00:00:00 --- log: started forth/04.12.19 00:08:52 --- quit: Herkamire ("bed") 00:17:59 --- join: fridge (~fridge@dsl-220-253-77-81.NSW.netspace.net.au) joined #forth 00:28:21 --- join: todd (~todd@chcgil2-ar7-4-34-136-241.chcgil2.dsl-verizon.net) joined #forth 00:29:10 --- nick: todd -> SDO 00:51:26 --- quit: Sonarman ("leaving") 01:07:40 --- join: qFox (C00K13S@82-169-140-229-mx.xdsl.tiscali.nl) joined #forth 01:19:39 --- join: Topaz (jonny@spc1-horn1-6-0-cust217.cosh.broadband.ntl.com) joined #forth 01:27:37 wow, insane... the stupid FORTH i wrote ages ago still works ;) 01:35:42 --- quit: Topaz (Remote closed the connection) 02:18:52 --- quit: froop ("beware the colors of doom!") 03:40:54 --- quit: qFox (Read error: 104 (Connection reset by peer)) 04:43:35 --- join: Topaz (jonny@spc1-horn1-6-0-cust217.cosh.broadband.ntl.com) joined #forth 05:01:59 --- join: kuvos (C00K13S@82-169-140-229-mx.xdsl.tiscali.nl) joined #forth 05:02:02 --- quit: kuvos (Client Quit) 05:27:06 --- join: qFox (C00K13S@82-169-140-229-mx.xdsl.tiscali.nl) joined #forth 05:36:52 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 05:36:53 --- quit: tathi (Client Quit) 06:24:39 --- join: saon (1000@c-24-129-90-197.se.client2.attbi.com) joined #forth 06:30:35 --- quit: crc2 ("I'll") 06:49:25 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 06:49:58 Hi 06:52:53 hi 06:57:30 --- quit: tathi ("restarting...") 07:28:34 --- quit: SDO (Read error: 110 (Connection timed out)) 07:40:51 --- join: froop (~wossname@rn-v1w5a06.uwaterloo.ca) joined #forth 07:43:14 --- join: SDO (~todd@chcgil2-ar7-4-34-136-241.chcgil2.dsl-verizon.net) joined #forth 07:48:21 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 07:58:51 --- quit: tathi ("booting herkforth") 09:53:24 hi all 09:53:30 Hi, crc 09:53:36 --- join: T0paz (jonny@spc1-horn1-6-0-cust217.cosh.broadband.ntl.com) joined #forth 09:53:39 what's new? 09:53:54 --- quit: T0paz (Remote closed the connection) 09:56:55 Nothing at all. 09:57:17 no progress on your optomizing forth compiler? 09:57:31 I got tired of it 09:57:34 :) 09:57:36 * robert has little patience 09:57:42 Perhaps I'll continue later 09:58:04 * crc is exploring the idea of a hybrid block editor/interactive interpreter 09:58:32 It's pretty interesting :) 09:59:11 How does that work? 09:59:58 --- join: Herkamire (~jason@h000094d30ba2.ne.client2.attbi.com) joined #forth 10:00:09 It shows the block editor at the top, and an interpreter below it 10:00:28 Ah, OK 10:00:38 Block -1 is TIB, Block -2 is PAD in this model 10:00:41 QBasic style;) 10:00:47 yup 10:01:07 The editor is line oriented right now, which plays nicely with the interpreter 10:02:44 --- join: imaginator (~George@georgeps.dsl.xmission.com) joined #forth 10:02:46 negative block numbers.... interesting 10:03:02 block 0 = start of code space + $20000 10:03:09 So negative numbers are allowed 10:03:38 hmm, i need to hack on my forth some more 10:03:53 because of my immense weeniness, the optimising colon compiler is actually written in python ;0 10:04:05 Interesting 10:04:27 * crc has removed does> in the experimental codebase 10:10:40 You don't like it? 10:10:53 Topaz: :) 10:11:34 Nope 10:11:48 Why? 10:11:53 I've never used it except once, and that was to implement constant 10:13:12 How about other high-level constructs? 10:13:22 such as? 10:13:35 Structures, OO, etc. 10:13:46 I don't do structures, OO, etc 10:13:56 I have done things like: 10:14:09 create ^Acrlf 1, 13, 10 10:14:13 I meant if others want to implement it. 10:14:18 And then accessed the elements by hand 10:14:32 ^A ? 10:14:37 ACSCII 1 10:15:02 Needed in the ACTION part of PRIVMSG (irc client implementation) 10:15:41 Ah 10:16:30 my FORTH is a bit of a joke, it compiles to AVR assembly, with a built in emulator ;) 10:17:21 Joking is what the forth community is about, Topaz! 10:18:08 can I get a retroforth binary for x86 linux? 10:18:13 hmm, where's that article on implementing FORTHs? 10:19:05 Topaz: I don't suppose you could be a little more specific? 10:19:06 http://www.retroforth.org/dev/beta_releases/rf76-beta2-linux.tar.gz 10:19:06 brb, getting food 10:19:48 thanks 10:20:16 heh, sorry, there was just a 5-section (or so) article describing the internal operation of FORTH, along the same lines as brodie's 'starting forth', but discussing the implementation 10:21:50 Build your own Forth? 10:21:54 Moving Forth? 10:22:13 ah, hight well be moving forth 10:22:15 might* 10:22:18 I got links to both on robos.org 10:22:22 If google fails 10:22:39 ah, excellent, tis 'moving forth', thanks 10:30:17 --- join: thinfu (thin@bespin.org) joined #forth 10:30:59 Hi 10:31:58 hmm, so the runtime code associated with DOES> puts the address of the cfa on the stack before running the DOES> stuff 10:32:46 er, pfa, not cfa 10:35:17 Right. 10:35:19 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 10:36:39 so where is the actual code that does this?? 10:36:42 -? 10:37:56 * Topaz reads 10:37:59 it probably puts the address on the stack and then JUMPS to it.. that way code after DOES> is run 10:40:52 Topaz: what forth are you looking at? 10:41:18 well, i'm trying to work out how to do it in my own forth 10:41:29 (the usual case of trying to implement one before really understanding the language ;) 10:41:41 but mine is subroutine threaded, since it targets very teeny microcontrollers 10:41:49 --- quit: imaginator (".") 10:41:51 (so isn't really a proper FORTH, yadda yadda ;) 10:42:10 bah 10:43:17 it appears i'll just encode a subroutine-call in the cf, and then look back down the CPU stack to find the cfa (and thereby pushing the address of the pfa on the stack) 10:43:38 you doing does> ? 10:43:42 back 10:43:42 yeah 10:44:06 well i like STC the best of the threading models, although i don't see how targeting small microcontrollers results in choosing STC 10:44:07 i'm just not entirely sure how this pfa-address-acquiring code gets put there, does DOES> stick the same code at the beginning of every 'DO*' word, or is this done by EXECUTE? 10:44:19 thinfu: since it isn't interactive 10:44:26 (actually, ignore that) 10:44:27 pfa ? 10:44:35 pfa = parameter field address 10:44:37 parameter field address, whatever it's called 10:44:46 points to the actuall code after : blah ..... ; 10:45:01 does> puts the code inside cfa 10:45:09 and that code acts on pfa 10:45:11 afaik 10:45:19 pfa points to the data for a CREATEd word then? 10:45:23 hmm, i see your point ;) i'm doing STC because it's fastest actually, heh 10:45:54 What is the cell size? 10:46:33 16-bit addresses, but with a few 8-bit arithmetic words 10:46:41 (just to confuse matters) 10:47:02 you can't confuse us 10:47:06 we're hardened forthers 10:47:07 :P 10:47:12 so the items on my stack are variable in size... this is invariably going to bite me horribly at some point, but i'll see what happens 10:47:38 just so i can do trivial arithmetic without having to do slow 16-bit maths with the 8-bit CPU 10:48:48 (all this premature optimisation is probably bad for the health, but i'm being impatient) 10:49:40 My oldest AVR Forth was a mix between 32 and 8 bits 10:50:40 given that i usually write all my AVR stuff in straight assembly, i guess it makes sense for me to go to extreme lengths to make my FORTH fast ;) 10:50:49 (or i won't use it) 10:52:58 Make it generate video 10:59:19 --- quit: tathi ("booting herkforth") 10:59:22 lets not and say we did 11:00:19 Microsoft-style 11:02:00 you read that article kc5jta gave the url to? 11:02:55 I don't know if I got any URL from him lately 11:02:59 Which are you refering to? 11:07:55 http://www.euronet.nl/users/frankvw/rants/microsoft/IhateMS.html 11:16:17 have you read that? 11:16:53 * robert checks 11:17:32 oh boy, another very angry and hopefully humorous rant about microsoft 11:17:59 froop: nope its highly articulate 11:18:14 and makes many good succint points 11:18:37 kc5tja gave it two thumbs up.. i haven't finished reading it myself 11:18:40 i'll believe it when i read it :) 11:18:43 >~@ 11:19:30 froop LIKEZ MS 11:19:36 CUZ HE USED THERE PRODUCTZ 11:22:04 :D I LUV NOTEPAD 11:22:37 --- quit: onetom (Read error: 232 (Connection reset by peer)) 11:26:53 --- join: onetom (~tom@cab.bio.u-szeged.hu) joined #forth 11:39:53 --- quit: onetom (sterling.freenode.net irc.freenode.net) 11:40:47 --- join: onetom (~tom@cab.bio.u-szeged.hu) joined #forth 11:42:07 --- join: Raystm2 (~Ray@adsl-69-149-45-208.dsl.rcsntx.swbell.net) joined #forth 11:42:27 Hi 11:42:43 --- quit: Raystm2 (Client Quit) 11:43:10 --- join: Raystm2 (~Ray@adsl-69-149-45-208.dsl.rcsntx.swbell.net) joined #forth 11:43:24 --- quit: Raystm2 (Remote closed the connection) 12:16:03 --- join: bbls (~bbls@80.97.121.133) joined #forth 12:16:08 hello 12:16:10 anyone here? 12:16:15 Hi bbls 12:16:18 hi crc 12:16:24 i'm new to forth 12:16:28 cool 12:16:36 i've just opened a FAQ few minutes ago 12:16:40 forth is an awesome language :) 12:16:52 i'm impressed by the simplicity of the language 12:17:32 I am too, even after over six years of using it 12:17:58 anyway my main problem is trying to translate problems in postfix syntax 12:18:10 im used to normal infix of prefix 12:18:13 but not to postfix 12:18:33 that's the hardest part to learning forth 12:19:52 is there a way to implement IF-THEN-ELSE as forth words? 12:20:10 Most forths have if/then and often else 12:20:30 yes, but i mean can they be implemented/emulated as simple words? 12:20:34 1 2 = if ...do if true... else ...do if false... then 12:20:43 They can be, depending on the forth you're using 12:20:54 i'm speaking in the general case 12:20:59 how would you implement them? 12:21:11 In RetroForth, I do: 12:21:12 : <>if $063b 2, $adad 2, $74 1, here 0 1, ; 12:21:12 : =if $063b 2, $adad 2, $75 1, here 0 1, ; 12:21:12 : : >if $063b 2, $adad 2, $7d 1, here 0 1, ; 12:21:16 : ?if $72 1, here 0 1, ; 12:21:18 : then dup here swap - 1 - swap c! $90 1, ; 12:21:40 Which inlines the code for "cmp eax, [esi]" "lodsd" "lodsd" and a conditional jump 12:22:04 not asm 12:22:06 then patches the address of the jump to skip over the code if the condition isn't met 12:22:07 hmm 12:22:09 i mean in a high level way 12:22:28 ah, if and else programmed using high level constructs 12:22:32 In gforth: 12:22:33 : IF 12:22:33 (compile) ?branch >mark ; 12:22:38 yes, that's what i'm asking for 12:22:46 : ELSE 12:22:46 POSTPONE AHEAD 1 CS-ROLL POSTPONE THEN ; 12:22:47 crc no that's again low level 12:23:13 How can a conditional be implemented with no low-level code? 12:23:22 that's what i'm asking 12:23:30 for example you might insert a boolean into the stack 12:23:32 as a 0 or 1 12:23:48 before this you insert two other operands 12:24:02 and you need a word to choose between those two operands 12:24:32 At some point you have to use low-level code 12:24:39 no, without low-level 12:24:53 do you know of any languages where that's possible? 12:25:03 you need to use some conditional already implemented in order to define IF 12:25:06 let me think 12:25:10 but IF is one of the most basic conditionals.. 12:25:22 well, IF can be implemented as an AND operator in circuits 12:25:39 AND is low-level though 12:25:47 forth has and 12:25:54 hmm 12:25:59 : IF AND JNZ ; ? 12:26:04 let me think, there should be a way to do it.. 12:26:06 JNZ is low elve ltho 12:26:15 low level* 12:26:50 bbls, btw we don't do ELSE anymore 12:27:03 some people do... 12:27:05 IF foo ;; THEN bar 12:27:14 * crc doesn't and thin doesn't though 12:27:49 instead of IF foo ELSE bar THEN 12:29:12 how do i access an element in the stack? 12:29:27 what sort of element? 12:29:32 if its a number, its accessible 12:29:33 which element? 12:29:47 if its an address then you can fetch from it or store to it 12:30:10 for example something like 0 ELEMENT returning the top, 1 ELEMENT returning the one before the top, etc 12:30:21 PICK 12:30:23 Some forths have "pick" for that 12:30:40 * crc dislikes pick as it encourages bad programming habits 12:30:42 its not considered good forth programming practice to use PICK however 12:30:45 heh 12:30:49 it is possible to delete some elements from the stack (which are not in the top of the stack) 12:30:54 ? 12:31:02 1 2 nip 12:31:05 removes the 1 12:31:11 and 2? what's about 2 12:31:18 That gets left behind 12:31:26 1 2 3 4 5 12:31:28 goes onto the stack 12:31:29 nip 12:31:32 gets rid of 4 12:31:41 and how do i remove 3? 12:31:41 nip gets rid of 2nd item on stack 12:31:48 : nip swap drop ; 12:31:56 to get rid of 3 12:31:59 rot 12:32:02 drop 12:32:04 i think 12:32:11 1 2 3 4 5 rot drop 12:32:17 yup 12:32:34 1 2 3 4 5 rot = 1 2 4 5 3 12:32:39 is it possible to write a conditional swap without using IF or CASE? 12:32:50 1 2 true CSWAP 12:32:56 will result in 21 12:33:02 1 2 false CSWAP 12:33:10 should result in 1 2 12:33:16 is it possible? 12:33:20 then you could have 12:33:22 when you said 21 you meant 2 1 right? 12:33:35 : cswap 0; drop swap ; 12:33:40 (In RetroForth) 12:33:47 where 0 = false 12:33:57 : true 1 ; : false 0 ; 12:34:03 what drop does? 12:34:13 drop removes the item from the top of the stack 12:34:16 gets rid of the top element on the stack 12:34:18 ok then 12:34:22 1 constant true 12:34:26 0 constant false 12:34:52 "1 2 condition cswap drop" is equivalent to IF condition THEN return 2 ELSE return 1 12:34:54 * crc prefers that to : true 1 ; 12:35:10 hmm 12:35:11 : PSEUDO-IF cswap drop; 12:35:27 : CSWAP 0; drop swap ; 12:35:30 hmm 12:35:37 0; is defined with IF :P 12:35:40 this might work after all :) 12:35:53 As far as I know only RetroForth and HerkForth have 0; defined 12:36:22 ok, then we can use 0 for false and 1 for true 12:36:49 let me think 12:36:49 we need a portable cswap 12:37:02 : 0; 0 =if r> drop drop then ; 12:37:11 then you'd better use IF to define your cswap :P 12:37:12 should work on most forths 12:37:21 you mean 12:37:29 : 0; 0 = if r> drop drop then ; 12:37:36 yeah 12:37:43 : 0; 0 = if r> 2drop then ; 12:37:53 For forths other than RetroForth and HerkForth 12:41:30 isn't there a more "mathematical/logical" way, that is completely portable? 12:41:45 Nope 12:42:04 there is XOR and AND and OR 12:42:19 but you need JMP, JNZ JEZ CMP etc 12:42:22 Forth implementations vary too widely for any code to be truely portable 12:42:23 which are all assembler definitions 12:45:45 how does forth mixes asm with high level code? 12:46:25 That depends on the forth 12:46:49 In RetroForth, all code compiles down to assembly, so you can do things like: 12:46:55 : foo [ $ad 1, ] ; 12:47:16 Which makes foo compile a "lodsd" instruction into its definition 12:47:34 (drop is implemented like that) 12:47:51 Other Forths have a "CODE:" prefix and a complete assembler in them 12:48:09 CODE: drop lodsd, ;CODE 12:48:33 Again, it's very specific to the implementation you're using 12:48:38 bbls, perhaps you could tell us about your goals? you seem to be concerned about portability ? 12:48:57 no, i'm rather concerned about a mathematical view of forth 12:49:15 how so? 12:49:30 i'm not good on math anyway :), i do it just because i'm interested into proof of corectness 12:50:18 hmm 12:50:40 and i'm looking to see if i can use a forth-styled language as internal representation 12:50:48 a pure functional forh 12:50:51 *forth 12:50:59 yeah there are some functional forths 12:51:19 do you know any i can check? 12:52:29 there might be some listed at http://www.forthfreak.net/wiki on the ForthImplementations page 12:53:04 plus http://factor.sourceforge.net 12:54:29 factor isn't really a forth though 12:55:07 slava (the creator/developer of factor) can often be found on #concatenative 12:55:14 (he's not there right now though) 12:56:04 --- join: Sonarman (~snofs@adsl-66-124-255-101.dsl.snfc21.pacbell.net) joined #forth 12:56:30 hmm 12:56:38 but is functional forths really forth in the first place? 12:56:39 :P 12:56:55 not sure, i'm not looking to code in forth atm :) 12:57:21 i rather need a good functional internal representation for my language 12:57:30 or maybe use forth 12:57:42 but is forth an imperative programming language? 12:58:02 because 2 3 + 2 * is rather functional 12:58:09 result:= (2+3)*2 12:59:43 forth isn't an imperative programming language 13:00:00 then it must be functional :) 13:00:03 its not fully functional either.. 13:00:07 hmm 13:00:15 it falls into a unique category of its own 13:00:29 which some ppl call "concatenative" 13:00:43 i'm interested into functional programming languages because building a proof of corectness is much easier for them 13:01:13 but doesn't goedel refute all formal representations? ;) 13:01:27 who is goedel? 13:02:36 as i said, i'm not a mathematician :) i'm just a coder 13:02:52 i'm not interested into this for math purposes but for coding purposes :) 13:03:03 ok 13:03:10 so why do you need a proof of correctness? 13:03:14 i'm not sure how thats useful 13:04:07 well i want to code an OS where all programs will be proof-carrying, but for this you need a language&compiler able to produce proof of correctness :) 13:04:14 so i 13:04:26 i'm searching for a base language 13:04:32 hmm 13:04:44 unfortunatelly most languages are unsuitable 13:04:48 yeah i was interested in coding an OS and thats what led me to forth also 13:04:56 me too 13:05:12 pure functional languages are the path of minimum resistance when trying to do proof of correctness 13:05:54 Perhaps a subset of LISP would work better than Forth for proving correctness 13:06:04 the main requirement for a language in order to make it suitable for this, i think that is code size (espressed in semantic tokens not syntactical characters) 13:06:12 forth is a fairly easy language to implement other languages on top.. or to implement functionalness 13:07:02 yeah i guess might be better to go with a minimal subset of lisp 13:07:12 would be functional out of the box.. 13:07:24 that was my main direction of search 13:08:07 but i've decided to see if there are languages where programs can be written even more compact 13:08:53 well its certainly possible to code a functional forth.. 13:09:11 maybe since forth is not imperative, it might be used per se 13:15:33 hmm 13:15:42 maybe a forth without I/O would be functional 13:16:03 it would be very limited in useability though 13:16:09 where would the input come from? 13:16:16 initial non-empty stack 13:16:17 and where would the output go? 13:16:30 the output would be the remaining values in the stack 13:16:33 hmm 13:16:50 this could be used to do some CGI on HTTP servers 13:17:13 would it be able to read/write to/from memory locations though? 13:17:19 no 13:17:24 just do stack processing 13:17:29 then it'd be very limited 13:17:35 why? 13:17:38 no variables, no constants, etc 13:17:46 well, you can define words as constants 13:18:00 Only very simple ones 13:18:01 i mean constants as worsd 13:18:07 *words 13:18:11 Where does the code to define a word come from? 13:18:32 from the source code i suppose 13:18:36 hmm 13:19:00 * crc can't see a language with no access to memory outside the stack 13:19:08 the limits would be great 13:19:18 hmm 13:19:30 can you give me an example where memory access would be needed? 13:19:33 strings would be sequences of chars on the stack, and hard to manipulate in any real way 13:20:05 hmm, i tought more about using more complex objects in the stack 13:20:18 maybe arbitrary theoretical sets 13:20:27 that ventures away from forth 13:20:41 yes, i know 13:20:43 If I did: "foo" 1 + 13:20:46 what would happen? 13:20:55 Or better yet: "1" 1 + 13:21:20 if we speak about an interpreter then you would an error 13:21:33 what about a compiler/ 13:21:34 ? 13:21:48 hmm, a compiler might be impossible to write 13:22:07 hmm, unless you have a proof of correctness that would guarantee that the code is right 13:22:18 this would be slow to run I'd think 13:22:41 slava's factor language checks data types and return types to make sure the generated code does what it's supposed to 13:23:58 does it have some kind of flags for each element in the stack specifying type? 13:24:09 yes 13:26:47 sounds like what you're thinking of is a minimal forth that you launch multiple instances of etc 13:29:56 i.e. does stack stuff then returns the results.. 13:30:57 well, i have to avoid I/O completelly 13:31:09 since programs with I/O are quite hard to validate 13:31:11 I/O has to take place at some point 13:31:23 yes, but it would be done by the OS in my case 13:31:30 hmm 13:31:40 so what can an application actually do? 13:31:48 a) it can't get input interactively 13:31:57 basicly the OS would display a bunch of windows, each running a pseudo CGI program 13:31:58 b) it can't display results on it's own 13:32:18 using some form of HTML-derivated interface 13:32:29 so the system gets all input, puts it on the stack, and the program returns the results on the stack? 13:32:36 yes 13:32:38 * crc wonders what happens if all the input isn't used 13:33:33 well, some predefined value could be put there, or the HTML might have a tag specifying that all input fields must be completed. there are some other workarounds possible i think.. 13:33:59 for example a program might be made of several parts 13:34:45 as a state machine 13:35:01 the current state would be send each time as parameter 13:35:43 then you could also have a function that checks of input validity 13:36:04 so there are ways to avoid bad input data 13:36:54 I mean, you send the input to a function, and if the function(s) don't use all the input, and some gets left on the stack along with the output, what happens then? 13:37:54 oh, the function must remove any unused input 13:38:12 hmm 13:38:17 just like how a word works, it removes input and it places the output value in the stack 13:38:39 except that words can leave existing values on the stack and just add more 13:39:36 yes 13:39:46 but i think that you got the ideea 13:40:49 I do now 13:41:05 I just fail to see how it's superior to block all I/O from a program 13:41:29 i don't do that because it's superior in general :)) 13:41:42 it's beter for the proof builder 13:41:56 But worse for the programmers 13:42:02 since this is the goal, i have to cut from somewhere 13:42:21 Who can't have any variables, etc 13:43:22 Your programs would have to do a lot of stack juggling 13:43:52 hmm, yes 13:44:38 in forth, the stack isn't meant to be use like an array 13:44:45 They couldn't create structures of any sort either 13:46:53