←2010-09-19 2010-09-20 2010-09-21→ ↑2010 ↑all
00:00:10 <zzo38> If they don't publish their methods, you cannot know what the experiment means at all, if in fact there is one, which there might not be!
00:00:12 <Sgeo|Empathy> Should I format the external HD?
00:00:15 <quintopia> and if 10000 independent groups all reproduced those results, all without publishing the methods?
00:00:22 <pikhq> 10,000 drunk hobos.
00:00:46 <quintopia> but i don't see how you can argue in that case that the result has not been independently verified
00:01:06 <pikhq> It's 10,000 completely different experiments which claim to be the same.
00:01:07 <quintopia> the basic structure of the experiment is the same, and it is that structure that makes the desired argument
00:01:20 <pikhq> None of which can be reproduced.
00:01:29 <pikhq> And, indeed, could be quite easily *made up*.
00:01:34 <pikhq> As such, 10,000 drunk hobos.
00:01:48 <fizzie> If you're talking about my plots the answer is "no, because I just went to sleep". (The features those plots used also need quite a lot of comments before they start to make sense at all.)
00:01:58 <quintopia> they can only be *made up* if 10000*x test subjects can all be convinced to lie
00:02:05 <Sgeo|Empathy> I think I'll format the HD first
00:02:12 <pikhq> Indeed. And perhaps they could.
00:02:17 <quintopia> thx fizzie
00:02:22 <pikhq> Or perhaps those test subjects don't exist.
00:02:29 <pikhq> Perhaps, in fact, it's merely claimed they exist.
00:02:42 <quintopia> ah, and they're all actually sock puppets?
00:03:06 <quintopia> well if you're going to argue that, then seeing the logs of the conversations will not convince you either
00:03:07 <pikhq> And I certainly couldn't test it with real people myself, because *I don't even know what they claimed to do*.
00:03:14 <quintopia> because it would just be someone talking to himself
00:03:27 <pikhq> See, with logs of the conversations just about anyone could *do it themselves*.
00:03:49 <pikhq> There could, in fact, be someone testing it and going "That's weird, absolutely *none of that* happened."
00:03:51 <quintopia> unless their subjects had also seen such logs and as such were not fooled
00:04:13 <Sgeo|Empathy> So don't let the subjects see the logs!
00:04:30 <pikhq> "We tested that gravity exists. Yes." This is as equally valid as this experiment. Namely, *it's a fucking hobo*.
00:04:59 <pikhq> Or perhaps you would prefer "We tested that gravity exists. In fact, it's God."
00:05:22 <quintopia> so, do you think that every published experiment that you have not, as yet, independently tried yourself, is nonsense?
00:05:31 <quintopia> i'm sure you take a leap of faith somewhere
00:05:49 <pikhq> No. However, it's complete and utter nonsense unless I could, in fact, *decide not to take it on faith*.
00:05:59 <zzo38> pikhq: Such a statement "We tested that gravity exists. In fact, it's God." is not even meaningful statement. It explains nothing about gravity or about God, or about the experiment you have performed.
00:06:10 <pikhq> zzo38: My point.
00:06:28 <zzo38> pikhq: OK.
00:06:41 <quintopia> pikhq: you could probably do that in this case anyway. just ask them what the methods used were on pain of death if you reveal them, and try them out
00:07:03 <pikhq> quintopia: But not necessarily.
00:07:04 <Sgeo|Empathy> External harddrives are huge
00:07:20 <pikhq> As such, it's not science, but A MOTHERFUCKING HOBO.
00:07:41 <quintopia> as such, you clearly don't care whether or not it is a hobo, since you haven't tried
00:07:50 <pikhq> I don't even care if every single thing said hobo says turns out to be true, ITS STILL A FUCKING HOBO.
00:08:03 <quintopia> personally, i care
00:08:13 <pikhq> I shall beat you with the full disclosure stick.
00:08:18 <quintopia> if a drunken hobo has the secret to a GUT, i want to know about it
00:08:24 <Sgeo|Empathy> The sky is blue. I know this because magical unicorns told me so.
00:08:38 <pikhq> quintopia: Learn. About. Science. NOW.
00:09:00 <quintopia> if there are enough reasons to believe that the hobo is not a crock of shit, i will go ask the hobo myself
00:09:10 <pikhq> quintopia: LEARN ABOUT SCIENCE NOW
00:09:30 <quintopia> boy did i get you wound up...
00:09:39 <pikhq> Yes. Morons do that.
00:09:57 <pikhq> Now learn about science or forevermore sound like a moron who thinks 2+2=þ
00:10:21 <Gregor> 2+2 does equal thorn.
00:10:24 <Sgeo|Empathy> quintopia: The sky is, at least sometimes, blue. Grass is often green.
00:10:27 <Sgeo|Empathy> I am God.
00:10:57 <pikhq> Gregor: Oh, right, I forgot about your 123þ counting system. :P
00:11:59 <Gregor> pikhq: Just part of my base-ð numbering system.
00:13:05 <Sgeo|Empathy> Hmm
00:13:14 <Sgeo|Empathy> I can store the logfile right on the new HD
00:14:09 <quintopia> i see the semantic difference between our definitions of valid now. you're talking about results being validated to the general science community, whereas i'm talking about an experiment being valid as a thing-in-itself. the latter is the precursor to the former with transmutation occurring upon publication of methodologies.
00:14:48 <pikhq> quintopia: Okay, so I'm talking about the useful definition and you're talking about the irrelevant one.
00:16:17 <Gregor> That's it. Time to make a stand.
00:16:41 -!- Gregor has set topic: Welcome to #esoteric, the international hub for esoteric topics in computing and programming languages | logs: http://tunes.org/~nef/logs/esoteric/?C=M;O=D.
00:16:47 <Gregor> A stand on a completely unrelated topic.
00:17:02 <Gregor> Or rather, a stand on a completely unrelated issue.
00:17:26 <Gregor> Henceforth I decree that #esoteric is for esoteric topics in computing, not just esoteric programming languages :P
00:19:04 <Sgeo|Empathy> Ok
00:19:12 <Sgeo|Empathy> Vorpal: you awake?
00:19:19 <Sgeo|Empathy> -n to get the easy stuff first
00:19:30 <Sgeo|Empathy> Then once that's done, -r3 to attempt the error-laden stuff?
00:20:42 <quintopia> hurray!
00:21:08 <quintopia> pikhq: it's a useful distinction to make nonetheless
00:21:47 <pikhq> quintopia: Yes, the distinction between a hobo saying correct things and a scientist saying correct things.
00:21:54 <pikhq> Please, don't encourage the hobos.
00:22:29 -!- augur has joined.
00:24:54 <quintopia> pikhq: not the distinction i was going for here, but an interesting strawman
00:27:56 <Sgeo|Empathy> Well, BRB
00:28:01 <Sgeo|Empathy> Wish me and my HD luck
00:28:09 -!- Sgeo|Empathy has quit (Remote host closed the connection).
00:29:29 -!- ais523 has quit (Remote host closed the connection).
00:30:04 <quintopia> to be clearer in the future, i shall refer to the yudkowsky experiments as "validatable" rather than "valid"
00:33:57 <nooga> i'm playing gramophone diagonally
00:34:10 -!- alise has joined.
00:34:17 <quintopia> i dunno what that means
00:34:18 <quintopia> hi alise
00:34:26 <nooga> 01:33 < nooga> i'm playing gramophone diagonally
00:35:31 -!- SgeoN2 has quit (Ping timeout: 245 seconds).
00:35:56 -!- SgeoN1 has joined.
00:36:10 <alise> 14:45:48 <Phantom_Hoover> Extrapolating Moore's law infinitely as evidence for a singularity seems... dodgy.
00:36:14 <alise> only Kurzweil does that
00:36:31 <alise> 14:47:25 <Phantom_Hoover> And the fact that the good old laws of thermodynamics have *already* started to cap processor clock speeds.
00:36:41 <alise> general consensus is that any sane seed AI is going to convert itself to nanotech asap
00:36:48 <alise> 14:48:28 <Phantom_Hoover> I mean, we can probably push quite a bit further, but not indefinitely.
00:36:52 <alise> you don't need infinite power.
00:37:04 <nooga> vinyl records are awesome
00:37:05 <alise> 14:49:26 <quintopia> adiabatic computing could go a long way toward stepping past current heat-related limitations
00:37:08 <SgeoN1> Stop it! Stop it! Stop it! Stop it!
00:37:08 <SgeoN1> Stop trying to boot from the HD!
00:37:09 * SgeoN1 cries
00:37:10 <alise> WARNING: Adiabatic computing is quackery
00:37:23 <nooga> psychedelic metal on vinym records is doubly awesome
00:37:35 <nooga> vinyl*
00:38:08 <pikhq> Anyone got any idea how "error: unable to fork: Cannot allocate memory" could happen with plenty of memory?
00:38:09 <SgeoN1> Argh
00:38:21 <alise> pikhq: too many procs?
00:38:27 <alise> limited memory space allocated to proc table?
00:38:55 <SgeoN1> Help me alise,!
00:39:04 <pikhq> alise: A mere 56 procs running.
00:39:10 <nooga> 56?!
00:39:16 <nooga> i've got 2 running
00:39:22 <pikhq> nooga: No, you don't.
00:39:30 <nooga> htop says that
00:39:37 <pikhq> Your system is *definitely* running more than 2.
00:39:45 <pikhq> Even if your current *shell* is not.
00:39:49 <SgeoN1> It's randomly refusing to boot into the USB!
00:39:49 <nooga> oh, 1 running
00:39:53 <alise> nooga: ps -A
00:40:00 <nooga> 308 total
00:40:06 <nooga> -.- ubuntu
00:40:16 <SgeoN1> Maybe I should boot from CD? But then, if Ubuntu, there's even more delay installing ddrescuee
00:40:18 <nooga> with apache stack
00:40:26 <nooga> n'shit
00:40:35 * pikhq turns off nested paging in VirtualBox
00:41:06 <SgeoN1> Help me!
00:41:15 <alise> SgeoN1: Hi.
00:41:25 <pikhq> Nope, that didn't do it.
00:41:27 <alise> Boot from CD.
00:42:05 * pikhq shall try qemu
00:42:24 <SgeoN1> Well, that could only have resulted in more damage to the hd
00:42:25 <alise> on TV program "I've Got a Secret": [[GSN ran a revival from April 17 to June 9, 2006 with an all-gay panel.]]
00:42:30 <SgeoN1> Putting in CD now
00:42:38 <alise> "Let's revive 'I've Got a Secret'... but with GAY."
00:43:05 <EOF> hmm
00:43:06 <quintopia> how many virtualboxen can one nest on a reasonable system before losing interactivity in the inner containers?
00:43:07 <EOF> no
00:43:26 <EOF> a few
00:43:30 <EOF> what OSes
00:44:11 <pikhq> quintopia: With the virtualisation extensions, "a lot".
00:46:17 <SgeoN1> Finally found the Ubuntu CD
00:46:19 <quintopia> aha
00:46:37 <SgeoN1> And the comp justsied
00:46:40 <SgeoN1> Died
00:46:51 <alise> 14:50:37 <Phantom_Hoover> We'll be right with some things, but wrong with many, many more.
00:47:01 <alise> which is why the Singularity is defined at precisely the point where we have no fucking idea what will happen
00:47:13 <alise> (by definition, when a seed AI surpasses human intelligence)
00:47:44 <quintopia> i think a lot of our guesses are ace though
00:47:53 <SgeoN1> I don't have the external mount
00:48:00 <SgeoN1> The HD is inside the cokputee
00:48:02 <quintopia> AI-controlled super-economies moving at the speed of light
00:48:08 <pikhq> qemu is rather slow without virtualisation extensions. And I'm too lazy to get KVM set up.
00:48:11 <pikhq> (that involves rebooting)
00:48:18 <quintopia> humans becoming happy complacent pawns in a machine they don't understand
00:48:20 <quintopia> it'll be great
00:48:50 <SgeoN1> I'll need to install ddrescue
00:48:50 <pikhq> Okay, this may actually be *too slow*.
00:49:24 <SgeoN1> Which requires putting in the wireless password, which requires x, unless you can give me a tutorial right now about doing that from the commandline
00:50:14 <SgeoN1> Alise, any lightning fast tutorial?
00:50:20 <alise> SgeoN1: Of?
00:50:24 -!- pikhq has quit (Quit: New kernel!).
00:50:39 <SgeoN1> Connecting to wifi from commandline
00:50:46 <alise> SgeoN1: Plug in an Ethernet cable.
00:50:53 <SgeoN1> N/m
00:50:59 <alise> It is basically impossible to do without a graphical tool.
00:51:52 <SgeoN1> This things ability to boot from USB is sporadic at beat
00:52:21 <SgeoN1> What do you mean 0 matching items???????
00:52:23 <alise> SgeoN1: You have no idea what parts of the disk are valuable, right?
00:52:37 <alise> package is "ddrescue"
00:52:58 <alise> quintopia: economies?
00:52:59 <SgeoN1> Whataoctwareaour e
00:53:01 <alise> In a post-scarce world? Why?
00:53:06 <SgeoN1> What software source?
00:53:08 <alise> *post-scarcity?
00:53:23 <alise> SgeoN1: Wait, it may actually be gone.
00:53:24 <alise> Nope.
00:53:28 <alise> universe
00:53:58 <alise> SgeoN1: How old is the drive?
00:54:03 <quintopia> alise: post-scarcity? wtf? there are always limitations on material, and a lot of material will be needed to fuel the AI overlords' need for memory and compute power
00:54:04 -!- FireFly has joined.
00:54:15 <SgeoN1> From 2006 or so
00:54:29 <alise> quintopia: The AI can sustain itself; nobody else needs those resources.
00:54:36 <SgeoN1> Ubuntu is set to spin down disks when unneeded. Wait, no it's not
00:54:49 <quintopia> alise: are you assuming a single AI?
00:55:03 <alise> quintopia: That is the only possible outcome.
00:55:14 <quintopia> i think even a single AI, sufficiently distributed, would compete with itself for resources
00:55:33 <alise> If there are two Friendly AIs, they will merge. If there are two un-Friendly AIs, we're fucked anyway. If there is a Friendly AI and an un-Friendly AI, whichever started first or improved faster wins.
00:55:44 <alise> After an AI has established power, it will be able to prevent another AI from being created.
00:55:54 <alise> SgeoN1: Is it installed?
00:55:56 <SgeoN1> Uh
00:56:03 <SgeoN1> Ddrescue is not autocompleting
00:56:12 <alise> SgeoN1: This is more important first.
00:56:16 <alise> SgeoN1: What media are you going to back up to?
00:56:16 -!- EOF has quit (Ping timeout: 252 seconds).
00:56:18 <SgeoN1> Ugh, do I really want DD_rescue?
00:56:18 <quintopia> alise: what if friendly AIs get lonely not having other higher intelligences to philosophize with?
00:56:21 <alise> Yes.
00:56:23 <alise> You really want ddrescue.
00:56:24 <SgeoN1> External hd
00:56:26 <alise> You really, really want ddrescue.
00:56:27 <SgeoN1> File on
00:56:30 <alise> SgeoN1: It is big enough, right?
00:56:43 <SgeoN1> And not _ ? Ok
00:56:46 <SgeoN1> Yes
00:56:50 <alise> quintopia: You are anthropomorphising. There is no reason an AI would have any desire to do those things.
00:56:59 <alise> SgeoN1: Mount the external HD. DO NOT MOUNT THE BROKEN HD.
00:57:33 <alise> SgeoN1: dd_rescue != ddrescue, btw
00:57:35 -!- pikhq has joined.
00:57:37 <alise> I'm assuming we're using GNU ddrescue here.
00:57:42 <SgeoN1> Gddrescue then
00:57:56 <alise> Whatever.
00:58:03 <SgeoN1> -n to get the easy stuff, -r3 for the rest, right?
00:58:06 <alise> SgeoN1: Am I right in thinking you have no idea where the valuable stuff is?
00:58:09 <alise> Hey, I'm trying to help you here.
00:58:18 <alise> Okay, are you in the mounted external HD?
00:58:31 <alise> SgeoN1
00:58:38 <quintopia> alise: you are deanthropomorphizing, there is no reason a human-created AI wouldn't have those tendencies grandfathered in
00:58:48 <SgeoN1> Physically, or directorywise?
00:58:54 <alise> quintopia: Coding an AI to feel loneliness would be a near-fatal mistake.
00:58:59 <quintopia> also, in general, there is no reason an AI would NOT have those desires
00:59:00 <alise> And it would certainly not be Friendly.
00:59:02 <alise> SgeoN1: Directorywise.
00:59:11 <SgeoN1> I have some idea
00:59:11 <alise> quintopia: Please read "Creating Friendly AI". Thanks.
00:59:16 <alise> SgeoN1: What?
00:59:28 <quintopia> alise: it's long. ijust started but i don't have time tonight
00:59:38 <SgeoN1> Documents and Settings, and one or two files in the root, I think
00:59:48 <zzo38> CWEB treats words like "elif" which are preprocessor commands as reserved words, but are they reserved words?
00:59:50 <alise> quintopia: Well, your opinion about loneliness is very wrong. I suggest finishing CFAI :P
01:00:02 <alise> SgeoN1: That's not helpful; you need to know sectors and shit. Okay, right, so, cd into the external HD.
01:00:05 <quintopia> in any case, there are other reasons an AI might want multiple versions of itself, or independent copies simultaneously running competing for resources etc.
01:00:21 <SgeoN1> You said directorywise
01:00:23 <SgeoN1> Ok
01:00:31 <alise> SgeoN1: I meant directorywise as far as cding.
01:00:37 <SgeoN1> Oh
01:00:39 <alise> SgeoN1: Okay, so,
01:00:43 <quintopia> loneliness was a stupid way of putting it. i was simplifying an uber-intelligences motivations for a human audience
01:00:46 <alise> SgeoN1: It is partitioned, right? The rescuable disk.
01:00:57 <alise> SgeoN1: Or not?
01:00:57 <SgeoN1> I think I know what to do
01:01:01 <alise> Hmm, I guess not.
01:01:06 <alise> SgeoN1: Erm, let's stick to the plan here.
01:01:08 <SgeoN1> Ddrescue -n oldhd imgfile logfile
01:01:09 <alise> You could permanently break your disk.
01:01:20 <alise> SgeoN1: It is sda, right?
01:01:21 <alise> The drive?
01:01:32 <alise> SgeoN1: ...
01:01:42 <SgeoN1> Ddrescue -r3 Oldfield mewing logfile
01:01:43 <SgeoN1> Rifht
01:01:44 <SgeoN1> With a semicolon in betweeen?
01:01:45 <alise> No.
01:01:52 <alise> SgeoN1: Please answer my question.
01:01:58 <SgeoN1> Not sure
01:02:02 <SgeoN1> Hold on
01:02:03 <alise> ls /dev/sd*
01:02:09 <alise> Run that, please.
01:02:12 <alise> I need to know if it's partitioned.
01:03:03 <alise> SgeoN1: ?
01:03:12 <SgeoN1> Crap
01:03:18 <SgeoN1> I typed stuff in, dammit
01:03:20 <alise> *CLAP CLAP CLAP*
01:03:29 <SgeoN1> I meant into irc
01:03:32 <alise> Oh
01:03:55 <SgeoN1> /Dev/sda1 /Dev/sea /Dev/sda4
01:04:11 <SgeoN1> Sda
01:04:15 <alise> o_O
01:04:21 <alise> Okay, let's just rescue the entire thing.
01:04:24 <alise> Your partition table looks fucked.
01:04:25 <alise> Wait a sec.
01:04:27 <alise> pikhq: ^
01:04:30 <alise> Just checking pikhq agrees.
01:04:40 <SgeoN1> What's wrong with the ddrescue commands I typed?
01:04:55 <alise> Well, definitely don't run them automatically after another, and the latter one is wrong.
01:04:58 <alise> Can you just be patient?
01:04:58 <pikhq> Your partition table looks weird.
01:05:01 <alise> Yeah.
01:05:04 <alise> pikhq: So let's just rescue the whole disk?
01:05:05 <pikhq> That's *valid* but weird.
01:05:08 <alise> Not the individual partitions.
01:05:10 <pikhq> alise: I'd say "yeah".
01:05:11 <alise> I mean, he has dropped the thing.
01:05:18 <alise> SgeoN1: Okay, type this very precisely and hit enter:
01:05:19 <pikhq> See what ddrescue gets.
01:05:29 <alise> sudo ddrescue --no-split /dev/sda imagefile logfile
01:05:33 <alise> Veeery precisely.
01:05:45 <alise> It will flag up a ton of errors, I bet.
01:05:47 <alise> That's fine.
01:05:57 <SgeoN1> I didn't put the euro in
01:06:01 <SgeoN1> I'm at a root prompt
01:06:08 <SgeoN1> Sudo
01:06:25 <SgeoN1> Doit.
01:06:29 <SgeoN1> ?
01:06:36 <quintopia> eh
01:06:39 <quintopia> i
01:06:40 <alise> SgeoN1: Euro?
01:06:44 <alise> Yes, do it.
01:06:48 <quintopia> i'ma have to leave now or forever hold my lameness
01:06:54 <alise> quintopia: wat
01:06:58 <SgeoN1> Gypped sudo
01:07:03 <quintopia> have work to do
01:07:09 <quintopia> need to not procrastinate anymore
01:08:10 <alise> SgeoN1: Is it going?
01:08:25 <SgeoN1> It's doing stuff
01:08:25 <SgeoN1> And errors are coming up
01:08:28 <alise> Good.
01:08:42 <pikhq> Expected.
01:08:51 <SgeoN1> How long will this take?
01:09:02 <pikhq> Anywhere from hours to days.
01:09:15 <pikhq> It's a function of how fucked up the drive is.
01:09:32 <SgeoN1> Ok then. What do I do when it's done? What was wrong with NY second command?
01:09:33 <alise> SgeoN1: And now, BOOKMARK THIS URL IN YOUR MOBILE BROWSER:
01:09:34 <alise> http://pastie.org/1169085.txt?key=cmgt84evfis2qjczs7cg
01:09:42 <alise> It has all the instructions for after that, very precisely written out.
01:10:05 <alise> This will get you a drive in your hand that has your old drive on it.
01:10:07 -!- augur has quit (Ping timeout: 240 seconds).
01:11:03 <pikhq> Except with more of the data readily accessible.
01:11:10 <SgeoN1> Can I proceed to use Ubuntu normally in the meantime?
01:11:41 <FireFly> Hm
01:11:48 <pikhq> Yes.
01:11:59 <FireFly> No Vorpal here?
01:12:05 <FireFly> Ah, sleeping
01:12:25 <alise> pikhq: Yes, just don't touch the drive.
01:12:27 <pikhq> Oh, and if you ever need to restart the job, just rerun the ddrescue command. It'll pick up from where it left off.
01:12:27 <alise> Erm.
01:12:31 <alise> SgeoN1.
01:12:34 <alise> What pikhq said.
01:12:37 <alise> But that's not how you deal with errors.
01:12:49 <alise> SgeoN1: Type this URL in your Ubuntu browser very carefully then put it on the desktop or whatever so you don't lose it:
01:12:52 <pikhq> Yeah, it's just how you deal with "I need to shut off my computer".
01:12:53 <alise> http://pastie.org/1169085.txt?key=cmgt84evfis2qjczs7cg
01:12:57 <alise> If you want a shorter URL i can get that.
01:13:07 <alise> SgeoN1:
01:13:12 <alise> http://bit.ly/bmW8YW
01:13:14 <alise> PUT THIS ON YOUR DESKTOP.
01:13:34 <SgeoN1> What's wrong with leaving it on the phone?
01:14:19 <alise> SgeoN1: Well, that works too.
01:14:26 <alise> I just thought it'd be easier to copy-and-paste from the main machine.
01:17:39 <SgeoN1> 1656 kB/s
01:18:02 <SgeoN1> This is going to take a while, isn't it?
01:19:25 <alise> SgeoN1: How big is the drive?
01:19:34 <SgeoN1> 100GB
01:20:07 <alise> SgeoN1: 17.59 hours.
01:20:13 <alise> SgeoN1: I assume it's still spewing errors?
01:20:23 <SgeoN1> Yep
01:20:27 <alise> SgeoN1: If so, then that won't be the end of it, since you'll still have the second and maybe even the third command to run.
01:20:32 <alise> And you might even have to increase the retries on that.
01:20:34 <SgeoN1> And the average speed keeps going down
01:20:47 <alise> SgeoN1: You can, as pikhq said, just terminate it and continue it later using *exactly the same* command.
01:20:53 <alise> (Check the logs if you do so do get it exact.)
01:21:06 <alise> Might be prudent to overnight it, though.
01:21:11 <alise> SgeoN1: How big's the external drive?
01:21:39 <SgeoN1> Hey, I can keep using the computer while it goes, so what's the problem? Just don't take the computer on my lap
01:21:46 <SgeoN1> 250GB
01:22:53 <alise> SgeoN1: Will you be awake in 17 hours?
01:23:02 <SgeoN1> No
01:23:04 <alise> Also, yeah, keep the computer SEATED AT A TABLE FOREVER.
01:23:07 <SgeoN1> Probably not
01:24:32 <alise> oerjan: i have an idea
01:24:42 <alise> let's stop enforcing Graue's ludicrous category policy
01:24:46 <alise> :P
01:25:37 <SgeoN1> Is having Ubuntus thing to spin down disks when posssble acceptable?
01:26:47 * oerjan swats alise -----###
01:27:02 <alise> oerjan: well, hey, it's an idea.
01:27:07 <alise> SgeoN1: don't worry. that disk will never get spun down.
01:27:13 <alise> and it being spinned down is good
01:27:30 <oerjan> alise: given how annoying i just found out it was _renaming_ a category, let's not
01:28:19 <SgeoN1> I seem to have access to a VMS system
01:30:22 <SgeoN1> Any ideas on what I can do to explore it?
01:31:11 <SgeoN1> Average rate went down to 707
01:31:19 <SgeoN1> 683
01:33:22 <oerjan> SgeoN1: you might try the help command
01:34:01 * oerjan recalls it was wonderfully hierarchical
01:34:48 <SgeoN1> For all I know, this is what runs the school's main ... styff
01:35:01 -!- SgeoN1 has quit (Quit: Bye).
01:35:24 -!- SgeoN1 has joined.
01:35:50 <oerjan> it even included individual commands for some programming languages
01:36:08 <SgeoN1> Cobol
01:36:42 * oerjan recalls there was a lisp
01:36:59 <SgeoN1> 568
01:38:09 <SgeoN1> 538
01:39:13 <oerjan> i think it was pascal that had the individual command thing
01:39:18 <SgeoN1> This is going to take a very long time, isn't it?
01:50:00 -!- FireFly has quit (Quit: Sweden is screwed).
01:55:29 <SgeoN1> About 3 and a half days
01:56:32 <SgeoN1> Most of the data I don't honestly care about
01:56:52 <SgeoN1> Any way to, say, use TestDisk and just take only what I really care about?
01:59:25 <alise> Maaybe. I'd just do this.
02:00:18 <SgeoN1> I wonder if my professor would be ok with me not bringing my laptop in on Monday
02:00:27 <SgeoN1> I can still do work from my phone
02:02:15 <SgeoN1> If the average goes down to 3.3 KB/s or so, may I give up?
02:03:00 <pikhq> Wow. It is estimated the LSD of marijuana is 1500 *pounds*.
02:03:08 <pikhq> Erm.
02:03:11 <pikhq> LD50
02:03:14 <pikhq> Not LSD
02:03:19 <pikhq> Amusing slip there.
02:03:21 <pikhq> FREUD!
02:04:21 -!- alise_ has joined.
02:05:43 <oerjan> so basically it's lethal only by crushing the victim?
02:06:00 <pikhq> oerjan: Or via carbon monoxide.
02:06:09 <SgeoN1> Afaict, so far, 99.5% of the data is being saved
02:06:13 -!- SgeoN1 has quit (Quit: Bye).
02:06:24 <oerjan> ah yes, suffocation
02:06:36 -!- SgeoN1 has joined.
02:07:06 -!- alise has quit (Ping timeout: 240 seconds).
02:07:47 -!- SgeoN1 has quit (Client Quit).
02:08:06 -!- SgeoN1 has joined.
02:11:09 <alise_> goodnight
02:11:09 <alise_> bye
02:11:10 -!- alise_ has quit (Quit: Leaving).
02:11:29 <SgeoN1> Awesome. Ubuntu thinks it's Puppy Linux
02:11:29 <SgeoN1> With Flash crashing every two seconds
02:11:29 <SgeoN1> Night alise
02:11:57 -!- iamcal has quit (Ping timeout: 272 seconds).
02:12:23 -!- augur has joined.
02:13:00 <SgeoN1> 4.3 days
02:13:28 -!- quintopia has quit (Ping timeout: 276 seconds).
02:13:34 -!- quintopia has joined.
02:16:07 -!- SgeoN2 has joined.
02:16:08 -!- SgeoN1 has quit (Read error: Connection reset by peer).
02:16:15 -!- cal153 has joined.
02:25:06 -!- SgeoN2 has quit (Ping timeout: 245 seconds).
02:27:46 -!- SgeoN1 has joined.
02:29:01 <SgeoN1> How long a wait is too long?
02:29:27 <zzo38> 42 years
02:29:42 <SgeoN1> It occurs to me that I have an ancient computer that I could devote to the task
02:29:46 <SgeoN1> Two, actually
02:31:19 <zzo38> You forgot the rule of two and a half
02:31:27 -!- augur has quit (Ping timeout: 252 seconds).
02:37:12 <zzo38> And also the rule of five!!
02:46:02 -!- tswett has joined.
02:46:10 <tswett> Who invented kilgame? Was it lament?
02:46:36 * oerjan thought he hadn't seen warrigal for a while
02:46:44 <tswett> Hi, oerjan.
02:46:50 <tswett> No, I don't think you've seen me for a while.
02:46:51 <oerjan> also, what's kilgame
02:47:28 <tswett> kilgame was this IRC game where you had a "kill" command so that you could shoot people, and you can only shoot once every 60 seconds, and once you're shot by two different people you're out of the game.
02:48:08 * oerjan doesn't recall ever seeing it
02:49:22 <oerjan> clearly you are an evil tswett from an alternate dimension
02:49:34 <tswett> No, definitely not!
02:49:48 <oerjan> you _would_ say that
02:50:01 <tswett> Would I do such a thing?
02:50:10 <tswett> I can't believe you'd accuse me of this.
02:50:49 <oerjan> well since kilgame doesn't exist in this universe (although there are other games by the same name), that is the only reasonable explanation.
02:51:40 <oerjan> in fact ""kilgame" irc" gives _no_ google hits
02:51:56 <oerjan> (nor does ""kilgame" site:tunes.org")
02:52:47 <tswett> "<ihope> oklopol: I see you've managed to get #kilbot."
02:53:04 <tswett> That was... whatever day 07.07.06 is.
02:53:22 <SgeoN1> WB warrihal
02:53:24 <SgeoN1> Warrigal
02:53:30 <tswett> Hi Sheo.
02:53:33 <tswett> >.>
02:53:37 <SgeoN1> Assuming it is you...
02:53:43 <tswett> Yep, it's me.
02:54:04 <tswett> I'm connected from arch06.cis.gvsu.edu; who else could it be?
02:54:06 * SgeoN1 is close to falling asleep
02:54:18 <SgeoN1> A hallucinogen
02:54:39 <SgeoN1> Brought about by lack of restfulbsleep
02:54:43 <tswett> Being close to falling asleep is a fun hallucinogen.
02:55:09 * tswett attempts to determine oklopol's real name.
02:55:13 <tswett> If Agora doesn't know, nobody knows.
02:55:17 <SgeoN1> I'm in a pizza place
02:55:25 <SgeoN1> And I want to just take a nap
02:56:24 <oerjan> tswett: ok that does help your case somewhat
02:56:39 * oerjan still has onokki somewhere
02:56:57 <oerjan> of course i haven't really deleted anything since then, so not surprising
03:01:01 <oerjan> tswett: see msg (he doesn't really want it in the open iirc)
03:05:27 <SgeoN1> It is currently projected to take 7.5 weeks.....
03:06:06 -!- SgeoN2 has joined.
03:06:11 <oerjan> SgeoN1: you should start getting _really_ worried when it says INSUFFICIENT DATA FOR MEANINGFUL ANSWER
03:06:18 <SgeoN2> Erm, 7 weeks and half a day
03:06:36 <SgeoN2> Lol
03:07:15 -!- augur has joined.
03:07:19 <SgeoN2> The guiding motive for the Singularity AI is recovering my data
03:09:41 -!- SgeoN1 has quit (Ping timeout: 245 seconds).
03:12:09 -!- augur has quit (Ping timeout: 252 seconds).
03:13:05 <SgeoN2> Is the speed at all a function of other stuff that I'm doing with the computer?
03:13:58 <oerjan> no, it just senses your fear of it stalling
03:14:58 <SgeoN2> I'm more afraid of it taking years
03:15:14 <oerjan> THEN THAT'S WHAT IT WILL DO
03:17:11 <SgeoN2> flash keeps crashing
03:17:21 <SgeoN2> I wonder if it's an old version
03:17:48 <SgeoN2> I.e. whatever Puppy Linux ships with
03:19:14 <SgeoN2> It's fun to crash in Sgeo's face! Come on just do segfault; crash in Sgeo's face
03:19:32 <SgeoN2> Ok, it's the newest version
03:22:39 <SgeoN2> Now that I closed a bunch of stuff, it's speedinbhp
03:23:18 <SgeoN2> Tremendously
03:24:42 <SgeoN2> I can just Ctrl-C this, right?
03:24:43 -!- SgeoN2 has quit (Quit: Bye).
03:25:17 -!- SgeoN1 has joined.
03:25:30 <SgeoN1> I appear to be unable to end it
03:25:43 <SgeoN1> There we go
03:26:29 <SgeoN1> ddrescue --no-split /dev/sda imagefile logfile
03:27:40 <SgeoN1> Oh look, everything works now
03:28:04 <SgeoN1> Ill reissue the command before I go to sleep
03:28:16 <SgeoN1> Is there a way to be certain that the HD is spun down?
03:28:39 <SgeoN1> Any easy way to check?
03:29:07 <calamari-> can smartmon tell you?
03:30:00 <SgeoN1> Command not found
03:30:54 <calamari-> oh, I was expecting you to look into it a bit more than issuing a command lol
03:31:11 <calamari-> I think the package is smartmontools
03:31:31 <calamari-> checking
03:33:28 <calamari-> oh, smartmontools is only for ata
03:33:34 <calamari-> not for sat
03:33:37 <calamari-> sata
03:45:09 <calamari-> SgeoN1: sudo hdparm -C /dev/sda
03:46:18 <SgeoN1> active/idle
03:46:38 <SgeoN1> I take it that that is not, in fact, spun down?
03:46:51 <Quadlex> Hey eso
03:47:03 <calamari-> yeah probably not
03:47:22 <SgeoN1> Um. I think ill take the HD out then just set up my old comp
03:47:25 <calamari-> you can do a -B 1 to have aggressive power saving
03:47:59 <calamari-> Check the current IDE power mode status, which will always be
03:47:59 <calamari-> one of unknown (drive does not support this command),
03:47:59 <calamari-> active/idle (normal operation), standby (low power mode, drive
03:47:59 <calamari-> has spun down), or sleeping (lowest power mode, drive is com‐
03:47:59 <calamari-> pletely shut down). The -S, -y, -Y, and -Z flags can be used to
03:48:00 <calamari-> manipulate the IDE power modes.
03:48:09 <SgeoN1> The reason I want it spun down is to avoid damaging it further
03:48:34 <calamari-> then you might as well unplug it
03:50:15 * calamari- tries -Y just for the hell of it
03:51:31 <calamari-> well it worked, but since the machine is probably writing tothe drive all the time, it just powered back up immediately.. cool tho
03:52:42 <SgeoN1> This thing will boot from USB only when it wants to
03:56:01 <SgeoN1> I can't figure out why it wont boot
03:56:22 <SgeoN1> Is there a way to have the Ubuntu LiveCD check the USB for data files
04:05:32 -!- augur has joined.
04:10:25 -!- augur has quit (Ping timeout: 276 seconds).
04:14:47 -!- augur has joined.
04:19:54 -!- augur has quit (Ping timeout: 240 seconds).
04:23:42 -!- augur has joined.
04:28:28 <zzo38> I don't like the \outer command
04:29:16 -!- augur has quit (Ping timeout: 276 seconds).
04:29:31 <trinithis> http://www.thathigh.com/story/2776983/
04:32:02 <zzo38> Something I found out about TeX, is that the \csname command will cause the named control sequence to be equivalent to \relax is that control sequence is currently undefined.
04:36:52 -!- augur has joined.
04:38:38 -!- mycroftiv has quit (Quit: leaving).
04:39:59 <SgeoN1> What the fuck?
04:40:11 <SgeoN1> How is it almost Monday already?
04:40:27 <trinithis> Very carefully.
04:40:37 <oerjan> what do you mean "almost", you bloody american
04:41:16 <trinithis> 3:20 hours left
04:41:19 <trinithis> till mon
04:41:40 <trinithis> ;)
04:41:40 -!- augur has quit (Ping timeout: 276 seconds).
04:42:01 * trinithis wonders if augur can forsee his own quitting
04:42:07 <SgeoN1> The fricking fuck. It felt like Friday or Saturday
04:43:05 <SgeoN1> I had hoped to take a shower on the wekend, (no hot water in this house)
04:43:32 <SgeoN1> I don't know I'd I even have clean clothing
04:43:35 <SgeoN1> Bibble
04:45:04 <oerjan> Sio
04:51:21 -!- GreaseMonkey has joined.
05:02:22 -!- augur has joined.
05:03:05 -!- augur has quit (Client Quit).
05:08:30 -!- bsmntbombdood_ has joined.
05:08:40 <SgeoN1> How do I see what processes are using a filesystem?
05:10:52 <SgeoN1> N/m
05:11:31 -!- bsmntbombdood has quit (Ping timeout: 276 seconds).
05:14:16 <quintopia> wuzzup?
05:16:16 <quintopia> oh, Sgeo is discovering the magic of lsof
05:16:30 <quintopia> you can also use it to find who's listening on which port you know
05:23:23 <zzo38> I have idea, of a new programming language with similar use as C, but different syntax, and that takes things from: TeX, C, Forth, WEB, and assembler.
05:23:50 <quintopia> explain plox
05:24:27 <zzo38> From C: Data types, structures, preprocessor macros, preprocessor include, unions, pointers, function pointers, storage classes, arrays.
05:24:51 <zzo38> From TeX: Category codes (and the ability to change them during compile time), and penalties (which you can insert anywhere in the program).
05:25:44 <zzo38> From Forth: reverse polish notation, immediate words, ability to create new control structures like Forth, ability to create parsers for reading a few extra words after a command.
05:26:33 <SgeoN1> Does casper-rw not work anymore?
05:26:35 <zzo38> O, and also explicit command to tell it you want to retrieve the value of a variable, by writing @ sign after the address (or name of variable, or something else).
05:28:23 <zzo38> From assembly language: Ability to enter direct CPU instructions/registers/etc, and things close to it but that might be more cross-platform.
05:28:44 -!- oerjan has quit (Quit: leaving).
05:30:46 <quintopia> is the goal to produce something really usable?
05:32:46 <zzo38> quintopia: Yes it would be usable.
05:33:05 <zzo38> And it would also have commands to have a lot of control over the optimizer and other things if needed.
05:34:07 <quintopia> could be interesting
05:47:23 <zzo38> When programming in TeX, I often find it useful to change category codes in the middle of the file, sometimes conditionally.
05:47:48 <quintopia> i, uh, just use the conference-provided templates >.>
05:48:12 <zzo38> quintopia: What conference-provided templates?
05:48:48 <quintopia> depends on the conference/journal
05:48:58 <quintopia> NIPS has a pretty nice one
05:49:10 <quintopia> APA use the same one for all their journals i think
05:49:12 <quintopia> AMS too
05:49:51 <zzo38> OK. I don't write TeX documents for conferences or journals, I write them for myself, so I just use Plain TeX, and then add whatever macros and other things are needed for the file I am writing.
05:50:21 <zzo38> I don't like to use LaTeX, because I prefer Plain TeX.
05:51:22 <zzo38> I sometimes find it very useful to change category codes temporarily for use with the \read and \write commands, and sometimes have a file include itself.
05:52:02 -!- mycroftiv has joined.
05:52:22 <quintopia> what's the minimum one would need to do to make plain TeX turing-complete?
05:52:51 <mycroftiv> add a brainfuck eval() via an escape sequence
05:53:04 <mycroftiv> maybe thats not the type of answer you were looking for though
05:53:35 <quintopia> that seems like far too much to add
05:53:47 <quintopia> since the thing you are adding is turing-complete in itself
05:53:58 <mycroftiv> its a deliberately degenerate solution
05:54:23 <zzo38> quintopia: TeX is turing complete. You just have to know how to program it!
05:54:28 <quintopia> sweet!
05:54:32 <quintopia> how?
05:54:53 <mycroftiv> postscript is turing complete also, isnt it?
05:57:20 <zzo38> I could write a brainfuck interpreter in TeX, or other things, if I wanted to. You can also see the yesweb codes for an example of a full program in TeX.
05:58:00 <zzo38> http://sprunge.us/BYYJ
05:58:44 <quintopia> what is yesweb supposed to do?
05:58:51 <zzo38> This program shows a lot of techniques of programming in TeX.
05:59:50 <zzo38> Here is an example file using yesweb: http://sprunge.us/aADQ
06:00:19 <zzo38> This example file generates two things: a printed document, and a C source code file.
06:00:48 <pikhq> mycroftiv: Yes, Postscript is
06:00:51 <pikhq> TC
06:01:04 <mycroftiv> i recall a cute story from a friend who had a job doing university IT
06:01:27 <quintopia> is yesweb some kind of macro system built on top of TeX?
06:01:32 <mycroftiv> some clever comp sci student decided to print out tickets to some event that had a lot of algorithmically generated using postscript
06:01:44 <zzo38> Actually this example file generates a printed document and two C source files, "hello.c" and "goodbye.c".
06:01:56 <zzo38> quintopia: Yes, yesweb is a kind of macro system built on top of TeX.
06:02:05 <mycroftiv> the IT lab printer seemed to be 'frozen' for hours and then suddenly came to life printing out tickets where each ticket had a huge amount of unique data procedurally generated
06:02:08 <zzo38> It is implemented entirely in TeX.
06:04:42 <zzo38> It does a few strange things, such as temporarily making the lowercase "n" into a comment character (like "%" is normally).
06:05:01 <quintopia> weird
06:06:07 <zzo38> The stuff between \z and @z is verbatim text, everything there treats all characters as normal characters.
06:07:09 <zzo38> The exception is the \webescape character, which must be doubled if you want to include that character verbatim. (Normally an at sign, but it can be changed using the \webescape command.)
06:07:55 -!- bsmntbombdood_ has quit (Ping timeout: 265 seconds).
06:08:34 <quintopia> how does looping work?
06:09:28 <zzo38> The stuff in between \z and @z will be printed in monospace text with a border around it, and will also be copied to the output files (which in this case have the ".c" extension, but they don't have to be C codes, they can be any text files).
06:09:31 -!- lament has joined.
06:09:38 <zzo38> Looping works in TeX using the \loop ... \repeat commands.
06:10:17 <zzo38> You can have like \loop\ABC\iftrue\XYZ\repeat so you can have codes both before and after the condition.
06:11:39 <zzo38> That is how you can do loops in TeX.
06:12:19 -!- sftp has quit (Remote host closed the connection).
06:16:23 <zzo38> You might like to know that loops are not built-in to primitive TeX. Plain TeX is a macro package implemented in primitive TeX, and Plain TeX defines the commands for loops.
06:20:01 <quintopia> but Plain TeX is written entirely in TeX?
06:22:52 <pikhq> Yeah.
06:23:04 <pikhq> It's the macros that TeX ships with to make it at least *somewhat* usable.
06:23:18 -!- bsmntbombdood_ has joined.
06:27:18 <quintopia> http://stashbox.org/729975/nooooo.jpg
06:27:38 <mycroftiv> oooooblubblub...blub
06:27:47 <SgeoN1> Primitive Tex doesn't have loops but loops can be made in it? Is this similar to if else then not being builtin to Forth?
06:29:32 <pikhq> SgeoN1: Very much so.
06:44:25 <zzo38> SgeoN1: Yes, like that.
06:44:51 <zzo38> (Some Forth systems do have IF ELSE THEN built-in, others do not. In the ones that do not, you can implement them.)
06:52:34 -!- augur has joined.
06:58:03 <Quadrescence> http://www.youtube.com/watch?v=hMtZfW2z9dw
07:00:02 -!- tombom has joined.
07:02:30 <zzo38> yesweb contains various strange things, but it is how TeX works. It changes category codes a lot, changes output routines, has macros that redefine themself and other macros, macros that call themself, macros that define macros that define macros that define other macros, the \expandafter command and \csname commands are used a bunch of times.....
07:02:51 <zzo38> Actually, I find the \expandafter and \csname commands useful when using TeX.
07:03:21 <zzo38> How often do *you* use those commands?
07:06:48 <Vorpal> <Sgeo|Empathy> Vorpal: you awake? <-- morning
07:06:52 <Vorpal> SgeoN1, ^
07:07:42 <Vorpal> zzo38, never I think
07:07:52 <Vorpal> zzo38, never used them in LaTeX
07:08:04 <Vorpal> (and I don't use plain text directly)
07:08:08 <Vorpal> plain tex*
07:09:43 <zzo38> Vorpal: OK. But I must say I find Plain TeX *much* better than LaTeX. I also find LaTeX bloated and stuff.
07:10:46 <SgeoN1> Meh, sleep now
07:11:09 <SgeoN1> I think I wanted to ask something about the HD issue
07:11:35 <SgeoN1> But right now, it's again out of the comp, will be using old computer to do the stuff
07:11:41 <SgeoN1> Ninight
07:11:43 -!- SgeoN1 has quit (Quit: Bye).
07:15:17 <zzo38> OK
07:15:19 -!- zzo38 has quit (Quit: zzo38).
07:34:40 -!- FireFly has joined.
07:48:09 <wareya> I'm a free market communist
07:49:46 -!- Sgeo has joined.
07:50:05 -!- Sgeo has quit (Client Quit).
07:50:34 -!- Sgeo has joined.
07:50:39 -!- tombom has quit (Quit: Leaving).
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:01:37 -!- cheater99 has quit (Ping timeout: 264 seconds).
08:01:56 -!- Sgeo has quit (Quit: Ex-Chat).
08:02:34 -!- FireFly has quit (Quit: swatted to death).
08:07:01 -!- bsmntbombdood_ has quit (Ping timeout: 276 seconds).
08:17:26 -!- cheater99 has joined.
08:23:26 -!- calamari- has quit (Quit: Leaving).
08:24:16 -!- nooga has quit (Ping timeout: 245 seconds).
08:32:47 -!- cheater99 has quit (Quit: Leaving).
08:51:46 -!- Phantom_Hoover has joined.
09:07:58 -!- augur has quit (Remote host closed the connection).
09:08:09 -!- augur has joined.
09:11:05 <Phantom_Hoover> 16:47:01 <alise> which is why the Singularity is defined at precisely the point where we have no fucking idea what will happen ← true, but you presuppose many things about it which aren't as certain as made out.
09:15:27 -!- lament has quit (Ping timeout: 265 seconds).
09:20:17 -!- jcp has quit (Ping timeout: 265 seconds).
09:24:38 -!- jcp has joined.
10:02:40 <fizzie> Whoa, why hadn't I seen the awesomeity of http://www.visual6502.org/JSSim/index.html before? (Caution: is probably not the fastest thing ever if you actually want to run something.)
10:06:21 -!- Phantom_Hoover has quit (Ping timeout: 245 seconds).
10:22:11 -!- Phantom_Hoover has joined.
11:35:06 -!- Phantom_Hoover has quit (Ping timeout: 245 seconds).
11:36:46 -!- Phantom_Hoover has joined.
11:40:01 -!- atrapado has joined.
11:43:57 -!- GreaseMonkey has quit (Remote host closed the connection).
11:55:34 <fizzie> Heh, here are typical sentences from alise and Vorpal:
11:55:36 <fizzie> >>> str(h_a.sampleSingle(50))
11:55:36 <fizzie> 'celt fstuooslk cikmwpe boaa,stblgcneinn, Ahu-onrts'
11:55:36 <fizzie> >>> str(h_v.sampleSingle(50))
11:55:36 <fizzie> "olsbioa1ce btkdus aeae boeee fh rie> w)ne 2n'l wec"
11:58:03 <Phantom_Hoover> To what end are you doing this?
12:02:31 <fizzie> Why, for fun, of course.
12:11:40 <Phantom_Hoover> To make alisebot?
12:12:50 <fizzie> No, for classificationary purposes.
12:13:00 <fizzie> Though it doesn't work really *that* well.
12:14:58 <Phantom_Hoover> Whom do you plan to classify?
12:15:02 <fizzie> Those same h_a/h_v models that generated that nonsense up there classified 61% of alise-comments as alise, and 71% of Vorpal-comments as Vorpal; that's not very much better than random chance. But at least it does something sensible for a single comment, unlike the book-authorship statistics, which need bazillion lines before they can make any sort of sensible decisions at all.
12:16:35 <fizzie> Could try with a lot larger models, there's certainly enough training data. Those were 14-state discrete-emission HMM's with this sort of constrained structure -- http://www.cis.hut.fi/htkallas/hmm.png -- and initial transition probs; and uniform-distribution emissions in each state before training.
12:16:58 <fizzie> I like the "Ahu-onrts" bit, though.
13:20:49 -!- jcp has quit (Ping timeout: 264 seconds).
13:22:11 -!- jcp has joined.
13:25:00 <Phantom_Hoover> Why is living in the future so boring??
13:25:25 <Phantom_Hoover> I mean, you would have thought someone would have made some kind of jetpack, but noooooo.
13:28:01 -!- jcp has quit (Ping timeout: 264 seconds).
13:31:41 -!- jcp has joined.
13:31:55 <fizzie> Phantom_Hoover: What do you mean they haven't made jetpacks? http://en.wikipedia.org/wiki/Jetpack -- "Jet pack T-73, flight time 9 minutes, max distance 11 miles, ~83 mph, $200,000 incl. training".
13:33:51 <fizzie> (Not that they actually sell one anywhere, but still.)
13:34:35 <Phantom_Hoover> Yay!
13:35:39 <Phantom_Hoover> Now, where's the moon base?
13:36:10 <fizzie> On the moon?
13:36:16 <fizzie> (What a silly question.)
13:39:49 -!- jcp has quit (Ping timeout: 276 seconds).
13:42:41 -!- jcp has joined.
13:43:36 <Phantom_Hoover> fizzie, where on the moon?
13:43:53 <fizzie> They don't tell that sort of info to us civilians.
13:51:47 -!- jcp has quit (Ping timeout: 272 seconds).
13:55:42 <Phantom_Hoover> fizzie, well, launching a mission to the moon is hardly low-key.
13:56:32 <fizzie> Ticket To The Moon.
13:58:11 -!- jcp has joined.
14:00:59 * Phantom_Hoover → walk
14:01:36 <fizzie> Walk To The Moon?
14:05:06 -!- Phantom_Hoover has quit (Ping timeout: 245 seconds).
14:07:27 -!- FireFly has joined.
14:16:01 -!- nooga has joined.
14:16:18 <nooga> gagag
14:44:52 -!- augur has quit (Ping timeout: 276 seconds).
15:14:58 -!- tswett has quit (Quit: Lost terminal).
15:16:51 -!- ais523 has joined.
15:17:01 -!- augur has joined.
15:25:34 -!- sftp has joined.
15:27:07 -!- Sgeo has joined.
15:31:05 <nooga> i just did my laundry, feels good
15:31:18 <nooga> that means i can program a washing machine!
15:34:22 <Sgeo> Aww
15:34:33 <Sgeo> XChat-GNOME doesn't automatically set me away when I go away from the thingy
15:35:45 -!- augur has quit (Ping timeout: 252 seconds).
15:36:28 -!- augur has joined.
15:44:48 <Sgeo> pikhq has jumped the sneeze
15:54:36 -!- webquint has joined.
15:55:26 -!- MigoMipo has joined.
15:57:06 -!- Sgeo has quit (Ping timeout: 240 seconds).
16:12:52 -!- Sgeo has joined.
16:13:09 <Sgeo> &fsckperl
16:13:30 -!- alise has joined.
16:14:30 -!- augur_ has joined.
16:14:36 -!- augur has quit (Read error: Connection reset by peer).
16:15:19 <pikhq> Well motherfucking hell. My car battery died.
16:15:39 <pikhq> I'll have to wait for parents to get home so I can actually get a new one.
16:15:52 <alise> My battery car died.
16:15:57 <alise> Now I can't drive my battery to places.
16:16:12 <pikhq> And I'm missing a test today. Here's hoping my prof. lets me do the test at a later date.
16:16:12 <alise> pikhq: Any luck with NixOS? I have the LiveCD working on USB.
16:16:16 <alise> So I can install it, theoretically.
16:16:22 <pikhq> alise: I've got it installed in a VM.
16:16:27 <pikhq> Not done anything with it, though.
16:16:29 <alise> pikhq: How boring.
16:16:35 <pikhq> The solution to all my problems was to use qemu.
16:16:56 <alise> pikhq: With qemu, it runs unbelievably slowly for me.
16:17:01 <alise> I guess you have virtualisation support.
16:17:28 <alise> 18:03:00 <pikhq> Wow. It is estimated the LSD of marijuana is 1500 *pounds*.
16:17:29 <Sgeo> This class is _far_ more convenient to do with Ubuntu than Windows
16:17:34 <alise> Most confusing phrase EVER.
16:17:43 <alise> Sgeo: Everything is etc.
16:17:44 <pikhq> alise: Yeah, I've got it using KVM.
16:17:46 -!- zzo38 has joined.
16:17:52 <alise> pikhq: Fuck you and your good hardware >:(
16:18:00 <pikhq> alise: You could use kqemu.
16:18:03 <alise> pikhq: Can I have your virtualisation module to plug into my processor?
16:18:06 <alise> pikhq: kqemu no longer exists.
16:18:12 <alise> "Because fuck you."
16:18:20 <Sgeo> Wait what?
16:18:20 <pikhq> Oh, Gentoo must be patching its qemu, then.
16:18:27 <Sgeo> What happened to kqemu?
16:18:31 <alise> Sgeo: Deprecated.
16:18:34 <pikhq> Sgeo: Apparently KVM.
16:18:37 <alise> Not on the site any more, not in the latest release.
16:18:39 <alise> Also, what pikhq said.
16:18:43 <alise> Which requires virtualisation support.
16:18:51 <alise> Because my life is worthlooooh, it can use Xen.
16:18:54 <Sgeo> Does Perl have first-class functions?
16:18:58 <alise> Can Xen do stuff without virtualisation support?
16:18:58 <alise> Sgeo: Yes.
16:19:01 <alise> &foo
16:19:09 <alise> Uh, wait.
16:19:10 <alise> No.
16:19:12 <alise> Um, yes.
16:19:13 <Sgeo> My professor didn't seem to understand what I was asking when I said "store subroutine in a variable"
16:19:14 <alise> But not like that.
16:19:15 <alise> Ask ais523.
16:19:24 <alise> Sgeo: well that's a stupid way of putting it :)
16:19:36 <alise> "Get a reference to a subroutine".
16:19:44 <ais523> Sgeo: it has first-class references to subroutines
16:19:58 <ais523> and you can store those in variables
16:20:17 <alise> ais523: what's the syntax?
16:20:19 <alise> not &foo
16:20:23 <ais523> and as subroutines are read-only (although you can create the things dynamically using lambdas), a reference to a subroutine is equivalent to the subroutine itself, except in terms of syntax
16:20:24 <alise> *foo?
16:20:26 <ais523> and \&foo
16:20:28 <alise> ahh
16:20:31 <ais523> for a reference to subroutine foo
16:20:37 <ais523> just like \$foo would be a reference to scalar foo
16:20:38 <alise> ais523: not \foo because that's sort of ambiguous, right?
16:20:45 <ais523> that's more meaningless than ambiguous
16:20:50 <alise> because foo() is just shorthand for &foo(), right?
16:20:58 <ais523> no, it passes arguments differently
16:21:14 <Sgeo> It's more than just whether you can do it before or after being defined?
16:21:17 * Sgeo headaches
16:21:19 <ais523> old-fashioned subroutine call: @_ = ("arg1, "arg2"); &foo;
16:21:27 <ais523> new subroutine cal: foo("arg1","arg2")
16:21:31 <ais523> *call
16:21:34 <Vorpal> <fizzie> Heh, here are typical sentences from alise and Vorpal:
16:21:34 <Vorpal> <fizzie> >>> str(h_a.sampleSingle(50))
16:21:34 <Vorpal> <fizzie> 'celt fstuooslk cikmwpe boaa,stblgcneinn, Ahu-onrts'
16:21:34 <Vorpal> <fizzie> >>> str(h_v.sampleSingle(50))
16:21:34 <Vorpal> <fizzie> "olsbioa1ce btkdus aeae boeee fh rie> w)ne 2n'l wec"
16:21:36 <Vorpal> err?
16:21:39 <Sgeo> I think she's showing us old-fashioned
16:21:40 <ais523> and @_ is managed for you, it even restores the original value
16:21:46 <Sgeo> Oh, wait
16:21:53 <alise> <ais523> old-fashioned subroutine call: @_ = ("arg1, "arg2"); &foo;
16:21:58 <alise> surely even old perl didn't make you do that
16:22:01 <Sgeo> No, just old-fashioned using it, not old-fashioned subroutine call
16:22:03 <ais523> alise: yes, it did
16:22:07 <alise> ais523: :-D
16:22:12 <ais523> in fact, using @_ may have just been convention
16:22:24 <alise> DB<1> @_ = ("Hello, world!\n"); &print;
16:22:25 <alise> Undefined subroutine &main::print called at (eval 6)[/usr/share/perl/5.10/perl5db.pl:638] line 2.
16:22:27 <alise> Bah.
16:22:29 <alise> How racist.
16:22:31 <ais523> print isn't a subroutine
16:22:34 <alise> I know.
16:22:35 <alise> How racist.
16:22:36 <ais523> it's a builtin function
16:22:38 -!- augur_ has changed nick to augur.
16:22:48 <pikhq> Of all the days, I had to miss class on the day of a test.
16:22:51 <Sgeo> Old-fashioned use of @_ within the body
16:22:54 <alise> pikhq: would NixOS paravirtualise properly?
16:22:59 <alise> pikhq: also, no public transport?
16:23:00 <ais523> even more fun, you had to scope @_ yourself
16:23:03 <alise> Sgeo: err
16:23:06 <alise> that's not old-fashioned
16:23:07 <pikhq> alise: It's the US.
16:23:11 <alise> that's how you do arguments in perl
16:23:14 <ais523> although it had a rather useful "local" keyword, which means "change the value now, restore it at the end of the block"
16:23:19 <alise> pikhq: lol @ your country sucks
16:23:24 <ais523> "local" still exists, but "my" is used more often, as it does actual scoping
16:23:28 <ais523> rather than just INTERCAL-style stashing
16:23:36 <pikhq> And yes, NixOS ought to paravirtualise properly. It's only the kernel that's involved in that.
16:23:44 <alise> ais523: I think Sgeo may need some therapy to realise that you manually shift @_ to do arguments in regular Perl.
16:23:54 <alise> pikhq: How hellish is setting up Xen?
16:23:59 <pikhq> Not very.
16:24:09 <ais523> alise: well, it's very flexible
16:24:11 <zzo38> It isn't only INTERCAL that stashes/retrieves variables in that way, dc also has a command to do something similar.
16:24:15 <alise> Sounds hellish to me.
16:24:16 <pikhq> Perhaps more-so on not-Gentoo, though.
16:24:27 <ais523> but if you stick to conventions, it's just the same as normal argument passing
16:24:32 <pikhq> I'd *imagine* the only painful bit would be setting up bridging.
16:24:35 <alise> pikhq: Nix's single-configuration-mechanism thing is /so cool/.
16:24:46 <alise> <pikhq> I'd *imagine* the only painful bit would be setting up bridging. ;; that's the thing, QEMU can use Xen
16:24:52 <ais523> sub foo {my $a = shift; my $b = shift;}
16:24:53 <alise> Somehow!
16:25:04 <ais523> tailcalls are done by setting up @_ by hand, then doing goto &foo;
16:25:11 <alise> ais523: heh
16:25:13 <alise> but most people write
16:25:14 <pikhq> alise: That's only for the virtualisation extensions.
16:25:16 <alise> my ($a, $b) = @_;
16:25:18 <alise> pikhq: osidjfoij
16:25:28 <ais523> alise: really? I haven't seen that very often
16:25:28 <alise> pikhq: Got a kqemu source package?
16:25:33 <alise> ais523: o_O
16:25:43 <alise> ais523: I have never seen a single person manually shift every argument like you've written.
16:25:51 <alise> I have universally seen my ($args, $here) = @_;
16:25:56 <ais523> at least, =shift is standard in TAEB, and it has some of the sanest Perl code in existence
16:26:02 <alise> Random example: http://www.cs.cf.ac.uk/Dave/PERL/node61.html
16:26:16 <alise> Random example: http://automatthias.wordpress.com/2007/01/19/perl-argument-passing-weirdness/
16:26:18 <ais523> occasionally it refers to elements of @_ directly, when it needs extreme performence
16:26:22 <pikhq> alise: Hmm. Use qemu pre-0.9.1 and http://wiki.qemu.org/download/kqemu-1.3.0pre11.tar.gz .
16:26:28 <alise> Random example: http://stuff.mit.edu/iap/perl/slides/sub_arguments.html (just one parameter, but still)
16:26:28 <ais523> *performance
16:26:51 <alise> ais523: if it even has functions it's not Ultra Clean Insane Perl ;-)
16:26:55 <ais523> (when it needs even /more/ extreme performance, it refers to the internals of objects in an encapsulation-breaking manner, but only once and with huge warnings in a comment next to it)
16:27:03 <alise> You need MooseX::Declare for that, which has actual method signatures with names.
16:27:14 <alise> pikhq: Or just install it on real hardware...
16:27:50 <alise> "Alex Smith (ais523)" --TAEB blog. I thought you were still paranoid about that.
16:28:17 <alise> It upsets me that TAEB plays Nethack better than I do.
16:28:29 <alise> At least with chess I can claim it's a genuinely very difficult game.
16:29:01 <alise> ais523: How far has Planar got?
16:29:10 <alise> When you say it gets into Sokoban, I presume it goes through the Mines first?
16:29:19 <alise> If not, HA HA I AM BETTER THAN TAEB
16:29:22 <ais523> alise: no, sorear and I were working on TAEB in parallel
16:29:32 <alise> ais523: err?
16:29:36 <ais523> he was working on a mines-completing version of the AI, I was working on a Sokoban-completing version
16:29:36 <alise> I meant the name/nick association
16:29:38 <ais523> so it never does both
16:29:40 <alise> oh
16:29:45 <ais523> alise: I'm guessing people have figured it by now
16:29:45 <alise> ais523: was
16:29:46 <alise> ?
16:30:01 <ais523> alise: well, neither of us have worked on it for a while
16:30:07 <ais523> but I'm not sure how deep sortaeb523 ever got
16:30:09 <alise> dormant?
16:30:25 <ais523> dormant, not abandoned
16:30:32 <ais523> dlvl 11, it sems
16:30:34 <ais523> (thanks Rodney!)
16:31:22 <alise> "Incidentally, Planar always asks for more information, if it’s available and doesn’t cost a turn, in order to analyse the situation as well as possible; for instance, if the material the golem was made of (relevant because it determines how dangerous it is) were undeducible from its colour on-screen, it would ask the framework to request NetHack to give more details, in this case by sending a farlook command to examine it remotely in detail."
16:31:23 <alise> is this automatic?
16:31:33 <ais523> in Planar, yes; in TAEB, no
16:31:39 <alise> i'd hide it all behind a VeryCleverNethackScreen interface
16:31:49 <alise> where all information is available, just sometimes it has to request (and cache it)
16:31:58 <alise> so you can pretend you have 4D vision, or something, and the depth tells you the farlook info
16:32:00 <alise> erm
16:32:01 <alise> 3D
16:32:29 <ais523> alise: I seem to remember that there was a huge flamewar on the subject a while ago
16:32:34 <alise> o_O
16:32:35 <alise> Why?
16:32:47 <ais523> part of the reason is that some TAEB AIs are content with guesses
16:32:48 <alise> pikhq: Incidentally, Alt-F9 on NixOS has an inexplicable Rogue game always started.
16:32:56 <alise> Why? Who knows, maybe they just like Rogue.
16:33:05 <alise> ais523: then they don't need to request the information
16:33:10 <alise> also, that's stupid
16:33:12 <alise> farlook isn't exactly cheating
16:33:30 <alise> "taking the worst-case scenario; Planar is just as scared of the Random Number Generator as any human would be"
16:33:32 <ais523> alise: it's not to do with cheating, it's to do with saving time
16:33:34 <alise> perhaps it should take risks upon occasion?
16:33:43 <alise> ais523: farlook doesn't take a turn or more than a few ms of realtime...
16:33:44 <Sgeo> Someone dislikes the "\$betty" that the professor is printing
16:33:52 <ais523> alise: TAEB's designed for online play
16:33:52 <Sgeo> And asks "Why not just print betty"?
16:33:59 <ais523> it takes a few hundred ms to farlook something
16:34:03 <alise> Sgeo: fail
16:34:05 <ais523> over telnet
16:34:13 <alise> ais523: meh
16:34:24 <alise> ais523: a proper nethack server would send down all such information with the frame :)
16:34:29 <alise> and have the local client display it upon request
16:34:53 <alise> "nHackbot owned the (known) high mark for score and dungeon depth for quite some time, achieving a score 18,432 and maximum depth of 10."
16:34:58 <alise> err, presumably as far as bots go, not players
16:35:02 <alise> :P
16:35:19 <ais523> yes
16:35:19 <alise> "Intrigued by nHackBot and itching to do something similar, Shawn Moore started nhbot in December 2005, using Perl. While nHackBot sought to eventually be a complete bot, nhbot had no such aspirations. It solved the screen parsing problem by not parsing the screen at all, opting instead to read just the top line. Rather than calculating where to move next, it walked around randomly. Eventually, it did gain the ability to read the bottom line in order to kno
16:35:19 <alise> w when to engrave Elbereth, but that was the extent of nhbot's sight. nhbot did remarkably well for being so near-sighted, posting a high score of 15,185. However, it did not descend very far, making it only to dungeon level 3."
16:35:23 <alise> Top line as in the first one of the screen???
16:35:25 <alise> That's usually blank
16:35:35 <ais523> it contains messages
16:35:39 <ais523> like "the jackal hits!"
16:35:41 <alise> haha
16:35:44 <alise> what a retarded bot
16:36:01 <alise> "GreyKnight" -- /that/ GreyKnight?
16:36:18 <alise> wait, i don't know a greyknight
16:36:20 <alise> thought i did
16:36:46 <Sgeo> I think I need a translator
16:37:06 <Sgeo> She fails to understand what I'm saying. She thinks I had no clue what was going on, when I just wanted to point something out
16:37:25 <Sgeo> alise, heraldic ensignia of Agora?
16:37:38 <alise> Sgeo: not the same guy, he had a wikipedia page
16:37:47 <alise> but if it is him
16:37:50 <alise> then he used to come here too
16:38:10 <alise> wait no
16:38:11 <alise> it is him
16:38:26 <alise> (he'd changed his user page; I found the copy I remember in the history.)
16:40:37 <Sgeo> How do you use a reference?
16:40:51 <alise> Cleverly.
16:41:08 <Sgeo> Is that a joke, or are you trying to give me a nightmare, or...
16:41:10 <alise> (I hereby defer horrible Perl questions to ais523, the poor thing. Actually, scratch that; it'll just scare him off.)
16:43:41 <Vorpal> alise, nice that your typical line is "celt fstuooslk cikmwpe boaa,stblgcneinn, Ahu-onrts'". A bit of fine tuning and we could replace you with a bot!
16:44:12 <alise> Vorpal: Ahu-onrts'! CELT FSTUOOSLK!!!
16:44:20 <Vorpal> alise, XD
16:44:51 <alise> (Translation: "Don't you EVER say that about my mother again, you [UNTRANSLATABLE] [UNTRANSLATABLE]!!!"
16:44:56 <Vorpal> alise, I have but one reply to that: lsbioa1ce btkdus aeae boeee fh rie> w)ne 2n'l wec
16:45:20 <alise> *ææ *bœee
16:45:27 <alise> Get your orthography right, you cikmwpe boaa.
16:45:32 <Vorpal> alise, no that isn't thypical of me
16:45:35 * Sgeo headaches at array coerced into scalar vs $#somearray
16:45:38 <ais523> Sgeo: extra sigil at the start
16:45:47 <Vorpal> I almost never use æ except when discussing the char itself
16:45:49 <Sgeo> It would be sensible if they were the same. They're not
16:45:52 <ais523> as in, $reference is a scalar holding a reference, $$scalar is the scalar it points to
16:45:54 <alise> Vorpal: You're always so bloody thypical, you stblgcneinn.
16:46:00 <ais523> likewise, %$scalar is the hash it points to, etc
16:46:07 <Sgeo> ais523, ah
16:46:08 <ais523> *$$reference, %$reference
16:46:29 <ais523> if $reference isn't a reference, or the wrong sort of reference, it causes a run-time warning
16:46:33 <ais523> Perl isn't big on the idea of run-time errors
16:46:45 <ais523> which is why you should always turn warnings on (-w command-line option, or "use warnings;" in the program)
16:46:49 <Vorpal> alise, hrrm. You might have spoken enough lines in here that fizzie could make a fungot language model for you
16:46:49 <fungot> Vorpal: the masamune! ride again!
16:46:58 <Vorpal> alise, not sure if I have spoken enough as well
16:47:02 <alise> Vorpal: I've spoken more than enough.
16:47:04 <Vorpal> fungot, oh? the sword?
16:47:04 <fungot> Vorpal: shall we get back to the present? he's been known. we reptites will rule the world in a mere door that keeps us bound, hand, foot...and tongue kid? ...oh, it's you, isn't this morbid? the great adventurer toma levine rests in a grave to the north. it's a great place for a picnic! heard that magus's place...
16:47:20 <Sgeo> Perl's random quirks are maddening
16:47:20 <alise> The YouTube corpus, for instance, is quite small. Because asiekierka wasn't smart enough to do anything more than manually copy and paste YouTube comments out.
16:47:21 <Vorpal> alise, right. It would be interesting :)
16:47:26 <alise> Sgeo: are English's?
16:47:32 <alise> Vorpal: It's already been done.
16:47:35 <alise> Grep logs for "virtuehird".
16:47:41 <Vorpal> alise, ah, about when?
16:47:47 <alise> About then.
16:47:47 <Vorpal> alise, I'm on laptop atm
16:47:48 <Sgeo> I've been speaking English as a little kid. I'm used to it
16:47:53 <alise> When I was still ehird, at least.
16:47:53 <Vorpal> alise, I don't have full logs locally
16:47:56 <Sgeo> It's a bit late to learn Perl as a little kid
16:47:59 <alise> Vorpal: DB.
16:48:09 <alise> select * from logs where type=0 and nick="virtuehird";
16:48:10 <Vorpal> alise, yeah but ssh has a latency > 2 seconds
16:48:14 <alise> Boo hoo hoo.
16:48:15 <ais523> Perl's quirks are mostly intentionally the same sorts of quirks you get in natural languages
16:48:16 <Vorpal> accessing it over ssh would be painful
16:48:22 <alise> Not many lines.
16:48:29 <Vorpal> alise, it works fine for irc but interactive stuff? no.
16:48:34 <ais523> Vorpal: ssh has a much lower latency than that
16:48:38 <ais523> people play roguelikes over it
16:48:44 <Vorpal> ais523, not when you are on EDGE
16:48:46 <alise> <Vorpal> alise, it works fine for irc but interactive stuff? no.
16:48:49 <alise> just quoted for posterity
16:48:59 <Vorpal> ais523, bluetooth to phone, which has only EDGE atm, not 3G
16:49:00 <alise> actually
16:49:00 <alise> `addquote <Vorpal> alise, it works fine for irc but interactive stuff? no.
16:49:10 <Vorpal> alise, irc because I run the client locally
16:49:12 <HackEgo> 225|<Vorpal> alise, it works fine for irc but interactive stuff? no.
16:49:17 <alise> Still funny.
16:49:29 <Vorpal> maybe
16:49:38 <alise> There are much less funny things in the QDB.
16:49:38 <Vorpal> alise, so you admitted I was funny
16:49:39 <Vorpal> :D
16:49:54 <alise> In the same way that things Bush says are funny.
16:49:59 <Vorpal> nah
16:50:04 <Vorpal> I prefer my original interpretation
16:50:47 <trinithis> (Vorpal should do "/me Sword")
16:50:56 <alise> */me sword
16:50:59 <alise> and no, he shouldn't
16:51:02 <alise> that wouldn't be amusing
16:51:06 <Vorpal> ais523, anyway yes ssh over local ethernet has less lag. So does ssh from university. Just not ssh over mobile tethered by bluetooth on GSM/EDGE.
16:51:15 -!- alise has changed nick to trinlthis.
16:51:18 <Vorpal> trinithis, fairly boring joke
16:51:18 * trinlthis hur hur
16:51:20 -!- trinlthis has changed nick to alise.
16:51:29 <trinithis> :D
16:51:32 <trinithis> I have a clone!
16:51:39 <Vorpal> had*
16:51:39 <alise> Stupidity:
16:51:41 <alise> ehird@dinky:~/Documents/#esoteric/logs$ grep --colour=force -ri 'virtuehird' | less
16:51:52 <ais523> trinithis: new? or a regular who's changed nick yet again?
16:51:57 <alise> new.
16:52:11 <Vorpal> ais523, besides the signal indicator on the phone is at 2 out of 4....
16:52:14 <ais523> nice to see new people
16:52:19 <trinithis> Hahah. As they say, imitation is the best form of flattery
16:52:21 <ais523> Vorpal: signal indicators are pretty meaningless...
16:52:40 <Vorpal> ais523, oh? just for phones or in general?
16:53:06 <ais523> phones in general
16:53:29 <Vorpal> ais523, yeah, it is pretty reliable for wlan (well I don't know if it is for wlan on phones, since my phone can't do wlan)
16:53:47 <ais523> Apple's original response to the iPhone signal strength issues was to release an OS patch to increase the signal strength meter
16:54:20 <Vorpal> hm
16:54:31 <Vorpal> ais523, so why can't they make the meter more reliable simply?
16:55:00 <ais523> Vorpal: phones normally know the exact signal strength in dBm, but don't display it to the user for fear of confusing them
16:55:16 <ais523> "why is my signal strength a negative number? why is it so much lower than my friend's, while I get better reception?"
16:55:36 <Sgeo> Does Perl have dynamically-scoped variables?
16:55:57 <ais523> Sgeo: lexical scoping: {my $variable; do stuff;}
16:56:05 <ais523> dynamical scoping: {local $variable; do stuff;}
16:56:25 <ais523> "my" is nearly always better
16:56:48 <ais523> unless you need the semantics of "local" for some reason, or you're trying to scope something like $/ where you need to temporarily modify the original $/ rather than create a new one
16:56:55 <Vorpal> ais523, why not just change the sign of it and call it something else
16:57:06 <ais523> alise: because higher numbers are better
16:57:12 <ais523> *Vorpal:
16:57:16 <ais523> it's just that the numbers are negative
16:57:26 <ais523> you could add a constant, I suppose
16:57:47 <fizzie> ais523: +100 sounds like a safe enough constant to add.
16:58:05 <ais523> fizzie: might need to be a bit more, strength can go below 100dBm on occasion with some antennas
16:58:16 <ais523> admittedly, it's hard to pick the signal up at that level, but it's possible
16:58:32 <fizzie> It's only a problem if you don't want it to ever be negative.
16:58:38 <ais523> what would be really nice would be for phones to display signal/noise ratio, but I'm not sure how feasible it is to calculate that on the fly
16:58:57 -!- webquint has quit (Quit: Page closed).
16:59:01 <Vorpal> ais523, hm higher numbers are better, yeah inverting sign wouldn't work, since that would make lower better
17:01:31 <pikhq> I need to freaking *remove the driver's side wheel* in order to replace the battery. WHAT THE HELL
17:01:35 <Vorpal> ais523, anyway it should be plausible to make a scale from worst to best with a bit of margin based on the antenna. Presumably you have a weakest-working-signal as well as a best-plausible-signal for a given phone. Then you could just add a bit of margin as required (probably mostly/only at the top end) and then make a scale from worst to best based on that?
17:01:44 <Vorpal> Either linear, or something more complex
17:01:51 <fizzie> ais523: Given how an absurd amount of computation there is in the audio codecs like amr-nb/amr-wb, I doubt a SNR estimate would really be felt at all. They might even already be doing that sort of stuff; especially for any VOIP codecs, for comfort noise generation and VAD.
17:01:59 <Vorpal> but that way you should get reasonable good indicator
17:01:59 <ais523> best-plausible-signal can be really high, though
17:02:10 <ais523> if you're right next to the mobile phone mast, and that's entirely plausible in some built up areas
17:02:37 <Vorpal> ais523, so show "it's off the scales" or something like that. Will get some laughs
17:02:45 <Vorpal> or just make it non-linear
17:02:54 <ais523> 0-9000 and OVER 9000?
17:03:02 <Vorpal> yeah XD
17:03:07 <ais523> it'd be non-linear anyway, it's the use of a logscale that makes the values negative in the first place
17:03:13 <ais523> but I'm wondering if you'd have to log twice, or something
17:04:43 <Vorpal> ais523, or just do it so that from a scale 0-100, then you give 98% for very-good-not-next-to-tower signal and then use the remaining bit for the extreme case
17:05:17 <Vorpal> ais523, I mean, non-linear scales to increase detail in some areas and reduce it in others isn't exactly uncommon. Just look at sRGB
17:07:12 <Vorpal> ais523, anyway phones doesn't provide instant response to signal strength changes. So doing log twice wouldn't be an issue, it isn't like it is done more than every few seconds...
17:07:20 <Vorpal> for the display that is
17:07:32 <Vorpal> perhaps it calculates dBm more rapidly internally
17:07:53 <ais523> doing log twice would lead to pretty much no change for very different signals at the low end, though
17:08:02 <ais523> maybe just sqrt(log(x))?
17:08:20 <ais523> logarithms take next to no time nowadays, especially if you use a lookup table
17:09:09 <Vorpal> I'm just suggesting a general idea, exactly how you scale it is a fudge factor and would require being able to test it to see how it works out
17:09:34 <Vorpal> ais523, anyway you need more than just signal strength, you need SNR as well
17:09:38 <ais523> well, it'd need to be scaled in a consistent way across phones for signal strengths to be comparable across phones
17:09:52 <ais523> and I mentioned SNR already as being more meaningful than strength
17:09:56 <Vorpal> yeah
17:10:05 <ais523> hmm, or what about measurements in theoretical maximum bits per second?
17:10:11 <ais523> that depends on the SNR
17:10:14 <Vorpal> ais523, but excellent SNR and weak signal isn't very usable either
17:10:21 <ais523> yes it is
17:10:26 <Vorpal> ais523, hm okay
17:10:30 <ais523> assuming you take internal noise in the phone into account in the SNR
17:11:20 -!- zzo38 has quit (Quit: NO CARRIER).
17:11:22 <Sgeo> Me: There are some languages that are more than just syntax
17:11:26 <Sgeo> Student: Like assembly
17:11:30 <Sgeo> Me: Like Haskell
17:11:37 <Vorpal> ais523, what if you have basically no noise? (unrealistic I know!) but a signal so weak that the electrical current or whatever from it can't be detected by the circuits in the phone?
17:11:37 <Sgeo> Student: Like I said, low-level languages
17:11:54 <ais523> Vorpal: if there is no noise, it can be detected, by definition
17:13:06 <Vorpal> ais523, could it not fall below the range of whatever sort of electronic thingy that the phone uses to pick up the signal from the antenna? A/D converter perhaps?
17:13:29 <ais523> Vorpal: just put an amplifier in there
17:13:38 <ais523> the noise from the amplifier /counts towards the SNR ratio/
17:13:49 <Vorpal> ais523, yes but if there isn't one, couldn't what I suggested happen?
17:14:02 <ais523> Vorpal: no, because then the SNR increases
17:14:13 <fizzie> >>> str(h_f.sampleSingle(50))
17:14:13 <fizzie> "iraa dsolh tesheshts c'eoe :ldeonshol: luece fh We"
17:14:13 <fizzie> (That's what I sound like.)
17:14:18 <ais523> because if you can't detect a signal, then it is below the noise floor by definition
17:14:20 <Vorpal> ais523, hm I guess you get quantum noise at some level
17:14:26 <Vorpal> so you can never get actual 0 noise
17:14:31 <ais523> correct
17:14:32 -!- Phantom_Hoover has joined.
17:14:37 <ais523> it's known as shot noise
17:14:42 <ais523> normally, thermal noise is much larger, though
17:14:55 <ais523> especially at room temperature
17:14:59 <Vorpal> ais523, which is good, divisions by zero in nature sounds so awkward ;)
17:16:01 <Vorpal> ais523, why shot noise?
17:16:17 <Vorpal> ais523, it seems so nonsensical... compared to quantum noise.
17:16:28 <ais523> because it's a particular cause of noise
17:16:35 <ais523> something to do with it tending to come all at once
17:16:59 <ais523> <Wikipedia> Shot noise is a type of electronic noise that occurs when the finite number of particles that carry energy (such as electrons in an electronic circuit or photons in an optical device) is small enough to give rise to detectable statistical fluctuations in a measurement. It is important in electronics, telecommunications, and fundamental physics.
17:17:17 <alise> http://dinnerinabottle.com/
17:17:18 <alise> MEATWATER
17:17:23 <Sgeo> I love how I didn't have to install anything to use the wifi
17:17:24 <Vorpal> hm
17:17:29 <alise> You can buy gelfite fish water. Seriously.
17:17:37 <Vorpal> Sgeo, how is that surprising?
17:17:40 <alise> <Sgeo> Me: There are some languages that are more than just syntax
17:17:40 <alise> <Sgeo> Student: Like assembly ;; wat
17:17:58 <ais523> Vorpal: well, you usually do on Windows
17:18:00 <ais523> unless it comes preinstalled
17:18:14 -!- moo8 has joined.
17:18:25 <Vorpal> aren't most languages "more than just syntax"? Depends on what you actually mean with it. It isn't terribly clear.
17:18:47 <Vorpal> well, I guess some esolangs might start out as only syntax
17:19:03 <Sgeo> The student was complaining about the way that this language was being taught
17:19:03 <Vorpal> and then you invent the other parts
17:19:11 <Sgeo> Ok, closing comp
17:19:16 <Vorpal> <ais523> Vorpal: well, you usually do on Windows <-- hm okay
17:19:19 <alise> ais523: windows does wifi fine by default ime
17:19:21 <alise> *IME
17:19:23 <ais523> well, asm doesn't even have a single standard syntax
17:19:27 <ais523> alise: preinstalled?
17:19:28 <alise> it'll depend on your hardware, of course
17:19:41 -!- Sgeo has quit (Quit: Ex-Chat).
17:19:46 <alise> ais523: yes, it has wifi support...
17:19:49 <ais523> also, the Windows desktop I use doesn't do wifi
17:19:50 <alise> since XP
17:19:52 <alise> maybe even 2002
17:19:53 <alise> erm
17:19:54 <alise> 2000
17:19:59 <ais523> although that's hardly surprising, given that it doesn't have a wifi card
17:20:08 <Vorpal> ais523, linux has way better hw support than windows. Though most manufactures provide drivers for windows to make up for that
17:20:18 <ais523> but Windows even takes time to install USB stick drivers when you plug in a USB stick, it takes around a minute the first time
17:20:31 <Vorpal> ais523, a minute? More like half a minute
17:20:38 <Vorpal> but I guess that depends on the computer
17:20:49 <Vorpal> ais523, also it sometimes takes time if you just use another usb port
17:20:52 <Vorpal> than last time
17:21:00 <Vorpal> which is just weird
17:21:08 <ais523> Vorpal: yes, it has to install the drivers separately for each manufacturer of USB sticks, and each physical port on the computer
17:21:16 <ais523> which makes me think abstraction fail
17:21:26 <Vorpal> ais523, and each port on an external hub. I tested that some years ago
17:21:27 <ais523> if you plug a USB stick in port 1 then port 2, you can plug it back into 1 again without waiting
17:21:28 <Vorpal> was on xp
17:21:30 <ais523> Vorpal: haha
17:21:44 <ais523> what if you plug the external hub into a different port, then reuse a port on the hub?
17:22:00 <Vorpal> ais523, doesn't it need to install drivers for the hub iirc?
17:22:06 <Vorpal> and um I don't remember what happened then
17:22:17 <Vorpal> and I don't have any computer handy to test with currently
17:22:32 <ais523> Windows always simultaneously frustrates and amuses me whenever I have to use it nowadays
17:22:34 <ais523> it's just so slow
17:22:48 -!- moo8 has quit (Quit: moo8).
17:23:13 <Vorpal> ais523, I survive when I have to use it mostly. Mostly because all the lab computers at uni are core 2 quad at 2.6 GHz or such iirc
17:23:25 <Vorpal> and run xp
17:24:18 <Vorpal> of course it has it's own quirks. Like profile not propagating between all labs (while "my documents" does).
17:24:34 -!- moo8 has joined.
17:24:40 <ais523> oh, I can survive it
17:24:43 <ais523> but it's still frustrating
17:25:12 <ais523> if you want some more fun, a network glitch at the University made directory listings take half an hour to load, during which you couldn't do anything else
17:25:14 <Vorpal> ais523, yeah just have to store the theme to a file and reload it if you get the default bg when logging in. Oh except that internet explorer takes a bit more time to get sane
17:25:20 <ais523> I find it hard to imagine what sort of OS glitch could manage that
17:25:25 <Vorpal> there is no firefox or such
17:25:35 <ais523> (we literally lost half an hour of a three-hour class due to that)
17:25:38 * pikhq hereby hates relatively recent American cars
17:26:17 <pikhq> THE HELL
17:26:18 <Vorpal> ais523, I think the rules require us to ask for permission to install software. But we are free to compile our own in visual studio for courses and such. So I can only conclude that it is okay to write your own browser and run it, but not okay to just install firefox :D
17:26:49 <Vorpal> <ais523> if you want some more fun, a network glitch at the University made directory listings take half an hour to load, during which you couldn't do anything else <-- had that recently
17:26:51 <ais523> does Firefox install in Visual Studio?
17:26:53 <ais523> *compile in
17:26:56 <Vorpal> no idea
17:27:05 <Vorpal> well, login took 20 minutes that day
17:27:48 <Vorpal> and um while I tried to move a file explorer.exe crashed with "<some error> at 0x00000000" iirc
17:27:50 <moo8> Hello
17:27:55 <fizzie> ais523: "For doing development on the CVS trunk (Mozilla 1.9 or higher), the standard compiler is Microsoft Visual C++, version 8."
17:27:58 <ais523> hi moo8
17:28:05 <Vorpal> ais523, msvc took a few minutes to compile hello world
17:28:07 <ais523> CVS?
17:28:12 <ais523> seriously?
17:28:18 <ais523> Vorpal: it does that
17:28:26 <Vorpal> ais523, well it doesn't normally
17:28:30 <Vorpal> ais523, it was just that day
17:28:34 <ais523> ah, OK
17:28:37 <Vorpal> fizzie, that's moz? not firefox?
17:29:09 <moo8> hello
17:29:18 <Vorpal> ais523, after all they are core 2 quad with 4 GB ram each, on gbit ethernet. Oh and internet was fast, it was just anything on network shares, and since almost everything is on network shares....
17:29:29 <fizzie> Vorpal: It's the whole project; but only 1.9, so up to FireFox 3. They've switched to Mercurial for 1.9.1/FireFox 3.5.
17:29:38 <Vorpal> ah
17:30:08 <moo8> hello?
17:30:19 <ais523> hi moo8
17:30:33 <ais523> it's possible that I can hear you but you can't hear me, I suppose
17:30:36 <ais523> but that's unusual on IRC
17:30:42 <moo8> hey
17:31:01 <fizzie> Their build system is, I think, a bit hairy.
17:31:04 <fizzie> "In addition to Visual Studio, you must install MozillaBuild, a bundle of software including just the right versions of bash, GNU make, autoconf, Mercurial, and much more."
17:31:41 <ais523> heh
17:31:48 <ais523> and they call this non-esoteric?
17:31:53 <moo8> I am bored
17:32:04 <ais523> hmm, now I want an esolang that has really precise dependencies on exact versions of things
17:32:08 <ais523> but I think gcc-bf might count already
17:32:12 <fizzie> The "official compiler" is still VC8 for FF 3/3.5/3.6/4 all, but apparently all those versions also build with VC9, versions <=3.5 in VC7.1, and versions >=4 in VC10.
17:32:51 <ais523> moo8: you could try writing an esoprogram or two, that helps avoid boredom
17:33:12 <ais523> or go to anagolf and try to solve some of the problems there (golf.shinh.org)
17:33:25 <Vorpal> ais523, actually I'm not surprised. I seen similar bundles of software in specific fragile versions before for windows. For other open source projects
17:33:31 <moo8> so bored
17:33:31 <Vorpal> I think supertux used to have one
17:34:14 <ais523> hmm, someone actually did a cheat submission for "cancel fractions", it seems
17:34:22 <ais523> and I have no idea how, clearly it isn't embed-style cheating or rand-style cheating
17:34:30 <Vorpal> ais523, and supertux didn't use msvc, it uses msys + mingw.
17:34:53 -!- moo8 has quit (Quit: moo8).
17:35:16 <quintopia> damn
17:35:19 <quintopia> i was gonna unbored him
17:36:06 <Vorpal> ais523, hm the obvious way to implement it would be with that <Greek guy, forgot who>'s algorithm which uses gcd
17:36:17 <ais523> Euclid's?
17:36:22 <Vorpal> ah yeah could be
17:37:02 <Vorpal> ais523, which one is the cheat one?
17:37:17 <ais523> the one where someone resubmitted a "nocheat" version, which was longer
17:37:33 <Vorpal> ah hm
17:37:35 <ais523> my Perl submission there, incidentally, semi-cheats by using a rationals library
17:37:44 <Vorpal> hm
17:37:45 <ais523> but I consider that to be genuine because it actually solves the problem
17:37:48 <ais523> presumably it's faster to do by hand
17:38:02 <Vorpal> bbl food
17:39:59 <ais523> ah, the other Perl submissions used the same library as me, but used eval rather than parsing by hand
17:40:03 <ais523> why didn't I think of that?
17:41:41 <ais523> wow at the Scheme solution (which three people found): "(port-map print read)"
17:41:47 <alise> what's the challenge?
17:41:47 * Phantom_Hoover reads the WP summary of the Epic of Gilgamesh and can't help but think of Turkish Star Wars.
17:42:06 -!- tombom has joined.
17:42:12 <alise> ais523?
17:42:24 <Phantom_Hoover> ais523, doesn't work in plt-r5rs for me.
17:42:26 <ais523> cancel fractions
17:42:28 <alise> "In order to determine how to accomplish the goal in question, strategic planning basically does routing in plan-space, again using Dijkstra’s algorithm." ;; genius
17:42:42 <ais523> alise: nowadays it uses a modified A*
17:42:46 <alise> ais523: i was about to say
17:42:49 <alise> ais523: hmm, might yours not be better?
17:42:51 <alise> your magical one
17:42:59 <ais523> plus my routing algorithm for when it actually corresponds to routing on the map
17:43:12 <Phantom_Hoover> What's Ais' Magical Routing Algorithm?
17:43:19 <ais523> it's... complicated
17:43:35 <ais523> it's like dijkstra's algorithm modified to cache really well
17:43:37 <Phantom_Hoover> Complicated is fun!
17:43:39 <quintopia> A*
17:43:52 <ais523> and beats A* after a while due to A* needing to be recalculated each time
17:44:01 <quintopia> hmm
17:44:03 <ais523> if you do enough routings on similar maps, and Planar does
17:44:08 <quintopia> any new data structures?
17:44:18 <quintopia> asymptotically better than dijkstra's on fib heap?
17:44:25 <Vorpal> <alise> "In order to determine how to accomplish the goal in question, strategic planning basically does routing in plan-space, again using Dijkstra’s algorithm." ;; genius <-- what are you talking about?
17:44:33 <ais523> Vorpal: TAEB::AI::Planar
17:44:36 <Vorpal> aha
17:44:43 <ais523> I only recognised the concept because alise was quoting me
17:45:41 <Vorpal> ah
17:46:00 <ais523> hahaha; the C solution I'm looking at to "cancel fractions" needs to use a string twice
17:46:16 <ais523> and instead of repeating the string or assigning it to a variable, it writes the string once, and ""-6 once
17:46:37 <Vorpal> ais523, is there a reason for that
17:46:37 <ais523> relying on the merging of constant strings that gcc does to make the trailing NULs of the two strings the same
17:46:44 <ais523> Vorpal: it's shorter, and it's a gold competition
17:46:51 <Vorpal> heh
17:46:56 <ais523> *golf
17:47:16 <ais523> who cares that it's ridiculously fragile, ranking for golf competitions only depends on the program running correctly once, normally
17:47:18 <Vorpal> ais523, wait, gcc doesn't merge constants unless they actually are constant
17:47:23 <Vorpal> and you can't write to constants
17:47:23 <ais523> they are
17:47:32 <ais523> it doesn't write to it at all, just offsets from it
17:47:38 <ais523> "six characters before a null string"
17:47:44 <ais523> that doesn't modify the null string
17:47:51 <Phantom_Hoover> Damnit, why can't I watch Futurama in the UK
17:48:21 <Vorpal> ais523, how does it know it can write there at all
17:48:32 <Vorpal> ais523, chances are it might be in a read only page?
17:48:33 <ais523> it isn't writing there
17:48:36 <ais523> it is in a read only page
17:48:40 <Vorpal> eh okay
17:48:47 <ais523> you do not be able to write to what a pointer points to to subtract from the pointer
17:49:13 <ais523> "What's the memory address for 6 characters before the start of your stack frame?" "You can't write there!"
17:49:16 <ais523> non sequitur...
17:49:34 <Vorpal> ais523, I misunderstood you as it was writing there
17:49:37 <ais523> (admittedly the address might not /exist/ due to virtual memory and paging, but writability doesn't affect existence)
17:49:45 <Vorpal> ais523, I read "<ais523> it doesn't write to it at all, just offsets from it" as "<ais523> it doesn't write to it at all, just to offsets from it"
17:49:49 <Vorpal> note the extra word
17:50:16 <ais523> <hallvabo> import java.util.*;enum C{C;int a,b,n,d,t;{for(Scanner S=new Scanner(System.in).useDelimiter("/|\n");b<1;System.out.println(d>a?n/a+"/"+d/a:n/a))for(a=n=S.nextInt(),b=d=S.nextInt();b>0;a=b,b=t%b)t=a;}}
17:50:24 <ais523> that really says a lot about Java
17:50:49 <Vorpal> ais523, that you can make it as messy as you want but there will still be quite a few long function names in it you can't avoid?
17:50:51 <quintopia> what is that supposed to do?
17:51:03 <ais523> cancel fractions, like all the other submissions to the golf competition in question
17:51:06 <ais523> and it actually succeeds at it
17:51:12 <ais523> but it's a lot longer than the submissions in most other langs
17:51:23 <quintopia> didn't know about said competition
17:51:40 <ais523> haha, I see how the golfscript competition cheats
17:51:46 <ais523> *golfscript entry cheats
17:51:50 <ais523> it contains embedded Ruby
17:51:54 <Vorpal> ais523, huh?
17:52:15 <ais523> golfscript's interp is written in Ruby, and someone added an "embed ruby" command to the lang, presumably on a whim
17:52:19 <Vorpal> ais523, but exec was denied?
17:52:23 <Vorpal> oh okay
17:52:27 <alise> ais523: are there any plans to make TAEB less cautious?
17:52:27 <ais523> see, no exec
17:52:33 <alise> it seems that always taking the worst-case damage isn't a good idea
17:52:37 <alise> it'd be better to figure out the distribution
17:52:42 <alise> and perhaps pick the most common, plus a bit
17:52:47 <alise> (while taking into account the worst case)
17:52:52 <ais523> alise: no, because it's so easy to pretty much avoid damage, in most cases
17:53:08 <ais523> bots have no issue with spamming Elbereth everywhere, and it's probably the best strategy
17:53:21 <quintopia> hey guys. any of y'all have a thought on whether this is computable: http://esoteric.voxelperfect.net/wiki/Bigot
17:53:23 <Vorpal> alise, doing that feels a bit like cheating :/
17:53:26 <Vorpal> err
17:53:27 <Vorpal> ais523, ^
17:53:37 <alise> Elbereth is a bit lame
17:53:41 <alise> AceHack should remove it ;-)
17:53:54 <ais523> alise: not Ace, that's not compatible with its goals
17:53:58 <alise> BAH
17:53:59 <fizzie> Either Spork or Un (or maybe both) do something to Elbereth already.
17:54:02 <alise> AliseHack
17:54:16 <ais523> Spork nerfs Elbereth; Un plans to, but I'm not sure if it has yet
17:54:29 <ais523> it's interesting that they both decided to nerf it rather than remove it outright
17:54:32 <Vorpal> alise, nah, a better fix would be to remove it for writing in dust/engraving, that would make it more expensive. One charge from a wand of fire or such
17:54:35 <alise> "TAEB, on seeing this frame, makes a couple of damaging inferences:
17:54:35 <alise> The lower half of the level has turned into solid rock." :D
17:54:47 <alise> ais523: nerf?
17:54:54 <Vorpal> ais523, nerf?
17:54:56 <alise> also, un?
17:55:02 <alise> oh, unnethack
17:55:03 <ais523> alise: make a lot less powerful, it's a common gaming perjorative
17:55:17 <alise> I am not down wit da LINGO
17:55:21 <ais523> the reference is if you replace a real gun with a nerf gun
17:55:22 <alise> Well okay I know a lot of it
17:55:24 <alise> right
17:55:30 <Vorpal> <alise> "TAEB, on seeing this frame, makes a couple of damaging inferences: <alise> The lower half of the level has turned into solid rock." :D <-- eh what?
17:55:45 <ais523> Vorpal: if you try parsing the screen before the game finishes rendering
17:56:03 <ais523> it doesn't actually infer that the level is half-rock, but rather that it's fallen into another level with the same number in a different branch
17:56:06 <ais523> which is arguably worse
17:56:08 <Vorpal> ais523, you could just wait for the relevant line to be output?
17:56:15 <ais523> Vorpal: no you couldn't?
17:56:32 <Vorpal> ais523, well, wait for the HP lines and such to be output
17:56:36 <ais523> the game's rendering order isn't consistent enough / tagged well enough to do that
17:56:39 <Vorpal> aha
17:56:40 <Vorpal> I see
17:56:47 <fizzie> Right, Un had made Elbereth not work on major demons, but hadn't nerfed it any more; but they added that "ctrl-e engraves Elbereth" quick-key, which is a bit... much.
17:56:53 <alise> ais523: can't you just store the order as it is in the code? :P
17:56:54 <Vorpal> ais523, so it arrives out of order in the byte stream?
17:57:07 <alise> fizzie: ha
17:57:10 <ais523> fizzie: Ace is going to map Elbereth-writing to .
17:57:10 <alise> Ctrl+E Ctrl+E Ctrl+E Ctrl+E Ctrl+E Ctrl+E Ctrl+E Ctrl+E
17:57:16 <alise> ais523: noooo
17:57:17 <ais523> and then reflavour it, to be less arbitrary
17:57:19 <alise> fuck that
17:57:24 <alise> Elbereth should be painful to do
17:57:29 <alise> as punishment :P
17:57:31 <ais523> alise: the idea is to keep the same gameplay mechanics as NetHack, but make the interface better
17:57:34 <alise> ais523: I use . to wait, anyway
17:57:36 <alise> although usually s
17:57:40 <alise> since it's more helpful
17:57:48 <ais523> so, keep using . to wait and have Elbereth protection in the meantime
17:57:59 <Vorpal> alise, I think my suggestion above that it would only work with burning, not with writing in dust/engraving should make it a lot more expensive
17:58:04 <fizzie> Aw, no-one did a better fraction-canceler in Forth than my somewhat crappy quickly done obvious gcd+output solution.
17:58:07 <Vorpal> and perhaps even the burned one could wear out over time
17:58:08 <ais523> in practice, Elbereth-writing tends to be control-shift-V anyway
17:58:18 <quintopia> alise, ais523: what game/challenge does the previous discussion refer to?
17:58:27 <cheater> i use . to hack puddings
17:58:29 <Vorpal> ais523, shift?
17:58:30 <ais523> quintopia: http://golf.shinh.org/p.rb?Cancel+fractions
17:58:35 <alise> cheater: pudding farmer!
17:58:38 <ais523> just one of many on anagolf
17:58:42 <cheater> alise: and proud of it
17:58:45 <ais523> but I'm caring about it more than others as I submitted it
17:58:53 <ais523> alise: just look at his/her nick
17:58:58 <ais523> not that pudding farming /is/ cheating, just boring
17:59:01 <alise> cheater: ,-_|_
17:59:13 <alise> Wow, worst middle-finger ASCII art ever.
17:59:16 <cheater> alise: (__(_)__)
17:59:29 <alise> ais523: I love the quote --
17:59:35 <ais523> oh, I interpreted the _|_ as bottom from Haskell
17:59:36 <Vorpal> cheater, I find pudding farming gives too little, I tend to go for death farming instead
17:59:37 <quintopia> ais523: i was already looking at that, but i have no idea how elbereth, spork, etc. applies to it
17:59:38 <alise> "The DevTeam has arranged an automatic and savage punishment for pudding farming. It's called pudding farming."
17:59:40 <cheater> oh, i thought it was a unary operator evaluated on Bottom
17:59:43 <alise> ais523: that's why I added the ,- thumb :P
17:59:43 <ais523> as it's the normal ASCII representation
17:59:51 <cheater> so i gave alise an ascii art of a closure.
17:59:53 <ais523> alise: yes, I was trying to figure out the value of ,
18:00:02 <alise> heh
18:00:02 <ais523> that you were subtracting bottom from
18:00:11 <ais523> although I suppose the value's irrelevant, because the sum as a whole has no value
18:00:33 <fizzie> ..|. is a nice, minimalistic finger.
18:00:33 <alise> Vorpal: isn't death farming pointless apart from ascending with weird stuff?
18:00:42 <ais523> quintopia: http://golf.shinh.org/p.rb?C+style+constants has almost a day left, it's another one I set
18:00:54 <ais523> alise: death farming's the fastest way to accumulate score
18:00:55 <Vorpal> alise, well if you want to max score, or get negative score, it is also somewhat useful
18:00:58 <ais523> if you're trying to max out the score counter
18:01:04 <alise> boring
18:01:16 <ais523> most of the monsters that give higher scores are difficult to farm
18:01:34 <Vorpal> ais523, I managed to get exactly 0 for score once with death farming and some careful planning in general
18:01:40 <cheater> alise is bored by EVERYTHING
18:01:44 <ais523> Vorpal: on NAO?
18:01:50 <Vorpal> ais523, no. Too much lag from here
18:01:58 <Vorpal> ais523, it is basically unplayable for me
18:02:01 <ais523> hmm, do it in devnull some year, and troll all the people aiming for minimum score
18:02:06 <ais523> it has servers all over the world
18:02:07 <Phantom_Hoover> Wait, Haskell has an empty type as standard?
18:02:11 <ais523> Phantom_Hoover: yes
18:02:15 <Vorpal> ais523, maybe, when is devnull?
18:02:15 <quintopia> ais523: this isn't helping me understand where you are death farming and where the monsters lurk
18:02:19 <ais523> how else would you create a function that didn't return a value
18:02:23 -!- atrapado has quit (Quit: Abandonando).
18:02:24 <ais523> Vorpal: every November
18:02:25 <Phantom_Hoover> ais523, !haskell example?
18:02:28 <Phantom_Hoover> ais523, ()
18:02:30 <Phantom_Hoover> *?
18:02:34 <Vorpal> ais523, ah, I have university and such then, tricky
18:02:39 <ais523> well, you probably want to put it in a monad so it can actually /do/ something
18:02:44 <Vorpal> ais523, the game in question lasted about half a month
18:02:52 <ais523> !haskell :t putStrLn "Hello, world!"
18:02:59 <Phantom_Hoover> And anyway, lazy semantics make non-returning functions basically pointless.
18:03:03 <EgoBot> putStrLn "Hello, world!" :: IO ()
18:03:05 <Phantom_Hoover> () is not empty.
18:03:07 <ais523> Vorpal: that's rather a while for death farming
18:03:09 <Phantom_Hoover> () is the unit type.
18:03:17 <alise> <Phantom_Hoover> Wait, Haskell has an empty type as standard?
18:03:17 <alise> <ais523> Phantom_Hoover: yes
18:03:18 <alise> liar
18:03:29 <ais523> hmm, I'm muddling terminology, aren't I?
18:03:30 <alise> or rather, misinformer
18:03:31 <Vorpal> ais523, I was doing extinctionism as well
18:03:33 <ais523> that isn't lying, though, just being wrong
18:03:33 <alise> or misinformed
18:03:36 <alise> ais523: yes
18:03:36 <ais523> Vorpal: ah, I see
18:03:41 <alise> |()| = 1
18:03:42 <alise> |Void| = 0
18:03:48 <alise> data Void -- requires extension to H'98
18:03:56 <Vorpal> ais523, I think I missed water devils though
18:04:01 <ais523> oh, a type with no values at all, such that you know any function that returns it doesn't return?
18:04:10 <ais523> C should have one of those, IMO
18:05:00 <Vorpal> ais523, like a combination of void and __attribute__((noreturn)) ?
18:05:03 <Phantom_Hoover> ais523, void
18:05:09 <ais523> Vorpal: yep
18:05:12 <ais523> Phantom_Hoover: that has one value
18:05:20 <ais523> like a 0-bit number
18:05:23 <Phantom_Hoover> IIRC WP gives void as the closest analogue to the empty type.
18:05:31 <ais523> void's a unit type
18:05:41 <Vorpal> ais523, non portable solution: #define noretvoid __attribute__((noreturn)) void
18:05:47 <Phantom_Hoover> void stuff();
18:05:50 <Vorpal> or whatever you want to call it
18:05:54 <Vorpal> ais523, then use that as a return type
18:05:55 <Phantom_Hoover> x = stuff;
18:05:56 <Vorpal> should work
18:06:01 <Phantom_Hoover> Surely that's not legal C?
18:06:06 <Phantom_Hoover> *stuff()
18:06:17 <Vorpal> Phantom_Hoover, as ais523 said, void is an unit type
18:06:27 <Phantom_Hoover> What's the value>
18:06:29 <ais523> if it's "void stuff()", then stuff itself is of type void(*)()
18:06:31 <Vorpal> void
18:06:37 <ais523> which has quite a lot of possilbe values
18:06:38 <ais523> *possible
18:06:47 <Vorpal> ais523, that's pointer to the func though?
18:07:00 <ais523> Vorpal: the name of a function is a pointer to it
18:07:00 <Phantom_Hoover> ais523, that's the type of the function, not its value.
18:07:27 <Vorpal> ais523, well yeah
18:07:55 <Vorpal> Phantom_Hoover, yes I have to agree here
18:08:01 <fizzie> Current WP view on C void is: "Note that despite the name, in all of these situations, the void type serves as a unit type, not as a zero or bottom type, even though unlike a real unit type which is a singleton, the void type comprises an empty set of values, and the language does not provide any way to declare an object or represent a value with type void."
18:08:07 <Phantom_Hoover> And void x; gives compiler errors.
18:08:49 <Phantom_Hoover> fizzie, |void| = 1 or 0?
18:08:52 <ais523> void's defined as an incomplete type that cannot be completed
18:09:04 <ais523> so via the definition, it's a perfect encapsulation for some type with no way to access it
18:09:08 <Phantom_Hoover> If it has no values, it is an empty type.
18:09:21 <ais523> you can assume it's the type defined by the HQ9++ program ++
18:10:10 <ais523> Phantom_Hoover: void quite possibly has values, even multiple values, but there's no way to access them except via type punning
18:10:14 <Vorpal> ais523, doesn't gcc support void as a special type? Like for stack marker something or such
18:10:28 <Vorpal> ais523, you used it in gcc-bf iirc?
18:10:28 <ais523> and no way to create them except via type punning either
18:10:50 <fizzie> Well, C99 does say "The void type comprises an empty set of values; it is an incomplete type that cannot be completed." -- so if you go by literal verbiage, I guess it'd be empty, but then it doesn't really make sense as a return value type.
18:10:55 <ais523> Vorpal: oh, you can dereference a pointer to void inside a goto, to jump to that memory address
18:11:12 <ais523> but that's just because extensions need to, according to the standard, fit into stuff that's undefined behaviour in other interps
18:11:17 <Phantom_Hoover> ais523, code snippet?
18:11:22 <Vorpal> fizzie, maybe it is overloaded in a manner similar to "static"
18:11:53 <ais523> Phantom_Hoover: {void* x; label: ; x = &&label; goto *x;}
18:11:58 <ais523> infinite loop via gcc extensions
18:12:32 <fizzie> ais523: Wouldn't it be a more elegant as {void *x; x = &&label; label: goto *x;} -- now you keep setting x all the time.
18:12:33 <Phantom_Hoover> Extensions, though, hence not standard C.
18:12:52 <ais523> fizzie: it'd be more elegant as an actual noncomputed loop, so it could be optimized
18:12:55 <Vorpal> Phantom_Hoover, exactly, which is why it doesn't affect our argument
18:13:08 <ais523> gcc-bf's use of it is along the lines of "goto *(void *)0;"
18:13:25 <ais523> which is completely nonportable even within gcc, but I know that gcc-bf handles it correctly
18:13:46 <ais523> this is an example of the general principle that you can do what the hell you like in system libraries and headers, and generally have to
18:14:26 <fizzie> Another part of C99 says, of void: "The (nonexistent) value of a void expression (an expression that has type void) shall not be used in any way, and implicit or explicit conversions (except to void) shall not be applied to such an expression."
18:14:54 <Phantom_Hoover> So it's empty?
18:15:03 <ais523> fizzie: you can create an effective void value via type punning, though
18:15:11 <ais523> but then you need to type-pun /again/ to determine what it is
18:15:20 <ais523> and you just get the same result as if you'd done the type punning in one step
18:15:36 <fizzie> But really, it's a terminology quibble; you could as well argue that since "type funcname();" returns a value of type, and you can legally say "void funcname();" and have that return, it can't be empty.
18:15:57 <Vorpal> <ais523> this is an example of the general principle that you can do what the hell you like in system libraries and headers, and generally have to <-- no you can't but you have to because you can't
18:16:22 <Vorpal> ais523, with that I mean that to avoid messing up for user code you often have to that sort of weird stuff
18:16:24 <ais523> Vorpal: you can, because nothing in the standards constrains system headers/libraries, apart from semantics
18:16:42 <ais523> the standards don't even require the standard #includes to actually be files
18:16:56 <ais523> #include <stdio.h> could just set an internal compiler flag, for instance, in a hypothetical implementation
18:16:58 <Vorpal> ais523, you can't define stuff that could collide with the app, and you must be prepared to survive a lot of strange #define (up to a point)
18:18:47 <ais523> Vorpal: if implementing the headers in C, correct
18:18:55 <ais523> nothing forces system headers to be in C!
18:19:48 <Vorpal> ais523, hm... have you seen any such system?
18:20:18 <ais523> no
18:20:27 <ais523> there's quite a lot of latitude in the C standards that has probably never been used
18:21:40 <Vorpal> ais523, I don't know about GCC but iirc clang sets up the initial #defines by passing a number of strings like "#define __clang__ 1" to the parser before starting to parse the actual files
18:21:48 <Vorpal> rather than just setting it up directly in some internal table
18:26:57 <nooga> clang's cool
18:29:55 <alise> ais523: taeb's code is unreasonably well-architectured!
18:30:21 <ais523> alise: heh
18:30:36 <fizzie> "NetHack bot" sounds like something that ought to be a real mess.
18:30:52 <ais523> at least two of the devs were using it as a test that the principles behind Moose were useful in practice
18:30:57 <Phantom_Hoover> TAEB
18:31:00 <Phantom_Hoover> *?
18:31:06 <ais523> as in, a test of the concept, rather than a test of the implementation
18:31:10 <ais523> something that's rarely done
18:31:11 <ais523> Phantom_Hoover: yep
18:31:24 <Phantom_Hoover> What is TAEB?
18:31:48 <ais523> fizzie: some of the devs were complaining that my code was hacky, and I said that it was trying to emulate a hack in NetHack's code itself correctly, so there wasn't much else you could do
18:33:50 <Vorpal> <ais523> at least two of the devs were using it as a test that the principles behind Moose were useful in practice <-- Moose?
18:33:59 <ais523> Vorpal: Perl object-orientation library
18:34:21 <ais523> it's rather indicative of how Perl works that you /can/ have a library that adds a new OO system
18:35:48 <nooga> POOP
18:35:57 <Vorpal> I often get transmission errors when copying data from/to my phone over bluetooth, I wonder why. Often I get like half an image, and then it works on next try. Nowdays I always check both size once after transfer and sha512sum twice
18:36:03 <nooga> or a library that changes syntax
18:36:08 <Vorpal> but that really shouldn't be needed
18:36:14 <ais523> nooga: not really
18:36:25 <ais523> you can have a library that preprocesses the code, but that's not quite the same
18:36:43 <nooga> it's what lacks in other langs
18:36:46 <Vorpal> Phantom_Hoover, TAEB is what has been discussed the past half screen
18:37:09 <Vorpal> Phantom_Hoover, it should be clear from the context
18:37:27 <ais523> Phantom_Hoover: TAEB is a NetHack bot framework, written in Perl
18:37:35 <ais523> combining it with an AI gives you a NetHack bot
18:37:44 <Phantom_Hoover> Ahh.
18:37:46 <ais523> and the AIs are written as Perl libraries
18:38:40 <Vorpal> ais523, anyway what is the last thing that nethack prints when drawing the screen?
18:39:06 <ais523> I think it's the lowest-right character that actually changed
18:39:17 <ais523> not counting some characters that are drawn at other points in the sequence
18:39:26 <Vorpal> ais523, so it doesn't print the HP stuff unless it needs to?
18:39:37 <Vorpal> how curious
18:39:39 <Vorpal> why
18:39:49 <Vorpal> it should take less code to just redraw the whole thing
18:39:51 <fizzie> Vorpal: Speaking of which (not that we were speaking of anything, this is a non sequitur), I took some panoramish (more like just stitched wide-angle) things from this water tower: http://zem.fi/~fis/watertower.jpg -- there's a restaurant on top of it and so on.
18:39:53 <ais523> to save terminal bandwidth, I think
18:39:57 <Vorpal> haha
18:40:19 <Vorpal> fizzie, and where is that pano?
18:40:20 <ais523> Vorpal: there are bits of code which clearly aren't assuming that terminal bandwidth is near-unlimited, like it is on modern computers
18:40:25 <fizzie> Vorpal: Being stitched.
18:40:30 <Vorpal> fizzie, also: nice water-tower
18:40:34 <ais523> even an option to create time delays via excess NULs in output
18:40:42 <Vorpal> fizzie, very unusual design
18:41:17 <Vorpal> ais523, err what? that would waste bw?
18:41:33 <ais523> Vorpal: deliberately, in order to create a time delay
18:41:47 <fizzie> lynx or screen or some-such did some pretty heavy screen-scrolling optimization; sometimes when you scrolled down a whole page, if it happened to have enough repeated structure in it, it'd just move some regions around the screen, since you can do region-to-region copies cheaply. (The way it moved things around was pretty visible on the vt510 at 9600 bps.)
18:41:49 <ais523> it used to be the standard method
18:42:24 <fizzie> Vorpal: The tower seems to have a Wikipedia page, though it's a bit stub-y: http://en.wikipedia.org/wiki/Haukilahti_water_tower
18:43:22 <Vorpal> ais523, why the time delay? excessive null bytes just wastes bw...
18:45:03 <ais523> Vorpal: because there's a limit to how fast characters can be sent to a terminal
18:45:12 <ais523> 1000 NULs would create a noticeable delay on most old-fashioned terminals
18:45:20 <ais523> due to the time it took to send them all down the line
18:45:52 <Vorpal> ais523, so you mean the *pipe* wasn't the limit?
18:46:05 <Vorpal> but the internal buffer of the terminal at the end?
18:46:05 <ais523> the speed of the connection was the limit
18:46:13 <ais523> it tended to be a serial connection at 2400 baud or so
18:46:20 <fizzie> Vorpal: First set, out of three: http://zem.fi/~fis/hp1.jpg -- that's about 129 degrees, horizontally.
18:46:24 <ais523> sending a kilobyte at 2400 baud will take a while no matter what the data is
18:46:35 <Vorpal> ais523, then those NUL would just waste that bw of the connection!? this makes no sense
18:47:01 <Vorpal> if the terminal hardware wasn't the limit, but the connection was
18:47:05 <Vorpal> then the NULs make no sense
18:47:06 <fizzie> Vorpal: It's for when you actually want to make a fixed-time delay.
18:47:23 <fizzie> To animate a twiddly thing, for example, or something.
18:47:25 <Vorpal> fizzie, hm
18:47:30 <fizzie> (That you don't want to twiddle too fast.)
18:47:32 <Vorpal> fizzie, in nethack?
18:47:51 <ais523> Vorpal: there are fixed-time delays in NetHack, yes
18:47:54 <fizzie> I guess there it could be the other animations.
18:48:07 <ais523> things like shift-direction
18:48:12 <ais523> which shows you at each location you move past
18:48:38 <Vorpal> ais523, oh hm I never noticed that
18:48:55 <ais523> there's an option to turn a) the delay, b) the intermediate rendering off
18:55:42 <Vorpal> fizzie, nice pano
18:55:51 <Vorpal> fizzie, (took a while to load)
18:56:27 <Vorpal> fizzie, no way to get to the absolute top and do a spherical pano?
18:56:29 <Vorpal> XD
18:57:54 <fizzie> Vorpal: It's the upload pipe again. The second one is at http://zem.fi/~fis/hp2.jpg and it's a bit larger (2.2M vs. 1.3M in bytes) so it'll probably take even longer.
18:58:47 <fizzie> There's a pile of antennas and such up there in the middle, so I'm sure it's technically speaking possible to get there, but it's not open to the public, no.
19:00:32 <Vorpal> fizzie, hm the antennas would block the view
19:00:40 <Vorpal> would need to remove them while taking the pano
19:03:51 <fizzie> In hp2.jpg there's the Otaniemi (where TKK's... sorry, Aalto University School of Science and Technomancy's campus is) water tower sort-of visible, though you pretty much have to know what you're looking for. It's a bit strange octagonal-ish shape too.
19:05:01 <fizzie> http://www.muuka.com/finnishpumpkin/towers/2/img/vote5.jpg is a google-searched image of it.
19:05:20 <fizzie> I guess that has a bit more than 8 sides, but polygonal anyway.
19:05:55 * Phantom_Hoover realises he has never seen a water tower.
19:06:24 <Vorpal> fizzie, wait, is hl2 not from the water tower?
19:06:49 <Vorpal> Phantom_Hoover, whaaat?
19:07:11 <Phantom_Hoover> We just... don't have them in Scotland.
19:07:23 <Vorpal> Phantom_Hoover, so they use powerful pumps instead?
19:07:24 <fizzie> Vorpal: They're both from the same place, just different directions.
19:07:31 <Phantom_Hoover> At least, not in the bits I've seen...
19:07:33 <ais523> Vorpal: much of Scotland is rather mountainous
19:07:36 <ais523> including most of the water sources
19:07:41 <Vorpal> ais523, ah I see
19:07:44 <ais523> so I imagine they can mostly just use gravity
19:07:54 <Phantom_Hoover> Yeah, Edinburgh has the Pentlands right next to it.
19:08:03 <Vorpal> ais523, so they just put reservoirs uphill?
19:08:05 <ais523> yep
19:08:08 <Phantom_Hoover> Uhu.
19:08:25 <Phantom_Hoover> I remember doing a school trip to one ages ago.
19:12:17 <Vorpal> there is one city near here that as a water tower that looks a bit like a mushroom in shape. It's officially named "The mushroom" (Svampen). There is some cafe near the top
19:12:23 <Vorpal> http://sv.wikipedia.org/wiki/Svampen,_%C3%96rebro has an image of it
19:12:44 <Vorpal> the trees are actually quite near
19:12:51 <Vorpal> it is taller than it looks from the image on wikipedia
19:13:54 <Vorpal> says it is 58m tall
19:14:39 <Phantom_Hoover> I feel like I'm missing out.
19:14:53 <Phantom_Hoover> Wait, I thought Scandinavia had mountains aplenty?
19:15:06 <fizzie> Vorpal: http://zem.fi/~fis/hp3.jpg -- and there's the third (last) one. It has a bit problematic exposure, mostly towards sunlight, but anyway. (Also shaped a bit differently, took it in two rows of five pics.)
19:15:16 <fizzie> Finland is very lacking in the mountains department.
19:15:27 <Phantom_Hoover> I must say, all of these water tower designs remind me of Edinburgh Airport's control tower.
19:15:50 <fizzie> There are some fells up north, but they're rather... lackluster, so to say.
19:16:05 <Phantom_Hoover> http://upload.wikimedia.org/wikipedia/commons/d/d1/Edinburgh_Airport_Control_Tower.jpg
19:16:13 <Vorpal> <Phantom_Hoover> Wait, I thought Scandinavia had mountains aplenty? <-- not uniformly
19:16:18 <Vorpal> and yeah Finland lacks that
19:16:34 <Vorpal> Phantom_Hoover, and it's very flat in the part of Sweden that I live
19:16:56 <fizzie> Norway's a bit more edgy.
19:17:06 <Vorpal> Phantom_Hoover, like a wormhole about to close
19:17:14 <Vorpal> Phantom_Hoover, I would be scared to travel up that ;)
19:17:40 <Vorpal> fizzie, nice last pano
19:19:01 <fizzie> Water towers here are usually just pretty boring "tall, slender cyliner, with a bit wider cylinder up top" designs; the Haukilahti UFO and Otaniemi nut-and-bolt styles are the exceptions.
19:19:06 <Phantom_Hoover> fizzie, your panoramas are much nicer than Vorpal's.
19:19:14 <Phantom_Hoover> Vorpal's break my computer
19:19:18 <fizzie> Oh, and there's a silly-looking one in Espoonlahti: http://4.bp.blogspot.com/_UnH8iNDl81s/SV4vLcBaRnI/AAAAAAAAAOY/jx2AkN5cZwM/s1600-h/IMG_0512.JPG
19:19:26 <fizzie> Phantom_Hoover: ITYM s/nicer/smaller/. :p
19:19:56 <Phantom_Hoover> Truly, Scotland has severely missed out on this font of architectural weirdness.
19:20:15 <Phantom_Hoover> So wait, do they use towers in England?
19:21:01 <fizzie> There's another UFO in Lauttasaari: http://fi.wikipedia.org/wiki/Tiedosto:Lauttasaari_water_tower.JPG
19:21:20 <Phantom_Hoover> That one looks more mushroomy.
19:21:33 <fizzie> It's been decommissioned, also; there's no water up there, or anything else.
19:21:37 <Phantom_Hoover> With the gill-like structures on the bottom.
19:21:50 <fizzie> They've planned all sorts of stuff, like a wall-climbing place, but nothing has materialized.
19:22:10 <Vorpal> <Phantom_Hoover> Vorpal's break my computer <-- that's your computer's issue
19:22:14 <Vorpal> not my fault
19:22:59 <Vorpal> <Phantom_Hoover> That one looks more mushroomy. <-- well, that depends on your point of view
19:24:02 <fizzie> Oh, oh, and the Hanko water tower has inexplicably a fish-shaped weather vane on top: http://upload.wikimedia.org/wikipedia/commons/f/f1/Hankoo_water_tower_July_10_2005.JPG
19:24:09 <fizzie> I don't know what's the story on that one.
19:25:08 <Vorpal> http://www.google.com/images?um=1&hl=en&biw=1614&bih=685&tbs=isch%3A1&sa=1&q=Svampen+%C3%96rebro&aq=f&aqi=&aql=&oq=&gs_rfai=
19:25:16 <Vorpal> Phantom_Hoover, there are some more pictures there
19:25:34 <Vorpal> plus for unknown reason the Örebro Castle
19:25:45 <Phantom_Hoover> It looks decidedly less mushroomy from that angle.
19:25:58 <Vorpal> Phantom_Hoover, from which one?
19:26:16 <Vorpal> Phantom_Hoover, I think the side view is quite nice
19:26:36 <Phantom_Hoover> All of the ones other than fizzie's.
19:26:40 <Vorpal> Phantom_Hoover, like an UFO parked on top of a giant concrete mushroom
19:27:32 <Vorpal> Phantom_Hoover, and the one fizzie linked was a different water tower
19:27:43 <Phantom_Hoover> Ah.
19:28:45 <fizzie> Yeah, mine was in Finland.
19:29:02 <fizzie> The Haukilahti UFO is also lit up so that it looks even more UFOy at night. I can't seem to find a nice high-resolution photo, but http://www.fonecta.fi/images/stillpics///files/343303.fonecta.client/1236243011351_1_large.jpg
19:29:45 <Vorpal> fizzie, they light up Svampen too. And sometimes in neon colours
19:30:05 <fizzie> Hah, and there's this one in Lahti, Finland: http://4.bp.blogspot.com/_UnH8iNDl81s/SVe3phDYq7I/AAAAAAAAAMo/l9xbqWAA5SI/s1600-h/IMG_0445.JPG
19:30:20 <Phantom_Hoover> For some reason I cannot comprehend, the relief map of Scotland I found on Google has "ALBANIA" written on it.
19:30:23 <fizzie> It looks like it's some sort of low-polygon approximation.
19:30:39 <Vorpal> I remember reading that Svampen was built top to bottom, using cranes to suspend the upper parts
19:30:41 -!- MigoMipo has quit (Ping timeout: 240 seconds).
19:30:46 <Vorpal> while the lower ones were being build
19:30:48 <Vorpal> built*
19:31:02 <Vorpal> <fizzie> It looks like it's some sort of low-polygon approximation. <-- indeed
19:31:35 <Phantom_Hoover> Apparently Albania is not in southeastern Europe as previously believed, but in the Highlands.
19:31:41 <Vorpal> fizzie, hm would fit right into darwinia!
19:34:22 <fizzie> Phantom_Hoover: There's a tiny little swamp (formerly a village, now there's something like three farms) near Lieksa (where our summer places are) called "Egyptinkorpi", lit. "Egypt's woods/wilderness/backwoods" (the "korpi" part is a bit tricky to translate).
19:34:34 <fizzie> I've always wondered about the name when driving past the road-sign pointing that way.
19:35:03 -!- MigoMipo has joined.
19:35:06 <fizzie> As close as I can determine, there's nothing whatsoever "egyptian" about the place.
19:35:33 <fizzie> (It's right next to the Russian border, but I don't see how that would be relevant either.)
19:35:39 -!- MigoMipo_ has joined.
19:35:49 <Vorpal> fizzie, what about the "Klockrike" somewhere near Linköping in Sweden?
19:35:53 <Vorpal> that's a pretty strange one too
19:36:32 <Vorpal> translates to something like "Country of Clocks" or such, a bit hard to translate the actual meaning.
19:36:38 <fizzie> OTOH, the named one village "Hörhö", which is the Finnish word for, well, "crackpot".
19:36:48 <Vorpal> or maybe "Country of bells"
19:36:55 <Vorpal> hard to tell, same word
19:37:00 <fizzie> "Hi, where are you from?" "Oh, uh, well.. the crackpot-land, actually..."
19:37:31 <Vorpal> fizzie, could be of Swedish origin?
19:37:42 <Vorpal> Hör = hear, hö = hay
19:37:50 <Vorpal> not very sensible, but better
19:38:03 <fizzie> I think that would be more likely on the western side, but I guess it's always a possibility.
19:38:38 <Vorpal> fizzie, "hörö" I would have said was a lot more likely as a Swedish place name, since ö = island
19:39:01 <Vorpal> and it isn't exactly uncommon for place names to end in -ö
19:39:24 -!- MigoMipo has quit (Ping timeout: 252 seconds).
19:39:37 <fizzie> There's quite a lot of -ö's in Finland, too. Especially since most of our islands tend to be in the more-or-less Swedish-speaking areas.
19:41:16 <Vorpal> hm
19:46:54 <nooga> i've got brown cheese from norway/1
19:48:41 <Vorpal> nooga, as opposed to norway/2?
19:48:49 <Vorpal> is that version number? arity number? division?
19:49:47 <fizzie> Okay, opened in openstreetmap approximately the middle third of the coastline between Helsinki and Hanko: it shows 24 place-names, and ten of those end in -ö. (Strömsö, Storramjö, Skämmö, Jakobramsjö, Vålö, Stora Fagerö, Svinö, Kalvö, Lilla Svartö, Träskö. And I guess you can count Kyrkogårdsön too. And there's Mallholmen and five "-landet"s. And only three that end in "-saari" (fi:island).)
19:50:10 -!- BeholdMyGlory has joined.
19:50:39 <fizzie> (I'm assuming osm is showing the "primary" names here; many have names both in Finnish and Swedish.)
19:51:50 <Vorpal> fizzie, the one ending in -sjö is "lake" not "island"
19:51:58 <Vorpal> (it doesn't make sense splitting it elsewhere)
19:52:40 <Vorpal> fizzie, -landet is a strange ending
19:53:06 <Vorpal> and -holmen translates to "island" too I think. Basically same idea anyway
19:53:25 <nooga> norway/1 = norway, doesn't it?
19:53:34 -!- wareya has quit (Read error: Connection reset by peer).
19:54:13 -!- wareya has joined.
19:55:14 <Vorpal> nooga, depends on what / is. And if it *is* division, what if it is floating point: who knows, especially on pentiums
19:58:26 <nooga> fizzie: outokumpu
19:58:48 -!- Wamanuz2 has joined.
19:59:03 <fizzie> nooga: Lit. translated: "weird bump".
19:59:26 <fizzie> (Or maybe "weirdhill", but I like the word "bump" more.)
19:59:54 <fizzie> Wiktionary says (of kumpu): "A hummock, hillock (small natural elevation of earth, generally smaller than kukkula (“hill”), larger than kumpare (“knoll”))."
20:00:12 <fizzie> (And "outo" is weird, strange.)
20:00:24 <fizzie> Strangehill sounds like someone's name.
20:01:57 -!- Wamanuz has quit (Ping timeout: 252 seconds).
20:12:28 -!- augur has quit (Ping timeout: 276 seconds).
20:14:08 -!- poiuy_qwert has joined.
20:20:03 <Phantom_Hoover> fizzie, my sister's P1 teacher was called Mrs Strange.
20:22:03 <fizzie> I just somehow was reminded of Mr Underhill.
20:24:15 -!- FireFly has quit (Quit: swatted to death).
20:24:53 -!- FireFly has joined.
20:29:24 <alise> bacl
20:29:56 <alise> ais523: I had a Nethack bot idea.
20:30:25 <ais523> oh dear
20:30:42 <fizzie> ais523: Are you saying "oh dear" just to be negative?
20:31:12 <ais523> it's a typical response to ideas
20:31:14 <alise> ais523: is that an "oh dear, I hope it doesn't manage to find a way to recursively self-improve and take over the world?"
20:31:23 <alise> because if so, thanks for the compliment!
20:31:48 <ais523> heh
20:31:58 <alise> ais523: are you interested in hearing it? it's okay if not
20:32:06 <ais523> yes
20:34:48 <alise> ais523: the idea is basically to be stupid. The bot knows how to move around, and what keys are actually valid actions; it knows how to respond to prompts, and it knows how to use menus and the inventory. But... nothing else. It has no idea that q quaffs, for instance. How does it find out? Simple! Everything is quantified, like in Planar, into a resource: HP, gold, etc. Things which are bad are resources that are always negative: for instance, every item w
20:34:48 <alise> ould have a negative inventory-space, every action would have a negative turns-gained, and so on. To start with, apart from when it wants to move somewhere, it acts completely at random. Every time it chooses an action, it notes some subset -- I'm not sure which yet -- of its environment, and the effects it had on all resources.
20:35:18 <alise> Then it becomes a case of "each turn, sort all possible actions by how richer they make me; if two items are equal, sort them randomly. Then pick the top action, and do it."
20:35:19 <ais523> hmm, you're trying to come up with a new AI algo, based on learning?
20:35:26 <alise> ais523: Pretty much. But the key here is:
20:35:28 <ais523> the issue is that what an action does depends on context so much
20:35:32 <alise> It would save this state in a file.
20:35:38 <ais523> well, that's obvious...
20:35:38 <alise> So it could die, and play again, and die, and play again, and learn.
20:35:46 <alise> ais523: well, no, because it's not obvious that it persists across games
20:35:49 <alise> which is the important thing
20:36:00 <ais523> learning algos normally persist...
20:36:13 <alise> It's also not obvious that it does any learning :)
20:36:14 <alise> Anyway, whatever.
20:36:31 <alise> ais523: I'm not sure what context to give.
20:36:43 <ais523> neither am I
20:36:47 <ais523> which is why it's a problem
20:37:06 <alise> I think perhaps a neighbourhood -- say two to three squares out? -- of cells, plus any special states: hunger, burden, and so on.
20:37:12 <alise> Oh, and for things like quaffing it would note what object it was, of course.
20:37:39 <alise> Of course, this is fundamentally probabilistic in the end; it may do something obviously stupid simply because it was never stupid before, and that's unavoidable. Still, stupid heuristics have surprisingly high success rates, I find...
20:38:15 <alise> ais523: psht, TAEB cheats at Telnet negotiation
20:38:24 <alise> it should clearly use a CPAN Telnet module which makes every part of it object-oriented
20:38:31 <alise> and tell that object what it will and won't do
20:38:38 <ais523> alise: IO::Telnet::HalfDuplex is on CPAN...
20:38:46 <ais523> admittedly, it was invented for TAEB, but it's still useful
20:38:52 <alise> however:
20:38:59 <alise> http://github.com/sartak/TAEB/blob/master/lib/TAEB/Interface/Telnet.pm
20:39:04 <alise> erm
20:39:05 <alise> http://github.com/sartak/TAEB/blob/master/lib/TAEB/Interface/Telnet.pm#L149
20:39:10 <alise> github.com/sartak/TAEB/blob/master/lib/TAEB/Interface/Telnet.pm#L149
20:39:14 <alise> it cheats at the actual negotiation
20:41:14 <alise> ais523: do you think the architecture has any hope of working:
20:41:16 <alise> *working?
20:41:23 <Vorpal> I got this strange thing from rosegarden "libgcc_s.so.1 must be installed for pthread_cancel to work"
20:41:24 <alise> obviously it'd have to be a melee fighter, not a wizard or anything
20:41:27 <Vorpal> that file exists
20:41:28 <alise> since that requires too large a range
20:41:29 <Vorpal> so wtf XD
20:41:54 <Vorpal> rosegarden actually works
20:41:57 <alise> I think if you ran it for a month or two, automatically respawning, it could do something sane.
20:42:00 <Vorpal> it just tells me that at exit
20:42:07 <alise> Oh, you'd have to have level-increase as a positive resource.
20:44:20 <alise> ais523: I think the chances of it ever ascending are zero, but I bet it could figure out the mines.
20:44:28 <alise> Sokoban... mmmmmmmnope.
20:44:38 <alise> Unless you cheated and had a built in resource called made-sokoban-closer-to-complete.
20:44:44 <alise> Then it might be able to learn.
20:47:23 <alise> hmm
20:47:28 <alise> I wonder if Moose is more comprehensive than CLOS?
20:47:34 <alise> It's inspired in many ways by CLOS, so...
20:47:53 <alise> It might just be equal.
20:48:22 -!- poiuy_qwert has quit (Quit: Leaving).
20:49:00 <Vorpal> <ais523> alise: IO::Telnet::HalfDuplex is on CPAN... <--- half duplex???
20:49:10 <Vorpal> can telnet operate in anything except full duplex?
20:50:28 <alise> Clearly.
20:50:42 <Vorpal> alise, I'm just so utterly astonished!
20:51:16 <Vorpal> alise, since tcp inherently provides full duplex it seems very strange
20:52:55 <Vorpal> alise, you tried gobolinux iirc?
20:53:05 <alise> Vorpal: Yeah. NixOS is cooler.
20:53:06 <Vorpal> alise, what was your opinion on it?
20:53:17 <alise> It's like GoboLinux but rollbackable! And awesome! And it has 64-bit support!
20:53:27 <Vorpal> alise, sure nixos is cooler, but I still wonder about gobolinux
20:53:29 <alise> And it doesn't use irritatingly sentence-cased directories!
20:53:50 <Vorpal> alise, sentence cased? As in "Binaries"?
20:53:51 <alise> Vorpal: Well... it's better than the usual organisation, but the community just isn't there to make it work nicely, you know? So it lags behind.
20:53:56 <alise> NixOS is more actively developed and, moreover, cooler.
20:54:01 <alise> Vorpal: /Programs/KDE-Libs/
20:54:07 <Vorpal> alise, have you tried nixos more than just livecd?
20:54:17 <alise> Vorpal: I'm going to install it if all goes according to plan
20:54:20 <alise> besides, it ships stock packages
20:54:28 <alise> and i already now how Nix itself works
20:54:32 <alise> so i already know what it is :P
20:54:35 <Vorpal> alise, ah, figured out issue with complex configs in etc?
20:54:39 <Vorpal> I didn't find from docs
20:54:47 <alise> configuration is done in configuration.nix
20:55:04 <Vorpal> alise, yes sure, and there was that sshd.start or whatever
20:55:09 <alise> and filesystems
20:55:18 <alise> i'm pretty sure
20:55:22 <alise> erm
20:55:24 <alise> forget that last line
20:55:31 <Vorpal> alise, but what if you want to set sshd configs like:
20:55:31 <alise> was writing something but train of thought got derailed
20:55:38 <Vorpal> Match User anoncvs
20:55:44 <Vorpal> X11Forwarding no
20:55:51 <Vorpal> AllowTCPForwarding no
20:55:59 <Vorpal> other complex options
20:56:02 <alise> Vorpal: I believe that will also be handled.
20:56:10 <alise> I can take a peek at the manual to see.
20:56:23 <Vorpal> alise, I checked the manuals on the website, and the wiki, and found nothing
20:56:29 <alise> Vorpal: It is probable that many packages still have regular configuration files somewhere.
20:56:42 <alise> Also, #nixos does exist, you know, and they are friendly. They're Haskell folk.
20:56:55 <alise> http://hydra.nixos.org/build/639152/download/1/nixos/manual.html "Example 4.1" has complex configuration
20:57:03 <alise> with a cron job, for instance
20:57:05 <Vorpal> alise, basically, I would consider the whole thing flawed beyond use if you can't do everything that you can do in normal sshd_config
20:57:11 <alise> "shows the configuration file for the locate service"
20:57:14 <Vorpal> sure you may say most doesn't care, I do however
20:57:22 <alise> of course you can. don't be an idiot.
20:57:32 <Vorpal> alise, good
20:57:37 <alise> :P
20:57:47 <alise> if you couldn't, I imagine they'd just make it use the default config file format, for instance
20:57:52 <alise> boot.extraModprobeConfig
20:57:52 <alise> Any additional configuration to be appended to the generated modprobe.conf. This is typically used to specify module options. See modprobe.conf(5) for details.
20:57:52 <alise> Default: ""
20:57:52 <alise> Example: "options parport_pc io=0x378 irq=7 dma=1\n"
20:57:52 <alise> there
20:58:01 <alise> so you can always, e.g., just say "oh and here's some strings to append"
20:58:33 <Vorpal> alise, I have custom ports and interfaces to listen to, and a lot more
20:58:38 <Vorpal> for sshd
20:59:48 <alise> https://svn.nixos.org/viewvc/nix/nixos/trunk/modules/services/networking/ssh/sshd.nix?revision=23872 appears to not have an append option, but i presume that's just because it's early days -- and nothing at all stops you from just making your own version of that file with such an option
21:00:26 <Vorpal> alise, there are a few config files that are often customised a lot and are very expressive. HTTPd and DNSds come to mind. As do radvd, ipsec stuff and sshd
21:00:51 <alise> s write your own service file that just writes the right config file :P
21:00:52 <alise> *so
21:01:00 <alise> from a config setting
21:01:11 <Vorpal> mhm
21:01:34 <Vorpal> alise, does the config do heredocs? sshd config would be non-trivial without that since it uses indention for blocks
21:01:46 <alise> '''...'''
21:01:51 <alise> erm
21:01:54 <alise> ''...''
21:02:00 <alise> as seen in https://svn.nixos.org/viewvc/nix/nixos/trunk/modules/services/networking/ssh/sshd.nix?revision=23872
21:02:08 <Vorpal> ah nice
21:02:26 <Vorpal> anyway yeah that file is too limited for me
21:02:57 <Vorpal> alise, for now I stay on arch, but nixos seems like an interesting distro, especially when it matured a bit
21:03:36 <alise> it's existed for two years or so :P
21:03:41 <alise> but yeah, early days
21:03:45 <alise> since it started out as academic
21:03:49 <Vorpal> arch is quite mature in a paradoxical way.
21:04:03 <Vorpal> On one hand it uses bleeding edge software, on the other hand the distro itself, the configuration stuff, and the community is quite mature
21:04:21 <alise> 19:17:11 <SgeoN2> flash keeps crashing
21:04:22 <alise> 19:17:21 <SgeoN2> I wonder if it's an old version
21:04:24 <alise> no, it just does that
21:04:32 <alise> Vorpal: the community is immature :)
21:04:35 <alise> sure mature as in age but the people...
21:04:41 <Vorpal> alise, depends which part of it
21:04:45 <ais523> <Vorpal> can telnet operate in anything except full duplex? <-- actually yes, but it doesn't work properly with curses
21:05:01 <Phantom_Hoover> Duplex?
21:05:08 <Vorpal> ais523, why? telnet runs over tcp, tcp provides full duplex
21:05:11 <ais523> but IO::Telnet::HalfDuplex is designed to take a full-duplex telnet connection
21:05:15 <ais523> and make it half-duplex
21:05:20 <Vorpal> err
21:05:21 <Vorpal> what?
21:05:26 <ais523> by working out when the other side has finished sending
21:05:30 <Vorpal> ah
21:05:42 <Vorpal> Phantom_Hoover, google
21:06:02 <ais523> Phantom_Hoover: "sending and receiving", in this concept
21:06:04 <Phantom_Hoover> Two-way?
21:06:06 <Phantom_Hoover> Ah.
21:06:09 <ais523> full duplex does both at once, half duplex does one at the time
21:06:10 <ais523> *in this context
21:06:22 <Vorpal> Phantom_Hoover, how tricky would it have been to use google
21:06:27 <Phantom_Hoover> What about null-duplex?
21:06:54 <Vorpal> if it had any meaning I think it would be "disconnected"
21:07:28 <Phantom_Hoover> Vorpal, oh what a tangled web we weave, when first we practice to use google
21:08:32 <Vorpal> alise, btw I looked at urbit and my first reaction was "wtf", the next one was "okay, the guy who designed that is a theorist, something like a+b looks like it would need O(n)..."
21:09:09 <Phantom_Hoover> Urbit?
21:09:15 <alise> yeah, well, you are manifestly not the target audience.
21:09:17 <Vorpal> alise, if I was about to design a language to actually be practically usable I wouldn't start out with lambda calculus and try to write an optimising compiler for it.
21:09:29 <alise> Phantom_Hoover: a manifesto in two parts:
21:09:30 <Vorpal> well sure it isn't exactly lambda calculus
21:09:31 <alise> http://moronlab.blogspot.com/2010/01/nock-maxwells-equations-of-software.html
21:09:32 <alise> http://moronlab.blogspot.com/2010/01/urbit-functional-programming-from.html
21:09:34 <Vorpal> but about as efficient
21:09:37 <Vorpal> as far as I can tell
21:09:38 <Phantom_Hoover> Vorpal, you have no VISION
21:09:39 <alise> Vorpal: it isn't even vaguely like the lambda calculus.
21:09:49 <alise> Phantom_Hoover: he's right in that addition is O(n) or something like it.
21:09:51 <alise> anyway, read the two posts.
21:09:57 <alise> the networking stuff is close to bang on. the rest is at least mind-expanding.
21:10:06 <Phantom_Hoover> alise, using the standard unary definition?
21:10:07 <Vorpal> alise, and multiplication is about O(n^2) at least
21:10:11 <Vorpal> if not worse
21:10:29 <alise> erm it's decrement that's difficult
21:10:34 <alise> Phantom_Hoover: it has integers built in; anyway, just give them a read :P
21:10:45 <Vorpal> alise, and yes, decrement I have no clue how to do
21:10:50 <Vorpal> keep a list of integers?
21:10:52 <Vorpal> or something
21:11:05 <alise> considering his challenge for basic competency is to construct decrement, you're definitely losing
21:11:07 <Phantom_Hoover> Vorpal, does it use Church numerals or their approximants?
21:11:17 <alise> Phantom_Hoover: it has native integers.
21:11:29 <alise> Vorpal: he explains how to do it efficiently with jets in the Urbit post.
21:11:34 <Vorpal> Phantom_Hoover, not church numerals, but it has a successor function but no addition or such
21:11:43 <Vorpal> no predecessor iirc
21:11:57 <Vorpal> alise, so yeah, doing it in the compiler basically
21:12:02 <alise> anyway, an innovator sees an idea and thinks "in what ways is this interesting?", not "in what ways is this flawed?"
21:12:03 <alise> Vorpal: no.
21:12:09 <alise> Vorpal: i will field no further loaded questions on the subject.
21:12:11 <Vorpal> hm
21:12:14 <Vorpal> okay
21:12:17 <Phantom_Hoover> Vorpal, \(_,x) -> (x, S x) is the easiest way in the pure LC.
21:12:39 <Vorpal> Phantom_Hoover, this is not LC. I only said it looked about as efficient for practical implementation.
21:12:47 <Vorpal> which is "not very"
21:13:35 <Vorpal> alise, loaded? huh. Whatever.
21:13:50 <Vorpal> alise, and I think it is interesting. Just utterly impractical :P
21:14:00 <Vorpal> alise, as an esolang it would be quite cool
21:20:55 -!- augur has joined.
21:25:47 <nooga> what's LC?
21:26:32 <alise> lambda c
21:26:36 <alise> the rest is obvio
21:26:36 <nooga> ah
21:26:59 <nooga> i've got a chair hat
21:26:59 <Vorpal> alise, Lambda C99
21:27:00 <Vorpal> hm
21:27:09 <Vorpal> alise, or... lambda C++!
21:27:11 <Vorpal> XD
21:27:55 <nooga> my friend asked me did I ever coded in A and B languages
21:28:09 <nooga> i told him that i don't know such langs
21:28:19 <ais523> B does exist, at least
21:28:32 <nooga> and then he started talking total bullshit about them
21:28:42 <nooga> probably he was completely making everything up
21:28:48 <nooga> ""
21:28:50 <pikhq> There is no A, though.
21:29:00 <nooga> "A is even harder than C man"
21:29:00 <ais523> there probably is, but it's likely unrelated
21:29:04 <ais523> there's also a D
21:29:08 <ais523> and various other letters
21:29:09 <Vorpal> B was preceded by something else iirc?
21:29:13 <pikhq> Vorpal: BCPL
21:29:13 <Vorpal> B..
21:29:15 <ais523> yep, BCPL
21:29:15 <Vorpal> ah
21:29:18 <Vorpal> yeah that's it
21:29:22 <ais523> the new lang was called B so as to save three bytes
21:29:27 <Vorpal> XD
21:29:33 <ais523> which I think was a joke, even though memory was tight at the time
21:29:47 <pikhq> B is probably more familiar to you than BCPL.
21:30:02 <Vorpal> ais523, good thing we didn't get the zero length string as a language
21:30:11 <Vorpal> that would have been awkward
21:30:14 <pikhq> (it looks like a K&R C with even less type safety)
21:30:16 <ais523> great esolang name idea!
21:30:26 <Vorpal> true
21:30:37 <Vorpal> ais523, but how would we put it on the wiki
21:30:51 <Vorpal> ais523, btw why did you name feather "feather"?
21:30:53 <ais523> under a different name, probably
21:30:57 <ais523> and because it's lightweight
21:31:01 <Vorpal> ah
21:31:01 <ais523> compared to Smalltalk
21:31:07 <Vorpal> um
21:31:08 <nooga> ah
21:31:13 <nooga> so feather is smalltalky
21:31:19 <Vorpal> ais523, I can't say I have dealt enough with smalltalk to know how heavy it is
21:32:18 <pikhq> Vorpal: Light, except that it is on a virtual machine.
21:32:50 <Vorpal> pikhq, you can make lightweight virtual machines
21:33:01 <Vorpal> pikhq, and is there any specific reason you can't compile smalltalk to native code?
21:33:38 <Phantom_Hoover> Perhaps the same reason as Lisp?
21:33:55 <Vorpal> Phantom_Hoover, you can compile some lisps
21:34:01 -!- augur has quit (Ping timeout: 264 seconds).
21:34:04 <Phantom_Hoover> But not others.
21:34:40 <Vorpal> Phantom_Hoover, yes, scheme for example because you could suddenly redefine + to mean something else on the fly, which would hugely complicate compiling
21:35:24 <ais523> pikhq: Smalltalk needs a heavy standard library to function correctly
21:35:31 <ais523> or at least, in a smalltalky way
21:35:37 <ais523> that can't be implemented in Smalltalk itself
21:35:46 <ais523> things like class classes, etc
21:38:11 -!- Sgeo has joined.
21:38:31 <Vorpal> Phantom_Hoover, hm I should put up an image just to see how you will react...
21:38:35 <pikhq> Vorpal: Smalltalk is kinda set up assuming that it *is the system*. You *could* compile it, but it's not very Smalltalky.
21:38:40 <Vorpal> Phantom_Hoover, can you stay for maybe 5-10 minutes?
21:38:54 <Phantom_Hoover> Vorpal, huh?
21:39:00 <Phantom_Hoover> You confuse me.
21:39:21 <Sgeo> I may have to drag benuphoenix in here
21:39:42 <Phantom_Hoover> Sgeo, who?
21:44:57 <alise> ais523: did you answer my question about my "AI"? sorry, i've lost track of backlog
21:45:41 <Vorpal> Phantom_Hoover, I have verified that this doesn't crash my browser, can't know if it crashes your of course. It shouldn't. http://sporksirc.net/~anmaster/tmp/test.png
21:45:59 <Vorpal> after all it is just 16 kB
21:46:14 <Vorpal> actually 13
21:46:22 <Vorpal> (was using du -sh not du -bsh)
21:46:26 <alise> What an impressively stupidly empty image.
21:46:29 <Sgeo> No crash here
21:46:31 <Vorpal> alise, indeed
21:46:36 <alise> Was that really what you sent Phantom_Hoover two vague messages about?
21:46:37 <alise> Really?
21:46:38 <Vorpal> alise, i just want to test a theory
21:46:49 <Vorpal> alise, he said he crashed on my panoramas
21:47:03 <Vorpal> alise, I want to test if it is large image file or large image size that does it
21:47:14 <Vorpal> alise, anyway it isn't empy
21:47:25 <Vorpal> alise, check the two outermost pixles on left and right
21:47:53 <Sgeo> >Sgeo< CTCP VERSION
21:47:53 <Sgeo> -Sgeo- VERSION xchat 0.26.1 Linux 2.6.32-24-generic [i686
21:47:56 <Sgeo> Hmm...
21:48:06 <Vorpal> though
21:48:11 <Vorpal> Phantom_Hoover seems to have timed out eh?
21:48:12 -!- Phantom_Hoover_ has joined.
21:48:19 <alise> Sgeo: what?
21:48:19 <Vorpal> Phantom_Hoover_, so what happened?
21:48:22 <Phantom_Hoover_> Vorpal, I HOPE YOU'RE HAPPY YOU BASTARD
21:48:31 <Sgeo> The weirdly low-looking version
21:48:44 <Vorpal> Phantom_Hoover_, a few questions: What OS? What browser (and version)? How much RAM?
21:48:46 <alise> Sgeo: that's xchat-gnome's version?
21:48:50 -!- Phantom_Hoover has quit (Read error: Connection reset by peer).
21:48:50 <Sgeo> Yes
21:48:52 <alise> Phantom_Hoover_: you don't have to answer that. </lawyet>
21:48:53 <Vorpal> Phantom_Hoover_, I'm trying to help you here, diagnosing the cause!
21:48:54 <alise> Sgeo: *.
21:48:55 <alise> not ?
21:48:59 <Sgeo> But someone might think it's plain XChat
21:49:08 <alise> Sgeo: Oh noooooooooooo
21:49:22 <Sgeo> And think it's an absurdly low version
21:49:32 <alise> Oh noooooooooooooooooooooooooooo
21:49:33 <Vorpal> Phantom_Hoover_, I can see how it would cause issues if you have 128 MB RAM
21:49:37 <Vorpal> but anything modern? No
21:49:40 <Phantom_Hoover_> Vorpal, Ubuntu 10.04 64-bit. Firefox 3.6. 3GB.
21:49:55 <alise> <Vorpal> AH THE PROBLEM IS UBUNTU
21:50:15 <Vorpal> Phantom_Hoover_, then your system is broken. It works fine on my ubuntu of same version, 64-bit, Firefox 3.6 with just 2 GB RAM
21:50:24 <Vorpal> and it works fine on my desktop with 1.5 GB RAM
21:50:46 <Vorpal> Phantom_Hoover_, I tested in firefox. Firefox got slow for about 5 seconds on my old sempron CPU then showed the image
21:50:50 <alise> Yes. Your system is broken. That is the only explanation. You are inadequate.
21:51:32 <Vorpal> alise, yes because it doesn't happen to anyone else and he has more RAM than me (and RAM is the only real potential issue). This leaves firefox or the system being broken.
21:51:44 <Vorpal> Phantom_Hoover_, perhaps it is some firefox addon that causes this?
21:51:51 <alise> it used to happen to me.
21:52:12 <Vorpal> alise, do you remember what changed when it stopped happening?
21:52:30 <alise> nope.
21:52:53 <Vorpal> hm
21:53:25 <Vorpal> Phantom_Hoover_, it might be worth trying firefox in safe mode on that page
21:53:43 <Vorpal> that would be: firefox -safe-mode
21:54:06 <Sgeo> Is xtrlock acceptable?
21:54:07 <Sgeo> Maybe xlockmore
21:54:40 <Vorpal> Sgeo, what do they do?
21:55:06 <Vorpal> Sgeo, x screen lockers? then I suggest using whatever the DE uses by default or slock
21:55:15 <Sgeo> Or maybe I should just set a password
21:55:28 <Vorpal> Sgeo, you should have a password on your user of course
21:55:32 <Vorpal> I presumed you already did
21:55:43 <Sgeo> I'm using a LiveUSB-like thingy
21:55:49 <Sgeo> Maybe I should switch to Knoppix
21:56:01 <Sgeo> Might be easier to boot from CD and read data from USB
21:56:06 <Sgeo> With Knoppix
21:56:12 <alise> Sgeo: slock
21:56:18 <alise> Also, no, it's not.
21:56:20 <alise> Keep using Ubuntu.
21:57:09 <Sgeo> What does Ubuntu have that Knoppix doesn't, besides GNOME and installability?
21:57:12 <Sgeo> BRB
21:57:17 <Vorpal> Phantom_Hoover_, did you hear the disk access a lot during the freeze?
21:57:27 <Vorpal> Phantom_Hoover_, indicating swap trashing
21:57:31 <Sgeo> Ok, GNOME's lock screen thingy doesn't work
21:57:49 <Phantom_Hoover_> Vorpal, the disk access light was off.
21:58:03 <Vorpal> Phantom_Hoover_, hrrm
21:58:07 <Phantom_Hoover_> Well, it blinked occasionally, but not often.
21:58:17 <Phantom_Hoover_> (At least, I assume it's the disk access light.)
21:58:23 <Vorpal> Phantom_Hoover_, did sysrq work? could you switch to a terminal with ctrl-alt-F1?
21:58:50 <Vorpal> Phantom_Hoover_, I assume you tried both
21:58:54 <Phantom_Hoover_> Yes to the first, no to the second.
21:58:58 <Sgeo> alise, answer please?
21:59:22 <Vorpal> Phantom_Hoover_, hm
21:59:38 <Vorpal> Phantom_Hoover_, anyway, can you wget it and open it in gimp without issues? That shouldn't crash.
21:59:43 <alise> Sgeo: what?
21:59:48 <Vorpal> as in, there is no plausible way it could
21:59:50 <Phantom_Hoover_> Sgeo, surely those two things are pretty good.
21:59:56 <alise> What does Ubuntu have that Windows doesn't?
22:00:02 <alise> There is no reason to arbitrarily pick an inferior option merely because it is different.
22:00:27 <Vorpal> Phantom_Hoover_, oh and what GPU? I'm considering if it could be some obscure hw double buffer size thingy limit, but firefox probably doesn't do zooming in opengl, so unlikely
22:01:31 <Phantom_Hoover_> Vorpal, a) I didn't try to zoom; b) my GPU is whatever crap Intel glued to the motherboard.
22:01:37 -!- Sgeo|irssi has joined.
22:01:42 <Sgeo|irssi> i3lock != slock?
22:01:44 * Sgeo|irssi now hates alise
22:01:53 <alise> i never said i3lock
22:01:54 <Sgeo|irssi> alise: There was no slock package
22:01:56 <Sgeo|irssi> Just i3lock
22:02:01 <alise> So fucking compile it!
22:02:05 <Sgeo|irssi> Which was supposedly an improved version of slock
22:02:06 <alise> it's one make!
22:02:11 -!- trinithis has quit (Ping timeout: 240 seconds).
22:02:12 <Vorpal> Phantom_Hoover_, works for me on intel crap so hm
22:02:15 <Sgeo|irssi> X seems to be frozen
22:02:15 <alise> Anyway, what's wrong with it?
22:02:19 <alise> You type your password in the screen.
22:02:20 <Sgeo|irssi> How do I unfreeze i3lock?
22:02:24 <alise> You type your password.
22:02:34 <Vorpal> alise, i3lock != slock
22:02:38 <alise> it's a fork
22:02:39 <alise> of slock
22:02:42 <alise> i just looked it up
22:02:44 <Vorpal> ah
22:02:54 <Vorpal> Sgeo|irssi, there is no password prompt, type password, hit enter
22:02:56 <Sgeo> There was never a prompt to type it
22:02:57 <Sgeo> Ah
22:02:59 <Vorpal> Sgeo|irssi, that's all
22:03:05 <alise> ugh, i3lock uses pam. and supports backgrounds. and beeps on wrong passwords (pointless). and uses PAM
22:03:08 <alise> how crappy
22:03:28 <Phantom_Hoover_> What's so wrong with backgrounds?
22:03:31 <Vorpal> alise, pam has the advantage of supporting stuff like ldap
22:03:32 <Vorpal> and such
22:03:41 <alise> but pam sucks :P
22:03:47 <alise> Phantom_Hoover_: slock is minimalist code.
22:03:57 <Vorpal> alise, sure, but there is no other alternative that exists today
22:03:58 <alise> i know you've never heard of minimalism, but there you go
22:04:03 <alise> Vorpal: not using LDAP
22:04:09 <Vorpal> alise, true, kerberos then
22:04:16 <Phantom_Hoover_> alise, I've heard of minimalism!
22:04:19 <alise> Vorpal: not using kerberos
22:04:20 <alise> erm
22:04:21 <alise> Phantom_Hoover_: *Sgeo:
22:04:26 -!- trinithis has joined.
22:04:26 <alise> no wait
22:04:28 <alise> *Phantom_Hoover_:
22:04:29 <alise> whatever
22:04:31 <alise> you all suck
22:04:40 <Phantom_Hoover_> I resent that remark!
22:04:40 <Vorpal> I think pam is moderately useful
22:04:43 <Vorpal> but that's all
22:05:26 <Sgeo|irssi> ~/exit
22:05:28 -!- Sgeo|irssi has quit (Client Quit).
22:05:28 <alise> Phantom_Hoover_: *resemble
22:05:29 <Vorpal> alise, I'm sure you wouldn't want plain shadow auth if you had to admin a network of 200 workstations with hundred of users where everyone could log in at any computer
22:05:32 -!- sebbu2 has changed nick to sebbu.
22:05:35 <Vorpal> alise, say, university network
22:05:43 <Vorpal> alise, then you might see the point of network auth
22:05:43 <Phantom_Hoover_> alise, *represent
22:05:44 <Vorpal> :P
22:05:48 <alise> Vorpal: yeah, well, sysadminning is hell.
22:05:51 <alise> which is why only the stupid do it
22:05:59 <Vorpal> alise, of course
22:06:10 <Vorpal> alise, but I'm just saying *network auth is sometimes useful*
22:06:15 <alise> blah blah blah robots
22:06:28 <Vorpal> alise, and *there exists no other good alternative to pam currently for this*
22:06:44 <Vorpal> I'm not saying I like pam, I don't
22:06:48 <Vorpal> I think it is a mess
22:06:57 <alise> robots
22:06:59 <Vorpal> but yeah, until someone written an alternative it is useful
22:07:03 <Vorpal> alise, robots how?
22:07:04 <Sgeo> DOes ActiveDirectory do something similar?
22:07:06 <Sgeo> *Does
22:07:09 <alise> botorobotatics
22:07:20 <Vorpal> Sgeo, activedirectory is iirc ldap + MS extensions
22:09:06 <Vorpal> alise, PAM is very much like unix: it isn't perfect, but it works (of course, in many other senses it is very different, such as "keep it simple", pam doesn't)
22:09:21 <alise> very much like Windows in that way too.
22:09:27 <Vorpal> alise, indeed
22:09:34 <Vorpal> alise, very much like x86!
22:09:47 <Vorpal> alise, heck it is very much like mostly everything mainstream
22:10:37 <alise> 19:48:09 <SgeoN1> The reason I want it spun down is to avoid damaging it further
22:10:42 <alise> you can't spin it down and rescue it at the same time...
22:11:09 <Sgeo> alise, I gave up on rescuing it for the moment
22:11:44 <Vorpal> I will leave further disk rescue help to you alise. I will have a hard time explaining the indention on my face matching the couture of the front of my desk.
22:12:13 <alise> i defer future disk rescue help to Vorpal
22:12:17 <alise> how convenient
22:12:18 <Vorpal> Sgeo, you have just been acting too stupidly through all this to get any further help from me.
22:12:22 <Vorpal> now night →
22:12:57 <Sgeo> o.O
22:13:32 <Sgeo> I have another computer that I think I'm going to use for this
22:16:15 <alise> 21:52:22 <quintopia> what's the minimum one would need to do to make plain TeX turing-complete?
22:16:16 <alise> it already is
22:16:19 <Sgeo> My understanding is that I could, in fact, pause it
22:17:55 <Sgeo> ...
22:17:56 <Sgeo> alise?
22:18:32 <alise> you can, but doing it across machines is dumb. and your drive is losing data every time you touch it, which is why you try and avoid doing anything but recovering it as soon as it happens.
22:18:37 <alise> also, i defer to Vorpal.
22:18:39 <Phantom_Hoover_> Hypothesis: all operating system ads are so obnoxious that the rage of people watching them could be used for power. Discuss.
22:18:49 -!- MigoMipo_ has quit (Read error: Connection reset by peer).
22:19:14 <Sgeo> I didn't realize it would take that long
22:19:30 <Sgeo> Leaving it would have meant needing to pause it before Mondays and Wednesdays, and resuming when I got back
22:21:04 <Sgeo> I'm going to get ready to go home, I think
22:21:07 <Sgeo> Bye for now all
22:21:29 -!- Sgeo has quit (Quit: Ex-Chat).
22:22:10 <alise> Phantom_Hoover_: There are several people rolling in their graves right now; that's more efficient.
22:22:15 <alise> I forget who my main idea was.
22:26:18 <Phantom_Hoover_> Is there a non-obnoxious OS ad?
22:26:38 <Phantom_Hoover_> i.e. one which actually makes you think "hey, I might install this."
22:29:09 <alise> no
22:29:21 <alise> well some of the proposed linux ones were funny but they didn't win
22:31:18 <Phantom_Hoover_> Link?
22:33:55 <alise> long lost it by now
22:36:48 * Phantom_Hoover_ → sleep
22:36:51 -!- Phantom_Hoover_ has quit (Remote host closed the connection).
22:47:09 -!- tombom has quit (Quit: Leaving).
22:53:35 -!- augur has joined.
23:01:40 <nooga> welll
23:03:31 <nooga> ia Atat a nice name?
23:09:53 <alise> for?
23:11:43 <alise> ais523: is "Student-Friendly Summary" allowed on Wikipedia?...
23:11:46 <alise> http://en.wikipedia.org/wiki/Standard_conditions_for_temperature_and_pressure
23:11:54 <alise> sounds like Here We Do Your Homework In The First Paragraph
23:12:10 <nooga> for a library
23:12:28 <alise> nooga: doing?
23:12:57 <nooga> regex
23:13:44 <alise> nooga: sane style or stupid style?
23:14:51 <nooga> stupid
23:15:15 -!- cpressey has joined.
23:18:15 <ais523> alise: that doesn't really sound like it fits the style guides...
23:18:24 <ais523> but I haven't read them for years
23:18:36 <alise> the link has it; I'm just pondering whether to remove it with a chainsaw or not
23:18:48 <alise> [[Standard conditions for atmosphere and pressure]]
23:38:07 <pikhq> *sigh* the US.
23:38:18 <nooga> budzilla
23:38:20 <pikhq> Where auto ownership is nearly mandatory.
23:53:33 <cpressey> And yet radio ownership continues to be optional. Hard to conceive, isn't it.
23:53:40 -!- cpressey has quit (Quit: leaving).
23:55:53 <ais523> actually, there's a proposed law in the US government (I'm not sure which branch, offhand) to make FM radios mandatory on mobile phones
23:56:01 <ais523> I doubt it'll get very far, it's so ridiculous...
23:59:36 <pikhq> ais523: The RIAA was advocating that the FCC mandate it.
23:59:49 <pikhq> Yes, the RIAA, bastion of retardedness.
23:59:50 <ais523> heh
←2010-09-19 2010-09-20 2010-09-21→ ↑2010 ↑all