...

00:00:06 screw this 00:00:20 no application does aac tagging compatibly! 00:00:21 why 00:00:32 I'm just going to transcode it to .ogg 00:05:15 -!- kallisti has quit (Ping timeout: 265 seconds). 00:31:48 i keep the master copy of my music collection in a simple "artist/album/nn - track.ext" directory structure 00:31:55 because multi-container tagging is a pain 00:32:25 -!- kallisti has joined. 00:32:25 -!- kallisti has quit (Client Quit). 00:32:30 i have a script to transcode a subset of that to ogg vorbis and tag it, for my portable player 00:34:11 -!- pikhq_ has joined. 00:34:28 -!- pikhq has quit (Ping timeout: 244 seconds). 00:53:32 -!- invariable has quit (Remote host closed the connection). 00:53:36 Do you like Ogg containers? 00:56:38 who doesn't? 01:02:52 I do not like greens oggs and hogs 01:02:58 *green 01:03:19 Do you like the Decompose type I have made up in Haskell? 01:06:58 -!- variable has joined. 01:08:15 I do not like them, o-er-jan 01:09:13 -!- pikhq has joined. 01:09:22 -!- pikhq_ has quit (Ping timeout: 252 seconds). 01:15:17 Do you like CodensityAsk (and CodensityAskT)? I have figured out many things about it. 01:20:21 (CodensityAsk (Store s)) is also MonadPlus. 01:21:17 hm on-board sucks so badly. I just compared it to my phone, my phone has way better bass than my desktop's onboard sound (using a pair of studio headphones in both cases). My SB Live beats both of course. 01:22:14 After edwardk told about F-algebras to make free monad and so on, I realized that making the free comonad from Maybe and from the Copeanoid class it is also a F-coalgebra for Maybe so it is the same thing. 01:22:37 Even though I can already see that it is a non-empty list comonad in both cases, now I can see the similarity, too. 01:28:52 -!- HackEgo has quit (Ping timeout: 265 seconds). 01:29:24 -!- HackEgo has joined. 01:29:36 zzo38, wow that went completely over my head 01:29:53 should learn more category theory I guess 01:34:29 TWO _HUNDRED_ THIRTEEN _THOUSAND_ ONE _HUNDRED_ FORTY_ FOUR 01:34:46 Today's puzzle: come up with #defines so that that expression works as expected. 01:35:00 The first few seem easy enough: 01:35:38 #define ONE 1; #define TWO 2; #define FOUR 4; #define THIRTEEN 13; #define FORTY_ 40 +; #define _HUNDRED_ * 100 + 01:35:53 The hard part is _THOUSAND_. 01:36:35 I guess the best I can think of is #define _THOUSAND_ ) * 1000 + ( 01:36:43 Then (TWO _HUNDRED_ THIRTEEN _THOUSAND_ ONE _HUNDRED_ FORTY_ FOUR) will work correctly. 01:37:00 I think. 01:37:12 #define TWO 213144; #define _HUNDRED_ + 0; #define THIRTEEN + 0; ... 01:37:29 kmc: I said "as expected". That's not how I would expect it to work. 01:37:35 oh 01:37:40 i wouldn't expect it to work at all 01:38:00 if one may assume that an expression like TWO _HUNDRED_ alone should also work, then there are more problems... 01:38:25 Nah, there'd be a separate symbol, _HUNDRED, for that. 01:38:31 oh 01:38:59 -!- DHeadshot has quit (Read error: Connection reset by peer). 01:39:03 <$ <$> <* <*> *> 01:39:06 -!- DH____ has joined. 01:39:15 well TWO and FOUR need to work alone, which means they cannot contain any unbalanced bracketing 01:40:27 Then FORTY_ must work next to FOUR, so it can't have unbalanced bracketing, either. 01:40:35 -!- DH____ has quit (Read error: Connection reset by peer). 01:40:42 Then you can conclude the same for _HUNDRED_, then for _THOUSAND_, right? 01:40:49 So you need some real trickery. 01:40:54 which tells me we need to use precedence to get the grouping correct, not brackets 01:42:20 it might be worth looking into more operators than + and * 01:42:20 is C++ allowed? ;) 01:42:36 i still don't understand the problem statement 01:42:39 The C preprocessor is supposed to be Turing-complete, isn't it? 01:42:45 No. 01:42:50 It's explicitly designed not to be. 01:42:54 what does it mean for it to work "as expected" 01:42:54 Oh. 01:43:00 cpp is turing complete if you allow a file to re-include itself 01:43:04 more or less? 01:43:04 No, it's not. 01:43:10 kmc: make it so that the written-out form of any number evaluates to the correct number. 01:43:11 We had this debate a while ago, it's still not. 01:43:12 tswett: C++ might make a lot of difference to C, since you can overload operators 01:43:13 though it is, with a certain nonstandard feature 01:43:21 i assume implementations have limits on sizes of things 01:43:23 cpp is designed to guarantee termination, no? 01:43:26 comex: which feature? 01:43:32 pragma push_macro 01:43:34 Well, sure, it could be with nonstandard features, but even with no limit on inclusion depth, it's not TC. 01:43:37 you can use it to do recursive macros 01:43:56 why is it not turing complete if you allow re-inclusion 01:43:57 You can't simulate state changes and without that sort of special macro you can't do recursion. 01:44:15 -!- pikhq_ has joined. 01:44:23 Or rather, sorry, you can't simulate unbounded memory. 01:44:28 -!- pikhq has quit (Ping timeout: 248 seconds). 01:44:28 You can only do a finite number of states. 01:44:34 tswett: i think we really need to know is this is supposed to work with just C... 01:44:46 Because you can't really define macros in terms of previously defined macros. 01:44:48 anyway, that wouldn't really help here because there's no enclosing macro 01:44:53 It seems easy to do it with C++. Just find some operators with the required precedence (you only need four) and overload them. 01:45:13 you actually only need one 01:45:28 So let's say it must be done in plain C. 01:45:55 If you were allowed to require that they be prefixed by some NUMBER_START macro it'd be easy >_> 01:46:05 Or, equivalently, suffixed. 01:46:10 (Or both of course) 01:46:13 #define NUMBER_START ((; #define NUMBER_END )) 01:46:18 And then my scheme works. 01:46:25 Gregor: what's your scheme? 01:47:03 tswett: struct numberthing { int value; struct numberthing (*add)(int v); struct numberthing (*mul)(int v); }; 01:47:19 You just need to be able to get the value back out at the end. 01:49:54 -!- DHeadshot has joined. 01:50:53 hm... 01:51:16 struct numberwang 01:51:20 -!- Dovregubben has quit (Ping timeout: 246 seconds). 01:51:52 Vorpal: OK then learn 01:52:07 You could look at Wikipedia article about F-algebra 01:52:08 -!- Dovregubben has joined. 01:53:45 hm what about the 3[a] trick, can that help? 01:54:45 * oerjan is browsing http://en.wikipedia.org/wiki/Operators_in_C_and_C%2B%2B#Operator_precedence 01:54:52 ... you brilliant, brilliant psychopath. 01:55:17 All you need is an array of size 1000, mapping x to 1000*x... oh wait, that operator's precedence is no more convenient than * though. 01:55:18 WHY THANK YOU 01:55:40 What are you trying 01:56:05 Gregor: except you can _combine_ it with a real + for the thousands? 01:56:17 TWO _HUNDRED_ THIRTEEN _THOUSAND_ ONE _HUNDRED_ FORTY_ FOUR 01:56:17 Today's puzzle: come up with #defines so that that expression works as expected. 01:56:28 3[a] trick? 01:56:42 hm the problem is you still need something else to join the part after it... 01:56:43 oerjan: I think it's salvageable, I just haven't figured out how yet. 01:56:50 comex: In C, a[3] and 3[a] are equivalent. 01:56:56 oh 01:57:45 is there a fake C facts blog 01:57:46 oerjan: Maybe EVERY op uses [], except that THOUSAND, MILLION etc use [...]+0 01:58:21 Oh, except then you couldn't get both TWENTY_ TWO and TWO to work, at least not without _TWO being a thing. 01:58:37 bitwise OR would have been nice if you tried doing this for binary :P 01:58:43 (|) 01:58:44 oerjan: Yeah, I thought of that. 01:58:44 "arrays in C actually contain one extra element at the end" is a commonly believed fake C fact 01:58:48 what if 01:58:50 but not that funny 01:59:05 #define TWO (two) 01:59:07 and ditto all the other ones 01:59:13 where two is a function that returns a function pointer, etc. 01:59:13 tswett: Is there a reason you've chosen to underscore-attach things exactly like that? Can't have TWO_ _HUNDRED_ _THIRTEEN_ _THOUSAND_ _ONE_ _HUNDRED_ _FORTY_ _FOUR? 01:59:42 comex: That was the gist of my struct solution (although yours is more elegant), but the problem is you need a suffix. 01:59:43 kmc: there is a technically true fact slightly similar - it is not undefined behavior to increment a pointer to that address 01:59:53 Otherwise you end up with a function pointer instead of data :( 01:59:55 but dereferencing it still is 01:59:57 yeah i think it stems from that 02:00:01 Gregor: pointers automatically cast to integers :) 02:00:05 lol 02:00:16 So the (two) function will just CONVENIENTLY be at location 0x2? 02:00:31 probably an error introduced when the C specification was translated from the original Aramaic 02:00:35 nah, but it would detect when you've come to the end and give you (void *) result 02:00:49 “detect” 02:00:49 How? 02:01:45 because you basically have an interpreter 02:02:08 Except that when you're finished is /exactly/ when you don't have an interpreter. 02:02:20 kmc: Is that fact believed because you can say int x[5]; and then compare &x[0] to &x[5]? 02:02:28 oh, I guess it doesn't know what you end with 02:02:31 Oh, yes, oerjan said that. 02:02:36 (Or something similar.) 02:03:17 never mind :) 02:03:38 (I mean, you could still do it by conveniently locating functions in such a way, or trapping SIGSEGV or whatever) 02:06:28 #define TWO +0<0?0:two() 02:07:14 use global state 02:07:31 comex: i was just pondering combining , and = :P 02:07:47 oh durr, , would be easier 02:08:13 +0,two() 02:09:59 http://www.uesp.net/dagger/dagbug.shtml 02:10:09 Wow, and they say modern Bethesda games are buggy. 02:11:52 Fun fact: In JavaScript, (0,eval)(x) is very different from eval(x) 02:12:32 shachafun 02:12:39 and then there's something in cpp about joining identifiers, which i only know exists 02:12:47 shachaf: nice 02:12:53 or tokens, or whatever 02:13:00 shachaf, wait, how does that even work? 02:13:18 @google javascript global eval 02:13:23 http://weblogs.java.net/blog/driscoll/archive/2009/09/08/eval-javascript-global-context 02:13:23 Title: Eval JavaScript in a global context | Java.net 02:13:26 Oh, wait, (x,y) is a sequencing operation in C-likes. 02:13:34 Indirect eval is global eval. 02:13:42 It's so that scope-modifying evals are statically knowable. 02:13:59 fun fact 0 = 1 02:14:10 <3 02:15:53 Yes, both < 3 02:16:00 oh hm 02:16:28 in c, are global variables and record fields in the same namespace? 02:16:43 otherwise there might be a way to use -> or . ... 02:17:21 * shachaf wonders what people are trying to do here. 02:18:08 * oerjan is just vaguely pondering the TWO _HUNDRED_ THIRTEEN _THOUSAND_ ONE _HUNDRED_ FORTY_ FOUR problem 02:19:06 without actual global state, it seems hard to use anything lower precedence than +, but if something high could be used, _THOUSAND_ could be * 1000 + if there was just 02:19:20 *-if there was just 02:22:52 oh hm FORTY_ TWO _HUNDRED is not necessary to support, i assume. which means _HUNDRED _could_ use the 3[a] trick for the left part 02:23:24 Yes. 02:23:34 The problem with the [a] trick is making both FORTY_ TWO and TWO work. 02:23:50 (And FORTY_ TWO _THOUSAND, of course) 02:24:25 That's why I asked tswett whether a solution with more connectors would work, e.g. FORTY_ _TWO_ _THOUSAND. 02:24:33 Then the [a] trick would give us everything we need. 02:25:02 (It also has the benefit that it'd be much easier to remember how to use it) 02:25:31 a -> or . trick could make FORTY_ TWO and TWO work assuming record fields are in a different namespace 02:25:38 What does C do when you give it a bunch of whitespace-separated digits, out of curiosity? 02:25:53 Phantom_Hoover: Cry a lot. 02:26:14 oerjan: That was my suffixing problem. 02:26:24 oerjan: You can make it work like that, but need a suffix to get it back to a number. 02:26:55 Gregor: or some initialized structures and variables, surely? 02:26:58 > 1 2 base 8 which is 10 base 10 and you take away 3 that's 7 02:27:01 1 02:27:22 wowchaf 02:27:32 oerjan: I don't see how having initialized structs and vars helps you avoid the suffix issue... 02:27:36 -!- Phosphenes has joined. 02:27:48 -!- Phosphenes has left ("PHOSPHENES HAS LEFT THE CHANNEL"). 02:27:50 wheegan? 02:27:59 Whoops, where were we. 02:28:18 dude 02:28:23 Gregor: if you can have two be both a global variable set to 2 and a field name in your struct... 02:28:23 every word in that sentence starts with 'w' 02:28:32 or TWO, even 02:28:32 Gregor: well, yeah, the idea is that each underscore represents the presence of an operator. 02:29:25 oerjan: Ohh, I see your game… but you'd still end up with a . problem with the underscores as presented. 02:29:36 tswett: If I'm allowed to throw underscores everywhere, it becomes quite achievable. 02:30:06 With my scheme, (FORTY_ _THOUSAND) would process to (40 + ) * 1000. 02:30:31 Uh, except no, because _THOUSAND cannot suppress the trailing parenthesis. 02:30:44 Okay, (40 + ) * 1000 + (0). 02:30:57 Since FORTY_ is 40 +, whereas FORTY is 40. 02:31:47 #define _HUNDRED_ [hArray] -> 02:32:25 or ., whatever 02:32:44 * oerjan has never actually used -> he thinks 02:33:24 I'mma build the [a] trick version. 02:33:54 #define FORTY_ FORTY -> 02:34:11 -!- Phantom_Hoover has quit (Read error: Connection reset by peer). 02:34:22 oops 02:34:35 that won't work 02:36:06 int FORTY = 40 02:36:33 -!- augur_ has joined. 02:37:08 NEGATIVE_ B _PLUS_OR_MINUS_ THE_SQUARE_ROOT_OF_ THE_QUANTITY_ B _SQUARED _MINUS_ FOUR _TIMES_ A _TIMES_ C _ALL_OVER_ TWO _TIMES_ A 02:37:13 #define FORTY_ array_or_field_40 02:37:25 um wait no that was the problem 02:37:34 -!- augur has quit (Ping timeout: 246 seconds). 02:37:43 getting FORTY_ to work both with and without HUNDRED_ preceding 02:38:24 um wait no that's fine 02:39:46 dammit searching for C namespaces gives all those hits about how to do user-definable ones 02:41:42 tswett++ 02:42:00 pretty sure you can have a field and a variable of the same name 02:42:03 linux does this all over 02:42:20 Certainly you can have two fields of the same name. 02:42:41 oh right 02:43:46 http://sprunge.us/KbiF Tada 02:44:17 The format isn't exactly as tswett described, but is in fact much, much simpler. 02:45:38 oh that's what you wanted the +0 for 02:47:48 Yeah, it would be unnecessary if I just required that the first word after THOUSAND or MILLION or whatnot not be prefixed. 02:47:49 But bleh. 02:48:23 * soundnfury wonders if you can pull some trick with ## 02:48:43 Um hello it's already done I win guys. 02:49:40 um, it's ugly? 02:49:54 and requires preinitialised tables? 02:49:58 :( 02:51:14 "It is also possible to concatenate two numbers (or a number and a name, such as 1.5 and e3) into a number." 02:51:22 so maybe... 02:51:26 tasty tasty token pasty 02:51:56 "However, it is an error if `##' appears at either end of a macro body. " 02:52:00 sadly. 02:52:03 The problem is you only have token pasting within macros X_X 02:52:08 They need to expand to something consistent. 02:52:22 so we'd have TWO##THIRTEEN##ONE##FORTY##FOUR, and then HUNDRED etc. would be noise keywords? 02:52:38 soundnfury: You couldn't even use that, because you can only paste tokens WITHIN macros. 02:52:41 or #define HUNDRED 0 // for eg TWO##HUNDRED##FIVE == 205 02:52:53 um, ok that sucks 02:53:20 Also, what the hell is TWO THIRTEEN ONE FORTY FOUR 02:53:29 213144. 02:53:33 Oh 02:53:35 ... obviously 02:53:40 Well for that you're just fucked in any system I think 8-D 02:53:54 Or possibly 2131404. But probably not. 02:54:07 Although if 213144 were a string of digits, I'd be likely to read it as "twenty-one thirty-one forty-four". 02:54:21 oh yeah, that would want to be ONE##FOUR##FOUR 02:54:25 (if we could) 02:54:30 (which we can't as Gregor notes) 02:54:53 MY SYSTEM IS BEST SYSTEM AND YOU'RE ALL JUST JEALOUS 02:54:55 anyway, why are we trying to do this? Is someone trying to make C feel more COBOLish? 02:55:00 lol 02:55:04 soundnfury: ask tswett 02:55:23 To be esoteric. 02:55:25 Ooooh, with my system you could eschew underscores for capitalization! FOUR becomes Four, _FOUR becomes four! 02:55:26 * soundnfury thinks Welsh television should have a show called Cobol y pwm 02:55:51 (it's funny because they have one called Pobol y cwm, whatever that means) 02:55:52 Gregor: fancy 02:57:06 "People in the valley" 02:57:16 it's about the wonderful art of padlock folding or, "padlockigami" 02:57:28 kmc: you are david mitchell AICMFP 02:57:45 :3 02:57:57 google seems to think "Cobol" still means people when switching the letters 02:58:02 sadly nothing for pwm 02:58:10 *+ translate 02:59:27 oh hm i recall p initially in celtic languages only exists in loanwords, so they may be rare... 02:59:32 see? they is his cousins! 03:00:05 or more precisely, that they dropped initial p at some point 03:01:08 also, the only native word in norwegian starting with p is the preposition på (which is a contraction of norse upp á) 03:01:12 -!- david_werecat has quit (Ping timeout: 244 seconds). 03:01:51 which i guess is cognate to upon 03:02:13 kmc: Is the mosh object synchronization thing meant to be used on its own yet? 03:02:22 http://pastie.org/4258682 - THIS TOTALLY WORKS, I IMAGINE 03:02:49 shachaf: not exactly, but it is reasonably well abstracted in the mosh source code 03:02:54 talk to KeithW about it 03:03:08 That's the eventual plan, though, right? 03:03:20 "plan" is a strong word 03:03:21 tswett: a most sublime imagination 03:03:35 hmm I'll actually be in Wales next week 03:03:42 is there something i should do there besides "take a train to London"? 03:03:56 -!- pikhq_ has quit (Ping timeout: 248 seconds). 03:04:01 -!- pikhq has joined. 03:04:32 kmc: obviously you need to visit llanfairpwllgwyngyllgogerychwyrndrobwllllantysiliogogogoch 03:04:50 obviously, i was planning to pick up some padlocks there 03:04:54 from a store or smith 03:05:09 obviously. 03:05:12 (wat?) 03:05:16 * kmc is now known as jonah 03:05:30 wat = http://www.youtube.com/watch?v=N-Fcy70fhds 03:06:06 * oerjan wonders why youtube things i'm using IE 7 03:06:10 *thinks 03:06:18 Are you using IE 7? 03:06:22 no, 8 03:06:41 and youtube is warning me to change since they'll stop supporting 7. 03:06:57 Are you using IE 7 compatibility mode? 03:07:15 "By Abergavenny, Merther Tydfil, and Llanfairpwllgwyngyllgogerychwyrndrobwllllantysiliogogogoch!" said Mr. Bargain-Price. "This is too much!" 03:07:23 - who knows what that's from? 03:07:44 oerjan: wow i thought you made that name up 03:09:47 kmc: you couldn't /make/ something like that up 03:09:53 -!- augur_ has quit (Ping timeout: 265 seconds). 03:09:54 well yes you could 03:09:59 but anyway it does seem like i will go through there 03:10:03 as it's on the rail line from holyhead 03:10:31 soundnfury: "The village's long name cannot be considered an authentic Welsh-language toponym. It was artificially contrived in the 1860s to bestow upon the station the honour of having the longest name of any railway station in Britain" 03:11:10 shachaf: hm no, although google+ likes to complain that i am :P 03:11:51 i suppose it might be the same cause... 03:11:57 -!- augur has joined. 03:12:11 kmc: sure, but it wasn't "made up" per se, it's authentic welsh, even if not an authentic welsh name 03:12:19 as in, it really does mean something 03:12:52 that's true 03:17:08 Ahh, Johnston Railway Sans... now there's a real corker of a font 03:18:13 what about dyddiadfymathoobellamaenarywbethygwaithanoddinierdigon 03:19:26 which i pulled out of a markov chain just now 03:27:18 In Soviet Russia, YouTube watches YOU!! 03:28:19 that is closer to an actual soviet russia joke than most 03:29:28 -!- edwardk has joined. 03:33:44 in soviet russia, worn-out meme repeats YOU!! 03:33:58 ;) 03:34:35 yeah 03:35:12 the meme'd ones lack the part where the reversal has a comprehensible but different meaning which comments on life in the soviet union 03:35:16 i.e. the humor 03:36:02 indeed. Although it's called humour 03:36:14 I RESPECTFULLY DISAGREE 03:36:24 I'M OFFENSIVE AND I FIND THIS WELSH 03:36:38 stop choosing to be offended, you mexican jew lizard 03:36:51 * soundnfury raises an eyebrow 03:37:09 ("Yes, actually. My mother was from Mexico and my dad was a Hebrew iguana, so...") 03:38:29 ("No, to be a jew your *mother* would have to be Hebrew, though it doesn't matter which one was the iguana.") 03:38:43 ("What are you quoting now?") 03:38:58 and that depends who you ask, but yeah basically 03:39:10 ("I'm not, I'm just irrespectfully disagruntling.") 03:41:52 -!- edwardk has quit (Quit: Computer has gone to sleep.). 03:42:20 so, what's your favourite font? 03:42:49 http://dotsies.org 03:44:57 soundnfury: Computer Modern because it has so many parameters which can be varied. But even some things about Computer Modern are not perfectly; but the source-codes is available so you can make improvements (as long as you do not call the new one "Computer Modern"), such as a parameter for "R" with straight lines, and so on. 03:45:07 * oerjan is _guessing_ that the warning he got against IE not working well on the site is why he doesn't see anything special there 03:45:40 oerjan: on dotsies? 03:45:44 yes 03:45:53 _or_ that could be the joke, i guess 03:46:13 i don't think this is a joke :/ 03:46:34 well it's a joke to me but i don't know if the author is in on it 03:46:42 ok. because i only see a boring normal font. 03:47:47 dotsies is here. 03:48:00 oerjan, it's a bunch of dots 03:48:05 The font is dots 03:48:07 it's an alterative way of writing the latin alphabet where every letter is a 5-bit column of black or white dots 03:48:18 because human vision is just a grid of pixels right 03:48:19 Sgeo: well so is a boring normal font 03:48:31 it's not like the brain has dedicated hardware for recognizing lines, curves, corners, etc. 03:48:52 ;P 03:49:02 like, this person clearly knows nothing about vision 03:49:06 not even the pathetic amount i know 03:49:12 which is "i get stoned with neuroscientists occasionally" 03:49:23 but you know, being a rails developer makes you an expert at every topic in the world 03:49:47 kmc, I... guess people could treat each word the way Chinese treat chinese words? 03:49:53 Or glyphs or whatever 03:49:56 that's the idea 03:50:03 but encoding it as pixels is still stupid 03:50:10 han characters have lines, curves, corners, etc. 03:50:29 so do joined arabic letters, hangul clusters, etc. 03:52:04 Note: Caps and lowercase differ by a dot above the letter 03:52:13 i would think that an attempt to revolutionize the latin alphabet should start with a thorough survey of other writing systems, preferably with psychophysical measurements from fluent users of each 03:52:30 and also a thorough understanding of human visual processing 03:53:50 but you know, all that research is invalid because it was performed by non-programmers 03:55:04 * kmc rage 03:59:13 -!- oerjan has quit (Quit: Good night). 03:59:32 .:| ::. ..: |.|... 03:59:46 (not actually meaning anything) 04:02:33 http://www.ccelian.com/concepca.html this is way cooler and also probably better 04:03:19 oh yeah and the dotsies guy didn't even put any thought into assigning better dot patterns to common letters 04:03:40 and has allegedly been working on this for 10 years 04:04:35 and braille already exists 04:05:01 and has real users who could tell you what's good and bad about it 04:05:04 kmc: well it's not at all like saying A = 65, 65 - 64 = 00001, 90 - 64 = 11010, then drawing the binary numbers as vertical stacks 04:05:19 it is a lot like that 04:06:16 it's not the binary counting sequence but it's another kind of regular binary sequence applied to the latin letters in order 04:08:04 * kmc rage 04:08:37 it got a fawning article in techcrunch too 04:08:55 Gah, dotsies looks like a *stupid* orthographic scheme. 04:09:08 but it was invented by a rails developer from SF!!! 04:10:54 "Keep up the crazy innovation, Craig! It’s guys like you who aren’t afraid to push the status quo that make the world better for the rest of us." 04:11:53 He managed to redesign the *least bad* part of English orthography, and didn't even make it that easy to figure out. 04:13:27 -!- ogrom has joined. 04:13:57 his design goals are absurd 04:14:15 which ones? 04:14:53 i would have thought you'd sympathize with this kind of crackpottery itidus21 04:16:29 it's tempting 04:17:04 but i have decided to try not to 04:17:55 "Dotsies is optimized for reading." this isn't clear 04:18:12 that claim would be more credible if the author displayed any understanding of how reading works 04:18:39 it is not a ridiculous idea to try to improve the latin alphabet and make it more optimized for reading 04:18:43 he's not saying it's faster 04:18:50 he's just saying it's optimized 04:18:53 Reading relies heavily on glyphs being distinctive. 04:18:56 but you would want to start by knowing something about the field 04:19:04 and surveying what's already out there 04:19:23 Aheegan. 04:19:32 this displays the classic crackpot attitude of "i'm so much smarter than everyone else that I don't need to understand what anyone else has done" 04:19:50 and he's concerned with how much space is taken up by text 04:19:51 plus the classic programmer attitude of assuming you know how something works in people because you know how it works in computers 04:20:09 itidus21: that makes sense, the time to read something is proportional to how quickly you can move the fovea across it 04:20:18 he does seem to know about the fovea at least 04:20:24 whachaf? 04:20:49 -!- pikhq_ has joined. 04:23:09 one thing i like about this alphabet is that the long bits on letters like ldbh is that you can see them even in blurry text 04:23:48 -!- pikhq has quit (Ping timeout: 248 seconds). 04:24:06 kmc: The Elian script there does look neat. 04:27:34 i don't think i would have such a problem with dotsies if it was just "hey, i made this, isn't it kind of cool / weird, what do you think" 04:27:47 but the author is so sure that it's better, and so, so ignorant of anything he would need to understand to make that claim 04:30:06 i mean i don't *know* that he's ignorant, but there are many things conspicuously absent from that page 04:30:18 anyway i think i have said everything i have to say on this subject 04:30:24 but it pisses me off and so i'm disposed to keep talking 04:36:53 plus the classic programmer attitude of assuming you know how something works in people because you know how it works in computers <-- technically humans are nothing more than neural networks. Complicated ones sure, but still just neural networks 04:37:21 ;) 04:37:39 humans are just atoms man 04:37:44 that too 04:37:48 so partical physicists are experts in psychology 04:37:49 and love 04:37:54 particle too 04:37:55 yes 04:38:13 kmc, what was the actual topic though? 04:39:02 which actual topic 04:39:36 the topic from which the line I commented upon was taken 04:39:43 I just read a few lines above and below 04:39:51 after getting back 04:39:58 also it's not like programmers who develop artificial neural networks have much insight into how the ANN has learned what it knows 04:40:02 it's just a black box 04:40:12 Vorpal: you could scroll up 04:40:16

#include main(){printf("Hello, world!\n");}

#include
main(){printf("Hello, world!\n");}