00:00:00 --- log: started forth/16.09.13 00:18:29 holy channel, what the hell you all unleash here? :-D 00:19:04 do you guys use vim or emacs or other editor? 00:19:59 yunfan, other. :) 00:21:39 vi and drracket ) 00:27:18 i am a vimer , but i am now consider to try emacs way 00:28:01 the only problem is there're too many ) and ( in elisp 00:28:18 i was wondering if there is a emacs using forth instead of elisp 00:38:43 emacs is kinda the opposite of forth thinking don't you think? within forth I always made a little block editor. 00:39:18 you could write in in ans forth tho 00:39:20 hue hue hue 00:39:28 make it portable ans forth 2 00:39:57 hahaha just damn... :) 00:40:02 What if there was some way to abstract the mathematical implementation of a piece of software 00:40:16 and say resolutely what that abstraction was and if a piece of software fit that model 00:40:30 and we could share these abstractions with each other and implement them in our own ways 00:40:47 But not like OML which is bad 00:41:24 for example there could be a logical and mathematical definition of the http protocol which could be implemented in any forth 00:41:59 you could create mathematical definitions for domain specific languages as well 00:42:11 but it is up to the programmer to implement that model themselves 00:43:56 some models would not really be mathematical though 00:45:17 we could have repositories of these models and try to make them as logically and computationally efficient as possible. 00:45:48 You could have different "distributions" of different models each with customized mathematical formulae 00:46:21 For example the people who know calculus could probably come up with better mathematical models than people who do not. 00:47:45 This would allow for ultimate efficiency first of all because each forth user would write everything from scratch using their custom forth as their own completely custom and ultimate plaform to solve their problem 00:48:58 John[Lisbeth], this is the kind of thing that Dijkstra liked to try to accomplish. 00:53:46 are you saying Dijkstra wanted to find the most mathematically simple algorithms for things? 00:57:01 Is there someplace where I can find lots of computer algorithms like these? 00:57:04 ones that I can use for free? 00:57:27 --- quit: nal1 (Quit: WeeChat 1.4) 00:58:16 --- quit: proteusguy (Ping timeout: 260 seconds) 01:04:31 I saw some small text editor in forth on github 01:05:06 fk dijkstra, COBOL ftw 01:09:35 --- quit: mnemnion (Remote host closed the connection) 01:20:43 I think it could be done 01:20:56 a mathematical notation to share forth logic 01:21:25 always working to the bare minimum simplest fastest mathematical definition. Something easy to implement in any forth. 01:22:16 This "mathematical notation" exists already. 01:22:25 It is called Krivine machine. 01:22:38 r u that guy from earlier 01:22:46 also sure maybe it is the krivine machine but it will take me time to investigate that 01:22:58 * ASau shrugs. 01:23:59 One week in a library easily saves a year of experimentation. 01:25:36 is this the machine? https://www.irif.univ-paris-diderot.fr/~krivine/articles/lazymach.pdf 01:27:17 Yes, I think it is the "original" paper. 01:28:33 Ok first I need to understand the four rules 01:28:48 and then I need to implement it in assembler 01:29:47 is this the first one? (M N, S, E) -> (M, (S,(N,E)), E) ? 01:33:10 If you want to follow that way, you can find some other paper easier to follow. 01:33:18 E.g. ones by Biernacka et al. 01:33:40 Can I already get a kriving machine for my operating system? 01:33:45 or do I have to build it? 01:34:08 --- join: proteusguy (~proteusgu@183.88.71.215) joined #forth 01:34:08 --- mode: ChanServ set +v proteusguy 01:35:16 See, e.g., "From Interpreter to Compiler and Virtual Machine: A Functional Derivation" by Ager et al. 01:35:32 It has some step by step explanations. 01:37:42 how many people in the world know about and use krivine machines? 01:38:08 Fewer than use forth. hahahaha :) 01:39:20 O'Caml builds upon it, so pretty any person who wanted to know how it work on the lower levels, should be aware. 01:39:33 O'Caml? 01:39:38 You can estimate the number of O'Caml programmers yourself. 01:39:44 I thought O'Caml had a terrible type system 01:39:59 Ocaml is a great functional style language. I don't know many that are aware of Krivine machines. 01:40:41 But with a krivine machine I can just start producing whatever I want mathematically 01:40:57 I simply need to notate my math 01:41:06 in krivine 01:41:20 John[Lisbeth], yes but you can without as well. :) It's orthogonal to the point. 01:41:36 so once I have an empty krivine machine, what do I put in it? 01:43:38 I guess what I want to do is patch together mathematical functions 01:43:55 first of all I want to notate them in a way that is purely functional and where the math can be proven 01:44:06 and second of all that hte math can be shared 01:44:15 and also so that it is easy to connect functions together 01:44:35 And so then I will start putting math functions in there and trying to combine them together and make them more simple 01:44:39 and math will be my programming language 01:45:25 John[Lisbeth], if this is your real interest - executable algebraic expressions & proofs where the syntax more closely the written form - I suggest you try haskell. In haskell, every "program" is actually a proof utilizing an extremely expressive type system. 01:45:57 is haskell implemented in this krivine machine? 01:46:35 Don't believe so but it doesn't matter. Krivine machines are an implementation detail. The expressive potential is the same. 01:47:14 Somehow in my mind I believe a krivine machine to be as easy to use as a forth 01:47:34 John[Lisbeth], http://haskellbook.com/ and https://www.microsoft.com/en-us/research/people/simonpj/ and https://github.com/pushcx/hpffp-resources 01:48:31 haskell woudl have to be implemented in a krivine machine in assembly 01:48:41 or I would have to be able to modify haskells so that it was 01:48:53 and it would have to be the most minimal assembly krivine machine it could be 01:48:59 --- quit: ASau (Ping timeout: 250 seconds) 01:49:23 The last link is a repo of all the referred original papers from the book in the first link. Those papers are basically all the first order concepts behind functional programming and type systems that you're interested in. Serious formal maths stuff and quite powerful. More academic than practical, however, unless your program is properly constrained. Forth takes a far more practical approach rather than being "pure". 01:50:45 What syntax does a krivine machine us when you get into it? 01:50:55 an empty one 01:51:13 Why do you insist on the krivine machine? Whatever syntax you like. Your higher order language determines that. 01:51:34 I could be wrong but I have to explore the possibility since forth is sort of imperative and in my heart of hearts I Think functional is better 01:51:42 and krivine is four rules 01:52:49 forth is concatenative. You can make it functional or imperative as you like. It doesn't enforce any programming style but it probably better supports the functional style than imperative when used properly. 01:54:03 John[Lisbeth], I think if you like to be able to derive your entire system from first order lambda calculus principles you should definitely go straight to haskell. If, however, your interest is a practical system that doesn't contain anything it doesn't absolutely need and still inherently supports functional style constructs then give forth a shot. 01:54:28 Either direction you go - I'd recommend playing with it for 3-6 weeks and seeing how you take to it. 01:56:43 I don't think haskell is the solution I want 01:56:59 Although now I can see I can get alot out of haskell's user created functions 01:57:17 But it sounds like I need to implement my own krivine machine at least and then implement forth in the krivine machine 01:57:29 krivine is simpler to build in assembly than forth 01:57:38 and from krivine I can construct forth 01:57:45 perhaps even purely functional forth 01:57:57 and I can use monads for side effects 01:58:02 John[Lisbeth], I don't believe that to be a true statement but I doubt the order of complexity is so much that it matters either way. 01:58:29 I still have ot investigate krivines more 01:58:34 but I Think krivine and forth can live in harmony 01:58:49 perhaps I will just implement rpn in krivine 01:59:11 John[Lisbeth], honestly I think you've been sold a bill of goods but the only way to really know is to try it out. Either way you will learn something useful I expect. 01:59:34 I have to choose something purely functional over something not 01:59:51 If I can do it in four steps then that's great 02:00:43 and then I guess once I had rpn in krivine I would make some sort of math package manager 02:00:46 John[Lisbeth], I advise avoiding "pure" anything. Those things rarely break out of the realm of theoretical into the realm of useful. 02:01:15 I'll just have to test the speeds 02:01:24 but krivine allows me to create a language out of math 02:01:49 John[Lisbeth], krivine is in no way unique in that respect. 02:02:10 I'll just have to see the speeds 02:02:33 assembly -> krivine -> postfix -> math package manager 02:02:53 good luck 02:03:35 Once I've got postfix I've got the power of forth, right? 02:03:49 implementing postfix on krivine is minimal if krivine is straight assembler 02:03:58 no. you need a data stack and a call stack as well and pointfree syntax. 02:04:18 and a libked list 02:04:22 Perhaps I would have to implement those functionallyf irst 02:04:28 and then postfix 02:04:31 I'd have to play around with it 02:04:37 but once in krivine it is just playing with math nothing more 02:04:48 and call threading 02:05:15 krivine -> unknown -> postfix -> unknown -> math package manager 02:05:54 Postfix is all I want. I hated haskell because of how complex it had become 02:06:12 composition using words and atoms is key 02:06:50 John[Lisbeth], then you're more interested in concatenative rather than functional. 02:07:04 purely functional concatenative 02:07:57 concatenative is the best syntax I could hope for 02:08:07 Perhaps one rivals it but I don't know what it is 02:08:08 it's not missing anything. 02:08:19 whats not missing anything? 02:08:27 concatenative syntax. 02:08:44 it's absolutely expressive and flexible and simple. 02:08:57 Are you against me implmeneting postfix in a krivine machine? 02:09:04 nope. 02:09:13 do you think it is foolish? 02:09:42 John[Lisbeth], depends on your goals. At the very least it's likely a worthwhile learning experience. 02:10:19 reverse polish notation takes lisp notation and makes it way more compact and efficient 02:10:19 If your goal is to implement a forth then yes - it's silly and actually probably antithetical to forth thinking. 02:10:41 --- quit: nighty (Quit: Disappears in a puff of smoke) 02:10:44 "forth" thinking is just as arbitrary as "goldfish thinking" 02:10:58 If you wanna implement a forth-like concatenative language with other properties that you believe a krivine machine offers then it's definitely worth checking out. 02:11:17 John[Lisbeth], no it isn't. It's actually quite strict. 02:11:36 Technically speaking I am a nihilist 02:12:01 Every assertion in my mind I recognize as having a probability of being true or false. Nothing I hold to be absolutely true or false simply probability. 02:12:10 John[Lisbeth], FWIW I almost NEVER code in forth. I'm always building little forth-like languages with interesting properties. I just find the pointfree, concatenative syntax a great starting point. 02:12:26 I do too 02:12:52 So I have nothing against doing something un-forth-like. Just be clear as to what your goals are and make sure your decisions are aligned with those. 02:13:53 Just understand - as soon as you start introducing anything that isn't absolutely necessary in terms of complexity then you've left forth thinking and abandoned much of its value. 02:14:11 I stand by krivine 02:14:22 It's that discipline which is required to succeed with forth and languages that share its foundation. 02:14:38 the foundation of forth is make it yourself 02:14:48 each person's forth is their own 02:14:52 John[Lisbeth], as long as krivine machines are giving you something that you can't do better or live without then you're on the right track. 02:15:23 John[Lisbeth], that's not the foundation, that's just generally a common conclusion. :) 02:15:38 krivine -> math package manager -> unknown 02:15:45 go for it. 02:15:53 stop chatting and start coding! ;-) 02:16:09 I am too busy thinking to code right now 02:19:57 Don't think too far ahead when you're still playing with first order principles. Get something tiny working and then test your theories on top of it. Will constrain your thinking to real options. If you hit a brick wall then step back and try again. 02:20:59 Besides small progress points will encourage you to proceed. If you make a big design up front it will quickly become so intimidating you will never see it happen. 02:21:16 So for since I discovered forth I've been breaking all the walls 02:21:31 I need to run and run until I smack into something and get a trapped in a wrong assumption 02:21:36 and then break my way out again 02:21:52 It's a process that has not failed me yet 02:23:30 good luck 02:24:09 if you don't want to talk that's fine 02:24:25 I learn well by asking questions of people and bouncing my ideas off of them 02:24:28 haha I didn't say that. I thought you were thinking. 02:25:21 John[Lisbeth], what kind of development have you done before? Curious what kind of project history you have so I understand your context. 02:25:40 Well let me just tell you about me since this is such a small audience 02:26:14 My name is John M. Beck and I live in Washington near Seattle. I moved up here 16 years ago from Texas because my dad got a job at Microsoft. 02:26:47 I've been homeschooled my whole life. I dropped out of college twice because I just couldn't get interested in my course work. Especially the second time because I was so enveloped in my computer work. 02:27:24 My dad gave me lots of electronics from when I was young and groomed me to be a programmer. Part of my schooling was to play kids games on the computer which were not only learning but designed to teach mouse interfaces 02:28:02 My first computer of my own was a windows xp computer and I spent the better part of a year just exploring the menus in this system and autistically clicking buttons just poking to see what they would do. 02:29:47 When I was 16 I was using a Windows computer and I had set automatic updates to turn off. Despite this windows restarted proudly announcing "restarting for updates" causing me to get a bad grade on assignment I had procrastinated on. I decided that if I was truly the master of my own computer it would not restart without my permission, and so I installed ubuntu 10.04 02:29:59 --- join: true-grue (~true-grue@176.14.222.10) joined #forth 02:30:29 I spent 2 years distro hopping and then I learned C++. I hated c++ so I learned bash. The people in #bash were so upset at me they told me to leave and talk to the people in #haskell or they would ban me. 02:31:03 When using haskell I realized I wanted to abstract all the syntax into polish notation so I wouldn't have to think hard to parse. That lead me to lisp for a few months 02:31:25 in that time I learned emacs and I am talking to you through an emacs prompt in tmux on a freebsd virtual machine in microsoft azure 02:31:46 I on a whim decided to look up reverse polish notation and so how much like lisp it was and how low resources it was. 02:31:56 So what systems have you actually written? What apps? Or just experiment with languages thus far? 02:32:13 THe only real project I have ever written was a dynamic finite state machine in javascript. It is pretty well documented. 02:32:41 I made the mistake of making it GPL II and not LGBL and so I mourne the fact that I can never program javascript professionally with it. 02:32:43 that's a tool, not an app. :) so I'd say experimented. 02:33:09 I could have stayed in c++ my first language but I am glad I made it to forth 02:33:17 It's your code. You can re-license it however you like. 02:33:31 I have to contact a lawyer before I do 02:33:41 but it's not very good anyway I would be able to re-implement it in forth now 02:33:45 John[Lisbeth], no you don't. You change the license and say "here". 02:33:54 it's irrelavent because javascript forth 02:33:56 I have to run now I'm afraid. 02:34:03 ttyl 02:34:03 l8er 02:34:45 lol maybe f# really was a good solution then and I was just dumb 02:38:34 --- quit: proteusguy (Ping timeout: 240 seconds) 02:45:56 http://www.kevinalbrecht.com/code/joy-mirror/jp-joyjoy.html ooh shiny 02:46:10 I wonder what it would be like to be a script kiddie with mathematical equations 02:47:39 The hard part is getting it optimized for my hardware which is doable in postfix. But here lies the tricky tricky question. Do I implement postfix and then optimize or do I optimize and then implement postfix. Either way I am implementing the math package manager first 02:48:46 I think the math package manager should be mainly http 02:50:04 well nah first I've got to have something that takes data and incorperates that data into the language 02:51:13 that data would have to take some format 02:51:26 in that format I can implement the optimizations for my hardware 02:51:52 the implementations of the optimizations create somethign approximal to a joy style postfix 02:52:22 Once I have postfix I have almost everything I need to feel complete in a language 02:58:02 assembly -> krivine -> strict ans perhaps 03:03:06 assembly -> krivine -> solution to problem being solved 03:04:13 Ah here is my mistake. assembly -> krivine -> repl -> solution to problem 03:05:26 assembly -> krivine -> postfix -> repl -> solution 03:06:27 this satisfies me providing krivine is the best solution to kernels and postfix is the best solution to mathematical notation. 03:06:57 but for postfix I need postfix components 03:07:36 assembly -> krivine -> postfix components -> repl -> solution 03:08:20 assembly -> krivine -> unknown -> postfix -> repl -> solution 03:12:13 the assembly should optimize krivine for my hardware 03:12:35 krivine should directly implement postfix 03:15:23 assembly to produce smallest krivine. Krivine to produce most minimal purely functional postfix. Most minimal purely functional postfix written mostly in it's self and optomized for my hardware. postfix accepts mathematical equations 03:15:35 math as a universal language can be shared into postfix 03:19:09 --- join: nighty (~nighty@s229123.ppp.asahi-net.or.jp) joined #forth 03:25:44 from postfix I can actually implement math problems in a sane way 03:25:53 and from there it's just all math. The open world of math is my package manager 03:27:31 most minimal krivine in assembly -> most minimal joy in krivine -> math including the maths to optomize my system 03:31:07 --- join: mnemnion (~mnemnion@2601:643:8102:7c95:b815:6961:102b:7f7) joined #forth 03:35:32 --- quit: mnemnion (Ping timeout: 250 seconds) 03:40:34 The number of people who know about Krivine machines is definitely large enough. 03:40:51 It'll solve my needs 03:40:56 KMs are to O'Caml what SECD to Lisp and WAM to Prolog. 03:41:31 If a person is not interested in internals of his programming language, he may not know about it. 03:41:51 If he is, he knows at least the term. 03:41:59 ...And CAM to Haskell. 03:42:35 ur back hooray 03:42:43 ok so 03:43:09 I need to get into a krivine machine 03:43:14 You write too much text that mostly follows Forth propaganda. 03:43:26 Propaganda and reality is not the same. 03:43:28 --- join: proteusguy (~proteusgu@14.207.40.215) joined #forth 03:43:28 --- mode: ChanServ set +v proteusguy 03:43:40 I need access to a krivine machine 03:43:47 do you know where I can get it? 03:44:11 otherwise I've got to learn some math 03:44:15 O'Caml should still have backend based on results of ZINC. 03:44:48 "Some math" you speak of is mostly algorithmic side. 03:45:00 Like you need to know what it is and how to implement it. 03:45:08 You don't need foundations behind it. 03:45:44 It's hard to find a good explanation of the algorithm in plain english or on notation I understand 03:47:04 If it helps you, you can start from SECD machines. 03:47:22 IIRC, there was a person who decided to do what you want. 03:47:39 He has implemented several versions of SECD machines and published the code. 03:48:16 On DSLs. 03:48:28 Forth is very bad choice to implement DSLs. 03:49:10 If you need demonstration, check how many useful DSLs are implemented in Common Lisp, Racket, JavaScript, Haskell, and how many of them are implemented in Forth. 03:50:05 Next. 03:50:22 You don't seem to understand why Forth has postfix syntax. 03:51:12 DGASAU, quantity has nothing to do with appropriateness. Forth is an outstanding language for building DSLs. ANY isomorphic concatenative language excels as DSLs. 03:51:14 Which is important part of understanding Forth technology. 03:51:51 proteusguy: it has everything to do with appropriateness. 03:52:07 There's a fundamental reason why algebra uses infix notation. 03:52:58 you know what... I'm pretty sure when it comes down to it, C was a bad choice for writing an OS 03:53:06 yet, it seems to have survived for four decades in that role 03:53:08 DGASAU, perhaps it has something to do with popularity of the language and not appropriateness? Otherwise everything would be written in java, right? 03:53:28 DocPlatypus: C _was_ a good choice for writing an OS. 03:53:35 DocPlatypus: _for_its_time_ 03:53:38 infix notation has its flaws 03:53:58 What is the fundamental reason that algebra uses infix notation? 03:53:59 postfix notation is different and alien... until you get used to it. 03:54:57 proteusguy: The most important operations are commutative and associative. 03:55:30 technically yes, though subtraction and division are important as well 03:55:57 You get those as inverse of addition and multiplication. 03:55:57 and technically they are just the inverses of addition and multiplication respectively 03:56:27 however... get into expressions where parentheses have to be nested multiple times, and this is where postfix notation starts to look better 03:56:44 DGASAU, still waiting to hear your claimed advantage. Are you saying it's easier to re-work or simplify the expression? 03:56:45 2 2 + looks like an odd way to add two and two... 03:57:09 proteusguy: yes, it is a lot easier to operate with expressions written that way. 03:57:25 until you're doing something like (-(b^2)Â±sqrt(b^2-4*a*c))/(2*a) 03:57:27 DocPlatypus, that's just symptomatic of instruction. Not inherent in the syntax itself. 03:57:28 DocPlatypus: technically, postfix notation is prefix notation written backwards. 03:58:43 DGASAU, so why are scientific calculators RPN? We do lots of algebra and calculus there. I actually find breaking up algorithms in to RPN as easy or easier than infix. 03:58:44 You can get accustomed to that style, but I'm afraid that verb-object order is a little bit more widespread. 03:59:21 okay, let's look at infix notation first 03:59:25 2+2 03:59:32 DGASAU, so you're claim is that it's culturally superior and not inherently superior in terms of usefulness? 03:59:38 take two, get ready to add something to it, which is also two 03:59:42 2 2 + 03:59:47 (postfix) 03:59:53 take two, take another two, and add them 04:00:10 or... in Forth terms... 04:00:26 proteusguy: for the same reasons as Forth exists: broken or limited platform. 04:00:35 push a two on the stack, push a two on the stack, take the top two items off the stack and add them, putting the result back on the stack 04:00:47 DocPlatypus: you're following Krivine machine's operation. :) 04:01:14 RPN is easier to follow than many infix expressions with parentheses 04:01:18 especially if properly commented 04:01:37 and it beats the hell out of using a bunch of different variables to store the intermediate results just for readability 04:01:59 Can you rewrite at least your (-(b^2)Â±sqrt(b^2-4*a*c))/(2*a) in RPN? 04:02:07 You have 5 seconds to do that. 04:02:36 get real, it's going to take a bit longer than 5 seconds 04:02:42 Why? 04:02:53 DGASAU, that's an interesting claim but you've yet to provide any evidence that infix is better than postfix for algorithmic expressions. I think it's more culturally intuitive, but as a useful tool both for human thought and computer implementation I've found postfix to be more practical and automatically leads you to more more normalized structure. infix leaves too many bad options that it takes a lot of experience to figure out the right ones. 04:03:08 working with anything that complicated is just going to take a bit longer, maybe a couple minutes 04:03:38 DocPlatypus: so, you admit that you have to _work_ on it to convert to RPN. 04:04:01 let me make sure I even have that formula right 04:04:11 it has been a while since I've worked with it 04:04:23 Did I ask to write correct formula?? 04:04:33 I asked to rewrite what you have written. :) 04:04:35 DGASAU, as a short hand notation infix is quite convenient. But no one actually works out the problem like that on one line. Those parenthesis are there because of flaws in infix (and prefix). They are not a feature. 04:04:50 proteusguy: bullshit. 04:05:07 A lot of people worked and continue to work in infix on larger problems. 04:05:20 DGASAU, I see you're full of opinions and baseless assertions but devoid of evidence. You used to be better than that. 04:05:27 DGASAU: no, proteusguy is spot on. in a complex expression, parentheses are a flaw of infix 04:05:45 and I got it wrong... it's -b not -b^2 04:05:45 anyway 04:05:48 DocPlatypus: you don't need parentheses for the most important operations. 04:06:07 (assume a, b, and c are Forth locals in floating point) 04:06:27 DGASAU, by time you've eliminated the parenthesis you've simply created an inefficient version of the postfix model. :) 04:07:00 It is not "by the time". 04:07:12 You always have it from the very beginning. 04:07:25 b fnegate b f**2 4e0 a c f* f* f- fsqrt f+ 2e0 a f* f/ 04:07:33 b fnegate b f**2 4e0 a c f* f* f- fsqrt f- 2e0 a f* f/ 04:07:37 there 04:07:44 about a minute and a half, not bad I guess 04:08:01 if you really want the wrong version I wrote from memory just throw in f**2 after b in both 04:08:06 Still not seconds that you spent to write original expression. 04:08:16 DGASAU: so fucking what? 04:08:21 So fucking that! 04:08:33 That is clear demonstration that RPN is a lot inferior. 04:08:44 who really gives a shit that it takes longer than 5 seconds? 04:08:47 To any person who isn't brainwashed. 04:09:12 What happens with expressions that take few minutes to write in infix? 04:09:13 okay, how much is 2+3*4/5+6-7? 04:09:25 DocPlatypus, if you were putting this expression into an RPN calculator would you really do it in one long line? 04:09:30 you have one chance to guess the right answer. 04:09:39 proteusguy: maybe 04:10:04 $ awk 'BEGIN {print 2+3*4/5+6-7}' < /dev/null 04:10:04 3.4 04:10:09 DocPlatypus, haha I wouldn't. I would simplify those terms and construct the expression as a combination of simpler terms. 04:10:22 DGASAU: wrong. 04:10:26 DocPlatypus: bullshit. 04:10:45 Let's go out to #math and ask there. 04:10:54 it's 3 04:11:00 because my calculator said so :-) 04:11:19 answer's 3.4 :) 04:11:27 I don't care what #math says 04:11:38 haha you got one of them integer only calculators? :P 04:11:52 now... on the other hand... 04:11:59 That's the point: you're brainwashed fanatic who rejects any opinion not aligned with yours. :) 04:12:06 2 3 4 5 6 7 - + / * + 04:12:16 always evaluates to the same result. 04:12:25 DGASAU, pot, meet kettle - black. 04:13:29 I probably put those in an entirely different order but I think the point is clear. 04:14:04 "I probably put those in an entirely different order" indicates that you are not even sure, if you've got your notation correct in the first place. :D 04:14:24 it's an example 04:14:31 Yes, it is. 04:15:06 If you can't deal with an example, how can we be sure that you can deal with real problem? 04:15:17 the point is clear, RPN is unambiguous and there's no squabbling over order of operations when the parentheses aren't there 04:15:42 DGASAU: I actually write Forth code. when I'm writing a real program I get it right 04:15:48 when I'm on IRC who really gives a shit? 04:16:20 DocPlatypus: again, if you can't deal with an example, we can't be sure that you can deal with real problem. 04:16:55 I can deal with the real problem. 04:17:13 The example above indicates that you can't. 04:17:28 it was an example to prove a point about RPN versus infix 04:17:30 If you don't care with quality of your arguments, why are you entering disputes then? 04:17:59 Yes, and it demonstrated that you can't even get your RPN correct. 04:18:03 okay... the RPN version of the example above, works out to be 5 04:18:30 The notation must be so much superior that you can't get it correct in the first place! 04:18:52 or... the infix equivalent of 2+(3*(4/(5+(6-7)))) 04:19:00 DGASAU, back in my school days I could work things out in RPN much faster than infix. 04:19:06 which is enough to cross the eyes of some people reading code 04:19:46 proteusguy: back in my school days I could evaluate really complex expressions with 2 digits of precision a lot faster than anyone using calculator. So what? 04:19:50 formulas always look terrible in 1 dimensional form anyway 04:19:53 especially matrix stuff 04:20:04 you would arrive at exactly 3 by: 2 3 + 4 * 5 / 6 + 7 - 04:20:07 there 04:20:08 But I used table of decimal logarithms for it. 04:20:34 nha_: agreed. one reason I regret never learning TeX notation 04:20:34 it speaks to your absolutely baseless claim that infix is superior to postfix 04:21:06 how many times has DGASAU been kicked out of this channel anyway? 04:21:29 I mean, I have not followed it continuously but I know it's been at least a few. 04:21:39 none by me. in fact he's still here only cause I asked that he not be banned. 04:21:46 * proteusguy sometimes regrets. haha 04:21:49 DocPlatypus: ten years ago you'd be kicked faster than me, but since the generation changed... 04:22:08 nha_: linear algebra is very good example how RPN is so much superior to infix notation. :) 04:22:16 I don't think I started hanging here until 2008 or so 04:22:50 DGASAU, um... no. in fact you are trolling. you're in a forth channel (the ONLY channel you're on btw) and knocking the fundamental notation. It's like going into haskell and saying types suck! Please stop. 04:22:51 I think sane people mostly left by then already. 04:23:00 DGASAU, you're welcome to go. 04:24:01 Yeah, I'll probably stop for now. 04:24:02 DGASAU: maybe they got tired of you and left? 04:24:26 Sectants' arguments have mostly ended with "get out of our lawn." 04:25:15 DocPlatypus: some have openly told that they got tired of Forth, actually, but you may think whatever you wish. :) 04:25:44 whats the single worst thing about forth? 04:26:19 nha_: to me, that most of the better and feature-complete implementations are proprietary 04:26:50 specifically, that GNU Forth (Gforth) is still missing a fair amount of stuff that I'd want to use 04:27:08 * proteusguy did walk into the #haskell channel when first learning and asked why they call their list structure a list when, in fact, it behaves much more like a stack. :) Was a firestorm for two days. I think we're a lot more open here. 04:27:23 lol 04:27:30 you guys 04:27:34 you are both fucking wrong as shit 04:27:44 for C, maybe i'd say its because it is too easy to accidentally hit undefined behavior 04:27:47 k-machine is the most minimal machine it is just four functions 04:27:53 does everything you need 04:27:54 however rpn is better 04:28:01 or that it claims to be low level but it's not really low level enough when you need control over the stack or tail calls or something 04:28:06 and you should implement rpn in a k-machine if you understand rpn 04:28:13 nha_: I would agree. it is way, way too easy to accidentally leave a gaping security hole in a C program 04:28:35 rpn is superior but forth is inferior 04:28:43 forth is just one version of rpn and in a sense any rpn is like a forth 04:28:49 it's a bit more difficult in Forth but still easier than I would like 04:28:52 John[Lisbeth], there is, in fact, a 3 instruction forth. :) 04:29:01 * proteusguy drops mic. 04:29:17 talk to me about mathematical functions here 04:29:20 I still want to write my own Forth, though... just to say I've done it before 04:29:32 but I have not been in much of a coding mood the last few months 04:29:50 what they are saying is rpn is superior and it clearly is 04:29:56 once you get good at it you can copy math about straight over 04:30:02 the problem is on a GNU/Linux the outer shell still needs to be written in something like C or assembler 04:30:05 and you can compile it in your head 04:30:18 but 04:30:24 John[Lisbeth], describe for me a COMPUTER instruction that isn't a mathematical function. Not entirely clear what your definition of math is. 04:30:24 forth is not the best environment 04:30:35 I am a nihilist math is arbitrary 04:30:39 it is probably not though 04:31:02