00:03:05 <hagb4rd> hi! i need some help with a ballistic calculation. i've this formula: Range = (InitialVelocity^2 x sine (2 x Angle)) / g ..now i need to modify it, so i can calculate the range so it works if shot from a specified height. can you help?
00:03:06 <hagb4rd> Read more: How to Calculate the Distance of a Projectile | eHow.com
00:03:32 <hagb4rd> ignore the read more thing
00:07:08 <quintopia> the flight time of a projectile is -.5*g*t^2+V_0*sin(theta)*t+h_0=0
00:07:32 <quintopia> so, apply quadratic formula, using the plus solution
00:09:26 <quintopia> the range is V_0*cos(theta)*t (where t is calculated as previously described)
00:18:08 <Sgeo> kallisti_, update
00:24:34 -!- Friendship has set topic: My Little Stallman: Freedom is Magic | So good to be in the Morning....In America! | http://codu.org/logs/_esoteric/ | http://esolangs.org/wiki/ has moved servers!.
00:35:15 -!- derdon has quit (Remote host closed the connection).
00:57:04 -!- smartinsrenato has joined.
00:57:10 -!- MoALTz has joined.
00:57:56 -!- smartinsrenato has left.
01:08:25 -!- cheater__ has joined.
01:09:56 -!- cheater_ has quit (Read error: Operation timed out).
01:24:42 <zzo38> Someone suggested for me to make roguelike game once; maybe I could make up "Extensible-Roguelike", using Haskell, and extensible-data and hampp. If you want to add rule variants, spells, kind of creatures, classes, items, areas, macros, etc, then you can just add a file.
01:24:50 <zzo38> And then recompile the program.
01:25:05 <zzo38> Rather than having to adjust the entire program to work with any changes you have made.
01:26:26 -!- zzo38 has quit (Quit: NO SEND).
01:28:04 -!- olsner has quit (Quit: Leaving).
01:40:10 <itidus20> to be honest, i don't like academization of what can loosely be termed scripting languages. I am just glad that you can't patent a grammatical structure in english.
01:40:36 <itidus20> An author could rise to prominence by being the only one allowed to split an infinitive.
01:41:20 <itidus20> You could buy a license to legally split infinitives from him.
01:41:59 <Friendship> He'd make millions from Star Trek royalties alone.
01:43:14 <oerjan> A different patent, getting royalties from Star Wars would require.
01:46:14 <itidus20> Now heres something to consider.
01:46:27 <itidus20> If you have a machine which translates english into algorithms
01:46:42 <itidus20> can you get away with coding all your software in english and dodge patents?
01:46:55 * oerjan demands royalties from itidus20 for violating his apostrophe omission patent
01:47:00 <itidus20> i guess not.. i am not thinking it through
01:47:40 <itidus20> or.. if you speak and a machine nearby with a microphone and voice recognition just happens to overhear you
01:47:56 <itidus20> and just happens to process your speech and perform some calculations
01:48:14 <itidus20> then if they outlaw that perhaps freedom of speech is being challenged
01:49:38 -!- Sgeo has quit (Read error: Connection reset by peer).
01:54:09 -!- Sgeo has joined.
02:03:47 <Friendship> kallisti_: Unlike Ubuntu, Debian sticks with the classique root-password system by default.
02:04:00 <Friendship> If you install sudo, you have to add users to the "sudo" group for them to have Ubuntu-like sudo privs.
02:04:18 -!- tzxn3 has quit (Read error: Connection reset by peer).
02:04:26 <Friendship> And no, of course it doesn't have compilers installed by default.
02:04:46 <kallisti_> actually I'm installing emacs first. :P
02:05:58 <kallisti_> this is good though because I'll probably pick up a few sysadmin things along the way
02:06:16 <kallisti_> I should run Arch or Gentoo or something for a crash course, I think.
02:09:11 <kallisti_> oh apparently I'm supposed to use visudo..
02:14:30 <pikhq_> Strictly speaking not necessary... visudo locks the sudoers file and performs syntax checks afterwards.
02:25:32 -!- augur has joined.
02:26:19 <kallisti_> Friendship: so basically I just add myself to the sudo group....?
02:28:24 * kallisti_ is curious as to what happens when he tries to edit sudoers with something other than visudo REBELLIOUSLY AGAINST THE RULES OF THE SUDOER FILE.
02:32:20 * pikhq_ notes that spaghetti carbonara = delicious
02:33:49 <kallisti_> yeah so I added myself to sudo but sudo complains that I'm not in the sudoers file...
02:36:30 <kallisti_> /etc/group confirms that I'm there...
02:38:23 <pikhq_> Your currently running login probably isn't in the sudo group.
02:38:32 <pikhq_> Groups are a per-process property.
02:38:53 <pikhq_> /etc/group, IIRC, is just used by /bin/login to set your initial groups.
02:39:22 <kallisti_> that means you can't change someone's group mid session...
02:42:58 -!- augur has quit (Remote host closed the connection).
03:14:13 <kallisti_> pikhq_: U MEAN LINUX NOT PERFECT OS?????
03:20:42 -!- oerjan has quit (Quit: Good night).
03:25:14 <Sgeo> kallisti_, tswett has been UIPDATED
03:26:08 <kallisti_> okay so I've added xmonad stuff to my .xsession
03:26:15 <kallisti_> but Debian apparently doesn't execute that by default?
03:28:04 <tswett> kallisti_: Sgeo has been WILCOED.
03:28:41 * Sgeo has no idea what that means >.>
03:38:29 -!- augur has joined.
03:45:20 <kallisti_> yeah so one of the tests of syb runs out of memory
03:48:33 -!- augur has quit (Remote host closed the connection).
03:53:48 <kallisti_> there's something strangely pleasant about configuring this all by hand
03:58:27 -!- lifthrasiir has quit (Ping timeout: 252 seconds).
03:58:59 -!- lifthrasiir has joined.
04:00:09 <kallisti_> would there be any reason not to move /root/bin to some other bin in my default PATH?
04:01:18 <kallisti_> if only elliott were here. he would tell me how horrible of an idea that is.
04:01:57 <itidus20> write it in recursive handwriting
04:03:25 -!- TeruFSX has joined.
04:05:28 <kallisti_> also would there be a reason not to copy my /etc and home directories from this install to my real install in order to speed up the setup time?
04:09:32 <kallisti_> I guess I could just copy my bash history and run most of that verbatim. :P
04:10:00 -!- monqy_ has joined.
04:10:18 -!- monqy has quit (Disconnected by services).
04:10:23 -!- monqy_ has changed nick to monqy.
04:11:59 <kallisti_> here's what I've been doing for the past 30 minutes:
04:12:10 <kallisti_> find out I'm missing some kind of *-dev package
04:12:17 <kallisti_> apt-get install that. wait for it install like 30 packages.
04:22:35 -!- MoALTz has quit (Quit: brb).
04:23:06 <kallisti_> must be like 5 lines of code in each one. :P
04:23:34 <kallisti_> oh wait I think this xmonad-contrib now. nevermind
04:24:19 -!- MoALTz has joined.
04:34:42 <kallisti_> normally, there's a .cabal directory
04:34:50 <kallisti_> but... I can't find cabal stuff anywhere.
04:34:53 -!- Jafet1 has joined.
04:35:02 -!- zzo38 has joined.
04:35:40 -!- zzo38 has set topic: My Little Stallman: Freedom is Magic | So good to be in the Daytime....In America! | http://codu.org/logs/_esoteric/ | http://esolangs.org/wiki/ has moved servers!.
04:37:45 <kallisti_> okay, so... apparently sudo installs things in /root
04:38:42 <kallisti_> ugh, how do I get Debian's sudo to run exactly like Ubuntu's. :P
04:39:47 -!- Jafet has quit (Read error: Connection reset by peer).
04:47:26 -!- calamari has joined.
05:10:12 -!- Jafet1 has changed nick to Jafet.
05:14:36 -!- atehwa has quit (Ping timeout: 252 seconds).
05:14:42 -!- atehwa has joined.
05:19:02 -!- augur has joined.
05:35:09 -!- Sgeo has quit (Ping timeout: 240 seconds).
05:43:54 -!- calamari has quit (Quit: Leaving).
05:46:55 -!- itidus21 has joined.
05:49:30 -!- itidus20 has quit (Ping timeout: 276 seconds).
05:53:00 -!- MoALTz_ has joined.
05:55:49 -!- MoALTz has quit (Ping timeout: 244 seconds).
06:03:43 -!- Sgeo has joined.
06:04:07 -!- PiRSquared17 has changed nick to area|away.
06:07:30 -!- augur has quit (Remote host closed the connection).
06:08:50 -!- calamari has joined.
06:16:07 -!- itidus20 has joined.
06:18:33 -!- itidus21 has quit (Ping timeout: 244 seconds).
06:22:01 -!- TeruFSX has quit (Read error: Connection reset by peer).
06:22:27 -!- TeruFSX has joined.
06:23:15 -!- itidus20 has quit (Read error: Connection reset by peer).
06:43:11 -!- calamari has quit (Quit: Leaving).
07:12:30 <zzo38> I was thinking about my ideas for Ibtlfmm programming language (I might eventually write a specification document, and/or a compiler, and/or a tag in Phlogjournal for stuff I occasionally write about it). One thing is that unlike Haskell, a class can be many things; it can be: a type signature of a method, a tag, a type family, a mathematical law, or a combination of classes.
07:14:52 <zzo38> If you need to make methods into separated classes, it already is; if you want to define superclasses afterward you can do so if it can be automatic; if something can be equivalently defined using one function or using multiple others, it can be added on later and work both ways (such as a monad using bind or join/fmap); etc
07:36:41 <zzo38> Do you have opinion about this?
07:53:44 -!- augur has joined.
07:58:20 <zzo38> Idea of kinds is * for ordinary inhabited types, # for raw types (using LLVM or whatever is going to be used), + for natural number types, & for constraints, @ for program modules, and -> as the function between kinds. User-defined kinds have alphanumeric names.
08:31:51 -!- MoALTz_ has quit (Ping timeout: 244 seconds).
08:32:43 -!- Phantom_Hoover has joined.
08:32:49 -!- Phantom_Hoover has quit (Changing host).
08:32:49 -!- Phantom_Hoover has joined.
08:35:39 -!- MoALTz has joined.
08:40:50 <zzo38> We can have many levels of meta-programming available.
08:57:28 -!- zzo38 has quit (Remote host closed the connection).
09:33:30 -!- Phantom_Hoover has quit (Remote host closed the connection).
09:38:17 -!- Phantom_Hoover has joined.
09:47:53 -!- NihilistDandy has joined.
10:08:37 -!- itidus21 has joined.
10:10:43 <itidus21> Basic Input/Output Commander guy is at work on his page :-D
10:10:50 -!- kmkr has joined.
10:12:41 <itidus21> he encoded a display of the lime slices in BIOC :-D
10:13:15 <kmkr> @PiRSquared17 i happened to read logs and to answer your question, i think i did that "!!!attention!!!" thing a few times to get people roused. although it was probably useless because i was largely here at wrong times and nobody was around.
10:13:45 <kmkr> damn. is @ doing something?
10:14:03 <kmkr> to PiRSquared17: i happened to read logs and to answer your question, i think i did that "!!!attention!!!" thing a few times to get people roused. although it was probably useless because i was largely here at wrong times and nobody was around.
10:14:25 <kmkr> i admit that it's annoying as f::: and i wouldn't do so today, 7 years later.
10:16:10 <kmkr> well, one learns better... anyway. now when i'm here, i want to thank everyone for making truth-machines, it's truly great to see so many of them in languages i didn't know almost anything about.
10:21:01 <kmkr> this is what i meant, i'm always timing my visits when nobody else is here... :\
10:25:37 <kmkr> anyway, i could just as well inquire if anyone has tried "hollow"? i find the language surprisingly expressive and making a self-interpreter is something that's on my list.
10:27:51 <kmkr> an esolanguage from 2005
10:29:03 <mRoman> Stlang has pattern matching, anonymous functions and sadly no constructors no destructors
10:29:17 <mRoman> although I can implement constructors in 20 seconds.
10:29:54 <mRoman> destructors are way more complicated.
10:30:19 <mRoman> python has destructors
10:30:29 <mRoman> and Objects are Pythonobjects, so that wouldn't be a problem
10:35:33 -!- Ngevd has joined.
10:41:45 <mRoman> http://codepad.org/5aGioOj3
10:47:19 <kmkr> by the way, has anyone managed anything in sortle? it's an esolang from some years back that i think might be worth some interest. i've tried to take a crack at it a few times but haven't managed anything worth mentioning.
10:48:56 -!- Vorpal has joined.
10:49:04 <kmkr> i have a feeling there might be ways to program in it even if it might get complicated like /// (which i still don't get...)
10:49:04 -!- NihilistDandy has quit (Read error: Connection reset by peer).
10:49:49 -!- itidus20 has joined.
10:52:23 -!- itidus21 has quit (Ping timeout: 255 seconds).
10:58:44 -!- Ngevd has quit (Ping timeout: 245 seconds).
11:05:41 -!- kmkr has quit (Ping timeout: 245 seconds).
11:09:09 -!- pikhq_ has quit (Ping timeout: 240 seconds).
11:09:18 -!- pikhq has joined.
11:12:46 -!- derdon has joined.
11:20:52 -!- monqy has quit (Quit: hello).
11:23:30 -!- Taneb has joined.
11:26:19 <mRoman> http://codepad.org/AWE5cttL <- good luck finding a hello world programm :D
11:28:07 <Taneb> Good luck being brickbrained by Phantom Hoover
11:45:17 -!- tzxn3 has joined.
11:53:32 -!- Taneb has quit (Quit: Page closed).
12:01:26 -!- olsner has joined.
12:05:54 <mRoman> hm. there's a bug anyway
12:19:35 -!- MoALTz has quit (Ping timeout: 252 seconds).
12:30:33 -!- derdon has quit (Remote host closed the connection).
12:33:42 <HackEgo> mRoman: 806) <Phantom_Hoover> BF derivatives are a cancer running throughout the fringes of the esolang community, and as the fringes vastly outweigh the core, we're screwed.
12:36:14 <fizzie> There seem to be no quotes about the bricks, but it's the usual response.
12:37:04 <fizzie> <ais523> <Star651, via email> Is this a spambot or something? I just got a user talk page message from Phantom Hoover about the BF-inspired languages, the "brick brain brochure."
12:38:23 <HackEgo> http://codu.org/projects/hackbot/fshg/index.cgi/raw-file/tip/paste/paste.23910
12:52:02 <Phantom_Hoover> mRoman, HW is trivial, BtW; you can pad the program out with empty loops to set the iptr.
12:53:13 <mRoman> I actually don't know what I've thought when I wrote that ;)
13:04:36 <mRoman> Friendship: Glass has no inheritance. At least nothing like that is mentioned in the article.
13:19:41 -!- hagb4rd has quit (Quit: Nettalk6 - www.ntalk.de).
13:20:46 -!- area|away has changed nick to PiRSquared.
13:26:58 <Friendship> It has no typing, so you can always duck it.
13:29:03 <mRoman> Stlang has also no typing, which is why you can cast a class in any other class
13:32:33 <mRoman> Stlang has no builtin classes
13:32:39 <mRoman> but a lot of powerful builtin functions.
13:35:25 <mRoman> But what exactly do you wanna know?
13:42:30 -!- augur_ has joined.
13:43:33 -!- augur has quit (Read error: Connection reset by peer).
13:58:12 -!- yiyus has quit (Ping timeout: 245 seconds).
13:58:52 -!- yiyus has joined.
14:07:28 -!- itidus21 has joined.
14:09:03 -!- Ngevd has joined.
14:09:23 -!- itidus20 has quit (Ping timeout: 240 seconds).
14:09:57 -!- Ngevd has changed nick to Taneb.
14:40:15 -!- Taneb has quit (Quit: Goodbye).
14:40:27 -!- Taneb has joined.
14:42:26 -!- yiyus has quit (Ping timeout: 244 seconds).
14:47:49 -!- yiyus has joined.
14:48:22 <mRoman> 134 builtin-functions yet, to be precisely.
14:48:45 <mRoman> so probably around 100.
14:53:35 <kmc> oh man, this is going to be impossible to debug
14:54:01 <kmc> i've written an interpreter for a tiny fexpr language in Scheme
14:54:13 <kmc> and then some library code for that language to make it enough like Scheme that it can run said interpreter
14:54:32 <kmc> but it's broken.
15:06:28 -!- MoALTz has joined.
15:13:32 <olsner> kmc: if you need to debug you have already lost
15:17:26 <kmc> otoh a metacircular evaluator serves as its own test suite *and* reference implementation
15:17:29 <kmc> so that's kinda nice
15:20:54 <kmc> i think i found the bug
15:22:40 -!- Phantom_Hoover has quit (Read error: Connection reset by peer).
15:24:09 -!- Phantom_Hoover has joined.
15:47:21 -!- kmc has quit (Quit: Leaving).
15:47:37 -!- kmc has joined.
16:03:14 -!- kmc has quit (Quit: Leaving).
16:03:33 -!- kmc has joined.
16:18:30 -!- Taneb has quit (Ping timeout: 240 seconds).
16:25:48 -!- hagb4rd has joined.
16:38:37 -!- yiyus has quit (Read error: Operation timed out).
16:41:29 -!- azaq23 has joined.
16:41:38 -!- azaq23 has quit (Max SendQ exceeded).
16:42:23 -!- azaq23 has joined.
16:43:54 -!- yiyus has joined.
16:47:56 -!- PiRSquared has changed nick to TogetherForever.
16:56:48 -!- TogetherForever has changed nick to WhnvrUNeedSmebdy.
17:01:37 -!- WhnvrUNeedSmebdy has changed nick to PiRSquared.
17:39:49 -!- calamari has joined.
18:00:58 -!- calamari has quit (Quit: Leaving).
18:15:57 -!- PiRSquared has changed nick to invex.
18:24:12 <kallisti_> Friendship: hi I just installed xinit how do I make Xorg happen?
18:24:40 <kallisti_> or whenever it's supposed to happen automatically
18:30:59 <Friendship> Idonno, depends on distro? On Debian the easiest way is to install a DM.
18:31:52 -!- elliott has joined.
18:32:00 <elliott> http://esolangs.org/w/index.php?title=Basic_Input/Output_Commander&curid=8301&diff=30609&oldid=30582 ...
18:37:11 -!- invex has quit (Quit: I CAN'T TAKE IT ANYMORE ARRGHH (ragequit ;) ... )).
18:39:34 <elliott> The text you wanted to save was blocked by the spam filter. This is probably caused by a link to a blacklisted external site.
18:39:34 <elliott> The following text is what triggered our spam filter: http://cheap-viagra
18:40:20 <elliott> mRoman: btw linking has always been case-sensitive
18:40:23 <elliott> just not in the first letter
18:47:37 <elliott> Phantom_Hoover: Well, I *got* it, I just didn't get it in any deeper sense.
18:47:49 <elliott> I wasn't sure whether there was any deeper sense I was missing in the first place.
18:48:20 <elliott> 23:02:51: <Friendship> <elliott> "I LIKE WRITING BIG, IT MAKES ME FEEL COOL" -- Friendship, 2005
18:48:21 <elliott> 23:02:55: <Friendship> I /probably/ said this.
18:50:24 <elliott> 23:36:51: <itidus20> wow that jhix guy.. <nooga> he's 14 and he's writing interpreters in haskell and he knows lambda calculus and everything!!!!!!
18:51:28 -!- kallisti_ has quit (Ping timeout: 244 seconds).
18:53:49 <elliott> 01:55:16: <kallisti_> wait, Debian doesn't have sudo?
18:53:49 <elliott> 01:55:49: <kallisti_> not by default it seems
18:54:07 <elliott> @tell kallisti Yes, Debian does have sudo by default, you just ignored my explicit and direct instructions on how to get it to use sudo by default.
18:54:18 <elliott> @tell kallisti And Ubuntu doesn't have gcc by default either.
18:54:33 <elliott> Friendship: Debian isn't root-password-only these days.
18:54:37 <elliott> The default installer mode supports setting up sudo.
18:55:30 -!- kallisti has joined.
18:55:30 -!- kallisti has quit (Changing host).
18:55:30 -!- kallisti has joined.
18:56:39 <elliott> 04:00:09: <kallisti_> would there be any reason not to move /root/bin to some other bin in my default PATH?
18:56:39 <elliott> How the fuck did you even get /root/bin?
18:56:57 <elliott> 04:05:28: <kallisti_> also would there be a reason not to copy my /etc and home directories from this install to my real install in order to speed up the setup time?
18:56:57 <elliott> Sure, if you want to break things faster.
18:57:11 <kallisti> elliott: turns out that by default sudo doesn't preserve your hold $HOME
18:57:11 <lambdabot> kallisti: You have 3 new messages. '/msg lambdabot @messages' to read them.
18:57:22 <kallisti> but if you add a line to sudoers that fixes it.
18:57:23 <Friendship> <elliott> Friendship: You did. // yeah, I found it X-D
18:57:28 <elliott> 04:37:39: <kallisti_> hm, oh...
18:57:28 <elliott> 04:37:45: <kallisti_> okay, so... apparently sudo installs things in /root
18:57:33 <elliott> You're not meant to run cabal-install as root.
18:57:33 -!- Phantom_Hoover has quit (Remote host closed the connection).
18:57:53 <elliott> kallisti: And that's a dangerous setting.
18:58:05 <elliott> (You're only meant to run cabal-install as root when you have global instalsl set up, and those aren't recommended.)
18:58:57 <kallisti> oh wait.... does slim use xinit or startx..
18:59:16 <elliott> And I have no idea what the fuck you're talking about.
18:59:29 <kallisti> well, slim starts X for me, right?
19:00:16 <elliott> kallisti: You should remove all ~/.cabal ~/.ghc you have.
19:00:21 <elliott> If you installed them as root it's horribly broken.
19:01:05 <kallisti> yes, with the setting to keep HOME
19:01:54 <kallisti> really, it should keep all of the env variables...
19:01:56 <elliott> You are meant to run cabal as your ordinary user.
19:02:01 <elliott> That's a massive security hole.
19:02:48 <elliott> You don't get a ~/.ghc as the result of doing that. Why do you think you would?
19:04:31 <kallisti> the main reason I want to preserve PATH is so that I can use sudo to install things that requires admin privs.
19:04:47 <elliott> 12:37:04: <fizzie> <ais523> <Star651, via email> Is this a spambot or something? I just got a user talk page message from Phantom Hoover about the BF-inspired languages, the "brick brain brochure."
19:05:04 <elliott> kallisti: And why would you want those to go in $HOME?
19:05:10 <elliott> root shouldn't own things in $HOME.
19:06:00 <kallisti> because that's how Ubuntu works and I like the way it works in Ubuntu.
19:06:09 <elliott> No, it doesn't "work" like that in Ubuntu.
19:06:15 <elliott> Ubuntu has never recommended you use sudo to install things into $HOME.
19:06:19 <elliott> That's completely nonsensical advice.
19:06:41 <kallisti> so there's no such thing as a package that makes use of $HOME in any way?
19:07:00 <kallisti> to say, create a hidden directory with config info or other data?
19:07:23 <elliott> I'm putting you on /ignore; the number of incoherent questions exceeds my ability to care about answering them.
19:08:15 <kallisti> elliott apparently has this delusion that people who ask questions actually have all the information they need to start with.
19:10:01 -!- Phantom_Hoover has joined.
19:10:48 <ion> There’s nothing wrong with using ~ for your personal stuff. No, not in Ubuntu either.
19:11:08 <kallisti> that's not quite what we're talking about.
19:11:36 <kallisti> I was under the impression that when I use apt-get to install packages, it also leaves directories in $HOME
19:11:56 <kallisti> or rather, it can leave a directory in $HOME
19:12:44 <ion> That would be utterly broken behavior.
19:12:58 <elliott> ion: Of course there's nothing wrong with that.
19:13:08 <elliott> ion: There's something wrong with using "sudo cabal" and wanting it to install into $HOME
19:13:18 <elliott> And, in general, installing things into $HOME using sudo for no apparent reason.
19:13:44 <Sgeo> elliott, ion is calling "sudo apt-get install making things in people's $HOME" broken behavior that doesn't exist.
19:13:51 <Sgeo> In case you missed what kallisti said
19:14:13 <elliott> Ah. Yes, I did, thankfully.
19:15:01 <ion> Well, apt-get install needs write permissions to stuff outside $HOME in any case, i thought root permissions were implied.
19:15:28 <elliott> Right, I didn't see any mention of apt-get.
19:16:06 <kallisti> 14:06 < kallisti> so there's no such thing as a package that makes use of $HOME in any way?
19:16:19 <kallisti> I see how you might have been confused.
19:18:16 <kallisti> but seriously that would have taken at most 2 replies to figure out what I was saying.
19:18:23 -!- kallisti has quit (Quit: leaving).
19:19:51 <elliott> Hey Phantom_Hoover, how do I avoid the awkward plural->singular transition in "General links to people should be to standard namespace articles on them, if one exists"?
19:21:28 <elliott> Phantom_Hoover: But surely "General links to people should be to standard namespace articles on them, if one exists, and to their user page otherwise." is awkward? It goes from plural to singular within the main bulk of the sentence.
19:21:42 <elliott> Blah, that sentence is a complete mess.
19:29:52 <Sgeo> Did kallisti fail to note that you had em on ignore?
19:32:05 <elliott> Phantom_Hoover: You should totally rewrite my sentence.
19:33:07 <elliott> Dammit. Now I'm legally required to make that policy.
19:35:21 -!- ais523 has joined.
19:36:30 <lambdabot> ais523: You have 1 new message. '/msg lambdabot @messages' to read it.
19:36:34 <lambdabot> elliott said 31m 40s ago: god bless
19:36:48 <ais523> hmm, the university lawyers have got back to us on the software license for this hardware compiler thing
19:37:03 <ais523> haha, it's one of those big EULA things
19:37:30 <elliott> ais523: <elliott> 12:37:04: <fizzie> <ais523> <Star651, via email> Is this a spambot or something? I just got a user talk page message from Phantom Hoover about the BF-inspired languages, the "brick brain brochure."
19:37:30 <elliott> <elliott> @tell ais523 god bless
19:37:55 <elliott> ais523: it's not going to be under a free license? :(
19:38:16 <ais523> it also contains a reservation of the reverse engineering right, which is funny as we're planning to supply source
19:39:15 <ais523> no, I might mention that it's vaguely ridiculous to put a reverse engineering clause in something that can't, by definition, be reverse engineered
19:39:30 <elliott> I meant, the license in general
19:39:41 <ais523> no, we get to supply feedback
19:39:44 <elliott> actually, is an EULA even a valid source code license?
19:39:55 <elliott> it sounds like they've completely misunderstood what is required
19:40:05 <ais523> it specifically permits derivative works under some circumstances
19:40:36 <elliott> *distribution of derivative works, presumably
19:40:37 <ais523> hmm, does what you'd want to do with it fall under the category of "academic teaching or non-commercial research"?
19:40:51 <ais523> it allows you to create them, but not distribute them
19:40:59 <elliott> ais523: err, creating derivative works is always permissable
19:41:02 <ais523> that also looks like a mistake
19:41:05 <elliott> licenses are about redistribution...
19:41:11 <ais523> elliott: it's a contract-based license, not copyright-based license
19:41:13 <ais523> those can go a bit further
19:41:36 <ais523> it's perfectly possible to make a contract with someone that they won't make a derivative work of something, in return for something else
19:41:36 <elliott> I don't suppose you'll be able to convince them to license it under something sane :p
19:42:07 <elliott> ais523: it would probably fall under non-commercial research, but I'll probably just try it out, assuming I can figure out a way to download it without agreeing to a contract
19:43:03 <ais523> get someone under 18 to download it?
19:43:16 <ais523> (I'd love to try that trick some day, as far as I can tell it's legally valid, but it's never been tested)
19:44:07 <elliott> ais523: coppro tried to get money back from a Windows license, IIRC
19:44:15 <elliott> despite agreeing to a contract saying he wouldn't try to
19:44:20 <elliott> because he was a minor at the time
19:44:31 -!- oerjan has joined.
19:44:34 <ais523> although, right, that'd be Canadian law
19:44:36 <elliott> (it came with a computer, I think)
19:45:13 <ais523> I was very impressed with Toshiba for putting the "you will not try to seek a refund for Windows" contract on the outside of the box, visible before I bought it, and colored in a way that it was prominent and visible
19:45:19 -!- oerjan has set topic: My Little Stallman: Freedom is Magic | So good to be in the Evening....In Europe! | http://codu.org/logs/_esoteric/ | http://esolangs.org/wiki/ has moved servers!.
19:45:54 <ais523> so I could actually make an informed decision about whether to buy the laptop based on the contracts that came with it
19:46:40 <ais523> (actually, in the UK, doing that may be required for it to be legally binding, based on some opinions by an important government lawyer that were made in a non-legally-binding way)
19:47:09 <oerjan> <kallisti_> also there's no cabal...
19:47:23 * oerjan wonders how many here remember that one
19:47:39 <ais523> <strike>[http://not.found.yet/bioc/c-compilers Original C/C++/C# compiler source code for Basic Input/Output Commander]</strike> (no source found yet)
19:47:52 <elliott> can we find an excuse to delete that thing?
19:47:57 <ais523> oerjan: not only do I remember it, I know the original context too
19:48:00 <ais523> (it predates the Web, IIRC)
19:48:09 <Sgeo> Delete what thing?
19:48:15 <elliott> ais523: that external link makes me think Esme
19:48:27 <elliott> like the author of the article is trying to troll us really badly
19:48:31 <elliott> Sgeo: http://esolangs.org/wiki/Basic_Input/Output_Commander
19:48:32 <ais523> sure it's not the same person?
19:48:57 <elliott> ais523: hmm, the rest of the article doesn't remind me of him
19:48:58 <ais523> elliott: hmm, I'm vaguely remembered of a case where someone who's known as quite a good troll deliberately trolled badly, as a form of metatrolling
19:49:01 <elliott> it actually has concrete details, for one
19:49:58 <elliott> heh, [[Esme]] isn't in [[Category:Languages]] or [[Category:Joke languages]]
19:50:04 <elliott> I almost went to fix that, then realised it's correct
19:50:35 <oerjan> is it in Category:Shameful?
19:50:41 <ais523> yes, despite that category not existing
19:50:51 <ais523> I think category:shameful was invented for it
19:51:19 <elliott> YOU'RE NOT MEANT TO LET THE WIKI OWNER FIND OUT ABOUT THAT CATEGORY
19:51:23 <ais523> meanwhile, how long is it going to be before mobile phone lawsuits hit singularity?
19:51:32 <ais523> elliott: but, wasn't it you who came up with it?
19:51:34 <oerjan> elliott: you created it, right?
19:51:50 <elliott> that doesn't invalidate the policy!
19:52:29 <ais523> that means we're not letting you find out about it, because you know already
19:52:39 <oerjan> maybe it should be mentioned in Category Talk:Shameful
19:53:34 <elliott> shouldn't it be bright by now, then?
19:54:12 <oerjan> <elliott> how is it still february <-- well it _is_ the groundhog month...
19:54:45 * ais523 is not really with it either
19:56:52 <oerjan> <zzo38> I was thinking about my ideas for Ibtlfmm programming language <-- i'm sorry but elliott will never be able to distinguish that from Itflabtijtslwi.
19:57:31 <elliott> i think i can remember "itflab", if i try real hard
19:58:02 <elliott> oerjan: btw keymaker comes in later :DDDD
19:58:37 <elliott> apparently he read the logs. which is even scarier a thought than when graue came in here a few days ago
20:00:17 <oerjan> they haven't added trollface to unicode yet, have they?
20:00:47 <elliott> can you or ais523 kick you thaanxekr
20:00:59 <oerjan> that will probably be the day elliott swears off the internet and become a cave dweller together with grigori perelman
20:01:38 <oerjan> ais523: i think my idea made his brain crawl
20:01:57 <elliott> ais523: he's saying a lot of confusing words
20:02:30 <elliott> he semese cenenegele desegesede
20:02:35 <ais523> oerjan: have you ever seen Just Connect? you probably haven't because it's on BBC4, and people rarely watch that even in the UK
20:02:52 <oerjan> elliott: it's for both lines
20:03:00 <ais523> it has a round where you have to reconstruct phrases which have had their vowels removed, and they're often really difficult
20:03:03 <elliott> ais523: i thought it was only connect ..... i haive ino idea why i know this
20:03:09 <ais523> elliott: oh, you might be right
20:03:12 <oerjan> ais523: even so i don't watch tv
20:03:14 <elliott> oerjan: no i mean too many vowels for me to stick in my head
20:03:20 <elliott> and keep track of position in
20:03:33 <ais523> oerjan: not at all? if so, I've watched more Norwegian TV than you have :)
20:03:36 <oerjan> actually i _have_ happened upon weakest link
20:04:03 <elliott> weakest link is our produest cultural export
20:04:06 <ais523> elliott: "he seems cunningly disguised"
20:04:07 <oerjan> ais523: well it has happened sometimes in other people's places
20:04:41 -!- pikhq_ has joined.
20:05:10 <elliott> oh no now i have to decide between the two chrome adblocking extensions again
20:05:34 -!- pikhq has quit (Ping timeout: 276 seconds).
20:05:48 <elliott> aaaaah my scherzo is unmeasurable
20:06:07 <oerjan> `addquote <elliott> aaaaah my scherzo is unmeasurable
20:06:15 <oerjan> i have _no_ idea what that means
20:06:19 <elliott> i was looking for some other german word
20:06:21 <HackEgo> 810) <elliott> aaaaah my scherzo is unmeasurable
20:06:24 <elliott> like oe of those words everyone says cant be translated into english
20:06:33 <elliott> but all i could think of was scherzo
20:06:45 <elliott> i thought of that one, its not that one though
20:06:50 <elliott> then i thought of schengen but its not that either
20:07:05 <elliott> then i typed in what it soudned like into googl ebut it was obviously wrong because it corrected me to arnold schwarznegger
20:07:16 <elliott> oerjan: yes that might be it
20:08:16 <elliott> Fahrvergnügen (German pronunciation: [ˈfaːɐ̯.fɛɐ̯ˌɡnyːɡn̩]) was an advertising slogan used by the German automobile manufacturer Volkswagen in a 1990 U.S. ad campaign that included a stick figure driving a Volkswagen car.[1]
20:08:35 <elliott> wir fahrvergnugen fahrvergnugen fahrvergnugen auf der autobahn
20:09:10 <elliott> ais523: fuck that. all vowels are teh same
20:09:19 <oerjan> (that came up in a newspaper quiz today)
20:09:41 <oerjan> elliott: so why do you protest when i leave them out?
20:09:41 <ais523> closest thing to ü in English is "er", although it isn't that close
20:09:49 <ais523> whereas German u is pretty similar to English u
20:10:13 <elliott> oerjan: theyre the SAME but they're not OPTIONAL
20:12:39 <oerjan> btw that reddit self-ad for fifthworldproblems creeps me out, especially after i learned what CMEPT means
20:13:04 <elliott> ais523: http://www.reddit.com/r/fifthworldproblems
20:13:08 <elliott> (there is no point trying to explain)
20:13:59 <oerjan> (ok i'm approximating with latin letters here)
20:14:12 <oerjan> which is rather backwards from germanic
20:14:54 <oerjan> and ge:Schmertz of course
20:15:52 <oerjan> i hesitated on that one
20:17:22 <elliott> ais523: does your operating system have a zsync package
20:17:29 <elliott> oh its ubuntu i guess it does
20:18:03 <ais523> gah, my boss has sent me an email that makes Evolution segfault when I try to read it
20:18:15 <ais523> or at least, close unexpectedly, not sure if it's SEGV in particular
20:18:15 <elliott> its' tryingt o protect you
20:18:25 <elliott> oerjan: does your operating system have a zsync package (this is a comedy)
20:18:31 <ais523> I'll see if I can read it through webmail, haven't used that for my university account for ages
20:18:46 <oerjan> ais523: a minor explanation, it _is_ of course weakly inspired by r/firstworldproblems. actually that may not help at all.
20:19:04 -!- monqy has joined.
20:20:05 <oerjan> or actually it does. firstworldproblems is about problems you only have because you have money in the first place
20:20:09 <ais523> meanwhile, Yahoo! have broken their POP certificate again
20:20:21 <elliott> Deewiant: does y- oh, arch, so yes
20:20:33 <oerjan> fifthworldproblems is about problems you can only possibly imagine having.
20:20:41 <elliott> does anyone use anything that isn't a debian derivative or arch or windows :P
20:20:54 <elliott> does solaris have "packages"
20:21:00 <elliott> isn't there a popular repository for it or something
20:21:04 <elliott> with a bunch of ported stuff
20:21:11 <Deewiant> https://www.opencsw.org/ maybe
20:22:07 <elliott> Deewiant: anyway whatever, do you have a convenient package for zsync
20:22:23 <elliott> i guess i should ask some bsd users if i want relevant answers
20:22:23 <oerjan> <elliott> oerjan: does your operating system have a zsync package (this is a comedy) <-- no, i just brush my teeth normally
20:22:34 <elliott> Deewiant: (if so, which version)
20:22:39 <elliott> ais523: (which version is yours?)
20:23:51 <elliott> that's the same version i had
20:23:54 <ais523> I'm wondering if I've missed the point of the question
20:23:56 <elliott> i thought it was on 1.6 for some reason
20:24:07 <elliott> ais523: i'm working on esolangs.org wiki backups
20:24:21 <elliott> my current plan is to distribute the mediawiki XML backup via zsync (= rsync over standard HTTP)
20:24:40 <elliott> because that means people can update it daily with like a kilobyte to download rather than 20 megs
20:24:46 <elliott> without me having to maintain a separate rsync server
20:25:06 <elliott> so i'm just checking that reasonable OSes have zsync
20:25:10 <ais523> elliott: btw, can I get rid of the encrypted wiki backup? I just remembered it's illegal in the UK to have encrypted files you don't know the password to (which is stupid, but there you go)
20:26:00 <oerjan> <mRoman> and Objects are Pythonobjects, so that wouldn't be a problem <-- THAT'S CHEATING
20:27:52 <oerjan> some day the uk parliament will pass a law that accidentally prohibits breathing, and that will be the end of ais523.
20:28:06 <ais523> oerjan: I'd break laws if the alternative were death
20:28:40 <elliott> what if the alternative was living inside a small metal box for the rest of your life
20:29:10 <ais523> elliott: there's a boundary somewhere, but I haven't quite worked out where yet
20:29:13 <ais523> and don't care to find out
20:30:15 -!- zzo38 has joined.
20:31:14 -!- sebbu2 has joined.
20:31:14 -!- sebbu2 has quit (Changing host).
20:31:15 -!- sebbu2 has joined.
20:34:35 -!- sebbu has quit (Ping timeout: 252 seconds).
20:35:33 <Deewiant> elliott: I'd need to update the repository due to opensolaris no longer existing, but there seems to be some kind of server problem... opencsw.org doesn't have it, anyway, and given that it's only in AUR it's unlikely to be in Solaris
20:36:06 <elliott> Oh well, sucks to be Solaris users
20:36:29 <Deewiant> Doesn't matter, there's never any updates or anything anyway
20:36:48 <elliott> i meant re not having zsyncgood, it's in freebsd
20:37:47 <elliott> but not netbsd oh well!!!!
20:38:41 <elliott> Deewiant: everyone forgot dragonfly
20:40:44 <elliott> Deewiant: did you know cpressey wrote dragonfly's installer????
20:40:53 <zzo38> Can you make the spam blacklist and abuse filters bypassed if and only if found inside of a <pre><nowiki> block and that the URL or text must be inserted by a autoconfirmed user? (So if not autoconfirmed, or not a <pre><nowiki> block, the blacklist will take effect and prevent you from saving the page) (Some people might invent "Casino Viagra" esolang; I have had this idea myself)
20:41:05 <elliott> oh dragonfly just uses pkgsrc
20:41:09 <elliott> SUCKS TO BEDRAGONFLY USERS
20:41:19 <elliott> zzo38: I can whitelist problematic entries
20:41:47 <elliott> probably the list frmo metawiki isn't helping all that much anyway
20:42:26 -!- MoALTz has quit (Ping timeout: 244 seconds).
20:42:29 <ais523> zzo38: that is technically possible, but I'd rather only do it on a case-by-case basis
20:42:37 <zzo38> I have not had any problems so far; I am simply mentioning a strange idea about esolangs
20:42:37 <ais523> (we can move the URLs from the spam blacklist to the abuse filter if necessary)
20:42:51 <ais523> (which can be set to allow specific users to add the URLs, or specific user groups)
20:43:01 <elliott> ais523: well in this case it was "cheap-viagra" that caused problems :)
20:43:11 <elliott> because it's on metawiki's master blacklist
20:43:20 <elliott> I just removed the http:// to get past it
20:43:35 <zzo38> Still, that is not the kind of problem I was talking about.
20:44:00 <zzo38> I was talking about an esolang where the source code might occasionally contain such blacklisted URLs
20:44:30 -!- augur_ has quit (*.net *.split).
20:44:31 -!- chickenz has quit (*.net *.split).
20:44:48 <elliott> zzo38: the problem with only allowing insertion by autoconfirmed user is that MW doesn't really keep track of insertion
20:44:54 <elliott> so such an extension would likely block all non-autoconfirmed users out
20:45:23 <zzo38> And, anyways, I was once on some wiki which blacklisted my own domain name (zzo38computer.cjb.net). I had an http:// URL; entering a gopher:// URL was accepted.
20:45:38 <elliott> probably it blacklisted cjb.net
20:45:41 -!- augur_ has joined.
20:45:41 -!- chickenz has joined.
20:45:49 -!- MoALTz has joined.
20:46:53 <ais523> elliott: AbuseFilter can be set to only block addition of links, not edits that preserve links that were already on the page
20:46:55 <zzo38> Yes, I guess so. But, cjb.net is used both as URL redirection and as dynamic DNS. When used as URL redirection obviously only http:// works; but used as dynamic DNS it can be any URI scheme so I suppose it works; if the URI scheme is gopher:// then you can know it is not a URL redirection!
20:47:30 <ais523> AbuseFilter is more flexible
20:47:36 <ais523> but SpamBlacklist scales better
20:49:04 <zzo38> (So if for whatever reason you run a server on a blacklisted domain, you can still use URI schemes other than HTTP to get through; it can be helpful. Of course spammers will never do that so it is still OK)
20:49:29 <ais523> zzo38: why do you think Gopher spam isn't more popular than it is?
20:50:44 <zzo38> ais523: Lack of support in Google and Internet Explorer; and also that they cannot receive payments, and that they cannot trick you with Flash and popups and hidden fields and all that stupid stuff
20:51:19 <elliott> hmm, this thing doesn't seem to work
20:51:40 -!- Phantom_Hoover has quit (Quit: Leaving).
20:51:53 -!- Phantom_Hoover has joined.
20:52:56 -!- Sgeo has quit (Quit: Leaving).
20:53:38 -!- Sgeo has joined.
20:53:51 <zzo38> Can I get a plain text version of CC0 so that it can be included in Haskell packages which are in the public domain? (This might solve the problem some people were complaining about when I created public domain Haskell packages, which is that public domain is not applicable in all jurisdictions)
20:54:58 <elliott> ais523: when you ask it to include image contents in XML dumps, MediaWiki encodes them in base64
20:55:08 <ais523> hmm, that /should/ compress well
20:55:09 <elliott> with DOS newlines, inexplicably
20:55:30 <ais523> aren't those standard for HTTP communication?
20:55:41 <zzo38> CRLF is the standard newlines on most internet protocols.
20:55:45 <fizzie> All MIME stuff is very CRLFy.
20:56:03 <elliott> ais523: well, in the file itself, I mean
20:56:12 <elliott> it goes from 164 megs to 18 megs gzipped
20:56:17 * elliott tries bzip2 out of curiosity
20:56:31 <ais523> I think bzip2 huffman-codes the individual bytes in the file
20:56:41 <zzo38> Sometimes they are called DOS newlines but it should be called printable newlines since CRLF is how it will print on a paper.
20:56:41 <ais523> so base64 and binary should come to almost exactly the same size bzip2ed
20:57:04 <ais523> although, I guess huffman-coding would work better on the binary if the original file had a very skewed frequency distribution
20:57:59 <elliott> -rw-r--r-- 1 www-data www-data 18M Feb 25 20:52 esolang.xml.gz
20:57:59 <elliott> -rw-r--r-- 1 root root 9.9M Feb 25 20:55 esolang.xml.orig.bz2
20:58:10 <elliott> bzip2 is really useless...
20:58:23 <elliott> I suppose it was useful for /something/ when xz didn't exist
20:58:29 <oerjan> <elliott> i did that at 13 :'( <-- yeah but you're elliott.
20:59:21 <ion> Did what at 13? (Do i want to know?)
20:59:31 <ais523> hmm, bzip2 is still the most compressy of the compression algos you typically get preinstalled on POSIXy systems
20:59:35 <ais523> and it's quite hard to beat
20:59:40 <zzo38> One idea I had for compression format, is for each file (or even part of a file said to be combined after decompression, or multiple files in a single block with an instruction to split after decompression), to have a reference to a compression program within the archive file. The compression program consists of reversible primitive instructions (reversible computing), with many common schemes as single instructions.
20:59:47 <ais523> I eventually got azip to do /slightly/ better than bzip2, but it couldn't compete even with lzma
21:00:13 <elliott> it's a rite of passage in england
21:00:17 <oerjan> ion: <elliott> 23:36:51: <itidus20> wow that jhix guy.. <nooga> he's 14 and he's writing interpreters in haskell and he knows lambda calculus and everything!!!!!!
21:00:37 <elliott> (That's an English euphemism for "eating a weasel".)
21:00:47 <fizzie> ais523: Also if you have a long repeated sequence, which any LZ-style compressor would encode as a back-reference, in the original binary, it doesn't necessarily translate to a repeated sequence in the base64 unless the distance between the starting points for original and repeated is a multiple of 3.
21:01:08 <oerjan> elliott: thus the expression "weaseling out" for people avoiding obligations
21:01:11 <zzo38> Is the idea I have about compression seems workable or anything else to you?
21:02:04 <zzo38> So that a compression scheme can be written in a reversible programming language designed for writing compression schemes.
21:02:05 <ais523> zzo38: it's a bit like using a self-extracting archive
21:02:11 <ais523> but with a more appropriate language
21:02:22 <fizzie> I can't really "reason" about what bzip2 does since it's got that BWT, which I don't really "intuit".
21:02:32 <ais523> fizzie: it's basically a sort by context
21:02:47 <ais523> you work out all cyclic permutations of the original file, then sort them
21:03:13 <zzo38> And possibly using for encryption as well, to require a password to decompress, if it uses instructions to request user input.
21:03:28 <ais523> and then record the /last/ character of each, in order
21:03:35 <ais523> the only really crazy thing about that operation is that it's reversible
21:04:18 <ais523> but you can expect the last character to tend to form long runs in a file with a lot of repetition, because if two cyclic permutations sort similarly to each other, they have a lot of characters in common, so the character before was probably the same as well
21:05:04 <zzo38> And then the user input simply becomes fixed data to the program if it is requested; that way it is still reversible; the user input is user input both on normal and reverse execution of the program.
21:05:54 -!- sebbu2 has changed nick to sebbu.
21:06:25 <kmc> (time (m-eval global-env '(m-eval global-env '(fact 5))))
21:06:25 <kmc> 598614 ms real time
21:06:52 <zzo38> ais523: I can understand like what you are saying about self-extracting; but in this case it would also be a binary code interpreted by a program which is uncompressor program, and the uncompressor program deal with options such as which files you want decompressed, listing the files, disassembling the program codes, etc
21:07:19 <zzo38> And that in this case it even is cross-platform instead of an executable file for a certain kind of computer or operating system.
21:08:37 <elliott> hey, somebody edit a page on the wiki
21:09:01 <ais523> for any particular reason?
21:10:19 <elliott> I knew that was going to happen before I loaded the wiki
21:10:26 <elliott> because I was looking through the access logs to check something
21:10:32 <ais523> I thought of putting actual spam there
21:10:46 <ais523> (I don't have many things that I could usefully spam; AceHack or Jettyplay, perhaps)
21:11:30 <zzo38> ais523: Perhaps if you want to mention AceHack or Jettyplay, use your own user page for that purpose instead.
21:11:36 <ais523> zzo38: but then it wouldn't be spam
21:11:58 <zzo38> I know. But just in case you wanted to mention them; it is not a requirement to mention them at all.
21:12:17 <ais523> elliott: did the test succeed?
21:12:28 <oerjan> <elliott> Phantom_Hoover: But surely "General links to people should be to standard namespace articles on them, if one exists, and to their user page otherwise." is awkward? It goes from plural to singular within the main bulk of the sentence.
21:12:54 <oerjan> ", if any exist, and to their user pages otherwise."
21:14:23 <elliott> oerjan: "if any exist" is nonsense
21:14:32 <elliott> oerjan: that implies you should link to a mainspace article if any articles on people exist
21:14:42 <elliott> ais523: that's secret! at least for the next five minutes or so
21:14:49 <oerjan> elliott: the alternative is to make the whole sentence singular, then
21:15:05 <ais523> hmm, a five-minute secret seems reasonable
21:15:11 <elliott> ais523: hey, where should I put a system-specific shell script to be called from a system user (= no homedir)'s crontab?
21:15:52 <oerjan> "A general link to a person should be to a standard namespace article on them,"
21:16:08 <ais523> elliott: /usr/lib subdir, right?
21:16:22 <ais523> "main namespace" is the technical term, I think
21:16:37 <oerjan> (choose your favorite gender neutral pronoun yourself)
21:16:46 <zzo38> Maybe, /usr/local/scripts/
21:16:51 <ais523> oh, I thought you meant arch-specific
21:16:54 <ais523> it's got to be in /etc, then
21:17:10 <zzo38> Or, /etc/local/scripts/
21:17:12 <elliott> I was thinking /usr/local/libexec/foo.sh
21:17:35 <elliott> oerjan: hm but the rest of the policies are plural :(
21:17:55 <elliott> ais523: surely /etc is only for system-local configuration of /existing/ programs
21:18:11 <oerjan> elliott: fine, rewrite it in chinese then
21:18:12 <ais523> elliott: /etc/cron.weekly?
21:18:24 <ais523> it's for configuration of cron, I guess
21:19:06 <elliott> ais523: ooh, I didn't realise that existed
21:19:15 <elliott> do you still have to put entries in crontab if you create something in there?
21:19:26 <elliott> also, is that just for root?
21:19:30 <ais523> no, the crontab runs everything in the directory at the appropriate intervals
21:19:54 <pikhq_> /usr/local/libexec or ~/local/libexec, I say.
21:20:14 <ais523> do system users normally /have/ home dirs?
21:20:27 <elliott> ais523: no, that's the point
21:20:32 <ais523> oh, huh, they mostly have home dirs in /var on this system
21:20:52 <elliott> /usr/local/lib or /usr/local/share, then, I suppose
21:21:04 <ais523> heh, "home dir" for a system user seems to be "dir most similar to what they do"
21:21:08 <ais523> e.g. "games" has a homedir of /usr/games
21:21:25 <zzo38> Do you have any comments about my ideas of Ibtlfmm programming language I was thinking about? (near timestamp 1330153950)
21:21:58 <pikhq_> Seems some distros prefer lib over libexec...
21:22:29 <pikhq_> In either case, /usr/local/ is maintained purely by you. The only things in there are things you personally installed.
21:23:35 <elliott> The /etc hierarchy contains configuration files. A "configuration file" is a local file used to control the operation of a program; it must be static and cannot be an executable binary. [4]
21:23:41 <elliott> hmm, are shell scripts binaries?
21:24:40 <elliott> pikhq_: FHS bans using */lib/* for executables, it seems
21:24:44 <elliott> including shell scripts, I think
21:25:13 <zzo38> And maybe instead of "return" and "join" I should just call it "eta" and "mu" like how it is usually in mathematics, or something similar;,like I don't know. Or the Maybe monad should be called a successor monad instead
21:28:36 <Friendship> elliott: $ file /usr/lib/gcc/x86_64-linux-gnu/4.6.2/cc1 /usr/lib/gcc/x86_64-linux-gnu/4.6.2/cc1: ELF 64-bit LSB executable
21:29:34 -!- MoALTz_ has joined.
21:31:33 -!- MoALTz has quit (Ping timeout: 265 seconds).
21:32:31 <elliott> Friendship: eys it deprereprerperpeprecated that
21:32:35 <elliott> also, also, ghc does hat too, also
21:33:15 <Sgeo> How well/poor would ABCL be for writing Minecraft plugins?
21:33:51 <lambdabot> 535754303593133660474212524530000905280702405852766803721875194185175525562...
21:38:27 <elliott> ais523: hey, is gzip more effective if it knows the file of the size beforehand?
21:38:55 <elliott> meanwhile, http://esolangs.org/wiki/Fool
21:39:46 <ion> `run </usr/bin/gdb gzip | wc -l
21:39:50 <HackEgo> bash: /usr/bin/gdb: No such file or directory \ 0
21:40:02 <ion> `run </usr/bin/python gzip | wc -l
21:40:12 <Friendship> < isn't a good way to go, it's still stattable.
21:40:13 <ion> whoops, i mean -c
21:40:30 <ion> friendship: That’s what i was going to compare it to.
21:40:33 <ion> `run </usr/bin/python gzip | wc -c
21:40:38 <ion> `run cat /usr/bin/python | gzip | wc -c
21:40:49 <elliott> *does* gzip stat if it's using stdin, though?
21:41:10 <Deewiant> `run gzip -c /usr/bin/python | wc -c
21:41:30 <elliott> oh, presumably it includes the filename
21:43:33 <elliott> the one time when i don't want vim to auto-decompress a .gz file...
21:45:01 <elliott> pikhq_: Any other ideas for filesystem location? :P
21:46:18 <oerjan> elliott: heh, trying to open /// from history completion on my browser address line opens // instead
21:46:36 <elliott> oerjan: hmm, what URL does it have recorded for it?
21:46:46 <elliott> Deewiant: Doesn't work once vim has already started
21:47:35 <oerjan> elliott: hm indeed it's http://esolangs.org/wiki///
21:47:47 <oerjan> adding an extra / fixes it
21:48:04 <ais523> elliott: I don't think gzip is
21:58:52 <ais523> whereas it doesn't matter for bzip2 either, because it waits for the EOF or an entire maximum-it-can-compress block before starting to compress either
22:02:45 <elliott> heh, ais523 won't like this: "Default cookie lifetime ($wgCookieExpiration) is increased to 180 days."
22:03:07 <ais523> not really; if I /want/ a cookie, I may well want it to persist
22:03:22 <ais523> there's a reason I approve/disapprove of cookies manually, rather than just blanket-blocking them all
22:05:17 <elliott> the active user count on http://esolangs.org/wiki/Special:Statistics is still broken :(
22:06:08 <zzo38> There should be a checkmark to not expire any cookie of this wiki, and make it only expire if keys change on the server.
22:06:23 <elliott> i don't believe cookies can never expire
22:07:40 <ais523> right, but you can tell cookies to not expire due to the passage of time
22:08:40 <ais523> you could just set an expiration date past Y2038
22:08:54 -!- Ngevd has joined.
22:09:50 <ais523> elliott: it is, because that's when the date rolls round
22:10:05 <elliott> cookies don't use 32-bit timestamps
22:10:21 <elliott> and modern OSes have used 64-bit timestamps for years
22:10:28 <ais523> but do modern browsers?
22:10:48 <elliott> they'll probably just use time_t
22:11:18 <zzo38> Do you like to use redundant userboxes that are redundant?
22:11:55 <ais523> elliott: time_t is 32-bit on 32-bit systems, usually, isn't it?
22:12:04 <elliott> zzo38: your site doesn't work via HTTP any more :(
22:12:25 <elliott> ais523: I don't think it is nowadays
22:12:49 <ais523> !c printf("%d\n", (int)sizeof(time_t));
22:13:05 <ais523> so that won't be too useful
22:13:28 <Ngevd> What if you set it to a date before 1970?
22:13:29 -!- itidus21 has quit (Ping timeout: 244 seconds).
22:13:41 <Ngevd> Is that crazy enough to work?
22:15:00 <ais523> $ echo -e '#include <time.h>\n#include <stdio.h>\nint main(void){printf("%d\\n", (int)sizeof(time_t)); return 0;}' | gcc -x c /dev/stdin && ./a.out
22:15:22 <elliott> nobody uses 32-bit any more anyway :p
22:15:31 <ais523> I use 32-bit OS on 64-bit hardware
22:15:32 <elliott> and the browser uses a 32-bit value
22:15:34 <elliott> then it'll just wrap around
22:15:35 <Ngevd> (I know nothing of the interior of http, nor the intererior of web browsers, and only basic web design)
22:15:38 <elliott> and thus be expired immediately
22:16:01 <ais523> "-x c" is a useful command-line switch
22:16:12 <ais523> C-INTERCAL doesn't have an equivalent, so I have a symlink to /dev/tty called tty.i
22:18:29 -!- itidus21 has joined.
22:18:55 * elliott still doesn't know where to put that cron job :(
22:19:36 <elliott> also, what's a good time to run a cron job that might slow things down slightly for a few minutes?
22:21:15 <elliott> Ngevd: I was thinking 5-7 AM
22:21:23 <elliott> either 5 AM, 6 AM, or 7 AM UTC
22:21:35 <Ngevd> That's like, when I'm most productive!
22:21:47 <ais523> elliott: I was going to suggest 6-8 AM UTC, because that's pretty early in the UK, and translates to night in the US too
22:21:54 <ais523> perhaps a little later, at 9AM UTC
22:22:14 <elliott> technically people might wake up at 7 am and browse esolang while eating their cornflakes :P
22:22:22 <elliott> Friendship: i thought you were east coast
22:22:35 <Friendship> I am, but I've heard that even after I moved, people continued to live there.
22:24:30 <elliott> Friendship: so out of 5,6,7,8 AM UTC which is nicest for you
22:28:25 <Ngevd> But I can live with any of them
22:28:48 <Ngevd> Actually, thinking about it, 7 is the best
22:29:05 <ais523> it probably won't matter much, but I know that 10am UTC I'd be quite likely to be using esolangs.org, 9am UTC I wouldn't
22:29:14 <oerjan> i think i prefer approx. 17:00 UTC *mad cackle*
22:29:16 <ais523> so that's what it's like for a Brit
22:29:25 <elliott> oerjan: oh i didn't realise youw ere here!
22:29:30 <elliott> OUT OF THE OPTIONS THERE which is best
22:29:41 <elliott> ais523: neither of those were one of the options! :P
22:29:51 <ais523> elliott: but it's an indication of where the cutoff is
22:30:06 <Friendship> Incidentally, 7BST would probably be better since most of us have silly daylight savings time ;)
22:30:07 <ais523> so I'm reasonably indifferent between those
22:30:11 <oerjan> elliott: i think they're about the same for me
22:30:26 * elliott has seen ais523 online at 5 AM
22:30:44 <ais523> right, most commonly on the Saturday -> Sunday boundary
22:30:50 <ais523> it may well happen tonight, in fact
22:31:14 <elliott> ais523: i suppose 5,6,7,8 are all equally bad from that perspective then
22:31:40 <ais523> elliott: or indeed, equally good
22:31:44 <elliott> (this is hard to change after-the-fact, btw)
22:31:54 <elliott> (because hypothetical backuppers will set cronjobs by it)
22:33:15 <oerjan> elliott: i don't think there's a big difference
22:35:25 * oerjan once again ponders why his /// quine had to use quoting tokens one character longer than the BCT interpreter
22:36:23 <oerjan> (it's because i couldn't manage to avoid having the sequence //\\ in the program proper)
22:36:27 <zzo38> Do you have any comments of my ideas near timestamp 1330153950?
22:37:06 <elliott> Friendship: Where would YOU put a shell script to be run from a crontab of a system user?
22:38:06 <elliott> Do users without /homes even *have* crontabs?
22:39:59 <elliott> ais523: is this grammatically correct? *General links to people should be to a standard namespace article on them, if one exists, and to their user page otherwise. When linking to a user page, the User: prefix should be visible on the link.
22:40:16 -!- Sgeo has quit (Ping timeout: 276 seconds).
22:40:25 <ais523> it's correct, although a little tricky to parse
22:40:54 <Ngevd> Esolang thinking time
22:41:04 <Ngevd> A queue-based functional language
22:41:11 -!- Sgeo has joined.
22:41:23 <Ngevd> It would have forward Polish notation
22:41:41 <monqy> what would that do
22:42:03 <monqy> what does that mean
22:42:13 <Ngevd> It puts 11 into a monad
22:42:36 <elliott> But seriously, <elliott> Do users without /homes even *have* crontabs?
22:46:47 <pikhq_> I don't think so, though POSIX permits it.
22:47:49 <pikhq_> User crontabs are in /var/spool/cron/crontabs, at least in vixie cron.
22:48:11 <elliott> So, erm, did we ever come to a consensus on where to put this bloody shell script?
22:48:24 <pikhq_> Most common crons are either vixie cron or forks thereof, so there you go.
22:49:02 <pikhq_> I think /usr/local/libexec or /usr/local/lib is your best bet.
22:49:27 <pikhq_> Say, /usr/local/lib/admin-scripts/foo.sh
22:50:16 <elliott> The /usr/share hierarchy is for all read-only architecture independent data files. [30]
22:50:16 <elliott> This hierarchy is intended to be shareable among all architecture platforms of a given OS; thus, for example, a site with i386, Alpha, and PPC platforms might maintain a single /usr/share directory that is centrally-mounted. Note, however, that /usr/share is generally not intended to be shared by different OSes or by different releases of the same OS.
22:50:17 <elliott> For example, the perl5 subdirectory for Perl 5 modules and libraries.
22:50:32 <elliott> Much of this data originally lived in /usr (man, doc) or /usr/lib (dict, terminfo, zoneinfo).
22:50:39 <elliott> I don't think that excludes using /usr/share to store shellscripts.
22:50:54 <elliott> /usr/lib includes object files, libraries, and internal binaries that are not intended to be executed directly by users or shell scripts. [22]
22:50:54 <elliott> Applications may use a single subdirectory under /usr/lib. If an application uses a subdirectory, all architecture-dependent data exclusively used by the application must be placed within that subdirectory. [23]
22:50:57 <elliott> Miscellaneous architecture-independent application-specific static files and subdirectories must be placed in /usr/share.
22:51:02 <elliott> seems to exclude putting shell scripts in /usr/lib to me
22:51:12 <elliott> (and by analogy /usr/local/lib_)
22:53:29 <elliott> Fuck it, hi there /usr/local/lib/cron/[CENSORED]
22:56:51 <elliott> pikhq_: Hey, did you say what time is best for you?
22:56:58 <elliott> Friendship: ais523: oerjan: Ngevd: I went for 6 am, btw
22:57:44 <ais523> elliott: [censored] is the actual name of the script? if not, it should be
22:58:00 <elliott> If I said yes, would you believe me?
22:59:12 <pikhq_> elliott: What are you going with right now?
22:59:42 <elliott> Bt the options are, like, 5-7 UTC.
23:00:29 <pikhq_> 05:00-07:00, or 17:00 to 19:00?
23:01:07 <pikhq_> Looks like the 05:00-07:00 range.
23:01:25 <pikhq_> Mmm... T0700Z, I suppose.
23:01:55 <Ngevd> In my UPCOMING QUEUE ESOLANG, things are very weird
23:01:59 <elliott> OK, I'll stick with 6 AM for now, then.
23:02:15 <elliott> I'm running the script now FWIW.
23:02:18 <elliott> Is the wiki going at normal speed?
23:02:34 <elliott> It should only go very slightly slower at the worst.
23:02:41 <Ngevd> That would become * 5 17, then 85
23:03:07 <Ngevd> And because a value is at the top of the stack, that is converted to a character
23:03:44 <Ngevd> Actually, it's a queue
23:03:49 <Ngevd> AN INVISIBLE QUEUE
23:04:07 -!- derdon has joined.
23:04:35 <Ngevd> The program is converted into a queue of functions on integers and integers
23:04:54 <Ngevd> I have no idea how + + + would work
23:07:53 <Ngevd> Maybe having two consecutive functions at the head of the queue would rotate the queue?
23:08:12 <Ngevd> So, + + + 7 8 would become + + 15?
23:08:47 <Ngevd> I think all the functions need to be overloaded
23:08:54 -!- zzo38 has left.
23:09:25 <Ngevd> Perhaps + is the functional thingy like Haskell's .
23:16:01 <elliott> ais523: http://esolangs.org/wiki/Esolang:Wiki_dumps
23:16:07 <oerjan> well you mean "op1 op2 op3 arg1 arg2 ..." becomes "op3 arg1 arg2 ... op1 op2"?
23:16:41 <Ngevd> A Hello World program would be
23:16:41 <Ngevd> 72 101 108 108 111 44 32 87 111 114 108 100 33 H
23:16:42 <oerjan> an alternative would be to let it evaluate "op3 arg1 arg2 ..." one step
23:16:59 <ais523> elliott: hmm, I might put that in anacron, if I can remember how
23:17:00 -!- augur_ has changed nick to augur.
23:17:13 <oerjan> and then if it starts with enough arguments, going to op2
23:17:35 <Ngevd> That takes away the queue nature somewhat, though
23:18:11 <oerjan> > map (chr . read) $ words "72 101 108 108 111 44 32 87 111 114 108 100 33"
23:18:23 <elliott> ais523: as long as you use zsync P:
23:18:31 <ais523> actually, I'll do it manually
23:18:38 <ais523> I'm usually offline, and I don't think there's a cron-while-online thing
23:18:50 <elliott> (OK, 19 megabytes a day probably isn't too bad, but (a) it's a pain for users (b) it multiplies)
23:19:10 <oerjan> Ngevd: my latter would be same as saying it always takes the leftmost function which has enough arguments
23:19:52 <ais523> elliott: hmm, that's a gzipped file, right?
23:20:02 <ais523> would you expect it to stay mostly the same between downloads, if it's compressed?
23:20:13 <ais523> oh, it's downloading the gzipped version, not the unzipped version
23:20:18 <ais523> but the unzipped one is what you specify
23:21:05 <Ngevd> I'm thinking of calling this "Fueue"
23:21:05 <oerjan> Ngevd: oh, your first has the nice property that the queue is essentially circular
23:21:24 <elliott> it generates the gzip file specially so that it can do incremental downloads of it
23:21:42 <oerjan> you just move the ip one step forward in the circle if there are not enough arguments to apply a function
23:21:43 <ais523> still, the file will be changing in the middle, rather than at the end
23:21:46 <ais523> I wonder if that'll confuse it
23:21:55 <elliott> ais523: no, it looks inside the file
23:22:01 <elliott> and makes zlib do alignment so that it works out
23:22:06 <elliott> I don't know the details, but it's fancy
23:25:19 * elliott emails pgimeno about the dumps
23:25:27 <elliott> Your Linode, solidity, has exceeded the notification threshold (1000) for disk io rate by averaging 3115.59 for the last 2 hours. The dashboard for this Linode is located at: <https://manager.linode.com/linodes/dashboard/solidity>
23:25:40 <elliott> heh, because of the backups
23:26:23 <ais523> elliott: what happens if you do too much disk IO on Linode?
23:26:30 <ais523> do they charge extra, or throttle it, or what?
23:26:53 <itidus21> you would hope they throttle it
23:27:09 <elliott> ais523: no, it's just a configurable warning
23:27:16 <elliott> to notify you of potential problems
23:27:19 <ais523> that lets you know that something might be going wrong
23:27:21 <elliott> you can set the rate it warns you for
23:28:23 <elliott> oh, zsync also does checksumming
23:28:58 <pikhq_> That's kinda a big part of the rsync algorithm. :)
23:29:17 <elliott> Note to self: "Mozilla/5.0 (compatible; YandexBot/3.0; +http://yandex.com/bots)"
23:29:25 <elliott> (a /really/ badly-behaved spider)
23:29:27 <mRoman> The esodump is 150mb big?
23:29:42 <elliott> mRoman: yes, it includes all files and revisiosn
23:29:48 <elliott> it's only 19 megs compressed, though
23:30:22 <pikhq_> elliott: What's so badly behaved about said spider?
23:30:29 <ais523> elliott: does it obey robots.txt?
23:30:39 <elliott> doesn't even look at /robots.txt, spiders *really aggressively*
23:30:55 <elliott> like, it's a considerable percentage of entries in the log file
23:30:58 <pikhq_> ... It doesn't read robots.txt?
23:31:12 <pikhq_> iptables the shit out of it.
23:31:24 <elliott> i was going to just block its hostname at the server level
23:31:30 <elliott> i guess iptables would use less bandwidth
23:31:30 <ais523> elliott: it claims to look at robots.txt
23:31:32 <elliott> but i have no idea how to use iptables
23:31:35 <ais523> and to obey Crawl-delay
23:31:41 <elliott> ais523: every bot claims to be well-behaved
23:31:58 <ais523> it also says that there are bots that imitate it, and suggest that you rDNS the IP
23:32:25 <ais523> Yandex own yandex.ru, yandex.net, yandex.com, and forward- and reverse-DNS correctly
23:32:36 <ais523> and they recommend you check the IP to make sure
23:32:39 <elliott> I stand corrected, it did load /robots.txt
23:32:41 <elliott> but I don't think it's obeying it
23:33:01 <elliott> ais523: this one is the real thing
23:33:14 <elliott> it's still spidering far faster than any good bot would, though
23:33:15 <pikhq_> elliott: How so? Doesn't look like you're disallowing /wiki
23:33:48 <pikhq_> I'd try adding a Crawl-delay for them.
23:33:51 <ais523> elliott: try putting in a Crawl-delay:1 just after the Disallow line
23:34:12 <ais523> Phantom_Hoover: that doesn't even make sense
23:34:14 <elliott> ais523: yeah, I'll consider it; I wasn't quoting that because I'm going to ban it for sure, anyway
23:34:18 <ais523> unless PHP needs an embedded JS interp for some reason
23:34:19 <elliott> just in case I did want to, later
23:34:37 <elliott> I love the "X, like Y" convention that user-agents have sprung up
23:34:53 <elliott> "CPU OS 5_0_1 like Mac OS X"
23:35:13 <Phantom_Hoover> Like, they had some speech-recognition-specific function in the standard library for some reason.
23:35:21 <ais523> elliott: this is what happens with something like user-agent sniffing which wasn't standardised in advance
23:35:29 <ais523> Phantom_Hoover: it's because most PHP users wouldn't be able to install libraries
23:35:33 <pikhq_> And was handled by morons.
23:35:37 <ais523> so they have to put it into the core to make sure it's always available
23:35:46 <Ngevd> I wonder if this is Turing Complet
23:35:49 <ais523> "yandex, like Googlebot"
23:35:51 <Phantom_Hoover> Anyway, this makes perfect sense: it lets you use Javascript instead of PHP.
23:35:55 <pikhq_> "Check for string 'Mozilla'... Okay, we can use HTML 3!"
23:36:14 <elliott> I like how everybody criticises PHP's library but loves Python's
23:36:20 <elliott> probably namespaces are the difference
23:36:31 <ais523> elliott: oh, I don't like the bloat in Python's either
23:36:37 <ais523> but I'm probably a bit of a special case there
23:39:39 <ais523> oh, something I discovered today; you can tell strings(1) to look for UTF-16
23:40:34 <elliott> elliott@solidity:~$ sudo grep -i yandex /var/log/nginx/access.log | wc -l
23:40:34 <elliott> elliott@solidity:~$ sudo grep -i Googlebot /var/log/nginx/access.log | wc -l
23:46:06 <tswett> In Python, if you define a function bar() inside of a class Foo, then Foo.bar becomes a method instead of an ordinary function, and instances of Foo get weird access to it.
23:46:20 <tswett> Does anything special happen if you define a class Bar inside of a class Foo?
23:46:30 <tswett> Or does it just create a class Foo.Bar and that's that?
23:48:24 <Ngevd> Fueue spec on wiki
23:48:48 <Ngevd> I don't think it's Turing-Complete
23:49:37 <Ngevd> Actually, is it too late to decategorize and move into my user space?
23:49:43 <Ngevd> No, I'm not doing that
23:50:03 -!- Ngevd has quit (Quit: Goodbye).
23:51:17 <elliott> ais523: hey, what's a good way of purely-functionally picking a random element from a tree that satisfies a predicate?
23:51:37 <elliott> (assume you have a "RandomGen -> Range -> (IntInRange, RandomGen)" function or such)
23:51:42 <ais523> elliott: random as in arbitrary, or as in equally weighted?
23:52:44 <elliott> right now I'm collecting all elements from the tree that satisfy the predicate into a linked list, picking a random number in [0,elemsThatSatisfyIt-1] and then getting that element from the linked list
23:52:49 <elliott> which sucks, because linked list indexing is O(n)
23:52:50 <ais523> traverse the tree, checking each element to see if it satisfies the predicate, adding the ones that do to an indexable pure-functional structure, then select a random index, then pick the element with that index from it
23:53:04 <ais523> I'm trying to figure out what purely functional structures are indexable
23:53:19 <tswett> Does each tree node carry around the number of leaves it has?
23:53:35 <tswett> If so, you can use that to descend to a random leaf.
23:53:56 <elliott> tswett: Yes, but it has to be equally-weighted among the elements that satisfy the predicate.
23:53:56 <tswett> If not, why not" Go to '3:19.
23:54:02 <elliott> ais523: "adding the ones that do to an indexable pure-functional structure, then select a random index, then pick the element with that index from it"
23:54:07 <ais523> elliott: are you expecting most elements to satisfy the predicates?
23:54:15 <elliott> there's 2,3 finger trees (aka Seq)
23:54:18 <tswett> No biggie. Select random elements from the tree, without replacement, until you find one that satisfies the predicate.
23:54:40 <elliott> but indexing is O(log(min(i,n-i)))
23:54:48 <elliott> and it has "just OK" constant factors
23:54:49 <tswett> Unless you know something about where the good leaves are, in advance.
23:54:59 <ais523> as in, if you select a random element, is it probably going to satisfy the predicate?
23:55:03 <elliott> tswett: FSVO equally-weighted, I presume
23:55:20 <elliott> it's most likely not going to
23:55:20 <ais523> the best algo would probably depend on that
23:55:27 <elliott> ais523: also, in the vast majority of cases, only *one* will
23:55:28 <ais523> hmm, traversing's probably better, then
23:55:33 <elliott> and I'd prefer to avoid the random selection at all in that case
23:55:34 <tswett> This algorithm will select every good leaf with equal probability. There's no bias anywhere.
23:55:46 <ais523> oh, in that case, your existing algo is probably best, except in pathological cases
23:55:47 <elliott> but it'll be really slow :)
23:55:53 <elliott> because of the linked list
23:56:04 <ais523> elliott: but a linked list is pretty good for small n
23:56:22 <tswett> But when (and if) you implement the "without replacement" part, you'll need to remove each bad leaf you encounter from the tree, and update the tree's metadata.
23:56:50 <elliott> ais523: not as good as I'd like :(
23:56:51 <ais523> tswett: removing an element from a tree is expensive in a pure-functional language
23:56:54 * Sgeo wonders if he should be alarmed that in Common Lisp, (car nil) and (cdr nil) both give nil
23:56:59 <tswett> Otherwise, the tree ((bad bad bad bad good) (good)) will be biased toward... um...
23:57:05 <elliott> ais523: not /that/ expensive
23:57:17 <ais523> right, it's only O(log n) if the tree is balanced
23:57:27 <elliott> it is sort of balanced, IIRC
23:57:40 <ais523> assuming that the language shares common subtrees, and without mutation it has no reason not to do that
23:57:42 <elliott> The implementation of Map is based on size balanced binary trees (or trees of bounded balance) as described by:
23:57:42 <elliott> Stephen Adams, "Efficient sets: a balancing act", Journal of Functional Programming 3(4):553-562, October 1993, http://www.swiss.ai.mit.edu/~adams/BB/.
23:57:42 <elliott> J. Nievergelt and E.M. Reingold, "Binary search trees of bounded balance", SIAM journal of computing 2(1), March 1973.
23:57:45 <Sgeo> elliott, is Common Lisp a bad language in your opinion?
23:58:15 <oerjan> hm you can select a random element from a list without traversing it more than once.
23:58:38 <oerjan> (including no need to calculate the length first)
23:58:47 <tswett> Maybe you could pre-optimize things by forming the tree into a tree.
23:58:49 <elliott> oerjan: that's not helpful
23:58:57 <elliott> (I can accumulate it as I traverse the tree)
23:59:15 <oerjan> elliott: no, that's not what i mean
23:59:31 <oerjan> for this to be efficient, the list must be produced lazily.
23:59:41 <oerjan> and consumed. then this could fuse.
23:59:46 -!- tzxn4 has joined.
23:59:48 <elliott> i highly doubt it will fuse
23:59:52 <elliott> at least I doubt Data.Map.toList does
23:59:53 -!- tzxn4 has quit (Read error: Connection reset by peer).