00:00:00 --- log: started forth/03.01.27 00:09:15 --- quit: onetom_ (leguin.freenode.net irc.freenode.net) 00:10:22 --- join: onetom (~tom@novtan.bio.u-szeged.hu) joined #forth 00:56:33 --- quit: Fractal (Read error: 110 (Connection timed out)) 00:58:18 --- join: Fractal (tqxkvm@h24-77-171-228.ok.shawcable.net) joined #forth 02:25:52 --- join: male (~male@cpe-024-033-030-124.midsouth.rr.com) joined #forth 02:38:04 --- quit: lament (Remote closed the connection) 03:02:49 --- join: Serg_Penguin (~Z@nat-ch1.nat.comex.ru) joined #forth 03:31:17 --- quit: Serg_Penguin () 03:54:43 --- quit: skylan (Read error: 104 (Connection reset by peer)) 03:55:15 --- join: skylan (sjh@207.164.213.132) joined #forth 03:55:23 --- join: Serg_Penguin (~Z@nat-ch1.nat.comex.ru) joined #forth 06:05:23 --- nick: arke|zZzZ -> arke 06:17:09 --- quit: Serg_Penguin (Read error: 110 (Connection timed out)) 06:24:49 --- quit: arke ("Client Exiting") 06:46:08 --- join: I440r (~mark4@ip209-183-83-51.ts.indy.net) joined #forth 07:06:10 Hey 07:06:49 hi :) 07:09:29 --- join: ASau (~asau@158.250.48.197) joined #forth 07:10:59 Good evening! 07:14:22 Evening, ASau :) 07:15:14 hi asau! 07:16:23 How do you think, why ANS Forth refused to continue tradition of hierarchial vocabulary structure? 07:17:05 I've found some pretty good properties of latter, but this can't be accepted 07:17:34 by ANS-fans as it require some additions to code. 07:17:36 asau what do you mean by "hierachial" ? 07:17:50 I mean tree. 07:18:23 erm- how can you have a vocabulary tree - doesnt that mean having vocabularies within vocabularies ? 07:18:35 i cant think of a reason for using that myself 07:18:45 Yes, I mean this. 07:19:01 You have root vocab. -- FORTH 07:19:26 actually, my root vocabulary is CALLED the root vocabulary 07:19:31 OOOH!!!!!!!!!!!1 07:19:38 all "voicabularies" are compiled INTO the root!!! 07:19:44 now i get it heh 07:19:46 morning 07:20:23 so the word "FORTH" which adds forth to contex is in the root vocab - so there IS a hierachial structure heh 07:20:23 just pinched a python/smalltalk facility, cast it into a forth word 07:20:48 hey bongo - thanx for that email dood - i never heared of that job site before :) 07:21:07 integrated word docs 07:21:18 I440r: ok. hope it helps 07:21:51 in less than 10 minutes hans blix is going to be giving his "speach" - he is probably going to have some sort of limp wristed "we want more time" crap 07:21:56 that jon i've sent you looked pretty good. i might have liked it wouldn't it habe been in the u.s. 07:21:56 Not all vocab.-s 07:21:57 FUCK more time - nuke iraq!!!!!!!!!!!!!!!!!!!!!1 07:21:57 HEH 07:22:21 nuke em till they glow then shoot em in the dark 07:22:44 Each vocabulary (new) is compiled into CURRENT vocabulary, which can be different from root. 07:22:52 but - venezuela is so much closer. 07:23:18 transüport costs are much lower 07:23:24 could even build a pipeline 07:23:33 I440r: Yes, nuke them all. Including every C++ and ANS forth programmer. 07:23:51 asau yes - if "forth" is current and i say "vocablary foo" the word "vocabulary" creates a word called "foo" within the ROOT vocab - not within the FORTH vocab (in isforth i mean) 07:24:21 current doesnt change tho - forth is STILL current 07:24:35 I440r: That is bad... Not as bad, but it is not good. 07:24:53 also, i dont have a word called ALSO - its not needed 07:24:53 It closes much more possibilities than opens. 07:25:24 asau if you are locked into the forth vocab and the forth vocab has all other vocabs in it YOU can unlock yourself. 07:25:43 "u.s depends on cheap oil like a junkie on the next shot" quote german minister 07:25:54 if however vocabs are all compiled into ROOT - and you are sealed into some other voc you cannot unseal yourself 07:26:00 --- quit: onetom ("leaving") 07:26:01 BTW, if you want to nuke someone, why not nuke world terrorist No.1, that is called USA? 07:26:54 asau thats such a fuckiing crock of shit statement. we dont attack un-provoked 07:26:56 --- join: onetom (~tom@novtan.bio.u-szeged.hu) joined #forth 07:26:59 Listen, unless the nuke runs Forth it is orthogonal to the conversation. 07:27:10 there are those that have (usa) 07:27:22 there are those that done. (almost everyone else) 07:27:27 those that have WORKED FOR IT 07:27:33 those that dont DIDNT 07:27:49 so - because we have and they dont they seek to destroy what we have worked for 07:27:55 As to vocabularies, I have to say that I do not see any advantages in non-tree vocabulary. 07:28:06 we just dont let them do so with impunity 07:28:27 asau have you looked at isforths vocabularies ? 07:28:30 ANTI's require much more buzz to change context. 07:28:38 the way ive done them is quite nice actually 07:28:56 I440r: No, elaborate, please. 07:29:07 only - remove everythign except root from context 07:29:26 forth/compiler/any-vocab - adds vocab to TOP of context 07:29:40 previous - drop top item of context 07:29:40 also 07:29:54 I440: That is done by FORTH call. (removing) 07:29:57 if context is root forth compiler <-- top) and i say "forth" 07:30:03 context becomes root compiler forth 07:30:28 What you call "adding on top..." is just climbing a tree. 07:30:36 a vocabulary adds itself to context if its not there already and rotates itself out to top of context if it IS there already 07:30:49 you also cant do context @ current ! in isforth 07:30:56 potential problem : 07:30:57 or the other way arrund 07:31:11 its not a tree, its a stack 07:31:13 you expect a voc to exist in the search list 07:31:25 now you make it current again, and drop it 07:31:32 you'd still expect it to be in the wordlist 07:31:38 --- join: Serg_Penguin (~Z@nat-ch1.nat.comex.ru) joined #forth 07:31:38 yes - if its NOT in the searh order its not searched 07:31:39 Hmm. And how should I exclude only 1 top voc.? 07:31:46 hi-i-i! 07:31:51 but as it has been rotated to top, it's gone now 07:31:52 Zdorovo, Serg! 07:31:57 you can have a vocabulary as current yet still not in context 07:31:57 thats not a problem 07:32:10 vocabulary foo foo definitions previous 07:32:22 foo is NOT in context but all new definitions go into foo 07:32:30 thats perfectly legal 07:32:30 and MIGHT be quite useful 07:32:36 hi serg! 07:32:41 You should look at STOIC, I440r. 07:32:51 Hey Serg_Penguin. 07:32:57 hie 07:32:59 STIOC uses a stack to store the vocabulary search list. 07:33:17 And is very similar to what you've just described. 07:33:22 say: forth foo compiler include bar ( does foo ... previous ) now it is forth compiler , no more foo ? 07:33:50 how's ur proj, Robert ? 07:35:18 i think i start to like the doc facility 07:35:30 : foo ... ; doc" foo doc" 07:35:46 ' foo doc type -> foo doc 07:35:53 serg 07:35:59 if context is 07:36:06 root foo forth compiler 07:36:14 and you say foo 07:36:14 you get 07:36:14 root forth compiler foo 07:36:21 previous drops top item so you get 07:36:29 foot forth compiler 07:36:36 so dont do "only previous" 07:36:36 heh 07:36:41 but that's bad ! 07:36:43 that seals you out of ALL vocabularies heh 07:37:22 i could add a test so that previous wont drop anything if thers nothing but root in context 07:37:44 how is it bad? 07:37:58 its only BAD if the user does something DUMB 07:38:03 and im not stopping him from doing so 07:38:16 cause you might drop vocs you don't intend to drop, just by making them contex, then dropping them again 07:38:30 correct 07:38:43 but thats YOUR fault, not mine for letting you 07:39:21 no, tis not my fault if i donÄÄ't use such a junk construct 07:39:28 no, tis not my fault if i donÄÄ't use such a junk construct 07:39:32 eh ??? 07:39:38 no, tis not my fault if i donÄÄ't use .. 07:39:43 such a junk construct 07:40:24 don't use ... 07:40:32 I guess that's why they call him Speuler. 07:40:36 x-chat behaves strangely ... 07:41:19 merges my line with part of previous line ... 07:41:41 brb - listening to hans blix 07:51:47 * Serg_Penguin stabs burning foul broom into throat of hans blix 07:56:11 --- join: gilbertbsd (~knoppix@67.97.122.17) joined #forth 07:56:27 Well, I've not seen arguments pro ANS vocabularies. 07:57:13 I have many arguments contra, for tree-nested vocabularies (fig-Forth-like). 07:58:30 i see little that makes me pro ANYTING ans related 07:59:47 Well, I can explain why I need tree-nested vocs. 08:00:14 fig voc disadvantage: only 1 context voc 08:00:37 no voc list 08:00:57 I found an interesting trick in my fig-F. 08:01:23 Speuler: You have current voc. and context voc. 08:01:43 true. current is the voc you compile to 08:01:51 context is search 08:02:04 Word is searched in context and down to the root, if nothing found -- in current and down. 08:03:02 right. makes it difficult to impossible to include two voc branches in the search 08:03:22 That is very useful trick, that you see everything down. 08:03:42 Speuler: Nowadays it is called "incapsulation" :) 08:05:07 I've found a way to write "simple precedence context-dependent grammar" extensions. 08:05:09 oo encapsulation still allow you to access encapsulated data through methods, returning the encapsulated data 08:05:40 ASau are you writing a forth for it? 08:05:45 Speuler: That is not used --- NO NEED! Listen. 08:05:56 * gilbertbsd connects his ear. 08:06:09 ASau: i listen to blix, but i can read what you type 08:06:11 gilbert: No, I have already written. 08:06:35 url? 08:07:13 Consider you make all sub-vocabulary words IMMEDIATE. 08:07:17 gilbert: ? 08:08:11 Then you can write words, that refine context _automatically_. 08:08:19 --- join: Kitanin (~clark@SCF61185.ab.hsia.telus.net) joined #forth 08:08:49 shudder 08:09:05 A first example, I wrote using such technique, follows. 08:09:05 ASau do you have an article on this? 08:09:13 gilbert: No. 08:09:20 Should I write? 08:09:24 SURE!!! 08:09:25 Where? 08:09:25 ;) 08:09:29 c.l.f 08:09:34 OR to some website. 08:09:38 but after this. 08:09:57 Hm? Who will publish? JFAR? 08:10:04 possibly. 08:10:06 or not. 08:10:14 post it to c.l.f sometime. 08:10:50 Should I continue explanation? 08:10:55 Here, I mean. 08:11:14 Well, I continue. 08:11:45 I'll change example to expromt. 08:11:58 good. 08:12:01 simplest voc scheme i've implemented as far was a flag per voc. true=searched, false=not searched. 08:12:10 no tree though 08:12:18 ASau. What you're describing is not new. 08:12:21 just a list of all vocs 08:12:29 linked 08:12:40 a linked list? 08:12:41 You know you can "get" various things, e.g. mail, news, web sites. 08:12:52 yes ASau. 08:13:06 couldn't control search order though 08:13:49 * gilbertbsd waits patiently for ASau's explanation. 08:13:52 But you can't have different "get" words at once, you have to write: GET-MAIL GET-NEWS GET-SITE 08:13:52 each voc was linked to the previously created voc 08:14:04 Let us look another way. 08:14:25 (not for search, just for being able to locate each voc) 08:14:38 We always can refine context, just as we do in oral speech. 08:15:10 The natural way is: "Get mail." 08:15:23 mail get sounds more natural to me :) 08:15:51 But we have only one way to distinguish: from _first_ word to _second_. 08:16:08 Speuler: You're right! 08:16:41 overload get ? 08:16:47 even oopsify it ?? 08:16:47 In some languages you can swap words, changing its forms maybe. 08:17:03 In Forth you can write phrases: 08:17:07 MAIL GET 08:17:13 NEWS GET 08:17:19 SITE GET 08:18:15 How this can be done even in compiling time: just make vocabulary switching immediate. 08:18:37 I.e.: VOCABULARY MAIL IMMEDIATE 08:18:59 MAIL DEFINITIONS : GET ( here you describe how to get mail ) ; 08:19:21 ( ... ) 08:19:45 VOCABULARY NEWS IMMEDIATE NEWS DEFINITIONS : GET ( ... ) ; 08:20:02 ( etc. --- EOE ) 08:20:25 In STOIC you say MAIL< GET > 08:20:34 Like I said. This is not new ;-) 08:20:36 Now let us see what can we do from such a technique. 08:20:57 the word "vocabulary" should automatically make "news" an immediate word - but why should a vocabulary be immediate ? 08:20:59 male let us see what can we do from such a technique from STOIC(al) 08:21:04 An impressive phrase is this: 08:21:39 TERMINAL PARAMETERS MONOCHROME 80 COLUMNS 24 LINES SET 08:21:57 More impressive ones can be imagined. 08:22:13 #! /usr/local/bin/stoical 08:22:13 % 04/24/02 - Jonathan Moore Liles. 08:22:13 % Library to approximate Rob Spruit's Cryptic Constructs, in STOICAL. 08:22:13 'cryptic branch 08:22:13 cryptic< definitions 08:22:14 '?{ : () if{ exec ; immediate 08:22:16 '}{ : () }else{ exec ; immediate 08:22:18 '?} : () }until exec ; immediate 08:22:20 '?}{ : () }while{ exec ; immediate 08:22:22 'xx{ : () loop{ exec ; immediate 08:22:24 > definitions 08:22:26 cryptic< 08:22:28 ;f 08:22:30 Translations: 08:22:32 Traditional Cryptic 08:22:34 ------- 08:22:36 if{ ... } ?{ ... } 08:22:38 if{ ... }else{ ... } ?{ ... }{ ... } 08:22:40 { ... }until { ... ?} 08:22:42 { ... }while{ ... } { ... ?}{ ... } 08:22:44 n n loop{ ... } xx{ } 08:22:51 Now cryptic< 1?{ 'foo = } > is legal. 08:23:23 male: That is just simple renaming/aliasing. 08:23:27 cryptic< is an immediate word that pushes the "cryptic" vocabulary onto the vocabulary stack. 08:23:42 No, ASau. I just used that as a handy example. 08:23:58 > is an immediate word that pops it. 08:24:48 Well, I continue. 08:25:06 ...I go on. 08:25:39 STOICAL's organization of the dictionary allows the programmer extreme 08:25:40 flexibility with his name space; By packaging groups of related words into 08:25:40 vocabularies he gains conceptual clarity, and increases code re-usability. 08:25:40 For example, one might create a vocabulary branch TERM< to contain a group of 08:25:40 terminal control words. These definitions could be stored in their own source 08:25:40 file, to be LOAD'ed by others without the concern of name conflicts. 08:25:42 In fact, the STOICAL distribution includes a library, named 'cryptic', that 08:25:44 implements an adaptation of Bob Spruit's Cryptic Constructs. This library can 08:25:46 be loaded, and its definitions exposed, by the following commands: 08:25:57 After I realized all this, I found good examples in my native Russian. 08:26:28 3.7.2. Vocabularies And Branches 08:26:28 A Branch is a word who's parameter field points to a vocabulary, and who's code 08:26:28 field points to instructions that push this value onto the vocabulary stack. 08:26:28 The BRANCH word is used to create these new vocabularies. The following example 08:26:28 creates a branch, FOO<, from the current vocabulary, pushes it on top of the 08:26:28 vocabulary stack, sets it to be the vocabulary for new definitions, defines the 08:26:30 word BAR into it, uses the word > to pop it off the vocabulary stack, and 08:26:32 finally, resets the current vocabulary. 08:26:34 'foo branch 08:26:36 foo< definitions 08:26:38 0 'bar constant 08:26:40 > definitions 08:26:46 Just as in natural language. We have imperative construction with verb at the end, 08:26:51 I'm suprised I'm not getting /kick'd for flooding. 08:26:53 ;-) 08:27:31 so we can express an _order_ just this appropriate "Forth" way. 08:28:09 I can't feel if such thing is unnatural (or sounds like that) in English 08:28:35 but I see no reason not to use it in Forth. 08:28:55 Another example: Forth inline assembly. 08:29:46 I consider x86 assembly, here I see that instructions with the same name do different things: 08:30:16 they operate different type operands. 08:31:06 Usual solution is to make mnemonic-word to check operand types, 08:31:36 but now we can make operands provide enough information about themselves 08:31:43 at earlier time. 08:32:28 Let us consider moving command: register->register, mem->rem, imm->reg. 08:33:14 Suppose we have ASM vocab., as "root" for asm. 08:33:28 Abbrev.: 08:33:41 : UNIT VOCABULARY IMMEDIATE ; 08:34:05 No, another abbrev.: 08:34:30 : UNIT VOCABULARY IMMEDIATE LATEST PFA CFA EXECUTE DEFINITIONS ; 08:34:54 read LATEST @ ;) 08:35:15 Hm... 08:35:20 --- join: semtex (~Speuler@mnch-d9ba4416.pool.mediaWays.net) joined #forth 08:35:22 Do you listen to me??? 08:35:45 (ping?) 08:36:12 --- quit: Speuler (Connection timed out) 08:36:28 re 08:36:36 Ok. 08:36:36 --- nick: semtex -> Speuler 08:36:50 heh 08:36:57 Well. 08:37:03 UNIT ASM 08:37:32 How would I write MOV command? It is interesting. 08:37:46 I want to write it this way: 08:37:55 AL CL MOV 08:38:18 AL addr ) MOV 08:38:31 imm # AL MOV 08:38:43 addr ) AL MOV 08:39:03 Yeah! PDP-11 forever :) 08:39:17 How this can be done. 08:39:49 create #registers 0 , 08:40:22 ASau do you use a pdp-11 ? 08:40:24 You see we can distinguish this 4 different commands dividing into 3 parts by first operand and 2 parts by second one. 08:40:40 : register 1 #registers dup @ -rot +! constant ; 08:40:51 register al 08:40:53 register bl 08:40:54 ... 08:41:26 So, I create in ASM three my UNITs: BYTE-REGISTER IMMED ADDR 08:41:29 : mov ( r1 r2 -- ) .. (build opcodes) c, c, ; 08:42:56 All byte-register words change context voc. to BYTE-REGISTER, # -- to IMMED and ) -- to ADDR 08:43:50 In that vocabularies I create sub-voc.-s BYTE-REGISTER and ADDR 08:44:20 male so have you done any more work on STOICAL? 08:44:22 And at there second level nested I create words MOV 08:44:37 Now, what I have. 08:45:01 If I miss an operand I get error: NOT FOUND! 08:46:06 i consider adding another word field 08:46:09 Actually I'm still working on the document/macro processor ;-) 08:46:18 a pointer to additional word data 08:46:30 like, source link, doc link etc 08:46:31 But I don't mind.. This is something I've needed for a long time. 08:46:52 possibly a header field 08:47:03 I never want to have to worry about formatting a document again ;-) 08:47:13 I don't need complex branching to distinguish different forms on the same command. 08:48:03 male: Use TeX. 08:48:09 or i gonna recyle the (unneeded) link field 08:48:17 Tex is equivalent to EMACS 08:48:18 TeX is a monster. 08:48:41 And what do you want? 08:48:50 Another sed or m4? 08:48:57 sed is good. 08:49:01 awk is better. 08:49:03 ed is best. 08:49:05 ASau, I just realized a perfect example of how I already accomplish what you're musing about. 08:49:51 roff ? 08:49:57 Oh, main idea in words: 08:50:30 Do not try to resolve context, make your words do it instead of you. 08:51:05 gilbert: Have you seen GEMA? 08:51:18 I am really impressed by it 08:51:25 In stoical the words :( and ): open and close the TYPES< vocabulary whichtranslates type names into their associated numbers. That way I can define a typechecked word with the following syntax 'foo :( string float ): ... ; 08:51:30 GEMA? did I dream that question? 08:51:38 --- quit: Serg_Penguin () 08:51:54 I am convinced I was asked that question yesterday. But I was not on the net yesterday. 08:52:32 But obviously I'm not the first person to ever try something like that. It has been the practice in the forth community for as long as I can remember. 08:52:59 Of course, gema does not replace sed or awk. But it is very good. 08:53:17 My processor is more like TRAC. 08:54:03 I can do substitutions with line breaking. Even in binary format. 08:54:25 male: What TRAC do you mean, which year :) 08:54:50 --- join: hp48nik (xru52729fj@1Cust138.tnt4.vancouver.bc.da.uu.net) joined #forth 08:54:50 Ha. More like TRAC-64. 08:55:05 hi all 08:55:33 Hmm. Rather scare language. 08:55:48 " : good-morning coffee shower irc ; " 08:55:51 But obviously I'm not doing most of the ridiculous things that Moores di. 08:55:54 Good evening. 08:55:54 Reverse polish lisp. 08:56:08 hp48 UserRPL is like Forth crossed with Lisp ;-) 08:56:16 I love RPL/2. 08:56:24 I just wish I spoke french ;-) 08:56:37 if Lisp had RPN, it would get rid of ((((((((((((((((((((((((()))))))))))))))))))))))) 08:56:53 male: pourquoi? 08:57:03 why do you wish that you spoke French? 08:57:17 http://www.makalis.fr/~bertrand/rpl2/ 08:57:19 Heh. 08:57:31 babelfishit male! 08:57:48 Eh, that doesn't work so well on C and Fortran ;-) 08:57:59 French is like BASIC... for the losing side of flame wars, and nothing more. 08:58:02 hp48 are you sure lisp would still behave LIKE lisp if it were RPLed? 08:58:07 hp48nik : A central idea in lisp is, believe it or not, lists which constitute all the parens. Converting to "RPN Lisp" or whatever would make it, well, not lisp. 08:58:23 Robert: Russian rules :) 08:58:30 ASau: Da ;) 08:58:33 Your functions are lists too, remember. 08:58:34 no. Binary rules. 08:58:35 You can have RPN without a stack fractal. 08:58:36 The second language that rules is German. 08:58:50 Nein! Schwedisch! 08:58:53 why? it hasn't ruled since the mid 40s/ 08:58:53 :)))) 08:59:11 English is spoken by a small percentage of the world's population, and that percentage has been in constant decline for decades (talking about native speakers: first language) 08:59:28 male : That's not the point. 08:59:30 EVERYONE in western europe speaks english. 08:59:41 Everyone in Eastern europe is learning to speak english. 08:59:43 ;) 08:59:55 gilbertbsd: I have dual EU/Canadian citizenship... what you say about english is very silly and untrue 09:00:05 thank you. 09:00:09 A my na vsjo plevat' hoteli :))) 09:00:19 hp48nik : Sadly it's what most americans actually believe. 09:00:34 Bah. Some would say that LISP is the language that shouldn't be. 09:00:36 ummm why not? 09:00:39 "Leave it to the Yankees to be totally ignorant about everything outside their country (and most things inside it, too)" 09:00:41 EVERYONE is gonna come here anyway!~ 09:00:54 And yall havta learn to talk american good! 09:01:14 * gilbertbsd is amused at being called a yankee. 09:01:26 Heh. 09:01:26 I live in miami, which happens to be a foreign country. 09:02:15 :( That "Leave it to the Yankee..." is false. 09:02:21 male : Modern incarnations of lisp are some of the most powerful, expressive languages available... 09:02:26 lisp with RPN would simply be ... lisp with RPN 09:02:37 Yankee have to be tought from time to time. 09:02:52 Like 11 Sen shows. 09:02:53 hey the innernet was invented in America. 09:03:00 ummm IRC was invented in America 09:03:04 it ALL comes from here! 09:03:05 IRC: Finland 09:03:08 nope. 09:03:10 gilbertbsd : Not true. IRC was invented in finland. 09:03:10 yahoo. 09:03:21 I was kidding! sheesh! 09:03:49 I told you, I LIVE in Miami, A foreign country in South eastern florida!! 09:03:54