00:00:00 --- log: started forth/04.03.30 00:36:14 --- quit: yeoh (Read error: 60 (Operation timed out)) 00:36:14 --- quit: madgarden (Read error: 54 (Connection reset by peer)) 00:36:42 --- join: yeoh (~yeoh@219.95.4.108) joined #forth 00:39:04 --- quit: yeoh (Client Quit) 01:11:29 --- join: yeoh (~yeoh@219.95.4.108) joined #forth 01:29:58 --- quit: ASau () 01:36:56 --- quit: yeoh ("User pushed the X - because it's Xtra, baby") 02:05:25 --- join: arke (~Chris@wbar8.lax1-4-11-099-104.dsl-verizon.net) joined #forth 02:52:22 --- join: crc (crc@1Cust61.tnt1.levittown.pa.da.uu.net) joined #forth 03:05:20 kc5tja: heh, you're on quite late tonight :) 03:10:29 * arke is CF-ing 03:29:46 --- join: ASau (~asau@158.250.48.196) joined #forth 03:31:08 Dobryjj den'! 03:31:26 Hello ASau 03:32:00 crc. I think this could be useful to you. 03:32:59 I don't know how it's in ANS Forth or in '83, but in FIG Forth I am able to change IN from console to point to pre-loaded Forth code. 03:33:09 I mean plain text. 03:33:35 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 03:33:39 Without line separators or something of that kind. 03:33:49 Just a large string. 03:34:08 One string. 03:34:42 I can describe what exactly I've done, if you're interested. 03:34:50 Yes, please 03:35:01 Well. 03:35:15 I've prepared a blocked file. 03:35:31 But I've no way to access it from Forth. 03:36:11 Then I've written R/W (a disk driver) in the way to work with RAM. 03:36:20 Like RAM drive. 03:37:13 But I was lazy to type R/W definition in every time on start. 03:38:12 hi 03:38:17 I've studied code and found that I can substitute memory address where WORD (and INTERPRET) gets input. 03:38:43 Actually, I substitute terminal buffer address. 03:39:22 In my Forth, address of area to parse from is calculated in this way (some details omitted): 03:39:29 TIB @ IN @ + 03:39:51 And IN is increased till the end of buffer. 03:40:34 But I can interrupt interpreting when I call QUIT (it asks user to enter something). 03:40:57 Or any other, indirectly call QUIT. 03:42:22 At the end, I've found solution to append to my binary file (COM) a Forth source with deleted CR LF and substituted HT to space. 03:42:53 Well, in fact, I substituted all control characters with spaces. 03:43:07 =D 03:43:19 Dobryjj den', arke! 03:43:42 terve alekseij 03:43:45 ja arke 03:43:46 My LIMIT word marks the end of image loaded. 03:44:12 I've found "magic command": 03:44:19 LIMIT TIB @ - IN ! 03:44:40 Guten Tag, mur! 03:45:52 crc. I don't know how you deal with input buffers, but you may consider this way as useful for compiling part of RF. 03:47:55 I'll look into using it to help improve the compiler. 03:47:59 If you've seen how postForth is done, you can combine postForth's approach to kernel and pre-compiling with approach to access text as described above. 03:50:18 And this is similar to the way, how MVP-Forth used to interprete command line. 04:00:40 * crc just got support for loading and saving a file working under Windows, now for the Linux version... 04:01:05 And I'm amused by this: 04:01:17 : FENCE! FENCE @ MAX FENCE ! ; 04:01:20 I guess I'll need to code a block editor now too. 04:01:35 What is FENCE 04:01:42 Take a look at FIG-Forth portable editor. 04:02:08 I've done block editor in 4 lines after that. 04:02:50 In just four lines? What functionality does it have? 04:03:12 Replace with input line. 04:03:19 Erase line. 04:04:16 It is sufficient for first time. 04:04:26 And it can be extended. 04:04:26 True enough. 04:09:34 It should be approx. this way: 04:25:26 : LINE >R C/L B/BUF */MOD R> B/SCR * + BLOCK + ; 04:25:37 ( n -- addr ) 04:25:58 Sorry, ( line# scr# -- addr ) 04:26:19 : E SCR @ LINE C/L BLANKS UPDATE ; 04:27:03 : TEXT HERE C/L 1+ BLANKS WORD HERE PAD C/L 1+ CMOVE ; 04:27:29 No. 04:27:32 --- nick: changler -> chandler 04:28:57 : P >R 1 WORD HERE COUNT R E R> LINE SWAP CMOVE ; 04:29:18 Do not include TEXT . 04:29:53 Well. 04:30:25 Main idea is to define a word that returns address of line in disc buffer. 04:30:45 n LINE C/L BLANKS should erase line #n. 04:30:55 This is called E 04:31:27 n P xxxxxx should place text into line #n. 04:32:02 Variable SCR is used to hold current screen number. 04:43:43 --- quit: crc ("Off to work...") 04:53:49 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 04:55:26 Hi 05:09:26 Dobryjj den', tathi! 05:16:06 hi Rob, ASau 05:35:00 How's your for today, tathi? 05:41:20 oh, pretty good. 05:41:53 I'm finally starting to do some serious work on my new assembler 05:41:58 How about you? 05:42:24 I'm migrating to blocks, it seems. 05:42:49 I've written good enough set of words to provide persistancy. 05:43:26 In week or two float numbers to go. 05:43:40 Do you write assembler in Forth? 05:46:20 No, in assembly language. 05:47:25 I thought about doing it in Forth, but decided it would be able to build itself sooner this way. 05:48:47 Hm. 05:48:57 Yet another FASM? 05:49:31 don't know fasm 05:49:46 http://fasmassembler.net/ IIRC. 05:49:58 ah, ok. 05:50:07 this is for PPC 05:50:25 Ah. 05:50:51 my Forth that I'm currently using _does_ do some assembly, but I decided I really want an assembler with Forth-like features 05:50:52 But anyway, wouldn't this be easier to do in Forth. 05:51:02 rather than just a Forth with an assembler on top. 05:51:10 yeah, probably. 05:51:21 Maybe you could adapt Gforth or any other one. 05:52:03 I'd like to be able to distribute this without requiring Gforth or whatever. 05:52:37 Hm. PPC asm should be much easier to work on. 05:53:34 I mainly want a replacement for the GNU assembler, which can be really annoying. 05:54:03 Plus I want forthy things like an interactive mode, and being able to define new "words" 05:54:49 How do you want to achive this? 05:55:17 This should be more like asm on top of Forth. 05:56:19 yeah, I'm reusing a lot of the code from my Forth. 05:56:50 But without the stack built in, and with a more complex parser. 05:58:11 You mean, you use single stack. 05:58:45 I mean, "words" take operands in registers, rather than on a stack. 05:59:06 Ah. 05:59:53 Though I expect it to be trivial to add Forth on top of this, and I'm sure I will. 06:00:32 One time I thought of two stage: source -> Forth-like intermediate text -> binary. 06:01:08 But I haven't tried this. 06:12:12 --- join: hovil (~hovil@CommSecureAustPtyLtd.sb1.optus.net.au) joined #forth 06:12:56 Guten Abend, hovil! 06:13:29 bonan nokton, ASau! 06:13:50 How is your Forth? 06:14:27 I don't have a forth 06:14:50 Well. 06:14:58 How is your Forth code? 06:16:03 simple 06:20:44 --- quit: arke (Read error: 60 (Operation timed out)) 06:21:02 installing dragonforth 06:25:42 --- quit: tathi ("leaving") 06:31:33 --- nick: chandler -> changler 06:35:43 hmmm... writing a morse translator seemed _too_ easy =P 06:38:56 well, a bad one was easy 06:41:27 It's pretty hard writing a good one. 06:41:39 That is, sound card input -> text. 06:45:15 hehe, I was only doing - . -..- - 06:45:55 Cheater. 06:47:19 having no dsp experience, I'd imagine it'd take a bit longer for me to write what you've proposed 06:48:17 Probably. 06:48:29 I don't know anything about that either, so.. 06:55:50 --- nick: changler -> chandler 07:15:46 --- join: ribbon (~ribbon@cpc1-bolt4-5-0-cust161.mant.cable.ntl.com) joined #forth 07:15:46 --- quit: networm_ (Read error: 104 (Connection reset by peer)) 07:16:13 I'm back. 07:16:22 Dobryjj vecher, ribbon! 07:16:51 Ogy ogy, ASau? 07:17:15 Sorry? 07:17:24 This language is unknown to me. 07:17:30 What does "ogy" mean? 07:19:37 It means, "I haven't got a clue what you just said." ;) 07:19:39 "Good evening." 07:19:58 And a good afternoon to you also ^^ 07:21:04 How's your Forth? 07:22:01 Non existent . . . 07:22:14 --- join: networm_ (~networm@L0652P18.dipool.highway.telekom.at) joined #forth 07:22:23 Then how's your Forth code? 07:22:35 Terve, networm_! 07:22:39 Similarly :) 07:22:51 Well. 07:22:59 You need to write one. 07:23:13 I'm currently just investigating the language... 07:23:24 Ah... 07:23:29 That's good. 07:23:39 So how do you find Forth? 07:23:57 As a language? 07:24:06 Simple enough, I suppose... 07:24:37 Well. 07:24:41 What about code? 07:25:19 I haven't written, or looked at any yet -- 07:25:30 Dug around the tutorials and what not... 07:25:38 Well. Take a look at something. 07:26:06 There's so many examples to look at that I can't recommend any. 07:26:46 There're... 07:27:16 My attraction to forth is the small size of it's core system... 07:28:21 That was my way to Forth too. 07:28:29 ...and through the, the ability to comprehend all of it ~ 07:28:46 Oh? What is your way now? 07:29:12 hi ASau! (before only xchat's auto-join joined..) 07:29:13 Now I'm trying to use it for bigger applications. 07:29:17 --- nick: networm_ -> networm 07:30:14 Are you using forth as your OS, or is it hosted under something else? 07:30:46 I'm using it hosted. 07:31:08 But I keep oportunity to configure it as standalone application. 07:31:32 I've experimented with such configuration. 07:32:18 So, what are you currently working on? 07:32:23 Actually, I'm thinking of more support for running standalone. 07:32:56 Currently, I'm writing floating point numbers package. 07:33:54 For fun? 07:33:58 It should be like primitive asm with FPN and control structures support. 07:34:03 For work. 07:34:08 * ribbon nods 07:34:28 I plan to use it standalone for numeric processing. 07:37:12 Were there no C libraries you could grab? 07:37:34 There're no C libraries for these tasks. 07:39:30 And I looked into GNU libc. 07:40:09 Just to be sure C does not fit my needs. 07:42:07 I thought there was a mathlib, or similar 07:42:09 * ribbon shrugs 07:42:14 Well. 07:42:19 There is GNU SL. 07:42:30 It lacks features I need. 07:42:51 Anyway, I have specific enough needs. 07:42:58 Fair enough. 07:43:46 I will be in need for more flexible language, as I don't have numbers only. 07:48:21 --- join: fridge_ (~fridge@dsl-203-113-230-141.NSW.netspace.net.au) joined #forth 07:49:21 What's the overall project? 07:50:13 It seem it should be expert system. 07:50:21 --- quit: fridge (Read error: 110 (Connection timed out)) 07:50:33 Wow -- what field? 07:50:36 I'm going step by step. 07:50:53 Chem. thermodyn. 07:52:00 And statistics. 07:52:00 Like that used by chemical engineers? 07:52:34 Yes. 07:53:10 But I'm not engineer, it won't contain everything possible. 07:55:03 That's a pretty big project ~ 07:55:08 You working on a big team? 07:55:26 I'm starting with my current needs. 07:55:35 (Or with an existing code base.) 07:55:47 * ribbon is reminded to think the forth way :) 07:55:50 That is modelling. 07:56:42 Current goal is multi-dimensional non-linear optimization. 08:00:15 What method(s) are you thinking of using? 08:04:00 For start, I need simple Hooke---Jeeves and/or Nelder---Mead algorythms. 08:04:14 That's for non-linear part. 08:04:54 Linear parameters I plan to find separately by elimination or LU. 08:05:55 I think that direct methods will be better. 08:06:26 My functions are complex enough to calculate gradient. 08:11:20 And, I tried multi-dimentional tangents. That does not work good. 08:13:11 I assume you've put all the maths down on paper first... 08:13:16 ...mixing maths and code... 08:13:19 * ribbon shudders 08:14:31 Well. 08:14:40 I've done this. 08:15:37 WLSS function is easy. 08:16:10 Weighted squares sum. 08:16:20 * ribbon okays 08:18:00 I have need to use nonlinear approximation by functions of such a kind: \sum_i w_i f_i(x) 08:18:52 The whole F is linear of w_i, but f_i depend on several parameters non-linearly. 08:19:30 I've programmed optimisation in C... 08:19:34 Brrrrrr. 08:20:34 Just think what you need to do when number of f_i is changed. 08:21:26 I'd need to see more of the maths before I could figure out what tricks I could play with it... 08:22:05 (I'm not saying that I want to see more maths, though ;) 08:22:51 Well, for first approx. I need f_i(x)=(t/sh t)^2, where t=a_i/x 08:23:32 a_i is a constant? 08:23:42 Second approx. is: F_i(x) = \int\limits_0^x f_i(t) dt. 08:23:49 a_i is parameter to find. 08:23:57 * ribbon nods 08:24:15 Sorry. 08:24:24 Second approx. is: 08:26:06 F_i(x)=\int\limits_0^{a_i} t^{b_i} f_i(x; p) dp. 08:26:38 F_i(x)=\int\limits_0^{a_i} t^{b_i} f(x; p) dp. 08:27:00 f(x; p) = ((p/x)/sh(p/x))^2 08:27:47 Actually, f is Plank-Einstein function. 08:27:54 F is Debye function. 08:28:26 Further is more. :))) 08:29:35 a_i = ^1A_i + ^2A_i x, b_i = ^1B_i + ^2B_i x 08:30:21 It's possible that one of functions is not P.-E. or D. :) 08:31:07 You've lost me :) 08:32:14 * ribbon goes to fetch paper and pencil 08:32:42 That's demonstration that you may go without "more of the maths." 08:34:17 You can't catch particularities of all the functions that can appear here. 08:45:17 I don't understand this notation: ^1A_i + ^2A_i x, b_i = ^1B_i + ^2B_i x 08:45:52 "A" with left upper index "1", right lower index "i" etc. 08:46:04 Usual TeX notation. 08:46:50 I thought so... 08:49:02 The most frightening inscription I've ever seen was 4 levels nested left lower index. 08:50:18 I wonder is that would need bracketing... 08:50:20 is=if 08:50:21 Something like _{_{_{_ia}\Delta}A}\xi 08:52:52 No in TeX you don't need bracketing, when you have single character. 08:54:06 I was thinking about whether the actual text needed the bracketing... I suppose not ~ 08:54:50 No. Integral has the lower precedence than multiplication. 08:55:17 Integrating... 08:55:26 How should it be? 08:55:40 ? 08:55:51 How should what be? 08:55:53 I'm not mathematician. I don't know right term. 08:56:02 Integral. 08:56:12 Thanks. 08:56:31 I never made it to mathematician status. 08:57:01 (Or 'integration'.) 08:57:11 Well. I never dealt with math. articles on calculus. 08:57:43 Integration is the act of integrating. Integral is either 1) the symbol used to denote integration, or 2) the result of said integration. 08:57:58 OK. 08:58:18 * kc5tja is in Calc-II, the course almost entirely dedicated to integration and its evilness. 08:58:36 kc5tja: Uni course? 08:58:43 ribbon: Yes. 08:58:57 I know Calculus only in native language. :) 08:59:05 * ribbon pities kc5tja 08:59:53 * kc5tja nods :) 09:02:20 Calculus is called "Math. Analysis" here. 09:03:27 Poor kc5tja 09:03:33 I may doubt in terms. 09:03:40 "Introduction to Analysis" is a 3rd year university course around here. Differential equations is considered a pre-requisite to take the class. :) 09:03:45 In English terms. 09:04:15 oh, I forget: it may not be 3rd year; it could be graduate level. 09:05:09 I'm not sure, whether I understand you right, I don't know your grades. 09:06:20 There are three Calculus-series classes that you start out with: Calc I, Calc II, and Calc III. Each takes a semester to complete. 09:06:24 Then there is Differential Equations. 09:06:29 Beyond that, I don't remember. 09:06:35 We have these grades: 1) school; 2) univ. (diploma); 3) "aspirantura" (kand. of some sci./hum./med./...); 4) doctor. 09:06:46 But after 4 years of education, you get a Bachelour's degree in whatever your major is. 09:07:01 Those are not grades in USA. 09:07:06 Those are institutions. 09:07:25 Grades are things like "You passed," "You got a C," or, "2.8 out of 4.0 Grade Point Average." 09:07:37 Hm. 09:08:04 That's internal to institution/university here. 09:08:51 After two years, you can earn an associate's degree. After four years, bachelor's degree. After six, master's degree. After eight or more, Ph.D. (doctorate degree). 09:09:17 This is all within the university scope. 09:09:51 Before the university level, you spend around 12 years in what you call "school": 6 in grade school, and 6 in so-called high-school. 09:10:04 We have no intermediate degrees in university. 09:10:21 Here it's: 3 for a bachelor's degree, 4 for a master's degree, and 7 for a post-doctorate degree 09:10:24 You either earn diploma or not. 09:10:56 (But we might be switching over to some other system Real Soon Now.) 09:11:45 ASau: We don't have intermediate degrees either (not unless you're willing to pay for them). When you enter college, you have to state whether you're matriculated for either a 2-year or 4-year degree. 09:11:46 That's called "specialist." 09:12:26 Then, after that, to go to graduate school, you have to take these wonky tests and whatnot, pay almost your entire life's savings, then work for sub-minimum-wage rates for some researcher to get your master's or doctorate degree. 09:12:51 :) 09:12:51 We don't have them at all. 09:12:52 Later. 09:13:01 --- quit: ribbon ("A life without meaning") 09:13:45 When you enter university you know study plan, 4 to 6 years. 09:13:50 I'm not sure what I would like to study when I go to university 09:14:12 You are to pass exams in time. 09:15:17 well, I have a few selected but haven't made a final choice yet 09:15:30 the plan was for my fiance to finish her degree, then I will start mine 09:15:48 so hopefully at no stage will we have to live like poor uni students 09:15:56 =) 09:16:07 At last year you are to perform research or engineer work and present results to special comission. 09:16:38 After that you get dimploma and become "specialist". 09:17:09 Engineer, researcher, doctor (when you finish medical inst.) etc. 09:19:04 And we have no selection of courses. 09:19:50 If you've choosen to be chemist, you are to pass exams on inorganic, organic, analytical, physical etc. chemistry. 09:19:51 ASau: So your university system is more like a guild system then. 09:20:56 And of course all general courses, like phylosophy, history, math. analysis, algebra etc. 09:20:59 Well, that's a neat way of handling those students who don't know what they want to be when they pass the exams. :) (Your degree is . . . Chemistry!) 09:22:03 Actually, you get enough knowledge to work in another field than you specialization. 09:22:13 Right. 09:22:41 So individual classes don't have things like final exams? You can't pass or fail individual classes? 09:22:43 There's optional course of pedagogic. 09:23:11 There're individual courses, of course. 09:23:35 They're for finer specialization. 09:24:37 E.g. I'm to become "wide field researching chemist." 09:24:59 More particular "physical chemistry." 09:28:44 whats the size of a typical stack? eg, how many items can you put on the stack before it'll bork up? 09:28:48 And as any "chemist" I've passed exams on chemical technology and several related courses. 09:29:11 qFox. Perform an experiment. 09:29:39 It depends on your system. 09:29:40 yes i can test it in winforth, but isnt there a standard for it? 09:29:47 No. 09:29:56 There can't be such standard. 09:30:20 not even a minimal? 09:30:27 You may hope there're no less than 4 cells available. 09:30:37 heh 09:30:41 ok 09:31:17 same for return stack i guess? 09:31:30 That's not a problem: in most cases you can configure your system to have depth needed. 09:31:44 Right. 09:32:02 Well. I should go. 09:32:06 Bye! 09:32:08 --- quit: ASau () 09:47:51 --- join: Mark4 (~Mark4@64.47.44.254) joined #forth 09:57:23 ANSI says that 8 cells is the minimum depth for a data stack. 09:57:30 Ditto for the return stack. 09:57:46 specifically why ? 09:57:58 i would tend to agree, that would be an absolute minimum heh 09:58:09 but why specifically 8 cells ? 09:58:13 I don't understand your question. 09:58:23 why do they state 8 cells specifically 09:58:32 why not 13 09:58:33 or 7 09:58:42 Because it's the smallest usable stack depth that can contain enough state for both a regular task's computation plus some room for interrupt handling. 09:58:44 whydo they need to specify ANY minimum ? 09:59:02 I just answered that question. 09:59:13 that really depends on implementation AND on the given processor doesnt it ? 09:59:27 AND the specific Forth implementation too. 09:59:29 But consider: 09:59:38 the smallest stack depth needed to perform computation without loss of state is 3. 09:59:42 That's ignoring interrupts. 09:59:48 So you figure the interrupt handler ALSO needs another 3. 09:59:48 ok 09:59:51 So we're up to 6. 10:00:12 Then we need to hold intermediate data pointers too, to preserve the state. That should be about 1 cell, so we are up to 7. 10:00:13 plus 2 more just in case ? 10:00:14 --- join: Herkamire (stjohns@h000094d30ba2.ne.client2.attbi.com) joined #forth 10:00:19 7 isn't a power of two, but 8 is. 10:00:31 errr 10:00:34 And in electronics, power-of-twos are overwhelmingly easier to implement in hardware than not. 10:00:36 immediate data pointers ? 10:00:45 What about them? 10:00:53 i dont understand what you mean by them 10:01:05 I never wrote the word immediate. 10:01:21 I wrote intermediate. 10:01:53 hrm 10:02:16 well i can live with 8 as a minimum even if i dont totally understand the logic behind it hehe 10:02:38 i would never cripple a forth that way tho unless i absolutely had to 10:02:49 ANSI only specifies a *minimum*, not a maximum. 10:02:53 yes 10:02:55 Please don't misrepresent what I'm saying. 10:02:59 i wasnt 10:03:15 nor am i critisizing ans for stating that minimum 10:03:18 If you want, you can have 4.2GB sized stacks. :D 10:03:29 lol 10:03:38 i think isforth has 4k each 10:04:00 which makes sense because its one page heh 10:05:07 FTS/Forth was 2K each, both packed in the same page. 10:05:32 ya my 8051 forth uses 128 bytes each, both in the same page (256 bytes) 10:05:36 But as FTS/Forth development has stalled in favor of the Kestrel, that is no longer of any concern to me. :) 10:05:46 heh 10:05:55 your not abandoning fs are you ? 10:06:00 pick it back up later ? 10:06:07 Speaking of which, I still need to find a good serial interconnect bus for expansion peripherals for the machine. 10:06:17 It's the single hardest part of the design I've ever encountered. 10:06:18 usb ? 10:06:36 Do you have any hint of a clue as to what is involved in supporting USB? 10:06:38 :) 10:06:45 And how much it'll cost me? 10:07:02 hmm the cygnal 8051's can support usb. 10:07:03 USB sucks ass compared to a number of available alternatives anyway. 10:07:10 Yeah, as a *slave* 10:07:12 Not as a host. 10:07:22 oh 10:07:57 * kc5tja is looking for a serial interconnect that: 10:08:07 1) doesn't cost an arm-n-leg to design into a device. 10:08:12 2) has adequate performance, 10:08:28 3) Doesn't require the CPU to babysit the bus (which USB kind of does in a way), 10:08:41 4) has minimum wires, 10:09:03 5) has an easy-to-understand, simple, and bug-free-by-design protocol on it. 10:09:19 Many of these goals are conflicting, of course. 10:09:22 now your asking for too much :) 10:09:45 And that's the reason I'm so hung up on it; I have several designs I can go with, but they each have their pros and cons, and I can't yet decide which is more important. 10:09:54 Commodore did it. I can do it. 10:10:02 :) 10:10:03 ys 10:10:04 yes 10:10:26 In fact, in many respects, Commodore's bus was *vastly* superior to USB. 10:10:46 SerialBus comes the closest to my ideal bus (SerialBus == IEEE-1394, BTW). 10:10:55