00:00:00 --- log: started forth/05.11.24 03:17:20 --- join: nyingen (n=jeeves@216-39-182-234.ip.theriver.com) joined #forth 03:18:32 --- quit: saon ("leaving") 03:18:59 --- join: saon (i=1000@unaffiliated/saon) joined #forth 04:15:32 --- join: akebono (i=fujiwara@218.231.249.169.eo.eaccess.ne.jp) joined #forth 04:15:32 Quote: Let us live for the beauty of our own reality. -- Charles Lamb 04:26:31 --- join: Jim7J1AJH (n=jim@221x115x224x2.ap221.ftth.ucom.ne.jp) joined #forth 04:28:21 anyone in here use colorforth? 04:29:10 i haven't 04:30:32 Interesting system, but I'm having some trouble figuring out exactly how to use it... 04:30:51 Not so well documented = ) 04:31:58 I'm a novice Forth user, I believe some other experts in this channel will answer you much better. 04:32:30 afaik, ColorForth is for Chuck's own chip. 04:33:01 Yeah, but the idea of a self-contained Forth running on the bare hardware is strangely compelling 04:33:30 But yeah, I'll check back later. It would be cool if some friendlier docs for colorforth were around 04:34:01 Jeff Fox's page may have some. 04:35:04 hmm, got a URL for that? 04:35:12 the link I found via google is a 404 04:35:32 http://www.ultratechnology.com/ 04:35:43 it should work. 04:35:51 thanks! 04:42:32 --- join: PoppaVic (n=pete@0-1pool47-140.nas30.chicago4.il.us.da.qwest.net) joined #forth 04:44:57 hi, poppavic 04:45:19 howdy 04:45:56 ohayoo = ) 04:46:03 as a novice forther in Tokyo, can I ask you to look at my five or six line forth codes? 04:46:11 ohayoo. 04:46:22 hmm, sure - I guess. 04:46:33 it's for gcd. 04:46:34 ..admittedly, the coffee is not working yet. 04:46:39 hehe 04:46:55 hmm. oh, good - I can take a pass then ;-) 04:46:56 I can't send you a cop. 04:47:02 cup. 04:47:21 s'ok, this is just the first cuppa 04:47:55 GCD: greatest-common-denominator? 04:48:04 : gcd ( a b -- g) 04:48:04 2dup < if swap then 04:48:04 2dupm 04:48:04 begin dup 0> while 04:48:04 >r swapd r> 04:48:05 2dupm 04:48:07 repeat 04:48:09 drop swapd ; 04:48:19 : swapd ( a b -- b a) swap drop ; 04:48:19 : 2dupm ( a b -- a b m) 2dup mod ; 04:48:47 the latter 2 lines should come first, I failed to paste the lines from my cygwin console. 04:49:07 how can I make it better? 04:49:38 hold on. I'm still getingboth eyes to read it 04:50:44 OK... what are you feeding it? 04:51:33 53 21 gcd should put 1 on the stack. 04:51:41 --- quit: aum () 04:52:20 321 123 gcd will produce 3 instead. 04:52:32 it's just a greater common demoninator routine. 04:52:37 hmm.. complete packages of 'b' within 'a'? 04:53:06 yeah, that'swhat I thjought... Yer just looking for the largest whole insert of lesser into greater 04:53:26 well, it's workin' 04:54:26 yes, I think it works. 04:54:40 but I'm not sure it's a good or bad forth program. 04:55:00 can it be smater> 04:55:01 ? 04:55:07 smarter 04:55:13 I'm not sure.. I have to reorganize my head for it. Not awake enough. 04:55:23 hahaha 04:55:33 it looks pretty tight to me, but I'm really new at forth ;p 04:55:41 10 30 gcd . 10 ok 04:55:43 it's four or five inthe us. 04:55:55 I used to do HP calculator RPN a lot, so I have some knowledge of postfix stuff 04:56:07 8am here - it's also dark as hell and 15 degrees w/o windchill. 04:56:19 ah New York you are. 04:56:21 6 am here...slated to be a balmy 78 degrees 04:56:29 shaddap! ;-) 04:56:40 oh, I wish it were cold...eternal summer gets old fast 04:56:58 well, A/C is a lovely thing - and I am NOT running out to snowblow 04:56:58 hp's calculater has limited stack size... 04:57:08 true 04:57:18 but it got me into Forth anyway 04:57:37 anyway, I'm glad to know my codes were terribly bad. 04:57:37 I dunno if I like the 2dupm word 04:57:55 how could you write it? 04:58:14 well, it's not very "reusable" 04:58:41 Ah, I see. Elements that are not reusable should not be factored. 04:58:46 you mean.... 04:58:50 I understood. 04:59:01 Yeah, generally it's a good idea 04:59:07 I see. 04:59:28 I'm still glaring at it, trying to recall how I used to manually determine this stuff. 04:59:57 I have another toy. 05:00:02 2 9 gcd == 1, and that bugs me. 05:00:40 okay, so the first swapping part should be removed. 05:00:55 iirc, I used to compare the denom and if I couldn't get a nice division, I'd double everything and use those values. 05:01:24 ..I keep thinking of measurements and tools.. 05:01:40 how about this? 05:01:50 : lcm ( a b -- l) 05:01:50 2dup * -rot gcd / ; 05:01:58 if I have 1/8" and 1/4" - I'd want to get them in the same scale. 05:03:17 ok.. lemme cogitate a mo... 05:03:48 which forth implementations are you using? 05:03:54 gforth for me 05:03:56 gforth now. 05:04:04 yeah, that's what I've got 05:04:08 just checking to see = ) 05:04:21 except for the latest version not installing well for me, it works pretty well 05:04:36 is there any easy way to call C libraries? 05:04:44 Not to interrupt lcm/gcd discussion... 05:04:46 from forth? 05:04:48 yeah 05:05:06 well, the latest gforth uses calls to do so, but I can't get it to build properly. 05:05:21 otoh, FICL is written in C and runs a forth. 05:05:24 In particular I want to use ncurses 05:05:27 and there is pfe 05:05:42 I was looking at pfe, but had trouble wrapping my head around it 05:06:11 so I'm totally open to recommendations, if any 05:06:15 yeah, PFE is a complex, complete lil' bastard - and I can't recall, but I seem to recall the docs were not all that friendly. 05:06:49 Of late, I'dhave to recommend FICL. Written in C to eval and use forth. 05:06:59 I'll check that out 05:07:10 gforth's manual has a section "Integrating Gforth into Cprograms". 05:07:16 I like forth a lot, but have had a little trouble figuring how how I could use it in my work 05:07:19 you can check it at google. 05:07:22 I've not had luck at all using/understanding FFI tools. 05:07:29 akebono: I'll check that 05:07:52 Right now I'm playing with Colorforth, which is a very weird and cool idea = ) 05:07:54 akebono: yeah - the newest gforth incorporates FFI, but I can't get doubles to work properly. 05:08:38 2-cell numbers are hard for use for novice users like me. 05:08:42 ;-0 05:08:47 well, they get used internally 05:09:10 but, you can also think of them as one-shot single-number pairs. 05:09:35 one better example is rational numbers. 05:09:45 brb.. nature is screaming... 05:10:13 and rational numbers easily require multiple precisions. 05:11:31 so, I need to write a set of operators handling multiple-length precision numbers. 05:11:46 to use rational numbers. ;-) 05:12:35 I hope Forth programming skill will be improved enough to write them soon. 05:12:45 my forth programming skill. 05:12:53 ;-) 05:18:21 oy.. geezus, dats better 05:22:21 using 100 or 1000 as the base. 05:22:34 hmm? 05:23:24 to implement bignum 05:23:30 ugh 05:23:49 I wanted to use rational numbers. 05:24:06 but I found I needed routines to handle bignum. 05:24:15 I've never had a need for such 05:24:54 but it costs too much. 05:25:06 yep. 05:25:40 If I needed huge numbers, I'd just suffer FP 05:25:50 ha. 05:26:16 but rational numbers have good things. 05:26:31 I can't even recall what a "rational number" is. 05:26:37 when one or two word numbers can suffice for the needed precision. 05:26:51 oh, yer talking fractions 05:26:58 123 1 = 123 05:27:03 yeah 05:27:17 somewhere, I recall seeing source for that sorta' thing 05:27:56 a bad thing is that each operation often needs reduction. 05:27:57 iirc, it was an old task for forthers 05:28:02 yep 05:28:14 rescaling, anyway 05:28:18 this requires lcm 05:28:23 which needs gcd. 05:28:42 it costs much even in single precision. 05:28:46 sure. It's not trivial 05:29:41 but only real numbers and irrational numbers can not be expressed by rational numbers. :-) 05:29:54 and, of course, this is why I prefer to think of forth/C/asm 05:31:13 I'm just a novice learner of Forth, but I think it's very different from C. 05:31:25 it is, in many ways 05:31:37 a different interpretation of the machine language. 05:31:42 otoh, I never liked Forth pretending to *BE* the OS 05:31:51 no 05:32:07 is it a minimalist approach to Forth? 05:32:13 whom? 05:32:43 forth as self-sufficient operating system. 05:32:51 You have to first recall there is no "Forth" - there is FIG-Forth, ANS-forth, Gforth, etc. 05:33:26 and there's no one single cpu. but there are many different machines. 05:33:33 akebono: well, FIG used to lay atop bios-calls, and most forths still think they are the end-all. Otherwise, the idea of FFI would be older and more widespread. 05:33:59 FFI? 05:34:02 sure. and, forthish "compiled code" doesn't usually care about the cpu 05:34:10 Foreign Function Interface 05:34:16 ah. 05:34:39 There are 2 popular FFI libs for C - and I can barely understand 'em. 05:34:54 ..getting them to play with forths is.... interesting. 05:35:43 This is usually why I think of using C to write/bootstrap a forth 05:36:12 ah, instead of machine codes. 05:36:45 sure: let the C-compiler generate the MC or asm (and then use the assembler): avoid the mc/asm 05:38:05 does it have anything to do with the issue of writing a forth interpreter in C? 05:38:16 hmm 05:38:37 well, I like C.. But, recall: Gforth REQUIRES gcc; and FICL uses ansi-C 05:39:09 it's not a pure forth approach. 05:39:25 Given an apropos lib and funcs, slapping together routines to manage a forthish-utility is easier. 05:39:39 yes. 05:39:58 The purist approach requires everyone to write everthing from scratch. 05:39:59 Sure it's not. And, unless yer willing to spend to get a commercial product, the metacompiling and stuff is sorta' out of reach. 05:40:18 akebono: yeah, which often means "damn the OS - I know better". 05:41:07 Now, historically, that was fine: forth used to be the first language on new platforms - maybe following an assembler. 05:41:31 ..it gets really tedious when you want to integrate and leverage code. 05:41:45 I feel the purist approach should reserve a place somewhere in this hasty busy enrionment in the computing industry. 05:42:05 environment - i meant. 05:42:11 it has a place - there must be at least 3 guys in here that asm their own distro. 05:42:23 :-) 05:42:47 ..I immediately avoid 'em, just because I feel C - at its worst - is still more portable and useful. 05:43:03 in a sense, we are using and buying forth products to reserve the place. 05:43:12 nonsense. 05:43:37 I mean there can be many other approaches. 05:43:42 there are 05:44:01 ..and the more restricted and deviant, the less shit is used. 05:44:34 I think both should be promoted. 05:44:43 whatever 05:45:30 but today many programmers need to first learn C or C++. 05:45:38 better or worse. 05:46:08 of course they first need to understand macine codes. 05:46:16 Nowadays they actually ridicule C, and use shit like Java, Python, Perl, Lisp, and .Net 05:46:23 not really. 05:46:39 Learning ASM is not a priority anymore. 05:47:06 but at least C is not such a high-level language as Fortran. 05:47:58 a novice learner have to wonder what the difference between an array name and its address. 05:48:08 is. 05:48:17 sure. 05:48:42 they can wonder all day long - they can buy a few books, spend hours and hours, and still do -S to get asm for reading. 05:49:18 C's array indexing syntax is very confusing. 05:49:28 doesn't matter 05:49:43 C's been around the horn many, many times. 05:50:48 C used to be a Pascal for ASM programmers. 05:51:07 Pascal or PL/1 05:51:29 I still piss folks off regular, because I prefer to consider C a "portable assembler" - and damn the verbatim shit. 05:52:09 not portable than FORTRAN. 05:52:31 because they rely on libraries. 05:52:39 huh? 05:53:22 FORTRAN had defined a strict specifications about its performance and behavior. 05:53:49 of course there were many exceptions, though. 05:54:05 I wish fortran on enemies 05:54:12 hahaha. 05:55:07 but treating C as a high-level language tends to fail. 05:55:21 Because it's systemic 05:55:44 Similarly, forth string-support makes C look magical 05:56:08 C doesn't support strings at all. 05:56:37 so, it's necessary to #include 05:56:46 true, but it's nonsupport has a lot of capabilities. 05:56:56 ..whereas, forth has far less. 05:57:10 I've been cogitating these issues for weeks. 05:58:01 I'm not an expert, but forth usually treats strings as counted strings or a pair. 05:58:08 afaik. 05:58:19 yeppers - then look at the internals. 05:58:31 it's not really pleasant 05:59:22 Anyway... Yeah, I like Forth - always have. I also live in C. I want my C to employ forthish and forthish to employ C. 05:59:50 in any language, string is a difficult thing. 06:00:09 because it may require GC. 06:00:45 not necessarily, but it CAN get shitty when you suffer wchar_t and uni* chars/strings 06:01:53 btw 06:02:03 can gforth have access to system calls? 06:02:59 it should 06:03:34 both shells/popen? I think 06:03:47 ah. 06:03:49 or, do you mean lower? I think they manage that, too 06:04:36 okay. I guess at least its Linux version will support most of all system calls. 06:04:57 Most likely 06:05:17 reading the Gforth source will leave you blind, bleeding and ragged 06:05:17 but I'm using cygwin for learning and playing with forth :-) 06:06:11 there are several forths for dos/doze 06:06:38 I have used pygmyForth 06:08:01 but old purist forths occupied a floppy disk. 06:08:15 accessing the disk directly. 06:08:32 yeppers "block files" 06:08:40 still convenient to support 06:09:10 some were not dependent on neither DOS nor BIOS. 06:09:15 I know 06:09:19 (I was there) 06:09:29 self-running 06:09:52 Actually, it just meant they wrote a driver/routine in asm/mc 06:10:25 yes, they wrote what OS/BIOS does today. 06:10:41 right 06:10:47 or the asm to access those 06:11:43 I feel forth is very flexible. 06:12:05 yep, it should be 06:12:10 and it's data type is very simple. 06:12:27 I'm thinking "too simple" - because it has more than one 06:12:57 it's like BLISS's approach. No type basically. 06:13:08 just words or bytes. 06:13:30 The problem is the interpretation 06:13:39 interpretation is up to you! 06:13:45 that's the prob 06:14:03 It's ok for interactive programming as a non-root 06:14:17 it's NOT ok for deployment. 06:14:28 it's like Fortran's EQUIVALENCE statement. 06:14:45 no idea - I've not seen or touched fortran in nearly 30 years 06:14:49 a word can be interpreted in any ways. 06:14:55 yep 06:15:22 signed, unsigned inter? or two 7-bit characters or 8-bit characters? 06:15:31 or two chars array? 06:15:53 right and ptrs and illegal addresses or misinterpreted or "strings" of chars, etc 06:16:14 It's actually at least as dangerous - as is - as asm. 06:16:32 one word can be executable. 06:16:35 Actually, let me retract and restate... 06:17:17 it may make the entire system halt. 06:17:22 I like the forth flexibility (and it can be improved), but that belongs in non-interactive, valid-code. It'd help if interactivity was a bit more paranoid. 06:17:51 I keep remembering F-PC 06:18:09 ..that sucker was stone-cold cool for experimenting and programming on dos 06:18:48 ..it was almost a 1:1 copy of the screen for turbo-pascal or turbo-C 06:19:24 or SmallTalk? :-P 06:20:35 So, yes.. I think that most forths are self-serving and generally foolish. I can easily envision better interactive-interfaces and I wouldexpect non-programmer or limited-programming environments to be a lot pickier.. Gforth is about as protective as I've seen. 06:20:51 I never suffered smalltalk - never wanted to do so. 06:20:59 I did try modula-2.. 06:21:19 IMPLEMENTATION MODULE 06:21:33 yeah, the cases and verbosity got old. 06:21:56 it's if then else syntax differed from that of Pascal. 06:22:54 it was simpler than Ada but more complicated than Pascal. 06:23:06 --- join: snowrichard (n=richard@adsl-69-155-177-154.dsl.lgvwtx.swbell.net) joined #forth 06:23:25 I believe C had advantages over those rigidly typed languages. 06:23:43 I sorta' agree 06:24:33 Problem with C is mostly a lack of ability to ascertain ABI, and the semi-sloppy method of separating API from code. 06:25:01 I say "semi" because, sure - you can fake it... But apparently kids are stupid, or their teachers need to be shot. 06:26:53 the simplicity of Lisp and Prolog look somewhat similar to forth. 06:27:13 though they were all very different internally. 06:27:15 well, Forth grew from the Lisp predecessors 06:27:56 because it's a linked list of executable cells. 06:27:56 Personally, I can't read lisp - and I can do better with forth 06:28:16 but they are very similar. 06:28:27 oh, several ideas - but yeah, the idea of a "list" maintains 06:28:57 and no absolute distinction between code and data 06:29:33 I personally believe there are issues with "parsing" and words, but I'm still trying to codify my views there. ANd, I have some issues with datatypes. 06:30:14 data type is one different theme. 06:30:23 akebono: well, the prob with that is: there are a multitude of distinct issues about code and data - and asm & os reflect this 06:31:29 think of user/group/root: and think of exec-data (code) and R/O data and r/w data, etc.. It all can come back and bite the proverbial butt. 06:33:24 language is not well seamlessly incorporated into the os environment. 06:33:35 Oh.. and recall: a freestanding forth may not care, but *nix sure will. 06:33:45 yeppers, more than true 06:34:55 also recall that - some places - you might need no headers, or you want the headers in memory different from the code/data: the variations HERE alone are less than amusing 06:35:34 may be it's achievable only on experimental systems like Plan 9 or self-running forth. 06:35:56 or lisp machine. 06:36:08 no.. I'm still tinkering.. It's all doable, but it can't be "purist" or ANS in the end. 06:36:18 file systems are much worse. 06:36:41 files and devices simply must rely on the underlying system 06:37:05 I should say: paths, files & devices 06:37:27 there are file systems accepting distinction of U&lc characters for file names. 06:37:37 Sure.. As FIG did, you could code yer own "standalone driver", but it'd need to reflect at least posix 06:37:46 other systems don't. 06:38:01 akebono: I use *nix on a powerbook.. I'm aware ;-) 06:38:10 but some of the other systems may support UTF-8. 06:38:18 that's the macintosh HFS. 06:38:27 I'd suggest a UTF-8 filesystem is not a great idea 06:38:43 otoh, if it was all integrated at the lowest levels, it'd work 06:38:57 Why do they support U&lc differentiation before they implement UTF8? 06:39:10 why don't -- sorry 06:39:15 akebono: well, I could install darwin linux, but macosx offers interesting possibilities I deal with 06:39:41 on the other hand, 06:39:51 there are many, many "hands" ;-) 06:40:14 the standard OSX doesn't have any text editor that can write in UTF-8. 06:40:24 it's terrible. 06:40:39 they have just a text editor supporting UTF-16. 06:40:43 TO be honest, while I can understand the idea behind utf-8, I don't believe I've ever used/noticed it. 06:41:10 I'm usually a "oh, fuck 'em all: I use ascii" guy. 06:41:28 UTF-8 is an ASCC-centric extension to keep the character code endian-free. 06:41:38 ASCII 06:41:40 and that's why I could suffer it 06:41:57 BUT: not as we currently consider "strings" 06:42:31 UTF-8 can be 6-byte length. 06:42:48 pay attention to the variable length nature. 06:43:35 it's not yet another expanded national character set. 06:44:16 Unicode is another tower of mysteries. 06:46:03 I'm not even sure how non-ascii shit would end up 'displayed' from a utf8 string into an ascii world. 06:47:26 I believe the 7-bit part is the same as that of ASCII. 06:47:54 so... how do apps/sys display the intrusions? 06:48:01 so, don't worry, as far as you don't try to go beyond the territory of ascii chars. 06:48:41 I've always liked the html mode of '%20' 06:48:54 at least one can type all that crap 06:50:07 I think the hig bit works. 06:50:11 high 06:51:02 Be interesting to see some tests - I've never tried to pervert my system. 06:51:11 it decides whether the following byte is also an ordinary ascii or not. 06:51:21 yeah. 06:51:30 Not a bad idea, then 06:51:41 ..explains the '%20' outputs 06:52:04 the first bit is always zero. 06:52:06 I still think "string" needs a total overhaul. 06:52:09 if it's an ascii. 06:52:40 yeah, the % was the important part. For html, spaces would be a nono and be treated similarly. 06:52:47 ..same for filepathspecs 06:53:39 if it's set to 1, it is the 2nd or 3rd byte of a 2-byte or 3-byte char. 06:54:19 if it's set to 110, it's the first byte of a 2-byte char. 06:54:36 Yeah, I glanced at a doc - never toyed with the stuff 06:54:44 if it's set to 1110, it's the first byte of a 3-byte char. 06:55:06 iirc, it's fixed at a max of 3-bytes (4?) though, right? 06:55:16 four bytes, i feel. 06:55:29 ok, that explains wchar_t and friends 06:56:19 Yeah.. I can now envision dealing with utf8. We need a far, far better interface. 06:56:57 UTF-16 is much simpler but it is a 16-bit code. 06:56:57 chars, strings, files, consoles, devices - access to members. We need an overhaul. 06:57:20 well, if you settle on a 32-bit int, yer pretty golden. 06:57:58 if so, UCS4 = UTF-32 is ideal for internal representation. 06:58:46 the prob is "internal" versus "external" - utf8 is prolly the best solution. 06:59:30 I edit a plain ascii-file; you see this as utf-8 extern, but use internals. 06:59:46 reversed, I'd see escaped-crap, no? 07:00:06 no. 07:00:12 hmm... ahh, therein is the rub 07:00:29 the local -> internal translation takes care everything. 07:00:49 ...if I write "foobar=", you could use Japanese (kanjii?) similarly? 07:00:52 and, the internal representation should be fixed length 07:01:17 yes. 07:01:38 on PCs, no escape sequences are used. 07:01:57 now wait - gimme' a "for instance" ala' Japanese. - we can't all use ASCII if we are to satisfy trolls. 07:02:02 instead, using Shift-JIS codes 07:02:17 shift wha-huh? 07:02:55 the genuine JIS codes are 7-bit/8-bit characters. 07:03:11 but they conflict with ascci and control chars without escape sequences. 07:03:13 oh-hah? hasu-desu? 07:03:21 ahhhh 07:03:24 I thought so 07:03:46