00:00:00 --- log: started forth/04.12.21 00:25:43 --- quit: yumehito () 01:02:23 --- join: snowrichard (~richard@adsl-69-155-177-153.dsl.lgvwtx.swbell.net) joined #forth 01:02:35 good morning 01:02:46 hi snowrichard 01:02:52 it's snowing there? 01:03:13 no not snowing. that is my last name :) 01:03:21 oh :) 01:04:19 http://veteransinaction.info 01:04:29 ((this computer )) 01:09:00 not too much to see 01:09:07 not yet :) 01:09:10 maybe put some of your work there :) 01:09:48 I had a small language I did at one time , but lost the code when my other site was shut down. 01:10:18 oh, did it had any interesting features? 01:11:01 was similar to forth except I used operator names from C, it had include files, and a way to execute bash commands. nothing fancy. 01:11:13 i see.. 01:12:31 there are a few forums set up there, but have just "welcome" messages in them now 01:12:38 :) 01:13:10 just got my static ip set up a few days ago. 01:14:06 how do you set a static ip? 01:14:23 I had to ask my ISP to give me block of them (I have 5) 01:14:35 how much does it cost? 01:14:52 added about 20 a month I think. 01:15:07 i see 01:15:07 have not got first bill to get itemized cost yet 01:15:45 but now if I add a hub I could set up a dedicated server or two 01:16:38 how much does all this cost? :) 01:16:46 you must have a quite high income :) 01:16:48 the dsl? 01:17:18 yes 01:18:03 plans listed at www.sbc.net for this region, you probably have different phone company. About $60 a month for the pro-speed static here. 01:18:29 I have a disability check 01:19:36 how much does this lowers the costs? 01:19:50 ? 01:20:00 doesn't affect the cost, but I can afford it 01:20:44 if you don't mind, what kind of disability? 01:21:01 I had a head injury when I was in the Navy 01:21:10 ah, sorry 01:21:16 doesn't seem to affect your intellect :) 01:21:40 it did for a while. my grades were lower aftward 01:22:09 I have learned how to compensate somewhat now. (take notes on computer, etc.) 01:22:59 the medications are better now, don't sedate me as much. 01:23:55 what medication did you took? 01:24:06 Zyprexa and Zoloft 01:24:17 hmm, i don't know about them 01:24:23 i took carbamazepine 01:24:27 used to be thorazine... makes you a zombie 01:24:43 :) 01:24:48 :) 01:25:37 if I had to pay for the medications I could not afford my bills ... the Veterans administration provides them. 01:27:48 so any way do you use forth? 01:29:54 i'm new to forth 01:30:09 i've discovered it while looking at some mac info 01:30:21 and i saw that openfirmware uses it and i wondered why :) 01:30:38 its been around for years. the sun's use the open firmware too 01:30:41 but the reason i study it now is the fact that i try to create my own language :) 01:31:17 lots of fun. what language do you code in if you are creating your own language 01:32:00 like C or assembler? 01:32:06 in what language i code my programs? or in what language i'm going to write the compiler? 01:32:19 to write the compiler 01:32:37 probably i'll write an interpreter in c or java then write the compiler in itself 01:32:49 ok 01:32:50 then i'll compile the compiler using the interpreter :) 01:33:35 from then on it could compile itself 01:33:58 had any courses in compiler design? 01:34:11 no 01:34:41 you might check a college bookstore or library for some references. 01:35:03 i'm familiar enough with programming to code almost anything :) 01:35:09 so coding a compiler is the least problem 01:35:19 the problem is defining the language 01:35:27 what other kinds of things have you programmed? 01:36:00 hmm, i remember trying to code an windows clone for a Z80 when i was about 13 years old :) 01:36:20 windows ?? on a z80? would be a little slow ;) 01:36:27 :) 01:36:50 i didn't maked it since i lacked docs on z80 arch (timers & such) 01:36:58 anyway now i'm working on another os 01:37:03 this time for x86 01:37:12 what year? my mother used to havea trs-80 model III which was a Z-80 01:37:38 go to intel.com and you can order a CDROM (or download) intel docs on thier chips 01:37:52 i've got the docs 01:37:55 ok 01:38:35 Morning. :) 01:38:40 hi robert 01:38:41 hi robert 01:38:59 Windows on a z80... interesting. 01:39:22 Protected memory seems to be the only thing which keeps Windows above the surface. 01:39:27 :) 01:40:05 if you don't have the hardware for protection you have to trust all the code which is not always possible. 01:40:19 About simple computer systems, did you see this? http://www.venturalink.net/~jamesc/ttl/ 01:40:32 Sometimes that works, though. 01:40:47 well, my ideea is to design a language for that a compiler can produce proof of corectness much more easly than current languages 01:41:01 then use those programs in an unprotected environment 01:44:31 That would be interesting. 01:49:34 the devil is in the details :) 01:50:35 Yes, reality is often a problem with such interesting projects 01:52:47 well, i've concluded that since there is no generic method of doing this 01:53:03 i'll have to build a large database of known cases 01:53:11 (cases known to be correct) 01:53:27 the cases would be parametrizable to catch as many subcases as possible 01:53:42 so the only barrier that remains is source code size 01:54:09 basicly because it become extremely difficult to check for corectness very large programs 01:54:22 thereofore i need a language that is as compact as possible 01:54:33 (from semantic point of view, no syntactic) 01:55:20 Sounds like that program won't fit in 1k 01:55:43 also the language has to be purely functional, to make it simple to process the code by the validator (since there are not side effects) 01:56:11 --- quit: swsch_ ("Leaving") 01:56:17 About correctness... what kind of correctness, just that it doesn't overwrite random memory, or that it matches some other specification? 01:56:50 it uses just the stack, so you wouldn't have problems with memory access 01:56:55 seems you would not only have to encode the program, but a specification for what it was supposed to do... might as wel run the specification 01:57:09 the kind of errors i think is divide by zero, index out of bounds 01:57:17 Sounds a bit like ADA I think 01:57:24 also there should be a word named assert 01:57:35 that would check if the top of the stack is true 01:57:41 if not it will return an error 01:57:58 the goal of the compiler is to prove that it's impossible to get an assertion to fail 01:58:14 based on a known input format for a function 01:58:24 Or SPARK 01:58:53 for example before each division, an assertion checking for divisor<>0 would be inserted at the initial stage of the compilation process 01:59:14 so the compiler has to prove that no assertion will ever fail 01:59:36 http://www.praxis-his.com/sparkada/ 01:59:52 you can add your own assertions and therefore implement any specification you want 02:02:13 well, that's just ada, it does't check for corectness in the way i meant 02:03:42 but if the value is from an input (user or file) how could you guarantee it would not be zero? 02:03:56 "SPARK source code is unambiguous and amenable to rigorous analysis. The language itself eliminates whole classes of common error and facilitates the very early detection of any that remain. A system of annotations embeds specification information in the code reducing the semantic gap between them. " 02:04:10 From the SPARK website 02:04:27 It's a commercial product though and probably not cheap 02:05:04 does it check for example for halting-related problems? 02:05:11 also for example 02:05:18 an inference engine, with built-in heuristics for common automatic strategies (for instance, infer A >= C from A >= B and B >= C [transitivity], infer X + N >= Y from X >= Y, N >= 0 [addition of inequalities], etc.); 02:05:21 I don't know the details; I've never used it 02:05:23 Heh. 02:05:25 which is obviously wrong 02:05:53 since a>0, b>0 doesn't necesarly mean that a+b>a 02:06:11 it's of course correct mathematically but not from a computer's point of view 02:06:49 --- join: swsch (~s@swsch.sustaining.supporter.pdpc) joined #forth 02:16:17 --- part: snowrichard left #forth 02:41:59 --- join: snowrichard (~richard@adsl-69-155-177-153.dsl.lgvwtx.swbell.net) joined #forth 02:47:06 --- quit: snowrichard ("Leaving") 02:48:38 --- quit: Teratogen ("SKYKING, SKYKING, DO NOT ANSWER.") 03:08:45 --- quit: froop ("the master of balance") 03:25:06 --- join: froop (~wossname@rn-v1w5a06.uwaterloo.ca) joined #forth 03:31:00 anyone here who knows a good forth IDE that works in text mode? 03:32:19 --- join: Serg_penguin (~z@212.34.52.140) joined #forth 03:32:26 Privet, Serg 03:33:39 --- quit: ASau (Read error: 60 (Operation timed out)) 03:34:27 robert what ide are you using? 03:35:02 I'm using an IDE? 03:35:15 Actually I'm using vim in Linux and edit.com in DOS. 03:35:37 disgusting :) 03:35:39 do you know any good interactive ide for forth? 03:38:53 i use any editor having programmable hotkeys 03:38:58 to run compiler 03:39:21 isn' 03:39:28 isn't supposed forth to be interactive? 03:40:47 I use that for testing and debugging. 03:59:55 --- join: aum (~aum@60.234.138.239) joined #forth 04:00:48 hi - need ideas - i'm implementing a forth, and need to define a syntax for 1-byte, 2-byte and 4-byte integers 04:01:30 using a stack of 8-bit values, but providing 8, 16 and 32-bit operations 04:02:46 what architecture/OS ? what 'bitness' is default ? 04:02:48 one thought - n$[0-9A-F]+ - eg, 1$c9, 1&242, 2$a042, 4$894f34ac 04:03:07 architecture is PIC microcontroller, stack is 8-bit 04:03:28 i'm writing an extension to PicForth to allow running programs resident in off-chip flash memory 04:04:15 using PicForth to build a Forth VM running on the PIC itself, reading and executing bytecodes from external I2C Serial EEPROM 04:04:20 aum: Cool... I used "123 lit8" or "123 lit" (for 32-bit integers) on my AVR Forth cross-compiler. 04:04:29 just make one word '$' cute 04:04:34 I made such a VM for the PIC as well. 04:04:43 It's purely 16-bit, however. 04:04:44 robert: got a link? 04:04:59 robert: does your VM execute code in off-chip memory? 04:05:12 '$ A1' is byte, '$ 12345678' is 32-bit 04:05:21 The AVR does (using DataFlash), the PIC version uses internal EEPROM. 04:05:56 robert: isn't that a bit tight, given that the 14-bit core chips only have 256 bytes EEPROM? 04:05:59 As those projects are a bit old, I haven't uploaded them to my new page. If you want them I'll DCC the old versions. 04:06:11 64 bytes on the 16f84 :) 04:06:15 Yes, it is. 04:06:33 Just made it for testing anyway. Adding support for external EEPROM shouldn't be hard though. 04:06:49 Serg_penguin: problem is when you want to specify the number 3 as 32-bit 04:07:04 robert: what did you write it in? assembler? picforth? 04:07:19 Why not use the 8-bit literal 3, and then 8->32 bit conversion? 04:07:22 asm 04:07:27 robert: can't DCC - behind fw - email is aum@freenet.org.nz 04:07:36 OK, I'll mail them 04:07:44 oh damn, i forgot this channel is logged to web 04:07:49 Hehe 04:07:53 (oh, well, I wasn't getting enough spam anyway) 04:07:56 I also sell email addresses to spammers. 04:07:59 ;) 04:08:03 heh 04:08:24 * aum would love to napalm the houses of people who actually /buy/ stuff from spammers 04:09:10 Haha, I think their non-working penis enlargement pills are punishment enough. 04:09:11 robert: have you seen picforth? 04:09:18 Seen, yes, but not used. 04:09:30 Damn it, that stuff is on the other box. 04:09:31 * robert boots 04:09:32 it's actually very good 04:09:40 aum: I've been using Picforth for a few weeks 04:09:40 So I've heard 04:09:57 got one of the best forth optimisers i've ever seen - it's really hard to hand-code in asm and beat the compiler 04:10:12 The annoying thing is every time I write something I find a new code generation bug 04:10:21 Heh 04:10:36 holden: you been writing to the list about those code-bank selection issues? 04:10:37 I think Samuel has got bored of fixing them now :-/ 04:10:40 Yes 04:10:47 ahh, k 04:10:57 i'm worried Samuel might be burned out from PicForth 04:11:01 he's been very quiet 04:11:49 i ported my main production code to CCS C, and it's been running bug-free ever since 04:12:12 You're David, right? 04:12:20 yep 04:12:40 I run the Gnupic mailing list BTW 04:12:48 ahh, hi! :) 04:13:03 you run the gnupic website as well? 04:13:07 No 04:13:29 Scott used to host the list himself but he had to stop for some reason and I offered to take over 04:14:07 * aum would love to capture the author of sdcc and chain him to a terminal with intravenous caffeine till he gets the PIC compilation fully up to spec 04:14:51 the lack of a trustworthy, capable, free, opensource language compiler for PIC is a major problem 04:14:53 You mean Sandeep? I don't think he had anything to do with the Pic port. Scott started that. 04:14:56 aum: Sent the mail 04:15:03 robert: thx 04:15:51 so, what's the verdict on 1$nn for 1-byte literals, 2$nnnn for 2-byte, 4$nnnnnnnn for 4-byte? ok or suck? 04:16:09 Personally I'm moving away from Pics except in the very low end stuff 04:16:21 holden: you mainly AVR now? 04:16:36 aum: Sounds good to me. 04:16:37 No, I'm moving to MSP430 and LPC2000 04:16:54 Hehe. 04:17:02 I'm going backwards. 04:17:06 holden: what's the deal with those? cheaper? more mem? nicer architecture? lower power? 04:17:17 Going to pick up "Automatic computing machines" from 1953 today. 04:17:55 Not cheaper. MSP430 is lower power, ARM is faster and has more memory. Both have a much nicer architecture. 04:18:16 gotta admit the PIC architecture drives me nuts 04:18:38 goddam RP0, RP1, PCLATH 04:18:46 Yep 04:18:57 nice concept to have all instructions as 1 word, but the price is very high 04:19:28 Especially when you're using a compiler that's supposed to take care of the bank switching for you but keeps getting it wrong :-( 04:19:39 oh, yes, that is painful 04:19:58 with picforth, i sprinkle in a few 'forth> no-cbank' instructions, and some bugs go away 04:20:06 also 'restore-bank' here and there 04:20:27 the real problem is that Forth is a horrible choice of languages for writing a cross-compiler 04:20:52 i can write stuff like that 12 times faster in Python than in Forth 04:21:05 Forth is great for target code though 04:21:20 Guess it depends on how simple the compiler is. 04:21:37 For a non-optimizing compiler, Forth is excellent. 04:21:58 painful for hacking text though 04:21:59 Picforth is far from simple- mainly because of all the fancy optimizations it does 04:22:20 Right 04:22:25 PicForth's core code mutilates my brain worse than a contaminated acid trip 04:22:34 aum: That's why you use the Forth's own parsing words. ;) 04:23:10 only problem then is constantly having to allocate chunks of RAM to store stuff 04:23:20 surprises me that gforth has no 'strdup' 04:24:11 CCS C is sweet-as 04:24:30 but it ain't real C 04:24:33 Hmm... 04:24:42 What do you need all the RAM for? 04:24:56 I think one of the Forth compilers I sent you uses like a few bytes of RAM. 04:25:33 got your email - thx - t'bird flagged it as spam, so i'm glad i was expecting it 04:25:48 Heh 04:26:14 Off to lunch - TTYL 04:26:17 Bye 04:26:21 thx for all that, cya 04:26:25 I hate this business with spam filters. 04:26:35 It's like the virus/antivirus deal 04:26:50 Only purpose is to prove how fucking stupid humanity can be 04:26:55 heh 04:27:21 like anti-spyware apps that install spyware 04:28:26 * robert shrugs 04:33:48 * Serg_penguin gets very few spam - coz i don't put my @ in trash places 04:35:59 :) 05:04:24 robert: how to achieve maximum band cover w/ fewest switched coils ? 05:06:46 --- join: Topaz (jonny@spc1-horn1-6-0-cust217.cosh.broadband.ntl.com) joined #forth 05:10:29 * robert returns 05:10:43 Serg_penguin: What do yiu mean? 05:11:37 see: u have 1-LC RX and a given range of vario-C 05:12:40 how would you organize L's to cover widest band w/ fewest L's ? 05:14:26 You could make a variable coil 05:15:19 variometer wastes Q, slide-contact makes more R and so wastes Q too 05:17:05 movable ferrite ? narrow range ... 05:17:28 Hmm.. So what are you thinking about? Switching between different coils? 05:18:41 i think of some weird arrangement allowing series and parallel plug of coils, having straight angle->frequency function, and fewest contacts/parasites too 05:19:19 like drum bandswitch, but w/ combinations 05:20:32 How about switching in capacitors? 05:21:56 Q limit, see: lowest C is 5-15 pF coz of parasites, highest - 300-500 pF coz Q=kL/C 05:33:47 once i invented funny circuit for tungsten lamps, but never built ^)) 05:35:53 --- quit: bbls (Read error: 110 (Connection timed out)) 06:33:11 --- join: yumehito (yumehito@giguz.turbo.nsk.su) joined #forth 06:43:07 --- quit: Serg_penguin (Read error: 104 (Connection reset by peer)) 06:48:57 --- join: bbls (~bbls@80.97.121.133) joined #forth 06:49:02 hello 06:49:13 does forth allow recursivity? 06:50:48 Depends on the implementation 06:50:50 http://www.public.iastate.edu/~forth/gforth_28.html 06:51:06 how often does someone runs into problems requiring recursivity? 06:51:31 In my experience hardly ever 06:51:43 i'm designing my own forth based flavour and i wonder if it would be ok to remove recursivity support 06:52:10 I wouldn't have thought that would be a big problem 06:53:20 i plan to add some static analisys tools, and it would be easyer if recursivity wouldn't be allowed 06:53:27 *easier 07:18:10 --- join: snowrichard (~richard@adsl-69-155-177-153.dsl.lgvwtx.swbell.net) joined #forth 07:18:16 hi snowrichard 07:18:16 hello 07:18:32 #richardisageek 07:18:41 can i paste you a few lines, something i've asked before you joined? 07:19:01 in my channel? 07:19:43 i wonder if anyone uses recursivity 07:24:39 well, it can make lots of algorithms very easy 07:24:49 but with a handy stack, you can usually implement around it anyway 07:25:04 yes, but it will make the job of implementing the static analiser much easier 07:43:40 --- part: fridge left #forth 07:56:07 hi 07:56:43 Topaz? 07:57:30 --- quit: froop ("gay and bored") 08:00:07 i'm back 08:00:33 lo 08:03:56 hi 08:04:17 how are strings implemented in forth? 08:07:06 arrays of characters 08:07:24 how are arrays implemented? 08:07:42 using asm? 08:07:51 or using the stack and common operations? 08:07:57 using create for memory alloc, and then pointers to it 08:08:09 ahh, calling external code 08:08:29 so it's not possible to create arrays using the basic words? 08:08:38 it is 08:08:51 lemme find a good text on it 08:08:55 ok 08:09:38 http://www.xs4all.nl/~thebeez/ForthPrimer/Forth_primer.html#tth_chAp3 08:09:39 there 08:09:43 thx 08:09:48 np 08:20:34 is the CREATE word used for array creation a primitive or it's defined in terms of other words? 08:21:01 http://www.xs4all.nl/~thebeez/ForthPrimer/Forth_primer.html#tth_sEc5.6 08:21:02 :) 08:21:58 create is a very very powerful primitive 08:22:44 would it be possible to define it into a library instead of main language if you don't have access to the assembler? 08:24:08 what exactly are you trying to do? 08:24:34 i wonder if it's possible to emulate arrays in a forth without support for arrays 08:25:06 i actuall want to create a kind of forth, and i'm unsure whenever there is a tehnique to emulate arrays 08:25:32 since if there is a way, then i would preffer to define the arrays in a librar 08:25:35 *library 08:25:42 and keep the core language as small as possible 08:27:35 hmmm 08:27:56 --- join: Herkamire (~jason@h000094d30ba2.ne.client2.attbi.com) joined #forth 08:27:57 i dunno, arrays are pretty core to a lot of useful stuff 08:28:05 the language is more of theoretical importance and i am not interested in performance 08:28:57 i intend it to use it for experiments related to proof of correctness 08:29:23 so i have to keep the core language small 08:29:43 hrmm, why not just throw forth out the window and just work from its design concepts instead? 08:30:02 i'm working from it's concepts :) 08:30:45 you could redefine the way create works i suppose 08:30:58 or design a new way to generate arrays entirely 08:31:15 maybe stack elements could contain lists? 08:32:00 the problem with arrays as they are defined in forth is that they are kept in a memory area separated from the stack 08:32:06 and i want everything in stack 08:33:20 yeah 08:33:43 but that's for the sake of speed and ease of implementation 08:35:11 keeping arrays in the stack would be quite difficult anyway 08:35:19 since if you would push an array 08:35:34 then a drop would have to remove the entire array and not just the top element 08:36:06 in other words, the array would have to be treated somehow as a single object 08:36:34 probably using lists as stack elements is the best way 08:37:02 --- quit: swsch (Read error: 104 (Connection reset by peer)) 08:37:18 yeah, that'd be difficult to do 08:37:53 --- join: swsch (~stefan@p5091E67A.dip.t-dialin.net) joined #forth 08:42:43 --- quit: onetom (Read error: 113 (No route to host)) 08:44:16 --- join: onetom (~tom@cab.bio.u-szeged.hu) joined #forth 08:49:04 --- join: qFox (C00K13S@82-169-140-229-mx.xdsl.tiscali.nl) joined #forth 08:50:57 --- join: Teratogen (~leontopod@intertwingled.net) joined #forth 09:02:01 --- part: snowrichard left #forth 09:46:02 --- join: arke (f2@bespin.org) joined #forth 09:46:11 hello 09:46:25 greetings 11:18:27 --- log: started forth/04.12.21 11:18:27 --- join: clog (nef@bespin.org) joined #forth 11:18:27 --- topic: 'SIB. Small Is Beautiful.' 11:18:27 --- topic: set by thinfu on [Sun Dec 19 17:10:12 2004] 11:18:27 --- names: list (clog saon Teratogen qFox swsch Herkamire yumehito Topaz aum madgarden ianp hrmpf I440r robert warpzero Fractal cmeme skylan holden @crc) 11:19:44 * Herkamire lives quite happily without flash support 11:20:05 flash is evil 11:20:40 if i want a movie, i'll get an mpeg or something 11:21:08 Right 11:21:29 Or the Forth source of an ASCII movie. 11:21:35 Oh, imagine that- MPEG banner ads with live audio. How much more annoying than Flash would that be? 11:22:10 holden: You wouldn't actually show them in the browser. :) 11:22:32 robert, funny you mention that 11:22:48 What, is that your latest project? 11:23:14 i'm starting to think of using aalib as my display system on the new text-only box i made 11:23:42 all i need is vim, lynx, irssi, and screen to get the true machine power! 11:23:58 or ed if i'm really adventurous 11:24:02 * robert pets links. 11:25:02 i could use the blinkenlights software too i guess 12:03:25 --- quit: saon ("Lost terminal") 12:03:38 --- join: saon (1000@c-24-129-90-197.se.client2.attbi.com) joined #forth 12:04:52 saon how'd you make your text-only box? 12:05:08 what sort of display is it? 12:05:55 oh, it's just freebsd 12:06:00 no x installed 12:06:19 --- join: bbls (~bbls@80.97.121.133) joined #forth 12:06:34 hello 12:09:48 Hi 12:11:56 hi robert 12:12:20 i've spoken with snowrichard before about a problem related to recursivity 12:12:50 basicly i want do design a forth flavour and i preffer not to have recursivity 12:13:11 therefore there should be a conversion layer that will transform all recursive calls into non recursive ones 12:13:39 now, in a normal language i would use an explicit stack 12:14:16 this stack would store data that normally would be local variables defined in the recursion function 12:14:57 but since forth uses the stack directly, how can i transorm a forth recursive call into a non-recursive one? 12:16:13 lol 12:16:22 Herkamire ? 12:16:51 bbls: how about you just don't define RECURSE 12:17:21 users probably should be allowed to use recursivity, just that it shouldn't exist in the internal representation of the program 12:17:44 if you don't like recursion then don't use it 12:17:53 no, i like the recursion 12:18:07 ? 12:18:16 just that i want to add some static analisys to the compiler, that would be much easly implemented if there would be no recursion 12:18:16 then why are you trying to restrict the forth to make it impossible? 12:18:38 no, i don't want to restrict it, just transform all recursive definitions into non-recursive ones 12:18:46 in a way similar to that used in other languages 12:19:04 what's the difference? 12:19:06 (using an explicit stack, like in backtracking without explicit recursive calls) 12:19:19 you want to eliminate tail-recursion? 12:19:27 no 12:19:30 you don't understand 12:19:31 so all calls (recursive or otherwise) push the rstack 12:19:55 do you know how to transform a generic recursive function in C into a non-recursive one? 12:20:07 (or in Pascal or other mainstream language) 12:20:26 you mean use a loop? 12:20:38 yes 12:21:16 this is not always possible (without making arrays or something) 12:21:19 any function can be transformed in a mechanical way 12:21:32 using an array that will hold the "local variables" 12:21:38 this array will act as a stack 12:21:46 that will emulate the recursive calls 12:21:46 what's the point of all this? 12:21:59 why emulate them instead of just calling them? 12:22:16 because the static analyser would be much more easly to implement 12:22:36 if there would be no recursion 12:22:57 what's the analyser do? 12:23:22 --- join: Sonarman (~snofs@ppp-66-124-254-107.dsl.snfc21.pacbell.net) joined #forth 12:24:00 that's more complicated to explain 12:24:09 basicly i'll have a predefined word 12:24:11 called assert 12:24:20 it will expect a boolean value in stack 12:24:34 if the value is false, then the program execution should be interrupted 12:24:51 now the analyser should check to see (as much as possible) that no assertion could ever fail 12:25:38 ahh 12:25:57 part of the reason I don't follow, is that these should be (at least funcionally) equivilent: 12:26:05 : foo ... tail-recurse ; 12:26:10 : foo begin ... again ; 12:26:28