00:00:00 --- log: started forth/13.02.02 00:32:02 --- quit: yunfan (Ping timeout: 245 seconds) 01:02:01 --- nick: tangentstorm -> tangentwork 01:25:54 I440r: Is there a good essay out there explaining why ANS Forth killed Forth? I'd rather just use an URL than rewrite the same pitch constantly. :) 01:26:14 ttmrichter, you must blog about it! 01:26:17 and then tweet it. 01:26:21 and then share it on g+ 01:26:23 --- join: Indecipherable (~Indeciphe@41.10.216.21) joined #forth 01:26:30 * ttmrichter doesn't do the Twit thing. 01:26:34 that´s the way to go about it these fine social networking days. 01:26:44 tweeter's for twats, no? 01:26:46 i agree. 01:27:08 I440r: Ping 01:27:19 I just figure that if the service is called Twitter, then the users are Twits, right, bjorkintosh? 01:27:31 They are twits. 01:27:40 it's an inside joke somewhere, i'm sure. 01:39:00 hi 01:39:39 ttmrichter, i dont think ans forth did kill forth. i just think ans forth has some major flaws all the academics drool over as advances 01:40:27 Indecipherable, how long ago did you ping me 01:40:53 its 3:40 am here, was about to go into power save :) 01:41:07 Don't have the timestamp on... ~5 minutes ago 01:41:18 you're in CST, I440r? 01:41:35 chicago 01:41:44 but about to move to missouri 01:41:51 why, is it better there? 01:41:58 last time i saw MO, i kept driving. 01:42:31 1: its not ILLANNOY which makes it something other than "the most morally bankrupt state in the union" 01:42:47 I'm in CST myself. 5:42PM. 01:42:49 i start a new job monday 01:46:18 pm, ttmrichter? hah. 01:46:22 oh congrats, I440r. 01:46:24 is it a dev job? 01:46:44 yes 01:46:54 embedded systems? 01:46:55 or php? 01:47:49 * tangentwork uses twitter. 01:48:07 tangentwork, but WHY? 01:48:09 embedded. 01:48:16 lovely. 01:48:19 doing forth? 01:48:23 not yet :) 01:48:32 cool, I440r 01:48:47 all their products use PIC's but nobody there likes them, they want to switch to avr;s or something else 01:49:15 bjorkintosh: because i want to reach people. same reason i hang out on irc i guess. 01:49:15 which PICs? 01:49:34 16's and 24's and maybe others 01:49:40 ah. 01:50:01 i think what killed forth is that there isn't really anyone advocating for forth. 01:50:12 i dont think forth is dead :) 01:50:22 its just a little sick because it has too low a user base 01:50:36 thats why i wrote isforth, as a teaching aid - so more people can learn 01:50:47 yeah you can never really kill a language :) 01:51:39 well not a computer language. human languages die all the time. 01:51:42 :/ 01:51:44 anyway. 01:51:48 hmm. 01:51:52 for the same reason. 01:52:15 cherokee is an intersesting example of a language that probably won't die. 01:52:38 because the people who speak cherokee decided to do something about it. 01:52:44 ah. 01:52:47 i am in oklahoma. 01:52:57 even ancient greek and ancient aramaic are still somewhat known languages tho there are some things lost 01:53:03 the casinos will ensure they don't die, i think. 01:53:22 i'm pretty sure they even worked with the rossetta stone people to make cherokee lessons 01:53:44 thatÅ lovely. 01:53:57 but... nobody's really doing that with forth. 01:54:07 even considering what I440r just said. 01:54:08 forth sources are open source. 01:54:14 that's really all that matters. 01:54:31 the truth is that, there is not a paucity of resources as there used to be... 01:54:58 yeah there's plenty of stuff out there... but there's no organization to it. 01:54:58 so forth is 'forthed' out of its niche. 01:55:28 but uhh, is there a particular thing that forth teaches that no other language does? 01:55:35 lisp has lambda and whatnot... 01:55:52 CM was a lisp coder. it was deficiencies in lisp that made him develop forth 01:56:02 well you could implement lisp in forth and forth in lisp without much effort. 01:56:03 forth is (mostly) imperative. and there is a lot of competition in that area. 01:56:24 read thinking forth. that has some good insight into forth 01:56:43 one thing forth has is the ability to extend the language in ways that other languages can't. 01:56:44 like forth being not a programming language but a system for developing application specific programming languages 01:56:49 because you have all these little tiny compilers.. 01:56:51 yeah. 01:57:12 --- join: kbmaniac (~dave@86.157.24.246) joined #forth 01:57:25 but OO and functional programming languages make the _same_ claim. 01:57:36 i think it's a really good way to show how to compose ideas 01:57:44 you build a DSL to suit your problem 01:58:49 there's no real difference between what these languages can do 01:58:53 hell, even prolog makes that claim. 01:59:09 it's all about style and community. 01:59:10 "you don't really program in genre x, you build a DSL to suit your problem". 01:59:31 there is nothing cherokee or aramaic can't do that english can 01:59:54 of course. 01:59:59 it is all about the casinos. 02:00:02 err. communities. 02:00:05 human languages describe the world... computer languages describe computers. 02:00:12 also.. 02:00:16 i don't think they describe computers. 02:00:17 bjorkintosh: I would actually submit that Prolog does, in fact, qualify as a language used to build DSLs. 02:00:26 an accounting app does not describe a computer. 02:00:45 not computer the thing on your desk but "computer"... the abstract notion. 02:00:48 sure it does. 02:00:49 ttmrichter, so do functional and OO languages. all make that claim. 02:00:55 bjorkintosh: https://bitbucket.org/ttmrichter/turing/src/6d9f6c8b89f8a2d1e016095df2c1370076c29f51/prolog/turing_test.pl?at=default 02:01:06 i use ledger-cli.org for my accounting 02:01:10 That's a toy Turing machine thing I wrote in Prolog. 02:01:17 Look at the definition of the "busy beaver" machine. 02:01:23 Then look at the mathematical definition of the same. 02:01:24 ledger has a REPL and everything 02:01:31 They're almost, but not quite, the same thing. 02:01:53 Anybody who knows the maths behind Turing machines can trivially write Turing machines for the Prolog simulator. 02:01:55 it's not a turing complete programming language, but if you look at the code, it's all about how computers implement accounting... 02:02:05 hmm 02:02:19 sort of like all books are about people. :) 02:02:24 ttmrichter, no doubt. 02:02:39 a book on accounting is to people what an accounting program is to computers. 02:02:40 tangentwork, in so far as they are currently written by people, for people. 02:02:50 Forth, I think, could also get similar levels of similarity to the maths. As could, likely, the MLs, Haskell, etc. 02:03:06 But I rather doubt Java could get even in the same city, not to mention ballpark. 02:03:12 sure it could 02:03:22 why not? 02:03:56 you'd just build up enough abstractions and you can make it look like anything. 02:04:04 the syntax might be in your way but you could do it 02:04:27 i was just thinking about PL/0 earlier... it has no arrays, pointers, or parameters 02:04:38 It's the syntax in particular, tangentwork, that makes me sceptical. 02:04:46 Java has too much ceremony to make a decent DSL. 02:04:57 tangentwork, wirth's pl/0? 02:05:01 yes 02:05:09 i implemented a pl/0 -> retro compiler recently 02:05:15 nice. 02:05:24 if i really had to do it, i could work out a way to make pointers, arrays, and parameters in pl/0 02:05:25 is it still an instructive language? pl/0? 02:05:52 it was an example language he used in one of his books, which had a chapter on compilers ( algorithms + data structures = programs ) 02:06:03 i know what it is. 02:06:12 how else would i have known it was wirth's language? 02:06:27 that book is out of print, but in later editions he dropped the chapter and made it its own book 02:06:43 hah. 02:06:45 when he did that he expanded it to a much bigger language called oberon-0 02:06:45 well. i own a copy. 02:06:52 i have the oberon book. 02:06:57 cool :) 02:07:18 hmm.. maybe i misunderstood your question. 02:07:21 project oberon, etc etc. and also 'into the realm of oberon'. 02:07:31 havene't read that one 02:07:45 tangentwork, i was trying to find out if pl/0 was still worth studying. 02:07:49 the oberon-0 book is called compiler construction 02:08:28 what i was trying to say is is that I think the PL/0 language is worth studying if you're making a compiler. 02:08:37 ah excellent. 02:08:45 PL/0 is intels language? 02:08:49 that became PLM? 02:08:49 i wrote a few programs in it to test my compiler. it was fun. 02:09:00 I440r, not at all. 02:09:06 I440r: no, this is a toy language by niklaus wirth. it's a very very tiny subset of pascal. 02:09:08 kk getting mixed up :) 02:09:12 aha 02:09:22 there is another language with that name though, related to PL/1 02:09:36 pl/0 simply stood for programming language zero. 02:09:38 why do academics all drool over niklaus wirth lol 02:09:40 which i think is what you're talking about. 02:09:46 I440r, because he was a tall thin man. 02:09:59 ... he worked from the board all the way up. 02:10:12 designed and built his own computers, just to teach how it was done. 02:10:13 i think he's kind of short. 02:10:25 he's also still alive. no need for the past tense yet. :) 02:11:21 anyway. point is if you really had to do it, you could make a bunch of variables like a1 a2 a3 a4 and so on to get arrays... then have a procedure with a giant IF block. 02:11:38 it would look absolutely horrible, but you could do it. 02:11:57 lol 02:11:58 and since it was so horrible, you'd probably write another program to generate it for you... 02:12:06 why bother, if your going to do that just use brainfuck! 02:12:15 and pretty soon you have the language you want. 02:12:17 tangentwork, here's the definition for tall thin man: http://research.microsoft.com/en-us/um/people/gbell/craytalk/sld002.htm 02:12:39 ohhh 02:12:57 gotcha 02:13:17 cray computers were also inaccurate 02:13:27 so they would come up with a WRONG answer faster than any other computer 02:13:41 I440r, were they? 02:13:42 anyway. point is all languages are equally extensible... so why do we even think of them as separate things at all? 02:13:53 tangentwork, because, people. 02:13:57 yes 02:14:08 why do forth people get annoyed when you make a forth that isn't like their forth? 02:14:09 I440r, well, wirth is a tall thin man. so was cray. 02:14:16 tangentwork, because, people. 02:14:35 if the point of forth is really that you can extend it and make any words you want... why get upset when someone does? :) 02:14:47 yeah.. people. 02:15:11 it's really all about finding a group of like minded people that do things the way you do, and creating a sort of community. 02:15:19 tangentwork, the point of forth is to define language and syntax with ease. with SIMPLICITY. no other language achieves this as well as forth does 02:15:28 but: people get old and die. 02:15:34 luckily new people are born. :) 02:15:46 some of the new people are smart. 02:15:55 most of them have never heard of forth though. 02:16:05 they have, however, heard about twitter. 02:16:12 nothing is forever 02:16:18 nothing man made is forever 02:16:21 i should say 02:17:11 not that they will ever come together to make a forth standard package or library like this: http://golang.org/pkg/ 02:17:19 but who cares? :) 02:17:20 right? 02:17:24 so... if there were someone who wanted a thriving forth community to exist in the world.... they would need to be on twitter and facebook and hacker news and reddit telling all the new people about forth. 02:17:47 tangentwork, yes, news about forth. but it is like telling them about assembly language. 02:17:59 slava prestov was such a person for a while... ( the guy who made factor ) 02:18:05 unless they are directly interested in working on bare metal, how will forth help them create the next vooza.com ? 02:18:27 nobody writes libs like that monster in forth because its a waste of time to do so 02:18:30 excellent point: where are the forth web frameworks? :) 02:18:38 thats coding shit now that you think you might need later 02:18:44 or... 02:18:50 but then when you need it it doesnt fit the situation and you code it from scratch 02:18:59 well... maybe you don't market forth as a java-killer. 02:19:19 maybe you present it as: here's the little language you can use on your own 02:19:19 writing canned libraries is what makes C so fucking horrible 02:19:32 to do anything useful in C you have to know 873654864287645823 gigs of header files 02:19:40 i could totally see a forth shell that was way better than bash, for example. 02:19:44 in forth you dont. you need DES encryption? fucking write it 02:20:01 yes. 02:20:05 i agree.. forth is a little language you can use for your own things. 02:20:18 so unless you are working on bare metal, why bother with forth? 02:20:20 i can also see forth replacing elisp in a text editor. 02:20:22 i think busybox should be rewritten in forth 02:20:33 most people do not know how to program at all. 02:20:39 they know how to google libraries 02:20:44 and butcher them into submission. 02:20:48 :) 02:21:04 so forth is not going to be their idea of a night out on the town. 02:21:19 yes. you take all the cake incredients that someone else made. mix them all up and put themin the easy bake compiler and out pops an application 02:21:27 look ma! im a computer programmer! 02:21:31 well I440r was saying the other day that he sat in this room for years and there was nobody here. 02:21:41 for about a year and a half 02:21:50 by then i had some regular fly'bys in here 02:21:57 well still.. now there's a bunch of people in here. 02:21:58 then a few people stayed. 02:22:11 mostly idlers who never talk and bots 02:22:18 i know for a fact that some of the people in here don't even know forth ;) 02:22:37 i do not know forth. but i am trying to know it. 02:22:59 bjorkintosh: are you a freak of nature? :D 02:23:15 like two heads or tentacles or something? 02:23:30 why do you say that, tangentwork? 02:23:48 because if not... perhaps there are other people out there like you who would also be interested in learning about forth, if they only even knew about it. 02:24:14 i have known about it for ages. 02:24:21 a really long time infact. 02:24:38 but not enough to really look into it. 02:24:42 but. i could get things done faster in ruby :) 02:24:45 or want to learn it. 02:25:00 oh i have desired long and hard to learn it. just as i have desired long and hard to bed j-lo. 02:25:01 forth is "in the noise" for most people. below the radar. they never even hear about it or it doest register as being worthy of examination 02:25:05 i always thought forth was just backwards lisp, personally... 02:25:27 no it's not. there are things one can do in lisp that i cannot see being done in forth. 02:25:32 well. 02:25:34 easily. 02:25:34 talking like taht can get you banned :P~ 02:25:36 not impossibly. 02:25:36 lol 02:25:39 yeah. my point, though... is that mixed in throughout the population is a small number who would be really open to the idea of forth 02:25:50 but it is likely because i do not know forth. 02:25:52 hell. 02:25:55 but anyway i need to zzz 02:25:57 it is ALL because i do not know forth. 02:26:20 the guys in suits who come knocking on your door talking about religion know that 95% of people are going to tell them to go away... 02:26:21 i believe that with the surge in interest in microcontrollers and tiny computers, it could really have an impact. 02:26:30 they're looking for the tiny percentage that are open to their message. 02:26:35 bjorkintosh, learn dup drop swap nip etc etc. then learn begin/while/repeat/until/for/next/do/loop 02:26:46 once you have those down writing code is not difficult 02:26:47 night I440r 02:26:52 I440r, me love drop swap nip long time. 02:26:58 bjorkintosh: The problem with that "surge of interest" is that microcontrollers these days are bigger and more bloated than many mainframes of the '70s. 02:27:01 i agree. microtrollers.... and also parallel computing. 02:27:17 forth is an ideal parallel computing language 02:27:22 well, ttmrichter the bloat is what forth can attack. 02:27:26 and perhaps, oberon :) 02:27:27 no surprise about greenarrays going that way. 02:27:28 but i digress. 02:27:36 I have an STM32 eval board here that probably performs about the same as the VAX 750 I used in university. 02:27:48 The chip is roughly the size of my thumbnail, in package. 02:28:06 --- join: yunfan (~jyf@unaffiliated/yunfan) joined #forth 02:28:14 the point is: there are people out there who would probably love to learn forth and just need a nudge. 02:28:21 There's so much space on it that people could use Ruby, I swear (not really, but you get my point). 02:28:47 And the eval board cost me less than $15. 02:28:53 and if you wanted to find those people... you probably ought to be on twitter and facebook google+ and reddit and hacker news and digg and youtube and.... etc etc 02:28:58 tangentwork, i think updating what i consider the forth canons would be a good start. 02:29:09 starting forth and thinking forth. 02:29:19 Updating them how? 02:29:20 one of these is now creative commons 02:29:29 i have it on my kindle. 02:29:58 i filed a bug report about the formatting of one of the pages in the book through the sourceforge bug tracker... not sure i ever got a response. 02:30:27 ttmrichter, what do people do with languages? they create things right? 02:30:36 what do they do with microcontrollers? they move things, right? 02:30:39 and make things blink. 02:30:56 so if it can be shown that a forth can help them do those things very easily... 02:30:59 then, voila. 02:31:05 i am not saying that people should do this or will do this... i'm just saying if someone wanted to see the forth community take off, this is how to do it. :) 02:31:45 tangentwork, people are on hackernews and reddit and g+, but i feel they are preaching to the choir. 02:31:48 i have ambitions like this for my own language, so that's what i'm doing :) 02:31:53 tell me, why today is forth relevant at all? 02:32:08 why is c++ relevant? 02:32:19 because you will always get a paycheck, if you know it. 02:32:25 you will never have to look hard for work. 02:32:31 no other reason. 02:32:55 i suppose this is something like my stump speech. when i give the speech in #pascal, i ask: why did C++ win, when for so many years pascal was just as viable? 02:33:09 pascal was not as viable. 02:33:16 and i am glad it did not do what c++ did. 02:33:22 sure it was. in 1995? 02:33:26 you know why? because Wirth the great did not make it as such. 02:33:35 even microsoft had a pascal compiler. 02:33:36 he wanted to use it as a teaching language. 02:33:46 but perverts corrupted his mission. 02:33:48 his holy mission. 02:33:59 now, when Lord Wirth, saw that things had changed, he changed his language. 02:34:05 it then became modula. 02:34:07 and then modula 2. 02:34:10 and oberon. and so on. 02:34:19 he did not remain married to pascal. 02:34:34 No, but other people did. 02:34:58 they should have listened to wirth. 02:35:16 If you were making apps in 1995, Turbo Pascal was definitely among the options at the top of the list. 02:35:23 and borland pascal and whatnot. 02:35:32 or quickpascal / microsoft pascal. 02:35:44 maybe 1992-93... 02:35:47 oh i liked tp 02:35:50 it was nice and fast. 02:35:54 after windows came out, not so much. 02:35:57 why didn't forth occupy that space? 02:36:01 --- quit: Indecipherable (Ping timeout: 246 seconds) 02:36:07 but TP back then was just as popular as python or ruby today. 02:36:12 forth did occupy that space. 02:36:16 in the 1980's. 02:36:23 i know. hell, i would kill for the tp bundle today. i loved it so much. 02:36:23 i learned that recently. 02:36:44 i didn't like the switch to the windowsy nonsense with owl and whatnot. 02:37:04 well... you can get TP5.5 for free from the embarcadero site, and also there's freepascal.org which is basically compatible. 02:37:26 in fact just yesterday i was working on this minesweeper game in retroforth 02:37:54 i think the last version i used was tp 7. 02:37:58 loved it. 02:38:00 minesweeper uses the flood fill algorithm to open up the area when you click on a cell that doesn't have a mine 02:38:07 hell, i also loved BASIC. 02:38:12 shame on me, right? 02:38:21 nah. GWBASIC was cool. :) 02:38:22 well. i could make it generate pictures and sound with a few lines. 02:38:32 --- quit: Nisstyre-laptop (Quit: Leaving) 02:38:34 anyway.... i wrote a flood fill implementation in 1994 in turbo pascal. 02:38:36 whereas i couldn't convince c to do the same for the life of me. 02:38:49 so yesterday I took that code and got it running under FPC. 02:39:03 fpc = free pascal compiler 02:39:28 http://imgur.com/gallery/sMsefsb 02:39:32 --- quit: kbmaniac (Ping timeout: 245 seconds) 02:39:44 nice. 02:39:45 i'm using emacs there, but fpc also comes with "fp"... which looks JUST like turbo vision 02:39:58 like what you got with TURBO.EXE 02:40:08 anyway.. this is off topic for #forth :) 02:40:44 but.. pascal and forth were both very popular in their day, but they didn't have anyone advocating for them. 02:40:56 no it is not. 02:41:04 you know what i liked about tp? 02:41:08 the libraries. 02:41:32 whenever i looked at the firmware on the old macs, all i saw was a screen full of words. 02:41:42 how the hell was i supposed to make sense of any of that crap? 02:42:01 back then, it pretty much had to be a company advocating for things... but with the internet, it's a whole different ballgame. ruby, perl, php, python.... none of these had strong commercial backing initially. 02:42:05 so things are organized as words. great. but one does not immediately get a sense of how any of it is organized. 02:42:53 well, the list of words. 02:43:03 yeah. we were looking at ncurses the other day in another channel... I tried to show someone the CRT unit... and he was like "how can you get anything done with just that handful of words?" 02:43:24 he probably didn't use the word words. 02:43:55 have you seen much about go? 02:43:57 but I think what was awesome about CRT (and what is awesome about forth-83) is that the word list is very small. 02:44:02 yeah 02:44:11 it reminds me of the old tp. 02:44:13 a lot. 02:44:20 kinda fun in that way. 02:44:24 so retro runs on this vm called ngaro. 02:44:30 ngaro is awesome. 02:44:41 in fact... ngaro is far more popular than retro. 02:45:02 at least according to what crcx (who maintains both) said. 02:45:29 i have a theory about this... i think people come in and see the 30 instructions in ngaro, and they can keep it in their heads. 02:45:44 but it makes no sense to me. 02:45:59 when i think of forth, it reminds me of the ultimate stack-oriented VM. 02:46:06 why put a vm on a vm? 02:46:09 pretty much the first thing I did when I started messing with it: implemented a little file called vt.rx 02:46:29 exactly. ngaro is very much like forth. 02:46:51 well so why put the same vm on vm. it feels ... unfactored . 02:46:54 it's a forth that ships with 30 words, spelled as numbers. 02:46:57 and redundant. 02:47:02 hah. 02:47:04 interesting. 02:47:13 back to machine code huh? 02:47:28 the numbers are treated as 32-bit integers. 02:47:37 well :) 02:47:47 so... each opcode is a 32-bit integer. 02:47:53 and these are all equally expressible as binary. 02:48:08 i have to look into ngaro. 02:48:20 therefore.. there are 4294967296 opcodes in ngaro. 02:48:44 but what is opcode 31? 02:48:48 i don't know 02:48:56 but there is a neat way to find out 02:49:04 see 31 ? 02:49:05 and this is what ngaro does... :) 02:49:38 what does ngaro do? 02:49:38 yep. it treats opcode 31 as : GOSUB 31 02:49:42 i see. 02:49:51 so it pushes the IP to the return stack and jumps to 31 02:50:32 so to me... ngaro or something very much like it is a spec for implementing the core of forth in any language you want. 02:50:50 30 seems like a lot. hasn't it been done in 3? 02:50:51 or less? 02:51:04 RodgerTheGreat has done something similar with his Mako VM, which also has multiple implementations, but I haven't had a chance to really look into it yet. 02:51:06 sure 02:51:24 I think 16 is a good number, personally... 02:51:58 there's a guy who's working on a port of retro called metro who uses 16 02:52:08 A 3-instruction forth for embedded systems work. 02:52:11 it exists. 02:52:36 3 is probably the minimum because of the structured program theorem. 02:52:48 brainf*k is 6 instructions 02:52:57 I've seen a 1-instruction machine somewhere. 02:52:59 --- join: kbmaniac (~dave@86.157.24.246) joined #forth 02:53:38 in theory you could do everything with only NAND or NOR but you'd still need some kind of way to route the results 02:53:45 well that's not theory.. .that's what actually happens. 02:53:52 on actual computers :) 02:53:53 yeah. there exists a OISC. 02:54:00 a simple move machine. 02:54:38 i don't really believe taht it's one instruction though. if it is, then they're doing something very clever to handle the routing and just not calling that thing an instruction. 02:55:18