00:07:59 zzzzzzz 00:09:04 -!- SEO_DUDE38 has joined. 00:24:39 bah, flag desacration is illegal in colorado 00:34:08 Feels weird to remove a hard disk from an old computer, plug it into my computer and then boot it in QEmu :P 00:34:17 Seems sort of insulting to the real computer when it runs faster in QEmu X-D 00:41:50 BrainGentlyCaress 00:48:17 * pikhq pisses on a flag, taking advantage of the federal law superseding state law in that instance 00:49:08 hooray constitution 00:49:20 GregorR: We can recompile it... we have the technology. Stronger, faster, better than before. 00:51:13 -!- sebbu2 has quit ("@+"). 00:51:55 * pikhq starts playing "harder better faster stronger" 00:51:57 -!- immibis has joined. 00:52:42 * RodgerTheGreat is listening to Around The World / Harder, Better, Faster, Stronger by Daft Punk from Coachella 2006 00:53:08 :) 00:53:10 -!- immibis has quit (Client Quit). 00:53:36 'tis the music that great programming is made of 00:54:06 * bsmntbombdood is listening to The Gathering / Infected Mushroom 00:56:43 that's music great acid trips are made of 00:58:01 haha. I have witnessed forum brilliance. 00:59:00 a guy named "With Your Ass" posted a thread asking suggestions for removing shards of a broken beer bottle from his backpack entitled "How do I get tiny glass shards out of my sack?" 00:59:39 putting it in the dryer probably 01:00:00 hm. That might actually work. 01:01:11 -!- rutlov has joined. 01:01:20 most people are suggesting duct-tape 01:02:04 that won't get the stuff embedded in the fabric 01:02:58 yeah- I think a dryer is the best idea I could think of 01:03:03 it would really suck to get tiny glass shards in your sack though 01:03:06 do you have experience with this sort of thing? 01:03:11 yeah, no kidding 01:03:46 i bet it would get infected 01:04:01 worse than glass shards: fiberglass 01:04:07 imagine the rash. :S 01:04:19 a la http://wiki.bmezine.com/index.php/Scrotal_infection (nsfw) 01:04:34 eeew 01:04:40 but yes 01:08:06 out of morbid curiosity, did you just type "infected scrotum" into google for that thing? 01:08:48 no, i've read that article before 01:09:16 although that is the first result 01:09:18 OH MY GOD 01:09:30 look at the images 01:09:52 I don' wanna. D:> 01:11:10 -!- rutlov has left (?). 01:13:06 "A Case Of Multiple Sebaceous Cysts Over Scrotum In A 35 Years Old Male" 01:14:22 that is just about the grossest thing ever 01:37:44 -!- puzzlet has joined. 01:37:45 -!- puzzlet_ has quit (Remote closed the connection). 01:51:58 -!- puzzlet_ has joined. 01:52:44 -!- puzzlet has quit (Remote closed the connection). 01:55:50 -!- moomoo has quit ("leaving"). 01:57:34 -!- puzzlet has joined. 02:09:35 -!- puzzlet_ has quit (Read error: 110 (Connection timed out)). 02:27:34 -!- bsmntbom1dood has joined. 02:38:40 -!- bsmntbombdood has quit (Read error: 110 (Connection timed out)). 03:24:26 -!- bsmntbom1dood has changed nick to bsmntbombdood. 03:24:50 -!- oerjan has quit ("Good night"). 03:44:22 * bsmntbombdood is listening to Pink Floyd - The Final Cut 03:53:55 * pikhq approves 03:55:55 i'm not sure i like 03:55:57 too uniform 04:00:47 Not listened to that album, so my approval is of the band. . . 04:00:52 ok 04:01:05 * RodgerTheGreat is listening to Emotion 98.6 by Mylo from Destroy Rock & Roll 04:35:04 -!- puzzlet has quit (Remote closed the connection). 04:43:20 -!- puzzlet has joined. 04:57:30 -!- puzzlet has quit (Remote closed the connection). 04:57:36 -!- puzzlet has joined. 04:58:55 -!- jix has quit (Nick collision from services.). 04:59:05 -!- jix has joined. 05:03:17 -!- puzzlet_ has joined. 05:15:22 -!- puzzlet has quit (Read error: 110 (Connection timed out)). 05:19:22 -!- EgoBot has joined. 05:54:54 -!- bsmntbombdood has quit (Read error: 110 (Connection timed out)). 06:17:27 -!- Arrogant has joined. 06:40:53 -!- bsmntbombdood has joined. 06:41:38 I made macros to make functions/calling easier in MISC :) 06:42:38 (cpp macros that is) 06:52:00 Now, I just need a backend for GCC ... 06:52:02 :P 07:36:16 -!- bsmntbom1dood has joined. 07:42:51 -!- bsmntbom2dood has joined. 07:47:30 -!- bsmntbombdood has quit (Read error: 110 (Connection timed out)). 07:54:06 I have macros for MUL/DIV/MOD now :) 07:54:08 -!- bsmntbom1dood has quit (Read error: 110 (Connection timed out)). 07:59:59 -!- clog has quit (ended). 08:00:00 -!- clog has joined. 09:17:46 -!- GreaseMonkey has quit ("Every time you screw up AWOS, *GreaseMonkey* kills a kitten."). 09:45:09 -!- ais523 has joined. 09:47:54 -!- molchuvka has joined. 10:00:25 -!- SEO_DUDE38 has quit (Remote closed the connection). 10:16:04 -!- SEO_DUDE38 has joined. 10:24:36 -!- ehird` has joined. 10:26:05 I'm annoyed at regular expressions not having enough computational class to do everything I want them to do 10:26:33 for instance, several times I'd have found it useful to have a regexp that matches a string with matched parentheses 10:26:52 so I'm working on a Turing-complete enhanced-regexp language 10:28:37 Perl 6 rules look pretty cool. 10:29:01 I haven't seen them 10:29:03 ais523: why 10:29:06 perl 5 regexps are TC 10:29:16 only by embedding Perl into them, I think 10:29:39 really? 10:29:41 i dont think so 10:29:49 i seem to remember with some bizzare hacks you can make it TC 10:29:54 e.g., that regexp that can solve sudoku 10:29:57 uses no perl 10:30:33 in my language, I can write ($$+$|'^'|) and get an expression that matches only strings with matched parens 10:30:37 how is that done in Perl? 10:30:55 (the $+ means 'match a copy of the containing group') 10:30:58 i do not actually know, sorry# 10:31:03 but i have heard a few times that they are TC 10:31:40 http://perl.abigail.be/Talks/Sudoku/HTML 10:32:28 also, I have things like (a*)b*:-c*:- to solve the famous n as, n bs, n cs problem, which I think is considerably shorter than the corresponding BF program 10:33:48 by the way, a non-TC program can solve Sudoku 10:33:56 (in theory, you could do it with a lookup table) 10:35:08 well 10:35:10 take a look at the examples 10:35:13 i sure don't grok them, 10:35:19 but i don't think they're using a lookup table 10:35:36 no, they're just placing a huge number of constraints on the problem 10:35:45 hm 10:35:49 so it isn't tc 10:36:11 traditional regular expressions can be compiled into a finite-state machine 10:36:19 http://www.google.co.uk/search?hl=en&client=firefox-a&rls=org.mozilla%3Aen-GB%3Aofficial&hs=KCL&q=perl+5+regular+expressions+turing+complete&btnG=Search&meta= Perhaps you will find something here. 10:36:20 and 10:36:25 perl 5 regexps are not traditional 10:36:32 I don't think Perl regular expressions are any better in that respect, but I don't know 10:45:06 * ais523 is reading through the Google results, and so far has found a lot of people disagreeing with each other 11:00:31 mathematical regexs can be compiled into FSMs -- I studied it at Uni 11:02:18 I'm reading through Perl5 and Perl6 regex syntax 11:02:38 Perl5 allows ?{} (embed Perl code), which obviously makes it TC, but that's cheating 11:03:29 it also allows ??{} (defered evaluation), which can be used to execute arbitrary Perl or which can be used to create a context-free grammar 11:03:35 I'm not sure if using that is cheating 11:04:07 Perl6's regexes are like Perl5s, but re-written to be more verbose and 'readable' to beginning programmers 11:04:31 if multiple 'rules' are given, it again creates a context-free grammar 11:04:54 but I'm not sure if a context-free grammar can match (a*)b*:-c*:- 11:07:24 it's interesting to see the parallel evolution of my regex language and Perl6's, though; for instance, they both contain constructs equivalent to Prolog's 'cut' and 'fail' 11:08:35 ...and we both changed . to also match newlines 11:28:22 back 11:28:49 i might make a regexp-like language 11:29:04 i've always wanted for ?[\\$]%(.=&)/ to match something 11:29:12 s/for// (How ironic) 11:29:56 what you wrote is a syntax error in my language, unfortunately (partly because I deliberately don't use square brackets for anything) 11:31:29 i've always wanted regexps that can match entire languages 11:31:30 ;) 11:31:46 a kind-of-regexp parser in one kind-of-regexp would be awesome! 11:31:49 I was designing my regexp language partly for that purpose 11:32:02 why don't any regexp impls have nested regexp groups? 11:32:17 what exactly do you mean by that 11:32:25 s/$/\?/ 11:32:29 (a) -> [['a']] 11:32:39 (a(b)) -> [['a', ['b']]] 11:32:43 see? 11:32:56 what do the square brackets represent? 11:33:01 you could generate an entire parse tree, with some additional commands (like "don't nest", "pop out" etc) 11:33:02 and 11:33:04 arrays in the language 11:33:18 from a match_str(pattern, str) 11:33:29 I think my language might be able to do that, but I can't remember the syntax offhand 11:33:32 so instead of (a) when given "a" returning ['a'] 11:33:37 it'd return [['a']] 11:33:49 so, a match is a list of text and submatches 11:33:51 the root list is too 11:33:57 so, it's a parse tree, really 11:34:18 it wouldn't do it by default, but you could write something like (a=\Ba\E) 11:34:21 if there is a way to say "append this to the result set, even if it is not in the string:", then you can parse an /entire language/ with one regexp0thing :D 11:34:28 where \B and \E change into [ and ] in the output 11:34:32 no no no 11:34:42 pretend we are in a programming language 11:34:44 NOT the regexp language 11:34:48 [x,y,z] is an array 11:35:00 [[x, y, z], [a, b, c]] is an array with two arrays as elements 11:35:08 matching (a) on the string a would produce [['a']] 11:35:28 think of the result of matching as something like [$1, $2, $3...] in perl 11:35:31 that's it: my regexp language outputs a set of nested arrays as its answer when \B and \E are used 11:35:44 except, $1, $2, etc. are ALSO match results 11:35:54 so just matching (a) on the strng 'a' gives [['a']] 11:36:01 the = causes its LHS to be 'replaced' with its RHS in a vague sense 11:36:07 (a(b))c((d)e) on 'abcde' produces: 11:36:21 [['a', ['b']], [['d'], 'e']] 11:36:44 so your output just mirrors the group syntax of the input? 11:37:48 kiiind of 11:37:55 except 11:38:11 there would be a modifier of some sort for () which says 'put this N levels upward in the tree' or 'pop out N levels' 11:38:16 or 'don't nest this match' 11:38:38 (a(b))c((d)e) in regular regexps in 'abcde' would be ['a', 'b', 'd', 'e'] which i do not think is nearly as useful for simple matching OR parsing 11:38:55 something like you suggest can be done reasonably simply with Perl5 regexps 11:39:06 you put an extra group around each section between groups, like this: 11:39:26 ((a)((b)))(c)(((d))(e)) 11:39:42 and then you simply have a template to fill $ns into, like this: 11:39:53 sure, i know 11:39:58 but its not as convenient, imo 11:40:07 you could do everything your regexps do with perl's regexp eval feature 11:40:12 that doesn't mean it's useful to do it like that 11:40:34 [[$2, [$4]], [[$8], $9]] (I'm going to finish writing this comment anyway, even though it's long after it's relevant) 11:40:52 the point is you could write a program to automatically compile regexps like you suggest into the existing form 11:42:06 sure, but add the other features of my regexp stuff i have in my head 11:42:11 which you can't compile short of eval 11:43:37 well, I like my language because I can do ($($1)$=\B=1\E|'^'|) 11:43:58 what does that do? 11:44:39 if applied to the string (a(b))c((d)e) it would return [['a',['b']],'c',[['d'],'e']] 11:44:58 in other words, it's parsing the brackets in the input string into the array notation of the language that you're using 11:45:38 -!- molchuvka has quit (Read error: 104 (Connection reset by peer)). 11:45:54 i'm pretty sure i could do similar 11:45:56 the nesting is not required 11:46:07 you can make a () pair not nest inline 11:46:10 to be exact, you'd have to put =^ at the end if you were using this as part of a larger regexp, but the system adds that automatically if it's not given anywhere 11:46:13 and you can manually add nesting levels with it 11:46:18 and add/remove elements to them 11:46:26 so, mine can do what yours does too 11:46:41 what would a regexp for doing that look like in your language? 11:48:10 i have no idea. i haven't written down my language yet. 11:48:13 i just have ideas for it. 11:48:17 that above would definately be possible though 11:49:24 * ais523 is still reading the Perl6 regexp syntax 11:51:53 ehird`: it seems that Perl6 does hierarchical capturing through capture brackets like you were suggesting above 11:52:12 (although it returns a nested hierarchy of Match objects rather than just arrays) 11:52:15 FUCK! I'm starting to think like Larry Wall. 11:52:23 arrays was kinda symbolic 11:52:31 the notation was a way to represent match data 11:53:40 so given the regex ( a ( b ) ) c ( ( d ) e ) you would get an output in which $1[0] was "d" (because the matches are 0-based) 11:54:51 cool. 11:55:02 can you do nammed submatches? 11:55:18 yes, in both Perl6 and my language 11:55:25 so, $r['g']['g2'] or similar 11:55:54 I don't think the named submatches are hierarchical, but I admit I'm a bit confused trying to read the document in question 11:56:02 http://dev.perl.org/perl6/doc/design/syn/S05.html 11:58:22 so here's a problem: suppose you have an Unlambda expression, and you want to convert it to another expression in which the arguments to every ` have been swapped 11:58:29 considering just s and k for simplicity, 11:59:19 I would write this as (`($1)($1)=`=2=1|s|k)=^ (the =^ is optional if this is the whole regex) 11:59:33 I wonder if it's possible to do anything like that in other regex languages? 11:59:37 bakc 11:59:39 *back 11:59:52 and hm 12:00:07 intuitively i think yes, but the regexp would be repetitive i think 12:00:10 also 12:00:19 another feature my regexps will have, is nested regexps 12:00:27 you can run other regexps on other strings, get their matches, etc 12:00:36 you can run a string as the regexp, even! 12:00:50 I was planning something similar, but hadn't worked out all the syntax 12:00:53 strings of course include groups, replacements on groups, etc 12:01:21 you could run a regexp on just what was matched by a group, even using that regexp to /change/ what that group matched from the point of view of the rest of the expression 12:01:39 or you could take what was matched by a group as the regexp itself, or both (or neither) 12:03:52 exactly 12:03:53 same here 12:04:03 the possibilities are endless ;) 12:04:41 'what was matched by a group' is the closest my language has to variables, as it can be changed retroactively (without modifying the original string; there are other ways to do that) 12:07:39 WD is GTP 12:07:47 aaaaaaaaaaaaaaaaaaaaargh 12:07:58 ?? 12:08:24 problems with my hard drive 12:08:30 the external one 12:08:35 serious problems 12:09:15 a good quote I just came across: 'we think the popular term "regex" is in the process of becoming a technical term with a precise meaning of: "something you do pattern matching with, kinda like a regular expression"' 12:09:22 the funny thing is that before some days a friend of me found bad sectors in his disk which is the almost the same model with mine. 12:10:01 is there a tool for surface scan in Linux? 12:11:56 ? 12:28:56 -!- molchuvka has joined. 12:30:36 ais523: :) 12:30:45 my language will let you assign arbitary things to groups 12:30:59 same here 12:31:01 so, combine the 'arbitary' and 'don't add to match results' modifiers, 12:31:02 and voila 12:31:02 variables 12:31:18 or, just use 'arbitary' to insert elements into various places in the match tree 12:31:22 I assume there's some way to generate an infinite number of them 12:31:23 -> easy parsetree generation 12:31:36 mine can, but only in a push-down way 12:32:28 turing complete parse-tree generation 12:32:30 bliss =) 12:32:38 no more fscking yacc or workalikes! 12:33:09 it seems you, me and Perl6 all have the same aim here 12:33:33 turing-complete parse tree generation? wow. 12:34:04 yes, RodgerTheGreat. 12:34:08 with regular-expression-alikes. 12:34:14 cool 12:34:21 have i mown your blind sufficiently yet? 12:34:34 one traditional problem in parsing a language: the construct (a, b):=(c, d) which assigns c to a and d to b exists in some languages 12:34:41 it's very hard to write those semantics in BNF 12:34:48 -!- Sgeo has joined. 12:34:50 why 12:35:14 because there have to be the same number of variables on each side, and the first variable on the LHS maps to the first variable on the RHS 12:35:15 in coco/r, it's straightforward to write this in EBNF 12:35:33 ll(k) compiler compiler 12:35:39 coco/r is horrid though, so shush ;) 12:35:54 the traditional solution was to accept any number on each side, and then check there were the same number on each side when translating the parse tree 12:35:56 ais523: good old a*nb*nc*n problem 12:36:09 yes, just in a different form 12:36:24 ehird`: please describe the word 'horrid', i dont have dictionary at hand 12:36:39 molchuvka: the fuel that powers the steam engine that creates nightmares 12:36:44 the a^nb^n problem can be solved by a push-down, with an expression like (a$+b|) in my notation 12:36:51 but it matches the last a with the first b 12:37:54 i wonder... 12:38:04 since these re-alikes are turing complete 12:38:18 and with mine at least, you can parse a full language into a parse-tree with only one single re-alikes 12:38:23 what expression parses perl 5? 12:38:52 Perl6 apparently has a built-in expression for parsing perl6; it's a constant that's just defined by the language 12:39:04 an expression to parse perl5 would take a while to write, I expect 12:39:28 but it would be useful because I've never seen a really good Perl syntax highlighter (that can handle regexps and the weird quoting syntax) 12:40:07 damn, just think how complex it would be 12:40:19 you'd be utilizing turing complete hacks all over the place to be 100% correct 12:40:53 for convenience i think my regexp language will include a special quine "variable" (group) that is just the string of the expression it is contained it 12:40:55 *in 12:41:02 this, incidentally, makes a quine very trivial :P 12:41:15 well, the null string's a quine in mine 12:41:23 heh 12:41:30 * ais523 tries to thing of a quine in their language that's at least one byte long 12:41:30 i can just imagine the docs for my quine 12:41:42 "This expression matches nothing into itself." 12:41:50 actually, anything 12:42:00 "This expression matches anything into itself." 12:42:41 you know, my regexp language will probably be really hard to parse (even in itself) and a mammoth task to implement 12:42:55 I'm having serious problems trying to implement mine 12:43:09 I'm writing the docs first and although they're unfinished, they're already several pages long 12:43:38 I do not think I will bother implementing mine, from the features I have in mind it'd be a project on the level of full-scale open source projects 12:43:57 Maybe if someone offers some help after I write a spec. :-) 12:44:21 perhaps you should focus on a simpler proof-of-concept 12:44:37 no, i want my amazingerific expressions in full ;) 12:45:06 heh. Or design an esolang that requires this feature in order to achieve turing-completeness. 12:45:29 that's simple 12:45:45 just make a program input and a regular expression 12:46:00 since they can perform turing complete calculations, voila 12:46:15 the match tree would be the output of the program 12:46:49 in fact you don't even need an input string 12:46:52 (it could just be empty) 12:47:01 since you can assign arbitary values to groups, and run regexps on groups 12:47:02 mine can generate output off no input at all, because it has features for inserting text into the input string 12:47:03 etc 12:47:22 ais523: mine can do that but only because you can explicitly modify and generate matches in the match tree 12:47:49 in mine, you can write (Hello\, world\!)^ to add Hello, world! to the input string 12:48:04 where the ^ 'negates' the preceding group 12:48:13 it has many more interesting uses as well 12:48:58 ais523: can yours be used to parse an entire language in one expression? 12:49:03 within reason that is, not using tons of TC-hackery 12:49:10 yes, pretty easily 12:49:13 because from what i have in my head mine could do that very easily and naturally 12:49:18 as in, not a hack a tall 12:49:21 just natural common usage 12:49:24 which i think is cool 12:49:33 PARSER constant in a language = XD 12:49:34 I have an = command that's not actually necessary, but makes such expressions more readable 12:49:52 -!- Arrogant has quit ("Leaving"). 12:50:13 i think mine will have "and" and "or" constructs 12:50:32 it depends on what you mean by 'and' 12:50:45 i.e. "if this (X) matches, then this (Y) must too, otherwise there is no match" 12:50:56 if you're just checking variables, you write the checks one after the other (possibly with a cut) and you get a short-circuit match 12:51:07 or, "if (X) does not match, but (Y) does, there is a match" 12:51:13 Perl6 has a 'two or more matches that must start and end at the same place construct' 12:51:15 and by x and y i mean two groups of some sort 12:51:31 as for or, that's just the (a|b) construct that's been in every regex language ever AFAIK 12:52:00 sure 12:52:05 but this could handle more complex constructs 12:52:13 and so can or 12:52:22 so it makes sense to ditch (a|b) and get the new syntax 12:52:27 actually, the syntax may be the same 12:52:28 who knows? 12:52:31 but it'll behave different 12:52:35 the syntax is the same in my language 12:52:45 and it behaves the same way, for esoteric values of 'same' 12:53:23 so (a|b) means 'match a, if that fails or there's a fail somewhere later try matching b instead' 12:53:39 &, i think, is a bit more interesting than | 12:53:42 if you don't want it to backtrack into the (a|b) group you can write (a|b)!- instead 12:53:51 anyway 12:53:52 you're right, & is definitely more interesting 12:53:56 with mine you can have different modifiers on a and b 12:53:58 etc 12:53:59 :) 12:54:21 obviously that's possible, you just put the modifiers inside the groups in question 12:54:31 i have no idea what the implications are, but or-ing an arbitary, in-output group with a non-arbitary, out-of-output group sounds... interesting 12:54:39 but, let's think about & 12:54:46 what are the implications? i get the feeling it could be very useful 12:54:50 but i don't quite see which 12:54:54 well, let's boil it down 12:55:21 A&B = if A, then B. else, A. 12:55:39 now - what are the implications of that in a regexp language? (Where the "if A" means "if A matches") 12:55:42 ehird`: for the or you can just write (a|b^), for instance, which either removes an a from the input string or failing that adds a b 12:56:45 for the and you've just defined, that's ab? (or (ab?)? depending on whether you want the not-a situation to succeed or fail) 12:56:59 and you can always just cut away any choicepoints that are left behind that you don't want 12:58:16 i'm pretty sure it isn't ab? 12:58:19 since 12:58:25 if a was true, and b was 12:58:29 only b would be "returned" 12:58:30 not ab 12:58:59 my regexes don't 'return' things, they just modify the input string and/or group-match memory 12:59:31 "return" means "add to the match tree" 13:00:09 then you want (a=x)(b=y)? in my language where x and y are the things to add to the output corresponding to a and b 13:00:09 but, you know what i mean 13:00:18 i'll demonstrate 13:00:36 x -> y -> z, meaning "expr x with input y matches tree z": 13:00:50 (a&b) -> "ab" -> [["b"]] 13:01:03 (a&b) -> "b" -> nope 13:01:14 (a&b) -> "a" -> nope 13:01:25 i think that's right 13:01:42 does nope refer to a fail, here? 13:02:02 in my language, because returning is explicit, you just don't return anything from a: (a=)(b=y) 13:02:37 i think it means a fail 13:02:39 ;) 13:03:11 all languages should have some backtracking construct IMO 13:03:29 how about: 13:03:29 because it makes it much easier to implement backtracking languages, and doesn't affect programs that don't need to use it 13:03:36 (a&b) -> "ab" -> [["b"]] 13:03:36 likewise for multithreading 13:03:38 (a&b) -> "ab" -> [["b"]] 13:03:48 you just wrote the same thing twice 13:03:49 (a&b) -> "a" -> [["b"]] 13:03:50 I know 13:03:58 (a&b) -> "b" -> [["a"]] 13:04:09 because that is "directly" if A then B else A 13:04:28 (a=b|b=a)? 13:04:53 incidentally, on single characters, you can write that 'ab=ba' in my language as shorthand 13:06:01 i dont know 13:06:14 i'm just trying to think of a useful way to implement an and construct 13:06:55 I think ab is probably the simplest sort of and construct; after all, both a and b have to match for ab to match 13:07:13 that sort of and has been around for ages, and nearly all regexes use it 13:07:20 sort of like and in Prolog 13:07:27 it's the 'and then' operator 13:11:01 no 13:11:16 in your programming language of choice, does a & b return [a, b] if both a and b are true? 13:11:19 no -- it returns b 13:11:29 so, ab is not a&b 13:11:51 in many languages, and just returns true or false, which was more the idea I was getting at 13:12:03 or in backtracking languages, succeed/fail 13:12:25 the 'and then' is clearly a separate operator from perl's and 13:12:39 "many languages" suck ;) 13:12:54 ok, let's take a lowest-common-denominator language 13:13:06 one that follows most patterns intuitively, but isn't very interesting : python 13:13:22 >>> 1 and 2 13:13:22 2 13:13:30 in just about every language i have used (that i didn't hate) that happened 13:13:59 I agree that all the best languages seem to implement it like that, apart from C-based languages (some of which are also good) 13:14:03 so -- 13:14:03 (a&b) -> "ab" -> [["b"]] 13:14:03 (a&b) -> "b" -> nope 13:14:03 (a&b) -> "a" -> nope 13:14:09 seems to be a reasonable construct 13:14:24 but in my language, nothing returns anything unless a return value is guaranteed 13:14:25 but the question is, is it useful like that? what is a reasonable usecase? 13:14:33 or is there another way to implement it that makes it more useful? 13:14:42 actually, in C, 1 && 2 == 2 13:14:49 no, 1 && 2 == 1 13:14:54 && always returns 0 or 1 13:14:59 ah, yes, silly me 13:15:17 lisp: (AND 1 2) ==> 2 13:15:18 you can write 1?2:0 for a Perl-style and (where the 0 is a constant) 13:15:43 BASIC: 1 AND 2 ==> 0 13:15:49 (because it always does it bitwise) 13:15:52 basic as an example? :P 13:16:09 anyway by 1 and 2 i didn't neccessarily mean 1 and 2 because of bitwise-ness 13:16:24 i just meant two things that have the truth-value of "true" but are not "true" themselves 13:17:00 I only picked BASIC because it did something different from the other languages 13:17:26 and I was demonstrating that for values other than true or false, what it did didn't just jump to set values nor take the second value 13:18:07 so i guess we're all stuck on trying to work out a use for my (a&b) semantics? :) 13:18:36 imagine a BASIC interpreter written in regexes 13:18:58 i don't want to 13:18:59