ehird: II've used a keyboard which rrepeats for the passt few weeksss and you don't ssese me complaining eveen though it'ss a pain in the butt at times :-P
21:28:04 <Deewiant> (I did order a new one, though)
21:28:18 <ehird> Maybe I should buy one of those fancy ThinkPad USB keyboards.
21:28:23 <ehird> They even come with the TrackPoint.
21:28:48 <ehird> These flat keys do seem to be straining my hands less than the deeper fare.
21:29:37 <Rugxulo> BTW, only operates up to 40 C (104 F), what happens then? melts???
21:29:52 <ehird> The circuitry stops functioning?
21:29:57 <ehird> Ooh, captain obvious to the rescue!
21:30:32 <Rugxulo> 104 F isn't that rare, but I guess the keyboard itself being that hot would be unlikely (Iraqi desert? Las Vegas?)
21:30:53 <ehird> For me, 30 C is unbearably hot.
21:31:00 <FireFly> 40°C sounds pretty hot, yeah
21:31:10 <FireFly> ~30-35 is normal summer for me
21:31:47 <ehird> The opposite of brr./
21:31:48 <Rugxulo> it's 30 C outside here now (86 F)
21:31:55 <ais523> wow, Debian/kFreeBSD now has first-class status in there (severe bugs there are considered release blockers)
21:31:57 <Rugxulo> of course, I'm inside with a/c on
21:32:08 <ehird> I like ~23C summers, ~17C normals.
21:32:27 <ehird> 0 makes me want to die.
21:32:28 -!- adam_d has joined.
21:32:28 <Rugxulo> ehird, you're used to colder weather than I am, I'd bet
21:32:36 <FireFly> ehird, you wouldn't like a swedish winter
21:32:38 <ehird> I live in Britain, so pretty much yes by definition.
21:32:54 <ehird> FireFly: In Britain we have this thing called central heating
21:33:31 <ehird> <FireFly> Clothes are overrated
21:33:43 <ehird> <FireFly> I just go out in the minus-celsius weather naked
21:33:44 <fizzie> It's been oscillating between 0 and 10 degrees Celsius around here last week.
21:33:51 <ehird> <FireFly> A BIT CHILLY BUT OH WELL
21:33:51 <Deewiant> It's around 8°C out there right now
21:34:00 <Rugxulo> Swedish women can go topless in public pools now
21:34:22 <ehird> I don't want to know why Rugxulo knows that.
21:34:53 <FireFly> My point was, I don't think they can?
21:35:05 <fizzie> "Current temperature at Otaniemi: 8.50 °C". I still look at the weather there at the university campus, since the address is so memorable ("outside.hut.fi").
21:35:23 <Deewiant> http://www.thefuckingweather.com/
21:35:52 * ehird WANTS FUCKING CELSIUS
21:35:58 <FireFly> I take it that's not celsius :P
21:35:58 <ehird> I AM NOT IN FUCKING ROCHESTER, NY
21:36:07 <ehird> There's a FUCKING CHECK BUTTON
21:36:11 <Deewiant> FireFly: Check "I WANT FUCKING CELSIUS"
21:36:21 <Deewiant> That's about 8 degrees, I guess.
21:36:29 <fizzie> 7°?! But it's still "alright".
21:36:31 <ehird> ITS FUCKING ....ALRIGHT
21:36:31 <ehird> I've seen better days
21:36:52 <ehird> I have the window open. Maybe I'm a cold cyborg.
21:36:55 <Deewiant> For 5-20 or so I think it says alright
21:37:02 <ehird> THE FUCKING FORECAST
21:37:03 <ehird> FORECAST:Partly CloudyPartly Cloudy
21:37:08 <ehird> In England our forecasts are so interesting.
21:37:17 <fizzie> Dubai, AE: "30°?! ITS FUCKING HOT" and then in a very tiny font: "So hot my dog developed sweat glands."
21:37:56 <Deewiant> ehird: 'Bout the same as here, then.
21:38:40 <fizzie> "Showers Early; Partly Cloudy". It's not exactly identical, but it's not much more interesting either.
21:38:41 * ehird has a reddit debate: <person> "I think foo." <me> "Physics disagrees." "Sure, but I think foo." "Physics disagrees."
21:39:32 <Deewiant> ( http://www.reddit.com/r/IAmA/comments/9rel6/i_am_a_cryonicist_when_i_die_i_will_be/c0e3j9l )
21:39:48 <ehird> http://www.reddit.com/r/IAmA/comments/9rel6/i_am_a_cryonicist_when_i_die_i_will_be/c0e3p6y
21:39:51 <ehird> Relevant, more concise summary.
21:40:09 <pikhq> ehird: Ah, Internet debate.
21:40:23 <ehird> "Yes, I know *physics* says that, but I never consented to physics!|
21:41:26 <Deewiant> ehird: FWIW that's a religious topic, physics says nothing.
21:41:59 <ehird> I am operating under the assumption of materialism
21:42:03 <Deewiant> You can't know whether it'll be "just like you were unconscious through the suspension".
21:42:19 <ehird> Yes, you can; your neurons didn't fire, then they fired again.
21:42:23 <Deewiant> ehird: Well, I'd state that assumption.
21:42:47 <pikhq> "X." "Y!" "X!" "YY!" "XXXX ZOMG!"
21:42:50 <ehird> Deewiant: I don't generally state "hurr! I am not operating on the assumption of supernatural bullshit!", but sure, if it goes for another cycle.
21:42:57 <pikhq> ^ Internet debate.
21:43:28 <Deewiant> ehird: The topic /is/ supernatural bullshit. "I just feel like I am the constant flow of energy in my brain."
21:44:01 <Pthing> "Yes, a continuous stream is how we feel, but this contradicts physics." is too, for that matter
21:44:05 <ehird> Deewiant: I interpreted that as meaning the electrical impulses.
21:44:05 <Deewiant> "I feel like <poorly defined concept>" is not really something you can argue with
21:44:12 <ehird> I guess interpreting it as retarded bullshit makes more sense.
21:44:43 <Deewiant> "I feel <foo>" is also not really something you can argue with.
21:45:44 <ehird> Well, they said think, not feel./
21:46:58 <ehird> Presumably his New Age feelings turned into rational, coherent thought by the second post!
21:47:33 <Deewiant> Or presumably he didn't want to say "feel" twice in a row so he said "think" the second time instead.
21:53:41 <Deewiant> fizzie: Have you taken any of the compiler courses, by any chance?
22:00:04 <ehird> So, conditional endorsement for the Cherry Stream XT keyboard: if all of them repeat keys like this, stay away. If they don't, this is one kinda fine keyboard.
22:00:54 <Deewiant> "If they all suck, they all suck. If not, all except the ones that suck are good."
22:00:59 <ehird> Also: Could do with less having a numberpad!
22:01:02 <ehird> Deewiant: EXACTLY.
22:01:16 <ehird> Actually I'm happy with this, it's only occasionally annoying! Well actually quite often, but it hasn't happened once while typing this sentence!
22:01:25 <Asztal> My keyboard used to repeat and drop Es too. (As well as dropping N and P.)
22:01:39 <olsner> my keyboard drops nothing!
22:01:50 <ehird> My keyboard drops your mom
22:02:00 <Deewiant> At first only one key chattered, then steadily more and more.
22:02:10 <ehird> Deewiant: Don't you use one of those expensive mechanical keyboards
22:02:21 <olsner> this keyboard only gets better with time - especially if you compare it to newly produced keyboards
22:02:49 <ehird> Buckling spring sux.
22:02:50 <Deewiant> ehird: Actually quite cheap, but yeah :-P
22:02:57 <olsner> nah, buckling spring awesome
22:03:06 <ehird> Mechanical keyswitches, sure.
22:03:10 <ehird> Buckling spring noooooooooooooooooooooo
22:03:16 <ehird> Deewiant: You probably class cheap keyboard = <$1,000 :P
22:03:29 <fizzie> Deewiant: I've done one, but they've been rearranging those around a lot, or so I believe.
22:03:44 <olsner> Deewiant: what's wrong with the layout?
22:03:51 <ehird> Deewiant: Why did you buy a keyboard from Japan?
22:03:55 <ehird> Anyway, that's expensive. :P
22:04:05 <Deewiant> That was cheap; the ass was in the shipping
22:04:12 <Ilari> Very rare keyboard: 122-key Model M (yes, those exist).
22:04:15 <ehird> Admittedly the das keyboard is rather more, at 12,895 yen in Europe.
22:04:34 <olsner> Ilari: 122? is that one of the terminal keyboards?
22:04:39 <ehird> http://www.recycledgoods.com/zoom_s_p_20604_1.jpg.ashx
22:04:43 <Deewiant> Of course I tried to send it back for repairs, which only improved the situation for a short time and also cost a shatload due to back-and-forth shipping :-P
22:04:43 <ehird> look at that + formation
22:04:50 <ehird> look at all of those useless keys!
22:05:25 <Deewiant> Mostly the US-type enter that you find on most buckling spring boards annoys me.
22:05:41 <Deewiant> ehird: From Japan because it's approximately the only place with acceptable keyboards.
22:06:00 <olsner> mine has a large (upside-down-L) enter key
22:06:02 <Deewiant> Korea is maybe another, haven't looked into it.
22:06:02 <ehird> Deewiant: I'm curious now as to what it was.
22:06:20 <ehird> Unicomp, Das Keyboard, ... all of those are sold in, well, not-Japan. All I can think of is the Happy Hacking keyboard, which you've dissed.
22:06:23 <fizzie> Deewiant: I've done "T-106.550 Ohjelmointikielen kääntäjät L", but I think that one is nowadays split into two parts (T-106.4200 + T-106.5450) and I'm not sure those two sum up to the same.
22:06:51 <ehird> Deewiant: Heh, you went to japan just to avoid a number pad? :P
22:06:55 <Deewiant> Das Keyboard is not numberpadless, Unicomp I can't remember exactly what I didn't like about them
22:07:13 <Ilari> olsner: 122-key Model-M for PC.
22:07:31 <Deewiant> fizzie: Yeah, it's split. Did you have to code a compiler in Java?
22:07:35 <ehird> I don't know what that implies, Deewiant .
22:07:44 <fizzie> Deewiant: In Java, for "MiniJava", yes.
22:08:06 <Deewiant> Bah, then I don't get to rant and rave about how we have to and you didn't.
22:08:12 <Ilari> olsner: Extremely rare.
22:08:24 <fizzie> Deewiant: Is it still targeting Sparc?
22:08:29 <Ilari> (I only have 102-key IBM keyboard).
22:08:45 <ehird> Deewiant: http://elitekeyboards.com/products.php?sub=filco_keyboards,majestouch_87key&pid=fkbn87zeb How does this differ from yours?
22:08:51 <ehird> It lacks 4 keys, but what are they?
22:09:20 <ehird> "**Fukka, pronounced foo-ka,"
22:09:25 <ehird> BAD NAME. BAD NAME!
22:09:42 <Deewiant> ehird: Implies: inverted-L enter, small backspace to make way for a key, smaller space to make way for two more meta-type keyss
22:10:06 <ehird> Shitty enter, unusable backspace, unusable space... wow, it implies a shitty keyboard.
22:10:25 <Deewiant> Small size does not a key unusable make.
22:10:38 <ehird> It does when it's backspace.
22:10:49 <Deewiant> Except when backspace is caps lock.
22:11:10 <ehird> Do you really backspace that often? :P
22:11:33 <Deewiant> Do you really caps lock that often? :-P
22:11:58 <Deewiant> I backspace more than I press any other key that I'd put there
22:12:04 <ehird> No, but I don't think backspace is the best thing to put there.
22:12:09 <ehird> I'd put tab there.
22:12:17 <ehird> For tab-completion, etc.
22:12:40 <ehird> OK then, how about....
22:13:05 <ehird> Deewiant: HOW ABOUT A MEDIA KEY
22:13:21 <ehird> You press the key and it pops up a dialog with the text "NO"
22:13:23 <Deewiant> The Das Keyboard had some key issues that were discussed on geekhack, I forget what they were
22:13:25 <ehird> It is thhte NO key
22:13:33 <ehird> Deewiant: Squeaking?
22:13:36 <ehird> Or in the results?
22:15:16 <Deewiant> Ah, Unicomp's boards have a groove on caps lock
22:16:20 <Deewiant> The one that doesn't isn't buckling spring.
22:16:22 <ehird> But yes, they use the exact same design as the Model M for the most part.
22:16:32 <ehird> The springs, the layout, the circuitry everything.
22:16:46 <Deewiant> Yeah, and that's unacceptable. :-P
22:17:07 <ehird> Cuz buckling spring sucks? IF SO I AGREE
22:17:14 <ehird> http://pckeyboards.stores.yahoo.net/en104bl.html has the One True Layout, though,
22:17:18 <Deewiant> Filco's keyboards are fairly close to optimal but too bad the Z series is a crapshoot
22:17:24 <ehird> Well, jusut the main section.
22:17:29 <ehird> Although the space could be bigger.
22:17:58 <Deewiant> Blue Cherries might be nice but those don't come in tenkeyless form. (Except for the one US dealer who had them special-ordered... in US layout only, of course)
22:18:24 -!- adam_d has quit (Read error: 110 (Connection timed out)).
22:18:35 <ehird> The thing with the tactile but non-clicky switches is that I'd feel compelled to bottom every key out so I can hear it.
22:18:54 <Deewiant> Nothing wrong with bottoming out IMO
22:19:24 <ehird> Clicky but non-tactile switches are funny; I think they exist.
22:19:37 <ehird> "Want a noisy keyboard? Don't want this to help you in any way?"
22:19:49 <ehird> I mean, non-clicky non-tactile, sure, the mechanical switches help... but why add useless noise?
22:20:48 <ais523> like mobile phone cameras that simulate a mechanical clicking sound when you use them?
22:20:59 <Deewiant> That's for legal reasons, isn't it
22:21:09 <ehird> Deewiant: They exist, though
22:21:11 <Deewiant> So you can't take a picture in secret
22:21:21 <Deewiant> ehird: Don't they all do that?
22:21:49 <ehird> I'm not talking about cameras
22:22:02 <ehird> For example, the Cherry MX Black is a non-tactile, non-clicky switch--which is to say it is linear and does not transmit a bump to the user's fingertip when pressed and it does not provide an audible click. The Cherry MX Blue, however, is both tactile and clicky. And the Cherry MX Brown is tactile, but not clicky. And all three require different amounts of force to actuate, the heaviest being the Black model, followed by the Blue, and then the Brown.
22:22:10 <ehird> The combinations are all sensible.
22:22:22 <ehird> I want my clicky non-tactile switches!
22:22:26 <Deewiant> You said you think they exist and now you say they exist
22:22:36 <Deewiant> I doubt they exist, personally :-P
22:22:38 <ehird> I don't know what you're talking about
22:22:43 <ehird> I thought they did though
22:22:49 <Deewiant> 2009-10-08 00:21:49 ( ehird) I'm not talking about cameras
22:22:56 <ehird> YOU ARE SO FUCKING CONFUSING SHUT UP.
22:22:58 <Deewiant> Then you had to be talking about the key switches
22:23:18 <Deewiant> It helps to not use "they" all the time.
22:23:34 <ehird> It helps not to use your mom all the time, Deewiant, but I don't see you heeding that advice.
22:24:04 <ehird> They're sane, but they are retards.s
22:24:22 <ehird> Personally I'd trust them over them, because they are just crazy, but they could alleviate that problem a bit with their help.
22:24:52 <Deewiant> Using "they" twice in the same clause referring to different entity groups is just fucked up, though.
22:25:14 <ehird> I have never made a single message talking about cameras before you accused me of doing so.
22:26:01 -!- kar8nga has quit (Remote closed the connection).
22:26:07 * ehird wonders why we still use staggered keyboard layouts.
22:26:16 <ehird> I mean, guize, we're not using typewriters here?
22:26:30 <FireFly> I'm using a Dvorak variant :)
22:26:35 <Deewiant> You said "they exist, though" after I posted about cameras; having previously said "they exist" re. keyswitches I figured "they" had to be cameras
22:26:59 <ehird> "They exist" wrt cameras didn't really make sense in context, so blah!
22:27:26 <Deewiant> It didn't make much sense in any context so I picked the most recent as most likely
22:27:27 <ehird> Also, why do people use Colemak? It's a total compromise.
22:27:58 <ehird> It's designed to be a mid-way point that's easier for QWERTY users to learn, and panders to WIMP applications' default keyboard shortcut set.
22:28:05 <ehird> That doesn't seem a sane long-term solution to me.
22:28:39 <Deewiant> It's designed as highly optimized while not differing from QWERTY where sensible
22:29:04 <Deewiant> Or more like "differing as little as possible"
22:29:11 <ehird> Why is that better than Dvorak, which is slightly more revolutionary but designed based on typing speed rather than the status quo?
22:29:23 <Deewiant> Because Colemak is supposedly better optimized.
22:29:43 <ehird> I haven't seen that being said anywhere.
22:30:07 <ehird> (Incidentally, some Colemak advocates are weird:
22:30:08 <ehird> "Colemak uses the home row 14% more than Dvorak, and 122% more than QWERTY
22:30:09 <ehird> On Dvorak your fingers move 10% more…and on QWERTY 102% more than Colemak"
22:30:13 <ehird> 102% less finger movement!
22:30:38 <Deewiant> X is 102% more than Y does not mean that Y is 102% less than X
22:30:38 <pikhq> ... Percents don't work that way, mmkay?
22:31:05 <ehird> i.e., it's oddly phrased at first sight.
22:32:26 <Deewiant> ehird: An argument I think I saw on colemak.com was that with Colemak they had the advantage of having computers (Dvorak being too old for this) so they could just run shit on lots of text and calculate what works better
22:33:00 <ehird> I highly doubt such a script would produce something so similar to QWERTY; they have clearly crippled it to retain similarity.
22:33:18 <ehird> Incidentally, that was done experimentally with a Dvorak typist on themselves, remember? End result was not that much changed from Dvorak, and not any faster.
22:33:19 <Deewiant> Sure, but it still beats Dvorak.
22:33:32 <ehird> So... Dvorak is pretty optimal by chance.
22:33:37 <ehird> Or, y'know, by reasoned design.
22:34:10 <ehird> Making a computer change shit according to typing patterns and analyse the results.
22:34:17 <Ilari> Remapping only the QWERTY letters region? That would retain lot of similarity...
22:35:01 <Deewiant> ehird: I didn't say a computer was the one changing shit
22:35:12 <ehird> Your mom changes shit. :|
22:35:29 <Deewiant> Without seeing the algo I can't comment; local maxima are likely
22:35:36 <Deewiant> If they started from QWERTY and got near-Dvorak then that's interesting
22:36:04 <ehird> They started from Dvorak, and the end optimal result was a little different from Dvorak, and not better enough to bother deviating from the standard.
22:36:22 <Deewiant> My guess is they got stuck in a local optimum.
22:36:27 <ehird> i.e., "Let's see how I can type better. Computer analyses, changes, he adjusts, types a bit, computer analyses, repeat."
22:37:03 <Deewiant> Anyway, interesting tidbit: http://en.wikipedia.org/wiki/Blickensderfer_typewriter had the keys DHIATENSOR on the bottom row, those being the most common letters in English
22:37:25 <Deewiant> Rearranging that to ARSTDHNEIO gives the Colemak home row.
22:37:33 <Rugxulo> the esoteric language ETA uses ETAOINSH due to popular usage
22:37:54 <ehird> Colemak obviously started with that due to its famousosity.
22:38:04 <ehird> (There's some disagreement on what the actual most common letters araer.)
22:38:26 <Rugxulo> what, never seen Wheel of Fortune? R S T L N E ;-)
22:39:01 <Pthing> etaoin shrdlu is the actual order
22:39:02 <ehird> guess who else never saw wheel of fortune
22:39:03 <Deewiant> Anyhoo, I didn't find out about the Blickensderfer until months after I started with Colemak :-P
22:40:36 <Deewiant> http://colemak.com/Compare has a Java applet for statisfaction
22:42:25 <ehird> "Java applet[…text not including 'not'…]satisfaction"
22:42:31 <Deewiant> ehird: As far as being crippled due to QWERTY similarity... I'm honestly not sure I'd want to move anything in Colemak, the similarity is mostly in rare letters
22:42:47 <Deewiant> ehird: statisfaction, not satisfaction
22:42:54 -!- oerjan has joined.
22:43:04 <Deewiant> As in, satisfaction through statistics.
22:43:07 -!- augur has quit (Read error: 110 (Connection timed out)).
22:43:14 <ehird> Well that's still a type of satisfaction.
22:43:23 <ehird> Maybe you mean the core dump of Java's statistics.
22:43:39 <Gregor> 4 October: http://lonelydino.com/?id=50 | 7 October: http://www.qwantz.com/index.php?comic=1568 | Coincidence? I THINK NOT.
22:45:07 <ehird> also, I think that's the most surreal Dinosaur Comic for a while
22:45:40 <Deewiant> You can probably find text in which Dvorak trumps Colemak according to those stats but it seems rare enough that I'd rather stick with Colemak
22:45:57 <ehird> 10:06:56 <AnMaster> to ehird when he joins: the power supply on my thinkpad turned out to be too weak to both charge the battery and sustain a heavy load (both cores at 95% or more, 3D graphics). It could sustain the load itself (so the battery didn't discharge, but nor did it charge)
22:45:57 <ehird> 10:07:09 <AnMaster> since he was considering a thinkpad he may be interested in this
22:46:06 <ehird> interesting, especially since I'll be using a 9/12-cell + 3-cell ultrabay battery
22:46:08 <ehird> I'll buy a beefy charger
22:46:12 <ehird> like 95W or something
22:46:13 <Deewiant> Messing with the number keys like the original Dvorak did could be cool, though
22:46:37 <ehird> Deewiant: Based on what, the statistical commonness of digits? :P
22:47:06 <fizzie> "aitnesloku" is the "etaoin shrdlu" phrase for Finnish; or at least the common one that's been going around, I have no clue whether it's actually from a sufficiently large corpus.
22:47:09 <Deewiant> http://en.wikipedia.org/wiki/Dvorak_Simplified_Keyboard#Original_Dvorak_layout
22:47:20 <ehird> 10:16:49 <AnMaster> Rugxulo, how long after that will 9.04 be supported?
22:47:20 <ehird> 10:16:56 <AnMaster> I hope at least a month or so
22:47:24 <ehird> Ubuntu 9.10 will be stable upon release.
22:47:39 <ehird> It is of the same stability stature as 9.04, so.
22:48:18 <Deewiant> Or flipping so that the digits need shift and () etc don't.
22:48:34 <ehird> 10:29:45 <oklokok> game theory is hard.
22:48:34 <ehird> 10:30:15 <AnMaster> oklokok, calculating limits is even harder.... *sigh*
22:48:42 <ehird> well not really, but i lol'd inside and grinned
22:49:11 <Deewiant> Limits tend to be fairly trivial
22:49:32 <ehird> I think in Sweden you're still on basic arithmetic when you enter university or something
22:49:57 <Deewiant> Limits are "basic arithmetic" now?
22:51:22 <ehird> Anyway, I'm joking
22:51:32 <ehird> But AnMaster *has* commented that Swedish mathematics education is ridiculously terrible, so.
22:51:39 <Deewiant> I'm aware he's in uni and I'm not sure what that has to do with anything :-P
22:51:43 <oerjan> Gregor: the third source comic link on that lonely dino strip is not working. also, why the heck are you going via google?
22:51:54 <ehird> i.e., he's gone past basic arithmetic in uni to limits
22:54:11 <Gregor> oerjan: Because the comic numbers don't correspond to the numbers in the file.
22:54:30 <Gregor> oerjan: Except in the rare cases when google won't do it, this method gives a much more useful comic number.
22:55:00 <oerjan> Gregor: rare? i checked and it happened for a source in the previous lonely dino too
22:55:04 <ehird> Just ask North to give a reverser script thing :-P
22:55:11 <ehird> oerjan: Rare? WHAT IT'S A WHOLE TWO
22:55:17 <ehird> How can it possibly be rare, it happened twice.
22:55:27 <oerjan> ehird: it's a 1/3 hit rate on my experimenting
22:55:50 <Gregor> 1/3 miss rate from your two experiments.
22:56:12 <ehird> fizzie: re: compression - why use a generic algo?
22:56:22 <ehird> 90% of the text is word, word, punctuation, space separated, right?
22:56:28 <ehird> Tally up the most common words, go from there.
22:56:33 <ehird> Shorthand, basically.
22:57:01 <oerjan> actually, the second one is not quite as bad, it does give the link but doesn't go directly to it
22:59:21 <fizzie> A shared dictionary for the messages might help, though I doubt that'd save so incredibly much space. It's basically what deflate would do (building the dictionary as it goes along) if I were to actually compress all the messages as a single block.
22:59:46 <Rugxulo> Deewiant: http://en.wikipedia.org/wiki/Wheel_of_Fortune_(U.S._game_show)
22:59:57 <ehird> fizzie: Eh! Link me; I'm reduce the fuck out of that size.
23:00:34 <Deewiant> I've played the DOS game and possibly seen parts of a Finnish equivalent
23:00:44 <fizzie> ehird: http://www.cis.hut.fi/htkallas/rfk86/ and messages.txt.
23:00:48 <ehird> [[Mr Berlusconi insists he should not be 'distracted' from governing
23:00:49 <ehird> Italy's Constitutional Court has overturned a law granting Prime Minister Silvio Berlusconi immunity from prosecution while in office.
23:00:49 <ehird> "We must govern for five years with or without the law."]]
23:00:56 <fizzie> What was Deewiant's final compression score on that?
23:01:19 <ehird> A million bux and a 1.7GiB of RAM?
23:01:46 <ehird> fizzie: logo.png; what's with the "Your job is []"?
23:02:00 <fizzie> ehird: The [] is the robot from the row above.
23:02:12 <fizzie> It's... maybe not so clear.
23:02:39 <ehird> fizzie: So, that text antialiasing on the name... that's done by flickering?
23:02:41 <fizzie> See, "you are robot"; and the robot is denoted with the [] character.
23:02:45 <ehird> I wonder if that actually, you know, works, visually.
23:03:03 <fizzie> It does, the screen response rate is so slow.
23:03:18 <ehird> Wouldn't that just make it look bold and blocky in slow alternation
23:03:49 <ehird> Clear to play, no?
23:03:52 <fizzie> FireFly: Actually it's now "clear to play".
23:03:56 -!- BeholdMyGlory has quit (Remote closed the connection).
23:04:03 <fizzie> ehird: The screenshot png there is an old one, which says "exit".
23:04:17 <ehird> Not a screenshot, is it, having greyscale pixels?
23:04:31 <fizzie> It's an emulator screenshot.
23:04:39 <fizzie> The emulators emulate the grayscale; it's so common.
23:04:57 <fizzie> But I think it does work quite well on the hardware.
23:05:15 <ehird> So, processing power for decoding doesn't really matter does it? I mean, the primary constraint is RAM. As long as it's instant on a modern computer...
23:05:50 <FireFly> RAM: 128 KB, 96 KB user-accessible <-- ow, a lot
23:05:55 <ehird> Nested quantifiers in regex; marked by <-- HERE in m/...thingy??? <-- HERE
23:06:17 <fizzie> Er, well, it's something like a thousand times slower (and restricted to 8-bit operations), so "instant on a modern computer" can stretch quite a lot.
23:06:34 <ehird> This will be simpler than deflate, anyway.
23:07:17 <ehird> Uh... Perl's function join does the opposite of split, right?
23:07:19 <ehird> Why am I using Perl.
23:07:24 <ehird> I'm writing this in Python.
23:07:30 <fizzie> FireFly: The "user-accessible" remark is pretty strange; what it means is that 96 KB is the amount which is used for the "filesystem" where the calculator stores all variables and programs; the remaining 32K is used for running programs (and the OS).
23:08:32 <FireFly> How much of it can ASM programs allocate? Still only the "user-accessible" part?
23:08:39 * FireFly wants a USB slot for his TI-82
23:08:58 <FireFly> 'cause TI-BASIC is UTTERLY SLOW
23:09:28 <Rugxulo> apparently Ander D'Nar wrote one for TI-8x already
23:09:51 <fizzie> There's one 16K page (page 1) that is mostly empty that an asm program can use as scratch space, and there are a few holes on the permanently mapped 16K page (page 0); there's something like 8 K in which it loads the program to execute, and some other places you can use more or less safely.
23:09:54 <ehird> This$ $is$ $the$ $tenth$ $key$ $you$'$ve$ $found$ $so$ $far$.$
23:10:31 <fizzie> (But the rest of the holes are a lot smaller; there's one kilobyte used by the graph screen copy, that's probably the biggest one of them.
23:10:59 <ehird> It's$ $a$ $Cat$ $5$ $cable$.$
23:10:59 <ehird> It's$ $a$ $U$.$S$.$$ $president$.$
23:11:01 <ehird> Same thing, really.
23:11:16 <ehird> I'll have to strip out those spaces and make them just part of the separator in the serialisation.
23:11:43 <FireFly> What are you doing, grabbing the [\w']+'s?
23:11:50 <ehird> messages = sys.stdin.read().split('\n')
23:11:51 <ehird> split_msgs = [re.split(r'([ .!?"])', msg) for msg in messages]
23:11:51 <ehird> print '\n'.join('$'.join(msg) for msg in split_msgs)
23:11:55 <fizzie> And of course if you're willing to code your program so that it can do a VAT (variable allocation table) lookup, you can put stuff in user-visible variables. But there's only 64K addressable on the processor; the memory is mapped in 16K pages, and it's a bit iffy to handle things that cross page boundaries.
23:11:55 <ehird> Going to optimise by word, basically.
23:12:16 <ais523> ^bf ++++++++++[>+++++++>++++++++++>+++>+<<<<-]>++.>+.+++++++..+++.>++.<<+++++++++++++++.>.+++.------.--------.>+.>.
23:12:21 <ais523> !bf ++++++++++[>+++++++>++++++++++>+++>+<<<<-]>++.>+.+++++++..+++.>++.<<+++++++++++++++.>.+++.------.--------.>+.>.
23:12:22 <ehird> Approaching$.$$ $$ $One$ $car$.$$ $$ $J$.$$ $$ $Followed$ $by$.$$ $$ $Two$ $car$.$$ $$ $M,$ $M$.$$ $$ $In$ $five$.$$ $Minutes$.$
23:12:27 <ehird> WTF is up with those $$s, I wonder?
23:12:29 <fizzie> Just remember to count your dictionary size in the compression results.
23:12:43 <ehird> fizzie: Naturally.
23:12:59 <ehird> fizzie: It could be included in the decompression code to save some cycles.
23:13:26 <fizzie> If it's not more than a kilobyte or so.
23:13:52 -!- coppro has joined.
23:13:54 <ehird> It's$ $a$ $blatant$ $plug$ $for$ $Ogg$ $Vorbis$,$$ $http://www$.$vorbis$.$com/
23:14:00 <ehird> Obviously those URLs are going to be non-optimal...
23:14:09 <ehird> fizzie: I'll omit the dictionary for fnords, obviously.
23:14:13 <ehird> (= Singular appearances.)
23:14:27 <ehird> So that'd likely come out with http://www, vorbis, and com/ literally.
23:14:45 -!- FireFly has quit ("Later").
23:15:19 <ehird> Ehh. Lossy compression is dead to me for new encodes.
23:15:43 <ehird> Also, Vorbis is probably not patent-free because nothing is.
23:15:47 <coppro> well, you should always have at least one lossless copy, but it's not what you want to be throwing around
23:16:04 <oerjan> ehird: whenever [ .!?"] matches two chars in sequence, you get an empty string as part of the split, i think
23:16:07 <ehird> (And CD doesn't count as a lossless copy, btw.)
23:16:09 <coppro> by "throwing around", I mean the Internet
23:16:12 <ehird> oerjan: Ah. You are indeed correct.
23:16:25 <ehird> CDs are only not lossless because it's so easy to damage them.
23:16:35 <ehird> Disks with backups and SSDs (failure mode doesn't lose data) are lossless.
23:16:54 <ehird> The$ $non-kitten$ $item$ $bites$!$
23:18:22 <AnMaster> ehird, anyway, digital isn't lossless
23:18:37 <ehird> Neither is any analog storage mechanismm.
23:18:51 <ehird> (And I hope you're not one of the "Vinyl sounds better than CD" folk...)
23:19:11 <AnMaster> ehird, no I'm one of those "live wins over both"
23:19:32 <ehird> Live music is hardly *more* polished than a recording.
23:19:56 <AnMaster> ehird, you don't have shouting fans at a classical concert
23:20:00 <ehird> If we go by authenticity, realising it in the actual platonic space of concepts and ideas is the best way to listen to music.
23:20:16 <AnMaster> and it is quite well polished indeed
23:20:30 <ehird> Yes, but a studio recording is hardly ever going to be less polished.
23:21:01 <Rugxulo> .mp3 is worse than CD quality, but nobody complains
23:21:47 <ehird> Are we talking 128kbps CBR mp3s?
23:21:57 <ehird> Because, sure, I can't tell the difference most of the time, but eww. Just eww.
23:21:59 <AnMaster> <Rugxulo> .mp3 is worse than CD quality, but nobody complains <-- I do.
23:22:14 <ehird> Rugxulo is using "complains" to mean "can tell the difference".
23:22:23 <ehird> (And it would be unwise to challenge that, as it's almost certainly true.)
23:22:44 <ehird> Well, sure, Frauhnhofer or whatever at 128 will be ridiculously bad.
23:22:52 <ehird> -V2 with a recent LAME is great, though.
23:23:01 <ehird> But really, lossless is fine.
23:23:19 <ehird> I use ALAC, but that's only because iTunes can't do FLAC non-retardedly.
23:23:28 <ehird> It's basically FLAC, but proprietary and larger.
23:23:40 <ehird> Well, I also have mp3s because pirates can't generally be picky.
23:23:52 <AnMaster> ehird, larger == fail for compression
23:24:02 <ehird> It's not larger than the original soucre.
23:24:06 <ehird> Just larger than FLAC.
23:24:13 <ehird> Which is to be expected, because FLAC is the king of lossless compression sizes.
23:24:25 <ehird> A few megs don't matter at these sizes. :P
23:24:58 <ehird> A$ $number$ $of$ $short$ $theatrical$ $productions$ $are$ $indexed$ $1$, $2$, $3$, ... $n$.$
23:25:08 <ehird> Now, I'm not sure having ", ... " as one symbol makes sense.
23:25:11 <ehird> Eh; probably doesn't matter.
23:25:23 <ehird> fizzie: Can the TI-86 handle non-byte aligned stuff?
23:25:33 <ehird> No reason not to have "a" or "the" be two bits, right? :P
23:26:02 <ehird> Admittedly it won't be very seekable. I could align at seek points.
23:26:59 <ehird> fizzie: Have you considered antialiasing victory.png? It is rather unsightly.
23:30:46 * coppro attempts to find a way to keep a window on top but make sure it doesn't get any mouse clicks
23:32:56 <ehird> Active window != top window.
23:33:13 <ehird> Use a sloppy-focus WM, raise said window, hover over another.
23:33:40 -!- ais523 has quit (Remote closed the connection).
23:33:42 <coppro> ehird: yeah, but I don't think KDE will do that
23:33:51 <ehird> kwin is just one window manager.
23:33:56 <ehird> You can replace it with another, temporarily or not.
23:34:03 <ehird> KDE doesn't depend on kwin.
23:34:09 <coppro> yeah, sorry, meant kwin
23:34:23 <coppro> but it's not worth my effort to find a wm that will do that and just use it temporarily
23:34:23 <ehird> Also, you could use the other WM and then replace it with kwin without moving the mouse.
23:34:33 <ehird> You already have it.
23:34:55 <ehird> I find your distro's packaging disturbing.
23:38:25 <ehird> Well, it's practically osmosified itself into general internet lingo by now, but yess.
23:41:03 <ehird> I love how Safari has an "uncrash" button
23:42:13 <Rugxulo> "advertising, advertising, advertising ... fix Vista ... advertising, adver ..."
23:42:27 <ehird> Holy crap, another Notational Velocity release?
23:42:43 <ehird> The world as we know it may be over. Say your goodbyes.
23:43:32 <fizzie> ehird: It can handle non-byte-aligned stuff approximately as well as any other byte-addressing processor. Anyway, the deflate-like stream is already composed of variable-length Huffman-coded symbols, that don't even pretend to be byte-aligned. I just align each separate message to start at a byte boundary.
23:43:47 <fizzie> And no, I don't think I'll bother antialiasing victory.png; it's animated, anyway, that sort of makes up for it.
23:44:13 <ehird> Animate it by just changing the colours of pixels; after all, the flickering will antialiase it differently.
23:44:16 <ehird> It's sort of like subpixel movement!
23:44:39 <ehird> Like, one frame just changes the colours of some of the antialiasing, the next moves the frame. It looks like a three-fame movement, but it isn't! Sort of.
23:46:13 <fizzie> Anyway, I'll probably try doing something like adding a dictionary of the most common substrings over the whole messages.txt, and changing the LZ77 repeat-pair encoding so that it can refer either to the output or the fixed dictionary; that should be a reasonable compromise method. But not today.
23:46:34 <ehird> I think a word-based approach will work best; it is English, after all.
23:47:19 <fizzie> A word-based approach is just a special case of a substring-based approach, really; if words are what are repeated, then words are what will end up in the dictionary.
23:49:12 <fizzie> What I like about LZ77 is that it pretty elegantly includes also run-length-encoding as a special case (with a length > 1 and distance = 1 pair); that makes a difference because I got bored of Z80 assembly and did line-wrapping with Perl, padding things with spaces so that they wrap properly in the message area.
23:53:58 -!- Rugxulo has left (?).
23:54:01 <fizzie> And replying to a comment ages ago in the backscroll, Ander D'Nar did not write one for "TI-8x", no matter how much it might say so on the news page; his version is for TI-83+ and TI-84+ (which really is pretty much identical) only, not the TI-86.
23:54:27 <fizzie> I wouldn't have been doing the port if it existed already.
23:55:08 <fizzie> Though the robotfindskitten.org people aren't answering my emails (well, email), so...
23:55:45 <ehird> It's been whole DAYS!
23:58:36 <fizzie> I shall be mailbombing them very soon now to teach them just who they're dealing with.
23:59:11 <fizzie> (In other words, I'll send a second email when I get the official rfk86 web page up.)