00:00:12 <AnMaster> olsner, it is low latency stuff. Anyway wine doesn't use it. It uses alsa directly. I have hardware mixing. SB Live 5.1
00:00:57 <ehirdiphone> arva c'frta deîmu xœt'yr • pala'ds fvrofvr ta'ãr
00:01:32 <olsner> jack replaces alsa? or goes around it somehow?
00:01:56 <ehirdiphone> Uses ALSA. ITs for professional audio work
00:02:02 <AnMaster> olsner, but it runs daemon with realtime priority and such
00:02:27 -!- ehirdiphone has quit (Quit: Get Colloquy for iPhone! http://mobile.colloquy.info).
00:04:02 <AnMaster> olsner, anyway, last I tried spotify didn't work well in wine at all. Hm.
00:04:16 <AnMaster> other apps, like portal, works fine in there
00:04:38 <AnMaster> olsner, didn't like it very much though
00:05:01 <AnMaster> olsner, I mean, sure it is innovative to a degree. But I really dislike the FPS concept
00:05:41 <coppro> currying always happens
00:05:53 -!- cpressey has joined.
00:05:55 <CakeProphet> the way it's compiled, is that always the same?
00:05:56 <AnMaster> olsner, I much prefer the third person perspective of for example nwn
00:06:24 <CakeProphet> it seems like for a typical function call GHC would avoid explicit currying.
00:06:26 <olsner> my setup for flawless spotify sound: disable hardware acceleration in spotify, choose alsa driver and "hardware acceleration: Emulation" in the wine configuration
00:06:53 <AnMaster> olsner, eh, hw acceleration in wine? don't remember that setting. What does it do
00:07:03 <olsner> there is a JACK driver too in wine though, I would guess that it either works better or was what caused the problems :)
00:07:18 <olsner> AnMaster: well, you disable it to make spotify work better, that's about all I know
00:07:25 <AnMaster> olsner, it doesn't seem to be compiled here
00:07:43 <AnMaster> olsner, ah but can that be set per app?
00:07:49 <oerjan> CakeProphet: i am pretty sure ghc doesn't curry if it knows all arguments at the outset
00:08:03 <olsner> the wine configuration or the spotify configuration?
00:08:39 <AnMaster> olsner, the applications tab seems to be just about the windows version
00:08:47 <AnMaster> olsner, some other app needs full hw accleration iirc
00:09:09 <AnMaster> olsner, would be annoying to switch all the time
00:09:24 <oerjan> CakeProphet: and it does lots of inlining optimization and stuff, which probably also can remove currying
00:09:47 -!- augur has quit (Ping timeout: 245 seconds).
00:10:01 <AnMaster> oerjan, does that mean large and fast code rather than small?
00:10:44 <AnMaster> oerjan, thus resulting in ghc being unsuitable for systems with little memory/space?
00:11:21 <oerjan> AnMaster: well laziness and lots of garbage collection would do that anyway
00:11:25 <olsner> AnMaster: ok, the wine setting didn't seem to have any effect on spotify anyway
00:11:58 <CakeProphet> hmmm... it seems like you could actually implement laziness via optimization.
00:12:07 <olsner> summary: spotify works fine in wine on linux
00:12:08 <CakeProphet> like constructing for loops out of maps and such.
00:12:16 <oerjan> AnMaster: i don't know that much about ghc really, it probably has several parameters for what to optimize for
00:12:23 <olsner> CakeProphet: implement via optimization?
00:13:01 <AnMaster> oerjan, then who is our resident expert on ghc internals
00:13:02 <CakeProphet> basically optimize lazy code into its eager semantics at compile time... if possible.
00:13:10 <AnMaster> according coppro there is likely to be one!
00:13:30 <coppro> or one of us can retrieve one quickly
00:13:34 <coppro> that was the second part of the theorem
00:13:52 <AnMaster> coppro, I thought you said "one of us becoming one"
00:13:55 <coppro> although you could also view them as two separate theorems
00:14:03 <coppro> no, I meant could fetch one and bring em here
00:14:13 <olsner> CakeProphet: I think that's what you commonly get, since laziness for iteration is such a common pattern... if you have a top-level loop that prints a list with many lazy values, you may end up with a for-loop that evalutates then prints
00:14:43 <olsner> where "loop" = mapM print values or something like that
00:14:50 <oerjan> CakeProphet: ghc _does_ construct loops out of maps, and has a "rules" mechanism for optimizing several forms
00:14:55 -!- cheater99 has joined.
00:14:56 <AnMaster> coppro, I don't remember much such fetching?
00:15:05 <olsner> as long as it can see far enough down the line of turtles
00:15:06 <AnMaster> coppro, or any at all to tell the truth
00:15:30 <coppro> AnMaster: yeah, it doesn't go on much. It was more academic to also include the fact that many of us also mingle in relatively academic circles
00:15:40 <coppro> for instance, I could retrieve an expert in cellular metabolism quite quickly
00:15:50 <AnMaster> coppro, and get him to join here!?
00:16:00 <oerjan> AnMaster: oh well i know _some_ random ghc internals, just not detailed enough for space vs. time optimization
00:16:23 <coppro> AnMaster: he hangs around on EFNet and is a good friend, I'm sure I could get him to at least join and then part ;)
00:16:24 <olsner> why would you want to ask an expert in cellular metabolism about ghc internals?
00:16:53 <AnMaster> btw did I ever mention that silly warning on the warning page of my electrical piano?
00:17:08 <AnMaster> you know the page with stuff like "do not put something with water in on top" and such
00:17:19 -!- Gracenotes has quit (Ping timeout: 252 seconds).
00:17:36 <AnMaster> did I ever tell about the extremely silly warning there?
00:17:46 <AnMaster> (much sillier than that about the water)
00:17:54 <oerjan> do not dance on top of piano?
00:18:07 <AnMaster> oerjan, "do not stand or jump on the piano"
00:18:10 <olsner> AnMaster: btw, if you're worried about code size in haskell you should be much more worried about the RTS and statically linking gmp and a bunch of other stuff into your program
00:18:46 <AnMaster> micahjohnston, strangely enough that one was missing. Should contact them and point out that flaw
00:19:09 <olsner> of course, dynamic linking gets rid of some of that, but if you're talking about embedding a single haskell program you'll still have to install that stuff :)
00:19:16 <AnMaster> olsner, indeed. Can you get a program smaller than 40 kB?
00:19:42 <olsner> haskell has them built in you know
00:19:43 <AnMaster> olsner, oh I confused gmp and mpfr
00:20:14 <AnMaster> olsner, the use case I'm thinking about is 16-bit and has no FPU
00:20:14 <olsner> yeah, or gmp with an exponent per bignum
00:20:24 <AnMaster> olsner, and after OS is loaded it has 56 kB free ram
00:20:42 <olsner> ehm, hmm, you'd have to *port* ghc to that first... *good* *luck*
00:21:03 <AnMaster> coppro, I would have to port llvm to it
00:21:03 <oerjan> i think they've added an alternative option to gmp recently, for license reasons i think
00:21:10 <olsner> maybe ehc could do something though, iirc it compiles to C and has a very small RTS
00:21:25 <coppro> AnMaster: Adding a new target to LLVM is pretty cool, actually
00:21:44 <coppro> I wonder if it has an MMIX target
00:21:53 <coppro> or whatever it's called right now
00:22:07 <coppro> AnMaster: LLVM is good C++, and also TableGen
00:22:42 <coppro> AnMaster: a project-specific code generator that takes a lot of the pain out of things
00:23:00 <AnMaster> coppro, I see. This platform is highly RISC and I need to output COFF as the binary format
00:23:06 <olsner> another way to use haskell for embedded development is to use the type system to produce a type-rich DSL that just generates code for the device
00:23:06 <coppro> originally designed and still primarily used for handling instruction tables for various architectures
00:23:16 <oerjan> olsner: i thought i read jhc is the one with small program size, if that's still developed
00:23:20 <AnMaster> coppro, I think that is about 50
00:23:34 <olsner> oerjan: hmm, I might be confusing them
00:23:36 <coppro> AnMaster: I'll see if I can find some examples
00:23:54 <AnMaster> coppro, if you feel up to the task ;P
00:24:02 -!- poiuy_qwert has quit (Quit: Leaving).
00:24:10 <oerjan> olsner: well i couldn't say ehc isn't also
00:24:10 <AnMaster> coppro, last gcc version to support it for coff at least was gcc 3.x
00:24:11 <olsner> AnMaster: indeed, JHC is the one I'm thinking of
00:24:38 <coppro> LLVM supports COFF for some targets, but I'm not sure how easily portable that is
00:24:57 <oerjan> AnMaster: whole program optimizing compiler for haskell, iirc
00:25:01 <AnMaster> coppro, also I need custom linker script. But I guess I will use the same outdated binutils
00:25:08 <olsner> http://repetae.net/computer/jhc/
00:25:12 <AnMaster> oerjan, I'm extremely surprised
00:25:24 <oerjan> AnMaster: no ghc is mainly separate compilation
00:25:42 <olsner> ghc doesn't do whole-program optimizations, just cross-module optimizations
00:26:05 <AnMaster> olsner, huh, so make the set of modules you optimise between be the whole program?
00:26:31 <olsner> ghc mainly does stuff like output bytecode (or something similar) for selected functions so they can be inlined in other modules
00:26:50 <olsner> other functions will be output as machine code in the .o and can't be inlined
00:31:23 -!- clog has quit (Ping timeout: 265 seconds).
00:31:23 -!- clog has quit (ended).
00:31:27 -!- clog has joined.
00:31:27 -!- clog has joined.
00:34:32 -!- relet has quit (Quit: Leaving.).
00:41:06 -!- augur has joined.
00:44:47 -!- FireFly has joined.
00:47:11 <CakeProphet> ever been an attempt to implement monads in C?
00:48:18 <CakeProphet> "Introducing Monads in C. Better known as Fun With Function Pointers."
00:51:52 <CakeProphet> so are arrows kind of like flowchart-monads?
00:52:44 -!- Quadrescence has joined.
00:53:03 <CakeProphet> I think you could use them with a DSP library.
00:57:42 -!- augur has quit (Ping timeout: 245 seconds).
01:09:16 -!- calamari has quit (Quit: Leaving).
01:13:16 -!- augur has joined.
01:17:31 -!- BeholdMyGlory has changed nick to Behold.
01:17:42 -!- augur has quit (Ping timeout: 245 seconds).
01:31:42 -!- zzo38 has joined.
01:32:41 -!- augur has joined.
01:44:58 <Gregor> AnMaster: Funny you mention it.
01:45:03 <Gregor> AnMaster: I was just thinking about ZEE yesterday.
01:45:28 <Gregor> AnMaster: I'm still stuck at the same point: I can't make progress until I have a story, I can't write a story myself because I'm not that kind of creative, and I can't find anybody to write a story for me.
01:47:35 <zzo38> Gregor: Did you see my story "Harper's Challenge"? It is hardly done much, and it isn't entirely my own either, it is group of people named by a pseudonym. And it has appendix.
01:48:46 <Gregor> zzo38: The problem with ZEE in the story department is that the nature of the game (and more to the point, the achievability of getting the requisite photos) has an enormous impact on the possible stories.
01:48:56 <Gregor> I keep getting great but unimplementable story ideas.
01:49:04 <zzo38> What does ZEE means?
01:50:28 <zzo38> *** INTRODUCTION *** This introduction intentionally left blank.
01:53:36 <Gregor> Zoom-Enhance-Extrapolate. It's an image-based maze game parodying those silly scenes from spy movies where they take an arbitrary image, zoom in on it, "enhance" it, and "extrapolate" random garbage out of it.
01:53:46 <Gregor> Or at least, it's supposed to be.
01:53:56 <Gregor> Right now it's most of an image-based maze game engine with no story :P
01:59:57 <zzo38> Maybe it doesn't need to have a story
02:08:58 -!- yiyus has quit (Ping timeout: 258 seconds).
02:12:53 -!- augur has quit (Ping timeout: 265 seconds).
02:13:02 -!- zzo38 has quit (Remote host closed the connection).
02:17:28 -!- yiyus has joined.
02:28:03 -!- CakeProphet has quit (Read error: Operation timed out).
02:28:17 -!- CakeProphet has joined.
02:33:02 -!- CakeProphet has quit (Ping timeout: 260 seconds).
02:44:58 -!- CakeProphet has joined.
02:49:47 -!- CakeProphet has quit (Ping timeout: 272 seconds).
02:56:18 -!- ski has joined.
03:09:03 -!- FireFly has quit (Quit: swatted to death).
03:09:04 -!- Behold has quit (Read error: Connection reset by peer).
03:13:28 -!- GreaseMonkey has joined.
03:18:25 -!- CakeProphet has joined.
03:19:57 -!- boily has joined.
03:31:07 -!- cal153 has quit (Ping timeout: 248 seconds).
03:34:18 -!- boily has quit (Quit: leaving).
03:46:40 <oklopol> Gregor: i've been thinking about the game too, recently, but, as you might guess depending on how well you know me, so probably not, i feel like you shouldn't use photos but render images on the fly. that would be more authentic, in any case i also have some story ideas, i would probably work on them more if this had been my idea
03:47:22 <oklopol> "<Gregor> I keep getting great but unimplementable story ideas." <<< okay i haven't gotten far enough to think about implementability :P
03:47:30 <Gregor> Rendering is actually /more/ unachievable than using photos. Well, depending on your required fidelity I suppose ... rendering sufficiently photorealistic images (even of scenes excluding humans) is ... complicated.
03:49:49 <oklopol> just iterate over pixels and see what's there, i don't really care how good it looks, you could just zoom if you can't make out what's happening. i don't really care about the parody aspect, so this puritanian version without actual images would probably not have humans in the story.
03:50:18 <oklopol> (i have ideas with humans too tho :D)
03:50:40 <Gregor> Not having humans is a plus.
03:50:57 <Gregor> Both storywise and achievability-wise.
03:51:05 <oklopol> well implementabilitywise at least, and i guess also the two other reasons
03:51:10 <Gregor> (The story would have humans obviously, but the images are more about finding clues then catching people in the act)
03:51:26 <oklopol> well it could end with you catching ppl in the act maybe
03:52:30 <oklopol> someone holding a bloody knife with a dick in his hand
03:53:03 -!- cal153 has joined.
03:56:58 <oklopol> well we can leave the blood out i guess
03:57:06 <oklopol> if you don't like violence
03:58:08 <oklopol> maybe some of the clues would involve you seeing some kind of leaf and then an expert tells you the tree only grows in a 10m x 10m area in X-town
03:58:26 -!- augur has joined.
03:59:29 <oklopol> in one of those shows there was a leaf somewhere and they suddenly know exactly where in us some meeting takes place or something
04:00:03 <oklopol> "we're in luck there are only two of this tree in the whole universe"
04:00:38 <oklopol> or maybe a law and order ending where the killer goes free
04:10:01 <Gregor> That comes down more to "achievability" :P
04:10:21 <Gregor> Getting a picture of somebody with a bloody knive in one hand and a severed penis in the other is tricky.
04:17:41 <CakeProphet> oklopol: plot twist: the expert is actually an alien.
04:21:27 -!- oerjan has quit (Quit: ZZZZZZZZZZZZZZZZZZZZZ).
04:25:20 <CakeProphet> (Kernel Orangutan-oriented Programming Style)
04:25:37 <pikhq> Dental fillings suck.
04:26:03 -!- bsmntbombdood has joined.
04:26:06 <pikhq> I had 2 today and am getting 2 tomorrow.
04:26:19 <coppro> CakeProphet: I want FROOTLOOPS
04:26:25 <pikhq> And the two fillings I got today are currently sensitive as hell.
04:27:08 <coppro> CakeProphet: lots of sugar
04:27:39 <oklopol> Gregor: oh umm i was thinking he killed someone else, and then masturbated because killing people is, as we all know, pretty hot
04:29:03 <CakeProphet> Fashionably Redundant Object Optimization Typography, for Learning Orangutan-Oriented PRogramming Style
04:30:08 <CakeProphet> A fashionably redundant object optimization topology, you say? Oh, well now I'm intruged.
04:30:21 <CakeProphet> But only if there is a paper about it. Must have an abstract.
04:30:41 <CakeProphet> otherwise I will lose interest and become distracted.
04:34:06 <CakeProphet> "Towards an understanding of orangutan-oriented programming through reprogrammable hydro-banana-morphisms"
04:35:13 <CakeProphet> I meed only add a few "keywords" to any phrase and it becomes academic, regardless of how ridiculous the other words sound.
04:35:56 <Gregor> I accept your challenge.
04:36:03 <Gregor> Your words are: "Getting a picture of somebody with a bloody knive in one hand and a severed penis in the other is tricky."
04:36:08 <Gregor> Add words to that to make it academic.
04:36:26 -!- augur has quit (Ping timeout: 265 seconds).
04:37:07 <CakeProphet> -pretends to dissect when really he is just tired as shit-
04:40:19 <CakeProphet> Getting a picture of somebody with a blood knive in one hand and a severed penis in the other is tricky, but can be readily reduced to a problem of optical complexity theory classes.
04:41:11 -!- augur has joined.
04:42:15 <CakeProphet> or, more generally, orangutan-oriented programming with optical complexity /heuristics/.
04:44:56 <CakeProphet> Though, sometimes a banana-morphism is more elegant
04:51:26 -!- coppro has quit (Remote host closed the connection).
04:52:04 -!- coppro has joined.
05:04:28 -!- augur has quit (Ping timeout: 265 seconds).
05:21:39 <Sgeo> Woo! Finished the Tower of Hanoi code
05:27:49 -!- MizardX has joined.
05:28:11 <Sgeo> Well, almost finished
05:28:15 <Sgeo> It's still a bit crashy
05:28:39 <Sgeo> Well, not crashy, but once anyone's won, clicking the third column is enough for anyone to win
05:33:00 * Sgeo watches someone fail to solve Tower of Hanoi
05:33:10 <Sgeo> He'd be on the right track if the goal was the second column and not the third :/
05:33:16 -!- augur has joined.
06:20:49 -!- augur has quit (Ping timeout: 265 seconds).
06:45:30 -!- augur has joined.
06:51:03 <coppro> man, I wish the flashblock guy wasn't such a jerk
06:51:31 -!- augur has quit (Remote host closed the connection).
06:52:22 -!- augur has joined.
07:02:41 -!- tombom has joined.
07:50:26 -!- tombom has quit (Quit: Leaving).
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:01:58 -!- pikhq has quit (Remote host closed the connection).
08:08:05 <coppro> also damn you inexplicable compulsion to find the trope that best describes my dating situation
08:20:13 -!- GreaseMonkey has quit (Quit: I'm using NO SCRIPT WHATSOEVER - Download it at file:///dev/null).
09:03:46 -!- josh_ has joined.
09:08:55 -!- josh_ has quit (Quit: Leaving).
09:44:16 -!- MigoMipo has joined.
10:30:48 -!- FireFly has joined.
11:03:34 -!- ais523 has joined.
11:30:13 <AnMaster> ugh the course literature for the autumn includes one by Bjarne Stroustrup. Yeargh.
11:30:52 <ais523> is he that bad at writing books?
11:31:09 <AnMaster> ais523, no idea. But surely you know what the topic will be of it
11:31:36 <AnMaster> he is after all the inventor of C++...
11:31:48 <ais523> yes, that doesn't mean everything he does in his entire life is C++-related
11:32:05 <AnMaster> ais523, well there is a high probability of it being C++
11:32:18 <ais523> you mean you don't actually know what the book is about?
11:32:50 <AnMaster> ais523, of course I do. But I meant in general, it should not have been a unlikely guess that it was about C++ for you
11:32:59 <AnMaster> the title is "The C++ Programming Language, Special Edition"
11:33:21 <AnMaster> I have no clue what the special edition bit is about
11:33:21 -!- oerjan has joined.
11:33:53 <fizzie> ais523: All his books -- http://www2.research.att.com/~bs/books.html [at his homepage] -- have the word "C++" in them.
11:34:00 <AnMaster> oerjan, since when did you become holly-wood-Australian?
11:34:13 <ais523> fizzie: hmm, an unfortunately narrow choice of topic
11:34:29 <ais523> I suppose if you're famous for something, it's an easy topic to convince people to publish your books on
11:34:39 <AnMaster> ais523, the same is true with the word s/ly narrow// applied
11:34:53 <oerjan> AnMaster: since i started making my greetings silly
11:35:01 <AnMaster> oerjan, and when did you start with that?
11:35:13 <ais523> AnMaster: why are you hanging around in #esoteric if you can't even handle C++?
11:35:14 <oerjan> AnMaster: a few years ago?
11:35:28 <AnMaster> oh and an 864 page book about computer networking...
11:35:38 <AnMaster> I can't think how anyone could write that much on computer networking
11:36:00 <AnMaster> on the other hand I long ago concluded that in English, the writers of course literature are paid per word
11:36:00 <fizzie> ais523: He has a wide(r) selection of topics in the full list of publications (interviews, conference papers, journal articles). It's just books that are rather focused.
11:36:18 <AnMaster> which is probably not the case for Swedish course literature
11:36:26 <AnMaster> it tends to be much more concise
11:36:34 <AnMaster> than English course literature
11:36:38 <fizzie> I would think computer networking is an easy topic to write a lot about. There's a lot of it going on, and none of it is especially simple.
11:37:32 <AnMaster> I'm not sure how to fit it in my backpack though
11:37:45 <AnMaster> as in, I can't imagine it would fit with the other large books
11:38:27 <fizzie> The special edition of the C++ book is supposedly 1029 pages.
11:38:51 <AnMaster> another title: "Data Structures and Algorithm Analysis in C++". That is like taking some of the most interesting topics of compsci (data structures and algorithm analysis) and dipping it in the tar of C++
11:39:57 <AnMaster> fizzie, yet that one was about half the price of the thick networking book
11:39:59 <fizzie> Your C++ antipathy is borderline pathological. It's not like you'd (probably) encounter the most warped corners of the language in a data structures book.
11:40:43 <fizzie> You'd prefer MIX assembly more, right?-)
11:40:58 <AnMaster> and huh "The essence of artificial Intelligence". I guess they want to get us interested in their master program (AI stuff). I'm going to move to another university for the master program personally
11:41:50 <AnMaster> it talks about data structures unless I misremember completely
11:42:10 <fizzie> Not *that* much, though.
11:42:19 <AnMaster> fizzie, "Effective C++" doesn't sound like a fun title either
11:42:36 <oerjan> Scary Concepts in Programming
11:43:12 <fizzie> CLRS ("Introduction to Algorithms") is one data structures/algorithms book that's also reasonably reasonable; the only language it contains is their own pseudo. And the book's website has the LaTeX package to typeset it, which is nice for homework exercises.
11:44:09 <fizzie> It doesn't go very deep into algorithm analysis or more esoteric data structures, but, well, it does say "introduction" right there in the title.
11:45:07 <AnMaster> fizzie, I much more look forward to the Swedish "Datatyper och algoritmer" which is bound to be a lot more concise. (see above wrt getting paid per word)
11:46:29 <fizzie> I'm not sure if paid-per-word is the reason, or just difference in cultures. Certainly the same distinction is there with English vs. Finnish textbooks. Especially the mathematics one.
11:47:26 <AnMaster> fizzie, I much prefer the conciser variant, it means I save hours.
11:47:27 <fizzie> And I'm not sure if that's so "especial", it's just that there's very little material in Finnish on our courses.
11:47:46 <AnMaster> fizzie, there have been about 50% so far, less during the year to come
11:49:02 <fizzie> You have a bit larger audience there, I guess it helps. I don't think I really have any computer-science related Finnish course literature on the bookshelf. There's some discrete maths and probabilistics, and then some electronics.
11:50:36 <fizzie> Oh, and a single, very narrow "C-ohjelmointikieli" (lit. "the C programming language"), though that's actually my wife's. (I haven't bothered to buy copies of all books.)
11:50:42 <AnMaster> fizzie, I have discrete maths, C, digital logic, electronics and database in Swedish. Yeah, that is what compsci turned into here :(
11:50:49 <AnMaster> and from what I heard, in many other places
11:51:07 <AnMaster> fizzie, your wife does CS too? I didn't remember that
11:51:38 <fizzie> Yes, though with a rather different focus (usability and communications and things like that) than I.
11:51:49 <AnMaster> fizzie, sounds like it might involve GUI
11:52:05 <oerjan> _her_ creations are actually usable
11:52:12 <fizzie> It also often involves people, which is my main reason to avoid it. :p
11:52:18 <AnMaster> oerjan, hey fungot is fully usable
11:52:19 <fungot> AnMaster: i just wrote?
11:52:56 <fizzie> oerjan: It's not as funny because it's so true.
11:53:11 <fizzie> But yes, who can ask for more than fungot; there's even a help command.
11:53:11 <fungot> fizzie: cmeme is a log bot afaik
11:53:54 <AnMaster> fizzie, which channel is it in?
11:54:03 <fizzie> [2005-10-11 13:57:19] < Gs30ng> cmeme is a log bot afaik
11:54:58 <oerjan> AnMaster: cmeme has been gone for years
11:55:10 -!- atrapado has joined.
11:55:34 <oerjan> which is a bit of a shame because its logs were much better formatted
11:55:47 <fizzie> oerjan: cmeme last quit from #esoteric in 2008-11-06, that's not so many years.
11:55:50 <oerjan> although that also meant they loaded slow
11:56:07 <oerjan> fizzie: well it's about half they years i've been here
11:56:56 <fizzie> There's still references in the tunes.org log-directory HEADER.html:
11:56:59 <fizzie> For a so-called "pretty" view of these logs, go to http://tunes.org/~coreyr/.
11:57:00 <fizzie> For even "prettier" (css'd, searchable, customizable, etc) logs, go to http://meme.b9.com.
11:57:15 <fizzie> But meme.b9.com -> ircbrowse.com, which does not exist.
11:58:00 <AnMaster> the logs are perfectly readable
11:58:34 <fizzie> The CSS thing was indeed prettier.
11:58:51 <fizzie> You could twiddle with what it looks like and all.
11:59:04 <AnMaster> fizzie, you could make it align nicely easily enough
11:59:56 <fizzie> Yes, and you could do any sort of formatting you want locally, but it was already prettified with several different styles out-of-the-box there.
12:00:09 <AnMaster> meh, who needs out of the box?
12:00:23 <fizzie> People who have other stuff to fiddle with, I guess.
12:00:36 <fizzie> I wonder if there's a LaTeX package for proper typesetting of irclogs yet.
12:00:43 <oerjan> yeah people who cannot whip up a quick prettifier in brainfuck don't _deserve_ to read the logs
12:01:06 <fizzie> oerjan: Yeah, those people can stay in the box.
12:02:33 <fizzie> My #esoteric log for 2003-2008 is 46 megabytes of text; that'd make a pretty nice set of books if bound into hardcover with attractive cover art, good typography and all that fluff.
12:02:47 <fizzie> Then I could quit my day job, and be a door-to-door ency^H^H^H^H#esoteric salesman.
12:03:52 <oerjan> slogan: it's better than fake persian rugs
12:04:01 <oerjan> AnMaster: YOU ARE NOT WORTHY
12:12:02 -!- BeholdMyGlory has joined.
12:26:32 -!- ais523 has quit (Remote host closed the connection).
12:50:11 -!- augur has quit (Remote host closed the connection).
12:53:32 -!- oerjan has quit (Quit: Later).
13:53:23 -!- bluebooblue has joined.
13:56:04 -!- bluebooblue has quit (Client Quit).
14:08:06 <oklopol> AnMaster: i've read that and i liked it even though i hate c++. but then again i like most books so i guess this is not very helpful... :D
14:11:18 <oklopol> books about algorithms that use an object oriented language are insane, they only have time to cover a few trivial ones because every tiny snippet takes pages and pages of code
14:11:56 <oklopol> (i'm not saying oo is inherently too verbose, i'm just grouping java and c++)
14:32:29 <AnMaster> oklopol, I fully agree with you when it comes to C++ and java
14:46:50 -!- coppro has quit (Remote host closed the connection).
14:48:07 -!- augur has joined.
15:00:36 <oklopol> i once read this 300 or 500 or something page book about algorithms in one afternoon, because it was always 5 pages of explanation of something completely trivial, followed by the same amount of code, the algorithms could've been explained much more concisely, because it was only the level of dijkstra and quicksort, but i guess they wanted to keep the text-to-code ratio sensible and added fluff (i guess i skipped many of the proofs t
15:02:59 <Deewiant> You really should configure your client properly / install a plugin so that you don't have to worry about splitting long lines manually
15:03:35 <oklopol> i should get nnscript or start using another client, mirc is horrible alone
15:04:02 <oklopol> but that's SOOOOOO much work
15:05:40 <oklopol> hmm okay i would have to get an older mirc
15:05:55 <oklopol> so maybe another client, but, well, i'm not THAT NEW an oklopol.
15:06:30 <CakeProphet> They should blink at me from the bottom of the screen
15:06:39 <oklopol> i find that really annoying too
15:06:59 <oklopol> well if it was once, but i mean if they just start blinking away like crazy
15:07:13 <CakeProphet> I wonder if I can change that in Gnome/Xorg/whatever
15:08:06 <Deewiant> It depends on your WM (so not Xorg)
15:08:33 <CakeProphet> well... no blink, but simply notify. The way Ubuntu does it is very subtle. Nowhere near as obnoxious as the blinking orange in XP.
15:08:51 <CakeProphet> Dunno about other distros. Haven't used them.
15:08:52 <AnMaster> <oklopol> that's the OLD oklopol
15:09:31 <AnMaster> <oklopol> well if it was once, but i mean if they just start blinking away like crazy <-- on my system they tend to blink twice and then stay in the highlighted variant
15:09:39 <AnMaster> oklopol, I find that works very well
15:09:51 <AnMaster> I don't think I manually configured it
15:10:14 <oklopol> i wouldn't like highlight either, probably, actually i probably couldn't stand it just to know something's happened in a window... i should be kept ignorant until i choose to look, unless it's really important
15:10:48 <AnMaster> oklopol, well it is just a shade of light blue at 20% opacity on top or such
15:10:53 <CakeProphet> well, most reasons for focus-stealing are important. new IMs, update manager finishes, new window opens, etc
15:11:09 <AnMaster> CakeProphet, I would hate that
15:11:27 <AnMaster> and pacman or apt-get never steals focus ;P
15:11:39 <AnMaster> (pacman is the package manager on arch linux)
15:11:55 <CakeProphet> AnMaster: yes. That is why I don't like focus-stealing and would prefer to switch all of it to highlighted notifications because they were nowhere near as intrusive to what I'm doing.
15:12:10 -!- Geekthras has quit (Ping timeout: 252 seconds).
15:12:10 <CakeProphet> the reason I came to this conclisuion was actually update manager though
15:12:50 -!- Geekthras has joined.
15:15:03 <AnMaster> CakeProphet, actually there is one thing that should steal focus. screen lockers
15:18:39 <AnMaster> bbl, going to do upgrades that need X not running
15:18:40 <CakeProphet> well then I'm going to reprogram GNOME to have "steal-focus" and "no-really-actually-steal-focus"
15:18:52 <CakeProphet> and then reprogram my screen locker to use the second onew.
15:27:11 <CakeProphet> AnMaster: Do screen lockers actually use the same mechanism to steal focus? They seem... distinctly different.
16:03:21 -!- relet has joined.
16:05:07 -!- Phantom_Hoover has joined.
16:06:32 <Gregor> R is larger than the set of computable numbers. Therefore I suspect that it is not.
16:07:13 <Phantom_Hoover> OK, then is it possible to biject two computable reals onto one?
16:07:39 <Phantom_Hoover> (Also, can't you apply the diagonal argument to computable reals?)
16:11:12 <Phantom_Hoover> Hence implying that they are also uncountably infinite?
16:43:42 -!- MigoMipo has quit (Remote host closed the connection).
16:44:06 -!- MigoMipo has joined.
16:48:12 <uorygl> The computable reals are definitely countably infinite.
16:48:17 -!- augur has quit (Remote host closed the connection).
16:49:21 <uorygl> The R^2 -> R bijection is computable if a real number x is represented by the predicate P(r) = r < x over rational numbers r. I think.
16:49:43 -!- augur has joined.
16:55:26 <uorygl> There are countably many Turing machines. Therefore, there are countably many computable real numbers.
16:55:55 <uorygl> The diagonal argument doesn't work because it's impossible to determine whether a Turing machine actually outputs a real number or not.
16:58:40 <augur> uorygl: reals are non-countable
16:58:53 <augur> rationals are countable but not reals
16:59:09 <Sgeo> But countable reals are a subset of reals
16:59:14 <Sgeo> erm, computable
17:00:14 <uorygl> augur: I never said they were.
17:00:50 <augur> oh sorry. computable reals. ok
17:03:54 <AnMaster> hm uncomputable reals. Chaitin's Constant and such right? Are they all like that? I mean, that we can't even have any clue about the value?
17:04:37 <uorygl> I'm thinking of an uncomputable real. I will gladly give you its entire decimal expansion.
17:05:05 <AnMaster> uorygl, is pi computable? I presume so
17:10:33 <Phantom_Hoover> The Tth bit of a real is 1 if there is a decay and 0 otherwise.
17:11:44 <uorygl> Anyway, I can imagine there being an uncomputable real number that we can approximate really well, except that we never actually know how close the approximation is.
17:13:19 <uorygl> Perhaps you're computing it, and the computation has lingered at 4.177187787026364558800429098840 for years, and then it suddenly increases to 6.
17:13:39 <uorygl> You can go, "Aw, we thought we were so close."
17:14:55 <uorygl> Heck, that's what Chaitin's constant is like, I think. You can run a computation that outputs an increasing sequence of numbers that approaches it. You'll never actually get there, and you'll never know how far it still has to increase.
17:17:42 -!- tombom has joined.
17:26:00 <cpressey> It's so much more convenient that way.
17:28:39 <Mathnerd314> uorygl: I think you can determine digits of it, in exponential time
17:29:30 <uorygl> That would be computing it.
17:29:39 <uorygl> Are you assuming that a Turing machine always halts in exponential time?
17:31:32 <cpressey> Well, Chaitin's "constant" is per TM. So some Chaitin's constants can be computed.
17:32:09 <cpressey> Maybe "coefficient" would be a less misleading term.
17:33:15 -!- wareya has quit (Ping timeout: 248 seconds).
17:33:29 <cpressey> I held two jobs developing C++, and I don't think I wrote a significant amount of C++ code at either of them. "Effective C++" helped me cope.
17:33:49 <cpressey> Oh. Actually 3. But one had some Perl and Python in it too.
17:40:43 <cpressey> C++ is ok, but only if you consciously forget that it is C++.
17:52:55 <Phantom_Hoover> cpressey, developing does not seem like an enjoyable profession.
17:53:27 <cpressey> Phantom_Hoover: Ha ha ha ha ha. It depends *a lot* on the place.
17:53:59 <cpressey> Never, EVER work at a company that makes their developers wear pagers. EVER.
17:55:02 <cpressey> Unless you enjoy stress-related illness, sure. Go for it. Have an ulcer for me!
17:56:34 <Phantom_Hoover> I hate UI programming in general, come to think of it.
17:58:14 <Sgeo> ROFL at bug in Tower of Hanoi code
17:58:18 <cpressey> Dip switches and panel LEDs all the way.
17:58:40 <Sgeo> It caused the puzzle to start with the disks the wrong way around
17:58:49 <Phantom_Hoover> (This principle should not apply to anyone writing software that I want to use)
18:00:03 * Phantom_Hoover loves orbital paths when G is proportional to the inverse of distance.
18:01:07 <cpressey> DIP switches and panel LEDs, yet still have a window manager that steal focus. Perhaps by flashing something distractingly nearby.
18:02:41 <Sgeo> http://i.imgur.com/P623C.jpg
18:04:26 -!- wareya has joined.
18:05:36 -!- sebbu2 has joined.
18:07:20 -!- sebbu has quit (Ping timeout: 265 seconds).
18:07:20 -!- sebbu2 has changed nick to sebbu.
18:08:31 <CakeProphet> for some reason my volume control disappeared upon upgrade to luci.
18:08:57 <Phantom_Hoover> CakeProphet, you're lucky weirder stuff didn't happen.
18:15:47 <cpressey> Doctor, I can't find this patient's pulse!
18:18:11 <cpressey> SELECT MEASUREMENT FROM VITALS WHERE MEDICAL_TERM='PULSE';
18:20:01 -!- wareya has quit (Ping timeout: 265 seconds).
18:20:50 <fizzie> Programmable dip switches, so that you can toggle them with software. And they should make a loud click whenever they toggle status.
18:31:02 -!- atrapado has quit (Quit: Abandonando).
18:40:58 <cpressey> It's a switch specially made for people who are dips.
18:42:55 -!- wareya has joined.
18:43:34 -!- Mathnerd314 has quit (Quit: ChatZilla 0.9.86-rdmsoft [XULRunner 1.9.2.3/20100401080539]).
18:43:45 <oklopol> Phantom_Hoover: no you can't apply the diagonal argument, who says the diagonal is wait are you talking about this i'll read
18:45:41 <oklopol> we can approximate many uncomputable real numbers well, take any uncomputable number and add a million random digits in the beginning
18:47:29 <oklopol> Phantom_Hoover: i read logs and just answer directly no matter how far up i am
18:47:50 <oklopol> but umm computable reals being uncountable
18:48:26 <cpressey> Um... every computable real is generated by some TM. And the TMs are countable.
18:48:30 <oklopol> you can't apply the diagonal argument, you take the surjection from naturals, and you find no contradiction
18:49:06 <oklopol> cpressey: yes, but that doesn't mean you can't ask WHY the converse can't be proven in some way
18:49:13 <oklopol> it's computable for obvious reasons?
18:49:21 <oklopol> i find them very much non-obvious
18:50:02 <oklopol> Phantom_Hoover: so there cannot be a COMPUTABLE surjection.
18:50:19 <cpressey> I'm finding something here very non-obvious.
18:50:20 <oklopol> if the surjection is not computable, there should be no contradiction
18:50:42 <oklopol> i mean i haven't thought about this, but i believe that's the answer
18:51:08 <oklopol> well obviously if you had a computable surjection, you could just iterate through nats and compute the finite amount of digits you need at that point
18:51:39 <oklopol> it would always take finite time because we're enumerating computable reals, which can be computed to any precision in finite time
18:52:50 <oklopol> maybe there's a concept of computable countability out there.
18:53:06 <oklopol> the computable reals are computably uncountable
18:53:34 <cpressey> I totally think there's something wrong here, but I can't put my finger on it.
18:54:05 <oklopol> i don't think there is, if you have a more specific feeling where the wrong is, i can try to elaborate
18:54:59 <oklopol> the point is there is no contradiction, you can't compute the real on the diagonal, because we did not at any point assume that we have an algorithm that lists computable reals, we just assumed a mathematical surjection from nats to creals
18:55:11 <cpressey> Say you have a TM-enumerating TM, call it K. Interpret the tape ("in the limit") of each of the enumerated TMs as a real. Then K generates all computable reals. Therefore the set of computable reals must be countable.
18:55:17 <oklopol> but you mean maybe we could just enumerate tm's?
18:55:24 <oklopol> yeah let's see what the problem is there
18:55:58 <oklopol> you can't know whether a tm outputs a computable real or whether it's non-halting, and only gives finite output
18:56:08 <oklopol> but i'll read what you said maybe
18:56:42 <cpressey> Well, a TM can still generate a computable real, and never halt. At least, that's how I interpreted Turing's paper.
18:56:54 <oklopol> why would you need a tm enumerating the tm's if you're proving they are countable?
18:57:22 <cpressey> You don't need it, except to graphically illustrate that it can all be done with one TM.
18:57:59 <oklopol> well okay, sure, but anyway it's trivial they are countable, the question was why diagonalization doesn't work
18:58:33 <uorygl> So in summary, I was right. :P
18:59:14 <oklopol> i guess that's as good a summary as any
18:59:15 <cpressey> oklopol: Granted. Actually that was why I introduced K. Enumerating another TM is kind of like reading off the diagonal and adding another row.
18:59:44 <cpressey> I say "kind of" because that's the part where I can't put my finger on it.
19:00:20 <cpressey> Maybe "computably diagnolizable"?
19:01:10 <uorygl> Hang on. I can't Fennicize my name as "Tänneri Svetti", because "svetti" isn't a type of place!
19:01:31 <uorygl> I'll have to make it one, I guess.
19:01:52 <uorygl> Why am I still thinking about this? :|
19:01:54 <cpressey> I dunno. I do like the idea of calling the "computably enumerable reals", though, and abbreviating it to "cereals"
19:01:58 <oklopol> cpressey: i don't think it's in any sense like reading off the diagonal and adding another row tbh :P
19:02:15 <oklopol> uorygl: fennicize your name?
19:02:47 <CakeProphet> cpressey: Now develop a concept a "dry" to "soggy" gradient for cereals.
19:03:09 <oklopol> uorygl: what's your actual name i forget
19:03:52 <oklopol> one of the few that ends with r
19:08:37 <uorygl> Hey, you made my Internet drop out.
19:09:02 <uorygl> Wow, and it means "field" or "ground", so it works as a surname!
19:09:32 <oklopol> it wouldn't surprise me if it was used as a surname, although i haven't seen it
19:09:45 <uorygl> Where's that Finnish name database?
19:10:33 <oklopol> or vestrekisteri or something or maybe something completely elseous.
19:10:59 <uorygl> Found it. http://verkkopalvelut.vrk.fi/Nimipalvelu/default.asp?L=3
19:11:29 <uorygl> Don't new surnames have to be unique?
19:11:32 <oklopol> so probably i have seen it
19:12:00 <uorygl> If that's the case, I can't change my surname to Tanner.
19:12:20 <uorygl> Svetti is open, as is Svettila.
19:12:46 <oklopol> http://fi.wikipedia.org/wiki/Sukunimilaki has the rules
19:12:48 <uorygl> But nothing can beat the sheer awesomeness of the word "Vorigali Turrila".
19:13:23 <uorygl> That would be helpful.
19:14:33 <oklopol> it can't be "cockville", it can't be "asdfgljhaoig", it can't be "everyman" (very common, i don't know what english surnames are most common), it can't be "john", it can't be "microsoft"
19:15:08 <oklopol> it says it can't be the surname of an existing family, maybe that means if they have like a protected surname, you can do that
19:15:25 <oklopol> because otherwise it's just a stronger everyman
19:15:55 <oklopol> so i do not know whether tanner qualifies
19:16:24 <oklopol> also sweat is hiki in finnish, you could use that a first name
19:17:03 <uorygl> I'll change that to Hikki. :P
19:17:45 <oklopol> my father's name is heikki
19:21:40 -!- calamari has joined.
19:23:09 <calamari> at jury duty, waiting.. reading a book about writing linux device drivers.. so yeah bored lol
19:23:49 -!- Phantom_Hoover has quit (Ping timeout: 264 seconds).
19:24:17 <uorygl> We shall all learn or teach Finnish while we wait.
19:27:22 <oklopol> suomenopiskelumateriaaliksi sinulle
19:27:26 <cpressey> Hm, they give you internet access in jury duty? Well, I guess if it's not sequestered and all...
19:27:53 <uorygl> Here onto you easy a sentence?
19:28:25 <calamari> well im ircing from my phone, but in the courtroom you have to turn phones off
19:28:45 <uorygl> Here, for you, an easy sentence.
19:29:33 * cpressey was almost picked for jury duty once, but they rejected me
19:30:57 <uorygl> suomenopiskelumateriaaliksi is "as Finnish study material"?
19:31:06 <calamari> yeah I've been in the courtroom once but was ultimately rejected
19:32:14 <calamari> I get called down every year, so I suppose eventually I may be selected
19:32:59 * cpressey wonders about the crime rate in Arizona
19:33:26 -!- calamari has quit (Remote host closed the connection).
19:33:38 <uorygl> Pending / New: 17. Active - Due: 42.
19:34:01 <oklopol> oh, you mean "erinomaista!"?
19:34:17 <oklopol> saying "erinomainen" is like saying "a great!"
19:34:29 <uorygl> So you use the partitive when yelling adjectives?
19:36:35 -!- calamari has joined.
19:38:23 <oklopol> also "liian pitk" can't really be used, yet you could definitely say "liian rumaa"
19:38:45 <oklopol> "liian pitk" or preferably a whole sentence
19:40:01 <oklopol> i don't know why, but short=lyhyt and pitk=long both sound weird used like that
19:40:28 <uorygl> Maybe it's like the English "fastly". It's never used, but there is no reason for this.
19:40:34 <oklopol> or well actually i think i do know, but err
19:40:47 <uorygl> Actually, I guess "fast" can be used as an adverb.
19:42:23 <oklopol> the "reason" is things can't be long in general, there has to be some object that's long
19:42:31 <oklopol> (but it can be an abstract object)
19:42:34 <cpressey> "fast" and "big" are extremely strange words in English, given how common they are.
19:42:48 <uorygl> Hm, "bigly" doesn't work, either.
19:42:57 <uorygl> Though I can't imagine why you would use that word. :P
19:42:59 <cpressey> "The show was a large hit!" No one says this.
19:42:59 <oklopol> indeed, i never even realized
19:43:20 <oklopol> yes that saying is largely unused
19:44:17 <oklopol> in finnish you can say you're "bigly happy"
19:44:40 <cpressey> You'd probably be better off saying "immensely happy" in English.
19:44:41 <oklopol> (not that it's all that idiomatic in that specific context)
19:44:50 <uorygl> How would you say "The dog is named Swarming"?
19:45:00 <uorygl> (That word was chosen to be as un-Finnish as possible. :P)
19:45:06 -!- coppro has joined.
19:45:08 <oklopol> koiran nimi on ...swarming
19:45:37 <oklopol> (was the point that i wouldn't translate? :P)
19:48:25 -!- calamari has quit (Quit: AndroidIRC 1.0).
19:49:10 <cpressey> Ooh, Android. And here I am with a 5-year-old Motorola Tracfone.
19:49:58 <oklopol> ah parveilla, i assumed there isn't a translation :-D
19:50:18 -!- oerjan has joined.
19:50:24 <oklopol> maybe Deewiant should teach you
19:50:35 <oklopol> oerjan and i have to talk about computable uncountability now
19:50:44 <uorygl> Everyone who knows any Finnish knows more Finnish than me!
19:50:56 <uorygl> Except for the mathematical impossibility of that.
19:51:02 <oklopol> yes, the computable reals are computably uncountable
19:51:13 <oerjan> uorygl: i sincerely doubt that
19:51:31 <oerjan> oklopol: well that's just diagonalization
19:51:36 <cpressey> oerjan: You may wish to read the logs and enlighten us. We're confused.
19:51:39 <oerjan> same as for the usual one
19:52:09 <oklopol> or well i guess cpressey is then :P
19:53:00 <oklopol> oerjan: yes diagonalization proves it, you consider this trivial, so i assume you took the same definition as i?
19:53:33 <oerjan> well i assume it doesn't depend _that_ much on definition
19:54:07 <oklopol> i'm wondering how this works for arbitrary sets
19:54:13 <oerjan> but if you have a computable function Natural -> CompReal it shouldn't be that hard to construct a computable real it doesn't hit with diagonalization
19:54:43 <cpressey> I'm basing "we're confused" on the observation that Phantom_Hoover and I seem to disagree.
19:54:45 <oklopol> well yeah the proof is simple, i just found it an interesting concept
19:54:56 <cpressey> At least one of us is confused.
19:55:15 <oklopol> cpressey: about what? phantom_hoover asked why diagonalization doesn't work, i explained, you were correct of course
19:55:39 <oklopol> we were just solving a paradox
19:55:53 <oerjan> <CakeProphet> windows should never steal focus
19:55:59 <cpressey> The cardinality of the of computable reals certainly looks countably infinite to me. But I don't see exactly what the flaw is in Phantom_Hoover's diagonalization argument.
19:56:21 <oklopol> i told you, the surjection isn't countable so you cannot compute the real on the diagonal
19:56:37 <uorygl> The diagonalization argument requires you to construct a computable real out of the diagonal of the list of all computable reals.
19:56:44 <uorygl> The thing is, no list of all computable reals is computable.
19:57:17 <coppro> Could you not create a Godel code to encode the algorithms to generate them?
19:57:34 <cpressey> Yeah, I thought I called that thing K...
19:57:34 <uorygl> You'd need to verify that everything in the list is actually a computable real.
19:57:57 <oerjan> coppro: no you need to solve something like the halting problem to check whether an algorithm generates a computable real
19:58:36 <cpressey> How to you "verify" anything about an enumeration machine?
19:58:48 <cpressey> The machine runs forever, generating more and more digits.
19:58:52 <oklopol> i think what the problem is
19:58:58 <cpressey> But the real it's generating is, therefore, computable.
19:59:01 <oklopol> cpressey: do you think diagonalization is used to prove a set is COUNTABLE?
19:59:22 <oklopol> you said something about adding a new element to the list
19:59:33 <cpressey> oklopol: No, it's used to prove a set has a higher cardinality than the sets making up the rows and columns, I thought.
20:00:03 <oklopol> okay then i still don't see why you want K, if you're proving the set of creals is countable, you don't need it
20:00:06 <uorygl> cpressey: given any Turing machine that purportedly represents a real number, there are three possibilities: one, it halts; two, it outputs infinitely many digits; three, it outputs finitely many digits and stalls.
20:00:30 <uorygl> Given a Turing machine, it is impossible to tell which class it falls into; this is equivalent to solving the halting problem.
20:00:34 <cpressey> uorygl: You could reduce that to 2 possibilities, if it's useful.
20:01:00 <oklopol> oerjan: is the existence of a computable bijection an equivalence relation on the class of all sets, is what i'm asking? well at least up to some classical cardinality
20:01:00 <cpressey> uorygl: Actually, no. We're talking about *enumeration* machines here, which never halt.
20:01:19 <uorygl> Okay, so one is not a possibility. Two and three still are.
20:01:24 <uorygl> And two and three still cannot be distinguished.
20:01:49 -!- oerjan has quit (Read error: Operation timed out).
20:02:12 <uorygl> oklopol: it satisfies R, S and T; that makes it an equivalence relation, right?
20:02:26 <oklopol> it does? how do you define it?
20:03:22 <cpressey> uorygl: I'm not quite convinced that they need to be distinguished. But I'll have to give thought to it later, unfortunately.
20:04:57 <uorygl> oklopol: I don't really know.
20:05:07 <oklopol> uorygl: but he isn't trying to prove anything about diagonalization, so K is not actually used for anything but another name for surjection
20:05:21 <uorygl> cpressey: of course they do. A list of computable real numbers can't contain something that isn't a computable real number.
20:05:23 <oklopol> he's trying to prove creals countable, so he doesn't use the enumeration part at all
20:05:29 <uorygl> Thus, you need to weed out the things that aren't computable real numbers.
20:07:21 <oklopol> cpressey: but assuming you wanted to know why diagonalization doesn't work (which you say you don't), then the reason they have to be distinguished is that while you can list all tm's, some of them never halt
20:07:22 <oklopol> but only output a finite amount of numbers, while you can, mathematically, interpret these as creals, a program that tries to get the nth digit (depending on how far in the diagonal we are), has to know whether the program actually ever outputs n digits, otherwise it has to use 0 or something
20:07:27 -!- oerjan has joined.
20:08:05 <oklopol> with R you can say a bijection has to map reals to something as a limit process, 2^R is so big you can't really even do that
20:08:47 <oklopol> because there's no way to get enough information in to distinguish all of those (well this is not very precise, maybe there's some fucked up way)
20:10:59 <oklopol> maybe we need to have not just sets but also some sort of representations for them
20:11:10 <oklopol> or maybe that doesn't really work
20:11:34 -!- oerjan has quit (Remote host closed the connection).
20:11:48 -!- oerjan has joined.
20:13:31 <oerjan> <oklopol> oerjan: is the existence of a computable bijection an equivalence relation on the class of all sets, is what i'm asking? well at least up to some classical cardinality
20:13:51 <oerjan> um i'm not sure the inverse of a computable bijection is necessarily computable
20:14:30 <oklopol> if the sets can be uncountable then...
20:14:33 <oerjan> for integers it is but for things where you cannot decide equality
20:15:01 <oklopol> but you can't even enumerate normal reals, and yet you can make functions from them; should we restrict to countable sets?
20:15:19 <oerjan> however if you include the computability of the inverse then clearly it's an equivalence relation
20:15:50 <oklopol> i guess it would be, clearly, if i knew the definition
20:16:09 <oklopol> i guess it's clear it's with any definition
20:16:55 <oerjan> i suppose you need a concept of computable set as well, otherwise it may not have meaning to have computable functions on them
20:17:28 <oklopol> so maybe we could assume our turing machines compute functions N -> N, what do the equivalence classes look like
20:18:28 <oklopol> but then inversibility is obvious
20:20:04 <oklopol> that is, a computable bijection has a computable inverse; but so what about injectivity and surjectivity, if we have an injection and a surjection, do we have a bijection?
20:20:17 <oklopol> (injection f, surjection g)
20:25:32 <uorygl> He means Banach analytical morphisms, of course.
20:27:07 <oklopol> okay i don't see the disproof but anyway at least cardinality would not be totally ordered then
20:27:24 <oklopol> now i think i am a bit confused.
20:27:38 <oklopol> i figured N would make things easier
20:28:46 <oerjan> oklopol: i suspect the classification would be extremely complicated
20:29:30 <oklopol> but do you see whether injectivity and surjectivity imply bijectivity? i doubt it
20:29:43 <oklopol> i mean it looks like it should be obviously false
20:30:08 <oerjan> since the sets don't need to be recursive - if they are then they are probably in such a bijection with the whole of N itself or {1,...,n} for some n
20:31:09 <oerjan> hm even if there is such a bijection with N, does it have to be recursive?
20:31:15 <AnMaster> anyone know what exactly the "Triaged" status means for ubuntu bugs?
20:31:31 <coppro> AnMaster: AFAICT, it means they've assigned a priority
20:31:43 <coppro> a function that is both an injection and a surjection must by definition be a bijection
20:31:57 <AnMaster> coppro, higher than normal or lower than normal?
20:32:08 <coppro> there's a separate listing for that
20:32:17 <coppro> probably assigned normal priority then
20:32:26 <coppro> it should be in the listing
20:32:27 <oklopol> we assume non-recursive, if you have inj and surj from an RE set to a RE set, then h is a bijection, i think, where given x, h computes surj(x), then for all y < x, it computes h(y) and if h(y)=h(x), it increments x and recurses onto itself
20:32:44 <AnMaster> coppro, well I don't know where the listing is indeed
20:32:58 <coppro> there should be a link in the email
20:33:01 <oklopol> oerjan: does that make sense?
20:33:14 <oerjan> oklopol: i think my brain refuses to think about this any more :D
20:33:15 <oklopol> we can prove by induction that h always terminates, because there's a finite amount of smaller y
20:33:29 <oklopol> x+1 means "such y that after x is enumerated, y is"
20:33:36 <oklopol> y<x means y is enumerated before x
20:34:25 <oklopol> (i mean especially as they are numbers...)
20:34:43 -!- Mathnerd314 has joined.
20:34:48 <AnMaster> coppro, I saw it in https://bugs.launchpad.net/ubuntu/+source/linux/+bug/524281
20:34:59 <AnMaster> coppro, there is that change to triaged
20:35:37 <AnMaster> coppro, but I have no idea about what priority it has
20:35:49 <coppro> AnMaster: right at the top "Importance"
20:35:58 <coppro> the fact that it's both Undecided and Triaged seems wrong
20:36:23 <cpressey> oklopol: OK, it makes sense now. The number down the diagonal is not a computable real, that's all. Doesn't actually tell you anything about the computable reals, except that there is some number that isn't one of them.
20:36:24 <coppro> oh wait, he added some tags there
20:36:38 <coppro> I've never quite understood the Ubuntu bug process
20:37:11 <oklopol> oerjan: maybe i should leave recursion out, assume f and g are the inj & surj, from X to Y, then h(x) first computes the whole list L = {h(y) | y<x}, then enumerates up to x on X side, and starts incrementing x (enumerating stuff after x that is) until it's no longer on the list.
20:37:21 <AnMaster> coppro, and now I'm considering downgrading to 2.6.31 kernel because of this
20:37:35 <AnMaster> coppro, however that means I have to use backported wlan driver
20:37:58 <AnMaster> oh and I would need to patch the kernel for a few other things too
20:38:03 <oklopol> we don't actually need the injection, so according to this infinite RE sets are in computable bijection iff there's a computable surjection
20:38:20 <coppro> AnMasteR: looks like it's an upstream bug
20:38:21 <oklopol> there's always a computable bijection anyway :D
20:38:26 <oklopol> xDDDDDDDDDDDDDDDDDDDDDDDDDDDDD
20:38:28 <oklopol> xDDDDDDDDDDDDDDDDDDDDDDDDDDdd
20:38:44 <AnMaster> coppro, yes but they should get something done about it. And it isn't clear where it is fixed
20:38:48 <AnMaster> Jun 22 18:29:09 tux gnome-session[19821]: WARNING: Unable to determine session: Unable to lookup session information for process '19821'
20:39:04 <AnMaster> coppro, so downgrade to 2.6.31 then?
20:39:22 <coppro> AnMaster: yes; it's a kernel bug and so you're stuck until the next kernel
20:39:27 <AnMaster> this is a pita because I need initramfs (/ on dm-crypt)
20:39:42 <AnMaster> coppro, there is no link to the kernel bugzilla though?
20:39:58 <coppro> I don't know how the kernel keeps track of their bugs
20:40:22 <AnMaster> a lot nicer than launchpad for bug tracking
20:41:13 <oklopol> oerjan: obviously a bijection is given by just checking how manieth element of the X x is, and then enumerate that manieth y in Y
20:41:13 <coppro> getting close to the point where I'm considering jumping ship to a stock Debian
20:41:21 <coppro> or even a more radical change
20:41:40 <AnMaster> coppro, well, I want things to work out of box on my laptop
20:41:47 <AnMaster> also converting would be a pain
20:42:09 <AnMaster> coppro, and if I jumped the ship it would be arch. But yeah that doesn't work very well.
20:42:26 <AnMaster> coppro, ubuntu works mostly out of the box on it. Just that you have to deubuntufy the desktop
20:42:53 <AnMaster> (no face browser and clearlooks theme)
20:43:12 <coppro> Ubuntu's just so horrible because there's no good support avenue. There's no where to complain with "My wireless is broken" because everyone will attempt to pass the buck on to the next guy 'oh, it's obviously a driver bug' 'oh, it's obviously a network-manager bug' 'oh, it's obviously a kernel bug' 'oh, it's obviously a hardware issue'
20:43:38 <coppro> and at each stage they pass it off, they expect you to file a new bug against the new product, rather than just reassining it
20:44:31 <oerjan> oklopol: um no. you _cannot_ necessarily check that if the set is not recursive.
20:45:08 <oklopol> oerjan: yes but given two RE sets there's a bijection between them
20:45:12 <oklopol> this is what i was proving
20:45:21 <oerjan> because you cannot necessarily decide which numbers less than x are in X
20:45:30 <oerjan> oklopol: computable bijection?
20:45:32 <oklopol> and i used the surjection from X to Y even though i can just make one from scratch.
20:45:37 <coppro> <Canonical> We are partnered with Sun. Also, we hate Java users.
20:45:43 <cpressey> But what I want to know is if there is a way to automatically translate primitive recursive functions to general-recursive but more efficient counterparts. I'm sure not all could be done, but I'd be happy with only 50%.
20:46:18 <oklopol> oerjan: y<x means y comes before x when you enumerate the elements of X, just choose some enumerating tm
20:46:24 <oerjan> oklopol: oh hm yes you're right, i thought you meant x's order in X among the natural numbers
20:46:29 <oklopol> i don't mean number inequality, as i mentioned above
20:46:46 <AnMaster> coppro, huh? I hate java, no big deal
20:46:48 <oklopol> and as i then added, it's a bit confusing given they are numbers :)
20:47:11 <oklopol> should've taken {0,1}^* or something maybe
20:47:44 <oklopol> or maybe not, not really very useful
20:47:53 <cpressey> Gee AnMaster, you hate both C++ *and* Java? :)
20:48:40 -!- EXPERIEN has joined.
20:48:48 -!- EXPERIEN has quit (Client Quit).
20:49:40 * oerjan guesses he can throw in C# too :D
20:49:56 <coppro> AnMaster: Java's still the best solution for online applets
20:50:06 <coppro> while I don't write it, I do use it
20:50:17 <coppro> HTML + JS is slow and not as fully-featured
20:50:31 <coppro> it's great for lots of things
20:50:53 <AnMaster> coppro, I offer jsmips as a counter example
20:51:07 <coppro> AnMaster: that doesn't solve performance issues
20:51:17 <coppro> it's a nice go at portability
20:51:42 <coppro> and once again, it lacks some features Java has
20:52:15 <AnMaster> I have neither flash nor java installed
20:52:22 <AnMaster> well I have java but not as browser plugin
20:52:39 <cpressey> The problem with JS is portability. I'm debugging a JS problem with IE right now.
20:52:40 <AnMaster> and to tell the truth, I seldom need either
20:52:57 <cpressey> Of course, proper use of jQuery would probably have avoided it, but still.
20:52:58 <AnMaster> cpressey, IE? "This page is best viewed with Chrome or Firefox"
20:53:31 <AnMaster> cpressey, on a clear moonless night
20:54:08 <cpressey> I've heard Java has portability problems too, but I've never seen them. Esp. for applets.
20:54:30 <AnMaster> cpressey, I would follow the spec to the letter
20:54:38 <AnMaster> if it is then broken in some browser it isn't my fault
20:55:20 <oerjan> and if cpressey then gets fired it isn't his fault
20:55:44 <cpressey> Well then, we need to all start calling it ECMAScript, you know.
20:55:49 <AnMaster> oerjan, psh, do academical work. It is a lot nicer.
20:56:05 <AnMaster> cpressey, jsscript != javascript
20:56:10 <AnMaster> I would go for javascript instead
20:56:27 <coppro> cpressey's correct; ECMAScript is the proper standard
20:56:32 <cpressey> AnMaster: My point was there is no spec for "Javascript".
20:57:05 <coppro> also, ECMAScript is a dumb language, which is another great reason not to use it
20:57:05 <cpressey> As for "jsscript", I'm not sure I've even heard anyone ever use that term...
20:57:49 <AnMaster> ah okay got an extra s there by mistake
20:58:12 <coppro> JavaScript and JScript were differing proprietary implementations of what became standard ECMAScript
20:58:30 <AnMaster> why is the file extension .js then?
20:58:37 <AnMaster> shouldn't it be .emca or something
20:58:41 <uorygl> The language itself is called SpiderMonkey.
21:00:02 <coppro> AnMaster: in modern usage, people don't know better and think javascript is equivalent
21:00:05 <coppro> .es is the proper extension
21:00:24 -!- Phantom_Hoover has joined.
21:00:38 <oklopol> i was rather wtf'd about "whispers innocently"
21:00:54 <uorygl> I guess I should figure out how to write Haskell bindings so that I can write a Haskell binding to (the most useful parts of) SpiderMonkey.
21:01:02 <coppro> formally, JavaScript now refers to the Mozilla implementation and dialect of ECMAScript
21:01:25 <uorygl> It will be called HaSM, despite the atrocity of this name.
21:01:34 <AnMaster> btw what did sun think about it being called javascript
21:01:40 <AnMaster> I can't imagine them being happy at all back then
21:01:46 <oklopol> it's an awesome mental typo
21:01:47 <cpressey> In windows, one of the context menu options for .js files is "Open with Command prompt"
21:01:55 <coppro> AnMaster: It was actually a marketing ploy
21:01:56 <uorygl> Hm. If I move to Finland, I'll probably be able to register hasm.fi.
21:02:05 <uorygl> AnMaster: I think Sun explicitly allowed it.
21:02:11 <uorygl> I have no idea why. :P
21:02:18 <coppro> yeah; it was to increase visibility of the name Java
21:02:31 <AnMaster> coppro, if javascript = mozilla's implementation then cpressey doesn't have to make it work in IE
21:02:38 <uorygl> It would be strange if JavaScript were actually Java script.
21:02:46 <AnMaster> after all, his employer should have said EMCAscript if that was what he meant
21:02:49 <coppro> it would be a better language
21:03:11 <uorygl> If I made a language called Foobar, then FoobarScript would in fact be a scripting version of the same language.
21:03:14 <AnMaster> wait a second, it is <script> tag using emcascript or javascript as the name for the language?
21:03:53 <oklopol> it would be a very logical scripting language
21:04:16 <uorygl> It's like Lojban, but with dynamic typing, and you can change the definitions of words on the fly.
21:04:34 <coppro> AnMaster: it should use ecmascrpit
21:04:36 <oerjan> wait lojban has static typing? >:)
21:04:39 <coppro> in practice, it uses JavaScript
21:05:42 <AnMaster> coppro, does emcascript work there?
21:05:58 <AnMaster> coppro, also I can't imagine that it should use ecmascrpit
21:07:57 <cpressey> ECMAScript is the scripting version of the language ECMA.
21:08:18 <cpressey> ECMA was a popular systems-construction language in the 70's and 80's. I believe it was a descendant of BLISS.
21:08:18 <uorygl> Don't tell me there's actually a language called ECMA.
21:08:23 <AnMaster> I have come to the conclusion that even 16-bit embedded systems are a great deal more elegant than any current desktop OS
21:08:40 <uorygl> Well, of course. They need to be simple, and simple = elegant.
21:08:44 <AnMaster> in fact any embedded system with limited resources is probably more elegant
21:08:53 <cpressey> uorygl: Oh, I'll tell you many, many things, if you let me.
21:09:12 <AnMaster> sure there might be a few nasty goto, but they are still elegant in a way
21:09:25 <AnMaster> since no one writes elegant desktop code
21:09:41 <AnMaster> well except for the lisp machines basically
21:14:16 <cpressey> Yes, the LISP machines write extremely nice desktop code. I admire them for that.
21:19:25 <AnMaster> cpressey, sad they are no longer around
21:19:41 <AnMaster> cpressey, or rather the people who wrote code for the LISP machines would very nice code
21:19:58 <AnMaster> it wasn't the machines themselves that did
21:23:28 <AnMaster> Phantom_Hoover, as opposed to embedded code or HPC code
21:24:31 <AnMaster> Phantom_Hoover, desktop or laptop I guess
21:24:52 <AnMaster> Phantom_Hoover, stuff designed to use on what "people" think of as computers I guess
21:24:59 <AnMaster> Phantom_Hoover, read the whole discussion instead
21:25:02 <AnMaster> cpressey, at times I wonder why I don't do something like: http://www.mastodon.biz/
21:25:10 <AnMaster> cpressey, well slightly newer than 2.0.x kernel
21:25:20 <AnMaster> 2.6.24 maybe? Somewhere around there
21:25:38 <AnMaster> then I realise how much work it would be
21:25:56 <Phantom_Hoover> Hmm, why don't we set up the #esoteric Lisp Machine Company?
21:26:30 <cpressey> AnMaster: Do something like what? Announce vapourware?
21:27:49 <AnMaster> cpressey, no, do a distro like that
21:28:13 <AnMaster> cpressey, anyway there is a download for it around somewhere
21:29:03 <AnMaster> cpressey, didn't work in virtualbox, did in qemu iirc
21:29:09 <cpressey> I'd rather do something like... well, whatever.
21:30:50 <AnMaster> I think it might be better to make an open source lisp OS aimed at common platforms
21:31:23 <Phantom_Hoover> Do we use C as a low-level systems language, or do we make our own Lispoid?
21:31:38 <Phantom_Hoover> AnMaster, Stallman tried that, he ended up with a crappy Chinese laptop.
21:31:52 <AnMaster> Phantom_Hoover, we need some asm anyway, twiddling control registers or such
21:32:12 <AnMaster> Phantom_Hoover, make our own open source hardware I meant obviously
21:32:32 <AnMaster> anyway not going to do it, not worth the effort
21:32:49 <AnMaster> Phantom_Hoover, I feel it should be more like scheme than common lisp
21:33:08 <AnMaster> Phantom_Hoover, no, that was taken by whoever made a computer with wire wrapping
21:33:15 <cpressey> Phantom_Hoover: Aren't we already?
21:34:30 <AnMaster> Phantom_Hoover, BLINKENLIGHTS!
21:35:14 <AnMaster> cpressey, it has a "single step" switch on the front ffs
21:37:37 <AnMaster> Phantom_Hoover, who are you talking to?
21:38:03 <AnMaster> Phantom_Hoover, I'm no expert at VHDL at all. heck I asked ais523 lots of VHDL questions recently
21:38:10 <AnMaster> he is the "expert" on that in here
21:38:14 <Phantom_Hoover> I stumbled into the top-down way of learning computing
21:38:43 <AnMaster> also I never managed to do anything ever with a top down approach. Bottom up for me
21:39:42 <AnMaster> like when implementing a new befunge interpreter, I start with implementing the stack and the funge space.
21:40:26 <AnMaster> Phantom_Hoover, wasn't it you that linked the goldberg computer?
21:40:32 <AnMaster> I think cpressey ought to see it
21:40:38 <cpressey> Clearly, in my dialect, "sick" has a subtle meaning that is lost here.
21:41:01 <AnMaster> cpressey, I'm no native speaker
21:41:13 <AnMaster> cpressey, does "sick" mean anything except the negative meaning?
21:42:15 <AnMaster> on the other hand, it is perfectly okay to say "jätteliten" in Swedish, literal translation would be "gigantically small" but meaning would be "extremely small"
21:42:44 <cpressey> Jacob's ladder for system clock.
21:43:17 <cpressey> The rest just... designs itself.
21:43:37 <AnMaster> Phantom_Hoover, a vacuum tube isn't really a full gate is it?
21:43:42 <AnMaster> isn't it more like a transistor?
21:44:23 <AnMaster> Phantom_Hoover, and a transistor doesn't have a truth table. It as a graph or plot or such showing the voltage as a function of the gate voltage or something like that
21:44:40 <AnMaster> I did have this on an exam some time ago XD
21:45:19 <cpressey> Well, a transistor, properly applied, is a NOT gate. Truth table follows.
21:45:32 <AnMaster> cpressey, which sort of transistor?
21:45:39 <AnMaster> cpressey, I'm thinking nMOS and pMOS here
21:45:54 <AnMaster> and as far as I remember you had two transistors for a NOT gate
21:45:55 <cpressey> AnMaster: Any sort of transistor.
21:46:10 <AnMaster> of course CMOS would have two transistors
21:53:35 -!- MizardX has quit (Ping timeout: 276 seconds).
22:09:16 <cpressey> I think the very concept of "Operating System" needs to be shelved. I prefer "Computing Environment".
22:09:32 <cpressey> So, I'm not going to build my own OS. I'm going to build my own CE.
22:10:13 <oerjan> this environmentalism has gone to far
22:10:47 <cpressey> It will be like a Lisp machine WITH ROCKET WINGS.
22:12:07 <uorygl> Yay, a place where people speak Finnish.
22:12:22 <oerjan> it's called finland. hth
22:12:59 <uorygl> It's difficult to /connect to Finland, though.
22:13:33 <uorygl> Or is that /connect European Union, /join Finland? I'm not really sure.
22:13:51 <oerjan> which is weird given that irc was invented there
22:19:00 -!- GreaseMonkey has joined.
22:22:26 <uorygl> What's "opetella" in the first-person singular present?
22:22:28 <cpressey> Phantom_Hoover: Yes, but for an entirely different purpose. I like to keep my opponents on their toes.
22:24:50 <cpressey> This includes most of Finland!
22:25:20 <cpressey> Because their language is much cooler than my own, and I am jealous.
22:28:17 <cpressey> Well, http://www.reocities.com/Area51/Vault/1790/qfin.html for one.
22:40:35 -!- ehirdiphone has joined.
22:41:10 -!- Phantom_Hoover has quit (Quit: Leaving).
22:41:58 <uorygl> I was listening to this podcast once where some French guy explained that our perception of a number depends on the lagaghism of a number.
22:42:11 <uorygl> "What?" said the interviewer.
22:42:32 <uorygl> Eventually, it was ascertained that he meant "logarithm".
22:43:33 <ehirdiphone> Also, I doubt our brains can efficiently calculate logarithms...
22:44:13 <cpressey> Yet we say "an order of magnitude difference" with ease.
22:44:47 <cpressey> Or "whatever fits my observations best at this moment", maybe.
22:47:28 <GreaseMonkey> i'm going to try doing stuff with the 1541 floppy drive using x64
22:48:29 <uorygl> ehirdiphone: well, he meant that if we see a certain number of things, the strength of the stimulus is proportional to the logarithm of the number of things seen.
22:48:41 <ehirdiphone> Well, okay, but that's just counting base 10 digits cpressey
22:49:12 <uorygl> Thus, if you go to the Pirahã, who don't use numbers, and show them a pile of one and a pile of nine, and ask them to construct a pile whose number lies exactly in between these two numbers, they'll generally make a pile of three.
22:49:29 <uorygl> What's right between : and :::::::::? ::: is, of course!
23:01:52 <ehirdiphone> One day the Pirahã are going to learn numbers and researchers are going to be SO fucked.
23:02:57 <Sgeo> "That feeling, that "I built that" or "I grew that" or "I fed that guy" or "I made these pants" feeling, can't be matched by anything the internet has to offer.
23:02:57 <Sgeo> Read more: http://www.cracked.com/article_15231_7-reasons-21st-century-making-you-miserable.html#ixzz0rccE1CeE"
23:03:16 <Sgeo> Wrong. That's why I program. That's the feeling I get when I make something that other people use.
23:03:25 <Sgeo> hi ehirdiphone
23:07:10 <GreaseMonkey> "So did we really need a study to tell us that more than 40 percent of what you say in an e-mail is misunderstood? Well, they did one anyway."
23:07:46 <ehirdiphone> I hypothesise that Sgeo has a brain powered entirely by pointless ego.
23:07:58 <Sgeo> Ducks? Thanks.
23:08:05 <ehirdiphone> GreaseMonkey: Were we running out of numbers?
23:08:28 <Sgeo> ehirdiphone, it was a reference
23:08:35 <GreaseMonkey> ehirdiphone: usually doesn't hurt to have another number
23:09:09 <Sgeo> Also, is there anything wrong with getting paid in recognition instead of money? You said that I shouldn't do this project unless I'm getting paid
23:09:26 <ehirdiphone> Thanks, antidisestablishmentarianism. Thantidisestablishmentarianism.
23:10:00 <Sgeo> Bless you, antidisestablishmentarianism. Blantidisestablishmentarianism
23:10:16 <ehirdiphone> Sgeo: Caring about recognition and the like is pointless and you will probably grow out of it. Unless the revolution comes you will always need money.
23:11:36 <cpressey> After the revolution, you will be paid directly in Recognition.
23:11:57 <cpressey> ehirdiphone: I'm stalled on my latest.
23:12:31 <ehirdiphone> (Whuffie from Down and Out in the Magic Kingdom, a book which I have not even read.)
23:13:18 <Sgeo> How can one ever outgrow liking the feeling of "I made this. People like it."?
23:14:22 <cpressey> ehirdiphone: Well I had one set of semantics, and implemented them, and tried to complete them, and realized that the theme of the thing suggested I have completely different semantics.
23:14:26 <ehirdiphone> Sgeo: By learning to only value the opinions of people with valuable opinions, and also by becoming more independent so that your value system is not compromised by other people's.
23:14:53 <cpressey> When I get back to it, though, I should have something Turing-equivalent, and nifty. Just bored of it right now.
23:15:19 <ehirdiphone> cpressey: How many times have you inadvertently invented Haskell?
23:15:45 <cpressey> ehirdiphone: Er, well, I don't usually do lazy evaluation. So very few, I think.
23:16:30 <cpressey> OK, *that* I have reinvented a few times.
23:17:25 <cpressey> I really want a language where I can just type in 10 HGR: HPLOT 0,0-100,100 to get a freaking line drawn.
23:17:40 <cpressey> These days it's all messing with GUIs and crap.
23:18:16 <cpressey> Eh, yes. For a particular thing I'm thinking of, a bitmap would be nicer though.
23:18:36 <cpressey> maybe UCB logo supports that now, I should check.
23:19:07 <ehirdiphone> It can export bitmaps afaik. At least, the MSWLogo port.
23:19:37 <cpressey> Can I say "rect xor 0 0 100 100" or the like?
23:20:21 <ehirdiphone> cpressey: Not the XOR part. And absolute addressing is discouraged in Logo, but you can do it.
23:20:53 -!- tombom has quit (Quit: Leaving).
23:20:59 <ehirdiphone> logo is nice because it's quite literally lisp without the patens plus a turtle
23:21:20 <cpressey> Kinda sorta need the XOR part. Oh well, BASIC didna support that eir.
23:22:26 * oerjan vaguely thinks oric basic had XOR. or maybe it was some pascal thing...
23:23:07 <ehirdiphone> for(x=startx; x<maxx; x++) for (y=starty; y<maxy; y++) gfx[x][y] = ~gfx[x][y];
23:23:33 <oerjan> well, not in the context of rectangles, presumably
23:24:06 <ehirdiphone> cpressey: Outputting TGA is a 15-line job, btw.
23:24:27 <cpressey> Apple BASIC had an XOR for drawing shapes.
23:24:39 <cpressey> But not for plotting poitns in the bitmap, iirc.
23:25:41 <cpressey> Well, the "shapes" were basically little logo or postscript programs, again iirc - arrays of draw-line and rotate commands. They could be XOR'ed onto the bitmap.
23:25:56 * cpressey looks for an Apple II emulator.
23:26:03 <cpressey> That's probably what I really want anyway.
23:27:45 <ehirdiphone> cpressey: My favourite doodling tool: Haskell. You represent an infinitely-detailed image as [0, 1] -> RGB.
23:28:29 <ehirdiphone> Checkerboard is "times N, round, black or white for even odd"
23:29:02 <ehirdiphone> i.e. (Double, Double) with range assumption
23:29:18 <ehirdiphone> Add Double argument for time: voilà, animation.
23:29:36 <cpressey> I did something similar with Prolog once. Er, well, actually Perl.
23:29:36 <ehirdiphone> Anyway, writing advanced helper functions is easy that way.
23:30:05 -!- MigoMipo has quit (Read error: Connection reset by peer).
23:30:11 <cpressey> Easy to write a Mandelbrot generator in that, I imagine.
23:30:17 <cpressey> I should probably try that someday.
23:31:38 <ehirdiphone> Lovely imaging trick: (x,y) is in the Sierpinski gasket iff x XOR y is zero.
23:33:24 <ehirdiphone> cpressey: It upsets me that nobody has perfectly rendered the Mandelbrot set.
23:33:51 <ehirdiphone> Since, afaik, membership is undecidable (requires an infinite computation to determine).
23:35:25 <ehirdiphone> cpressey: You and dbc should have an epic duel.
23:36:21 <cpressey> dbc and I are categorical duels.
23:36:27 -!- ehirdiphone has quit (Quit: Get Colloquy for iPhone! http://mobile.colloquy.info).
23:38:06 <dbc> Only one what?
23:44:09 -!- uorygl has quit (Ping timeout: 240 seconds).
23:46:04 -!- Warrigal has joined.
23:50:04 -!- sebbu2 has joined.
23:50:13 -!- Warrigal has changed nick to uorygl.
23:50:28 -!- sebbu has quit (Read error: Connection reset by peer).
23:50:36 -!- sebbu2 has changed nick to sebbu.
23:54:48 -!- relet has quit (Quit: Leaving.).