00:00:00 --- log: started forth/02.08.14 02:25:09 --- join: Serg_penguin (~Z@nat-ch1.nat.comex.ru) joined #forth 02:25:31 --- part: Serg_penguin left #forth 02:58:39 --- quit: Fractal (Read error: 110 (Connection timed out)) 04:11:56 --- join: Serg_penguin (~Z@nat-ch1.nat.comex.ru) joined #forth 04:12:45 --- part: Serg_penguin left #forth 05:13:58 --- join: Serg_penguin (~Z@nat-ch1.nat.comex.ru) joined #forth 05:14:28 --- quit: Serg_penguin (Client Quit) 05:39:33 --- join: Serg_penguin (~Z@nat-ch1.nat.comex.ru) joined #forth 05:40:21 --- quit: Serg_penguin (Client Quit) 06:01:35 --- quit: ChanServ (Shutting Down) 06:01:52 --- join: ChanServ (ChanServ@services.) joined #forth 06:01:52 --- mode: bear.openprojects.net set +o ChanServ 06:01:52 --- mode: ChanServ set +l 83 06:03:34 --- join: cleverdra (julianf@0-1pool36-188.nas2.florence1.sc.us.da.qwest.net) joined #forth 06:22:17 --- join: proteusguy (~irc@24-197-147-197.charterga.net) joined #forth 06:36:36 --- join: I440r (~mark4@1Cust88.tnt2.bloomington.in.da.uu.net) joined #forth 06:36:59 hello I440r. 06:37:06 hi :) 06:37:15 u seen my assembler so far ? 06:37:35 its still a work in progress and thers some things i need to figure out still but its a start :) 06:38:00 im going thru the opcode tables (intel dox) and sorting them into a more logical order 06:38:38 No. (How could I have seen it?) OK. 06:39:09 hang on :) 06:39:36 67.241.42.88 06:39:38 ftp to that address 06:39:43 i think its still in there 06:39:45 asm.f 06:40:03 ive done a LITTLE more work since then but that will show you what im doing 06:44:00 * cleverdra nods. 06:44:19 it all revolves arround a; 06:44:29 well 06:44:39 i had to do some trickery with overrides and prefixes tho 06:44:43 that was sort of annoying 07:01:04 --- join: proteus (~irc@24-197-147-197.charterga.net) joined #forth 07:01:56 hi proteus 07:08:56 --- quit: proteusguy (Read error: 110 (Connection timed out)) 07:12:16 --- quit: proteus () 07:12:17 --- join: proteusguy (~irc@24-197-147-197.charterga.net) joined #forth 07:12:27 I440r - hello. 07:13:08 :) 07:13:20 im procrastinating working on my assembler 07:13:43 My irc keeps losing its connection but the client doesn't notice so I just quite getting messages without warning... 07:13:45 i dont think its going to be a very good assembler actually 07:13:58 what client ? 07:14:05 mIRC under win32. 07:14:15 aha 07:14:31 i think mirc is one of (if not the) best irc clients there is 07:14:35 Do you know if the client or the server provides the ping cause I don't seem to be getting any and I recall that I used to when I used irc way back when... 07:14:39 pity there isnt a linux version 07:14:51 pity it can't keep me connected... 07:17:08 --- quit: I440r () 07:17:50 * cleverdra boggles that I440r thinks that. 07:18:07 --- join: I440r (~mark4@1Cust88.tnt2.bloomington.in.da.uu.net) joined #forth 07:23:26 --- join: kc5tja (~kc5tja@ip68-8-206-226.sd.sd.cox.net) joined #forth 07:23:54 welcome back, I440r. Good morning, kc5tja 07:24:31 :) 07:24:43 methunks im lagged 07:24:43 too 07:25:22 You're lagged. Badly. 07:25:23 kc5 07:25:38 15.135 seconds 07:25:41 dec rebister is 1111 111w 11 001 reg 07:25:43 I have a 3 second lag myself. 07:25:51 but thers an alternate encoding 07:26:00 1111 111w: mod 001 rm 07:26:19 erm no 07:26:26 It's actually the same encoding if you look at it carefully enough. :) 07:26:29 0100 1 reg 07:26:29 i mean 07:26:41 whers teh w bit in the 1 byte opcode for dec reg ? 07:26:48 There isn't one. 07:27:22 IIRC, you have 0100 0rrr and 0100 1rrr; I think bit 3 is the 'w' bit. 07:27:37 then how do you know if its (e)bx or bl your decrementing ? 07:27:40 Personally, I don't support the one-byte encoding. 07:27:57 40 == DEC AL, 48 == DEC AX 07:29:02 then for this opcode bit 3 is the w bit 07:29:27 If memory serves me correctly, yes. 07:29:28 typical intel, consistantly inconsistant :P 07:29:31 What's the single byte opcode for INC? 07:29:36 inconsistEnt ? 07:29:36 errr 07:29:38 me cant spellz 07:29:46 40 might be INC instructions. 07:29:53 Shit, I can't remember. 07:30:01 of corse, the documentation doesnt tell you that 07:30:04 0100 0 reg 07:30:09 OK, scratch that. 07:30:16 si no. bit 3 is the direction bit 07:30:16 40 = INC AX, 48 = DEC AX 07:30:21 Right. 07:30:29 I knew it looked wrong, but I couldn't put my finger on it. 07:30:29 rite 07:30:32 I just woke up. 07:31:11 maybe the alternate only works on 8 bit registers 07:31:38 I only supported the 2-byte form because the 1-byte forms are not supported in x86-64. 07:32:28 rite, i had already decided not to support them, they arent fully documented for a start 07:32:31 0100 xxxx is used as an instruction prefix in 64-bit mode for selecting the r8-r15 (yes, it has 16 registers! AT LAST!), and a couple of other things. 07:32:46 also, dec memory doesnt show the case where a SIB is needed 07:33:02 None of the instructions actually do. 07:33:05 in fact ALOT of teh instruction encodings dont show the case where a sib is needed 07:33:09 It's assumed based on the modr/m byte. 07:33:14 inc mem only shows the modrm case 07:33:39 which doesnt help me 07:34:20 It's an extension of the modr/m byte, and it is consistent (the only thing consistent about IA-32), so I have to agree that there's no real need to. 07:35:02 theres way too much thats either NOT documented at all or VERY BADLY documented here by intel 07:35:14 and its not making my attempt to write an assembler any easier 07:35:49 actually part of the problem is that i cant visualize the sib 07:36:04 is a sib only needed if you do scaling ? 07:36:20 If you use more than one register in an effective address calculation too. 07:36:24 or if you use TWO registers to get effective address 07:36:29 x[EAX] requires only a modr/m byte 07:36:32 [eax+ebx] 07:36:34 where ebx COULD be scaled or not 07:36:36 x[EAX+EBX] requires an SIB. 07:37:00 thats what i thunked 07:37:50 are there any cases where SIB encoding is illegal? 07:37:55 someone tries to 07:38:07 XXX [eax] + [ebx] 07:38:19 where xxx is some opcode or other 07:38:28 like jmp [eax+ebx] 07:38:35 --- join: Speuler (l@62.206.56.189) joined #forth 07:38:36 cept i know thats legal 07:38:37 No, it's perfectly valid there. 07:38:44 but are there any cases where its not legal? 07:38:49 I don't know. 07:38:50 g'day 07:38:59 i know it is there heh 07:39:08 i USE it in isforth :) 07:39:08 hi bongo 07:39:19 I don't care either. Again, if the user wants to shoot himself in the foot, let him. 07:39:33 hi markIV 07:39:40 rite heh 07:40:13 i wish they would order this opcode table with similar opcodes near each other instead of sorting on mneumonic 07:40:38 functional grouping ... 07:40:46 would save me the trouble of trying to do it myself 07:40:51 its a fucking bitch, i KNNOW im going to make mistakes 07:41:10 I wish they would just combine all the different instruction *forms*, then provide an indication for each instruction which form it adheres to. 07:42:01 kc5tja: wouldn't that require a more or less orthogonal instruction set where intel is notorably bad at ? 07:42:24 i dont like the mov /r or mov Eb,Eg style of showing teh opcode tables 07:42:24 i cant fucking read them hehe 07:42:29 they make no sense 07:42:49 Speuler: No, not necessarily. 07:43:05 But having an orthogonal instruction set is of course very helpful for that. 07:43:15 they dont help me to construct an opcode based on mneumonic etc 07:43:23 * kc5tja likes the 0-operand stack processor's instruction set: it only has 1 opcode form. :) 07:43:29 fucking google pisses me off 07:43:41 if your on a 33.6 like me and your downloading something google refuses to search for you 07:43:57 it times out in like 3 seconds flat and disconencts you 07:44:24 Timeouts are normally browser-induced things. 07:44:26 kc5tja: i'd consider ( x -- y ) as a different opcode form than ( x -- ) 07:44:35 Speuler: It's not. 07:44:46 That just specifies what it takes on the data stack, and what it produces. 07:44:57 The opcode encoding is still the same as every other instruction. 07:45:08 true 07:45:51 The only abnormality in a 0-operand stack instruction set is # or LIT, which must push an immediate value onto the stack. 07:46:33 or words, parsing input stream ? 07:46:34 But in the latest generation of stack CPUs, even that is made more regular by making explicit the (PC)+ addressing mode. 07:46:45 The CPU doesn't parse an input stream. 07:46:53 The software that runs on it does. 07:46:58 And what happens in software is anybody's guess. 07:48:18 you could work around literal, using a word, producing an address, containing the literal 07:49:34 That reduces your precision, so it'll take multiple instructions to produce a full-width integer. 07:49:43 i think markIV did something like that to his literals (they are represented as constants) 07:50:04 Many RISC processors do something similar. 07:50:22 as an option 07:50:31 They have a dedicated class of instructions that just loads a small immediate value into a register. 07:50:37 And another instruction to load the high-bits. 07:50:39 a const is like a constant but its state smart, it compiles a literal 07:51:06 Stack CPUs can do the same, I suppose, but if it uses a packed instruction layout (e.g., 5 instructions per memory word), then the (PC)+ method becomes more efficient. 07:51:08 iirc, novix had an opcode for literals 0...7 07:52:00 i.e. 8 opcodes, differing by bits 0:2 07:52:04 Right. 07:52:53 fucking idiots showing multi byte opcodes as "single byte opcodes" 07:53:05 every fucking opcode is a one byte opcode if you dont count teh modrm dammit 07:53:08 in general, other literals you might want to represent as constants anyway 07:53:41 That depends on what it is you're compiling for. 07:53:53 except with some twisted algorithms, using "magic" numbers 07:54:07 Define magic here? 07:54:17 like, determining day of week from a date 07:54:36 If you're targeting a native code compiler, CONSTANTs will actually slow you down (unless they're state-smart, like I440r's) 07:54:37 ... 185729749 3659276 */ .... 07:55:11 hmmm... overhead involved with literals should be comparable to that of constants 07:55:21 Those I definitely would treat as symbolic constants; remembering the function of those numbers is hard without having names for them. 07:55:58 Again, it depends on the processor architecture and the method of compilation, and how smart the compiler is. 07:56:03 you may have 5 or more of those "magic" numbers in a short algorithm. 07:56:22 A little RNG, for instance. 07:56:52 i'd tend to put literals into those algorithms, and blackbox it 07:57:16 For RNG, yeah, you don't need to know the purpose of the number. 07:57:19 i.e. label it "don't touch it, or you'll break it" 07:57:26 For computing weekday from a date, however, it's critical. 07:57:31 No. 07:57:32 Flat out no. 07:57:43 I've had to maintain code with exactly those types of comments in it. 07:57:55 I will never do so again. 07:57:57 well, don't touch it :) 07:58:06 it says so in the manual 07:58:08 You make it sound Oh So Peachy Keen. 07:58:15 The real world doesn't work that way 07:58:20 Sometimes, you HAVE to touch it. 07:58:26 or replace it 07:58:30 When it's broken, when you think it might be broken, the comment telling you only that you'll break it is no help at all. 07:59:07 why fix it if you can replace it against something better ? 07:59:08 One of those needs-total-rewrite-for-small-change designs. 07:59:23 Speuler: Because it might be the best solution already, if it weren't buggy. 07:59:31 putting lits in constants would make it worse 07:59:42 Speuler: Or maybe I'm trying to certify that it's already not buggy. 07:59:48 good idea 08:00:05 No, it wouldn't. 08:00:23 Speuler - how would that make it worse? 08:00:24 : ... [ DaysPerWeek WeeksPerYear * ... + ] LITERAL ... ; 08:00:27 some people would claim that reality is generated by the power of mind 08:00:56 kc5tja: no... those "magic" numbers do NOT represent DaysPerWeek or WeeksPerYear 08:01:09 they are just chosen so the alg produces the correct result 08:01:28 Speuler: See? If you'd used a meaningful constant naem for those magic numbers, I wouldn't have made the assumption/error. 08:01:43 but there's hardly a way to see WHY those numbers have that particular value 08:01:58 Speuler - are you referring to the kind of magic that a numerical analysist might make? Thresholds for approximation and such? 08:01:59 However, I'm willing to bet that each of those numbers is a composite of days/week and week/years.... 08:02:32 In my experience, there is always a reason for everything. 08:02:54 Magic numbers like that don't exist by pure happenstance. 08:03:12 oh sure, there may be a reason, but the reason may be "the result is only correct if you use exactly these values" 08:03:27 0001 constant foo-threshold \ iterate until iterations differ by .0001 08:03:49 Then, again, I'd make use of Forth's ability to interpret while compiling to get LITERAL-performance with symbolic naming maintainability. 08:03:54 [ foo bar baz * + ] LITERAL 08:04:10 Or, if it's long, and repititious, I'd do something like this: 08:04:17 : blort foo bar baz * + ; 08:04:17 That's probably the wrong thing to do, there -- a numerical analyst might work on your algorithm and then create a magic number -- the comment and the word would be the same, but the number no more comprehensible. 08:04:23 ... [ blort ] LITERAL ... 08:06:01 At any rate... 08:09:22 --- quit: proteusguy (Connection timed out) 08:11:13 * I440r is getting annoyed again with these opcode tables from intel grr 08:13:42 * kc5tja never used the opcode tables. 08:13:58 I just went through chapter 3, page by page, implementing the instructions I felt was convenient to have. :) 08:14:01 * I440r is starting to understand why 08:14:14 i want ALLLLLLLLL instructions available 08:14:21 including mmx, fpu and sse 08:14:29 I didn't bother with those. 08:14:31 I have no need for them. 08:14:36 but these tables are totally fucked up 08:14:38 Maybe later. 08:14:50 they are making life very difficult for me 08:15:11 that and teh fact that the more i get into this the less i understand.... 08:15:30 the more i get frustrated 08:15:42 the less likely it is that im going to complete the assembler 08:16:30 anyone knowing the "lewis carrol" dow calculation ? 08:16:48 Examples 08:16:49 1783, September 18 08:16:51 If you want, you can forget for now that opcodes bear any relation to each other and just implement them individually. Factor as you notice the need; start moving in-core CODE words to forthsrc words using your assembler... 08:16:57 Speuler: Nope; but I'm not sure I want to know with a name like that. :) 08:16:57 17, divided by 4, leaves "1" over; 1 from 3 gives "2"; twice 2 is "4". 08:16:57 83 is 6 dozen and 11, giving 17; plus 2 gives 19, i.e. (dividing by 7) "5". Total 9, i.e. 08:17:07 "2" 08:17:07 The item for August is "8 from 10", i.e. "2"; so, for September, it is "2 plus 3", i.e. "5" Total 7, i.e. "0", which goes out. 08:17:07 18 gives "4". Answer, "Thursday". 08:17:40 no magic numbers :) 08:17:53 clever thats not an option 08:17:55 Speuler - no, I don't. The DOWs that I see tend to pick an known date and then calculate by that. 08:18:05 http://acadprojwww.wlu.edu/vol4/BlackmerH/public_html/xliberty/history/carroll.html 08:18:15 I440r - why not? Eventually you'll get everything, but you'll get everything without choking it all down at once. 08:18:23 im not going to do one opcode for mov ax,bx and another opcode for mov ax, cx and another opcode for mov ax, dx 08:18:23 etc 08:18:38 cleverdra: that's the most intuitive thing to do, but there are shorter and quicker ways to do that 08:18:54 I440r - that's not I meant :-/ but nevermind. 08:19:07 And that algorithm just flat out does not make sense to me, and I would never consider using it in an actual product. 08:19:30 cleverdra k 08:19:42 Not without more explanation of what is going on inside the algorithm. 08:20:02 im just getting more pissed off with the whole fucking assembler, i REALY never wanted to write it in the first place 08:20:04 i tend to run a comparitive test of the whole date-range the alg is supposed to function when i examine a new one 08:20:24 but ive got no choice because theres not a single other person who even gives a fuck about coding isforth 08:20:26 if it yields the same results, it supposedly will do that whenever it is called 08:20:53 Besides, 19/5 is not 5. 08:20:54 :) 08:21:07 I can't prove it'll yield the same results. 08:22:13 (If the underlying arithmetic yields the same results given the same arguments, I don't see how a self-enclosed algorithm like that couldn't yield the same results given the same arguments.) 08:22:20 thats 19 7 mod 08:22:41 But the text said explicitly (ie., dividing by 5) 08:22:54 So, I assumed, you divided by five. 08:23:11 An algorithm is not supposed to be black magic. 08:23:27 Every step is documented; it HAS to be documented, in order to arrive at a good implementation. 08:23:45 like, genetic code ? 08:24:01 Yes, and look how long it's taking us to learn what the hell it does? :D 08:24:06 have the machine develop their own alg by selection 08:24:23 and nobody knows why the alg works ... 08:24:23 Sorry, but I refuse to follow that path as well. 08:24:52 * cleverdra . o O ( neural-net training ) 08:24:56 well, you eat, without knowing in detail how digestion works 08:25:01 And if you're referring to genetic algorithms, you'll note that the algorithm is constant, only the inputs to the algorithm change. 08:25:15 not necessarily 08:25:18 But I don't make or maintain humans for a living! 08:25:29 you eat for a living 08:25:33 No I don't. 08:25:37 Speuler - no you don't. 08:25:42 you eat for fun ? 08:25:45 I eat to survive, but I certainly don't get paid to eat. 08:25:52 In fact, I usually pay other people to eat. 08:25:54 ah 08:27:18 And in fact, as a person with irritable bowel syndrome, I really wish I didn't have to eat at all. It's quite a chore for me. 08:27:47 take the example of programmable logic, task: implement a low-freq oscillator. 08:28:41 there's a case of a design, where the circuit "designed itself" 08:28:51 Says who? 08:28:53 nobody knew how it worked 08:29:14 By what logic are you thinking this? 08:29:21 you know its difficult to implement low-freq oscillators just with gate logic ? 08:29:38 Yes, so I don't use them. 08:30:02 Alternatively, use very long strings of inverters. 08:30:22 and some chains of flip-flops 08:30:25 (e.g., to accumulate the propegation delays into a sufficiently long pulse duration for oscillation) 08:30:52 freq dividers would be legal too i figure 08:30:57 Gate logic implies purely combinatorial circuits. Flip-flops makes it sequential. 08:31:01 That's what I'd use. 08:31:16 High frequency base oscillator with a divide-by-N circuit to produce the low-frequency output. 08:31:38 a long chain of inverters could qualify as sequential too 08:31:51 Not strictly speaking, no. 08:31:54 It's still combinatorial. 08:32:08 after all, you use the accumulated propagation delay as the timing base 08:32:11 * kc5tja works for a chip design house... :) 08:32:41 it would be purely combinational with ideal propagation delay=0 08:32:44 The Digital Electronics industry has rather precise meanings for the words sequential and combinatorial. 08:33:03 We know and understand that nothing has a 0 propegation delay, so we have to draw the line somewhere. 08:33:26 Memory elements makes the circuit sequential (e.g., a state machine); lack of memory elements makes it combinatorial (data-flow) 08:33:59 In fact, it's true also in the mathematical sense as well. 08:34:16 Combinatorial circuits can be modeled as sets of combinators which operate on incoming data. 08:34:27 I'm almost positive, come to think of it, that that's where the name combinatorial came from. 08:36:22 What I hate more than anything else in the world, is a roommate who eats all of the food you've paid for. 08:36:26 * kc5tja sighs 08:36:31 Well, I guess I'll get ready for work now. 08:36:33 --- nick: kc5tja -> kc-shower 08:38:11 http://www.interlog.com/~r937/doomsday.html 08:39:01 short and pretty: http://www.ffmusiconline.com/FF10/D1/19-Sight_of_Spira.mid 08:50:14 hi 08:57:40 --- nick: kc-shower -> kc5tja 09:16:27 cleverdra: cant understand what is this doomsday thing 09:16:47 cleverdra: would u b so kind as 2 explain me what is that good 4? 09:18:27 --- quit: clog (^C) 09:18:27 --- log: stopped forth/02.08.14 09:19:40 --- log: started forth/02.08.14 09:19:40 --- join: clog (nef@bespin.org) joined #forth 09:19:40 --- topic: 'http://www.msmisp.com/futuretest/Forth's_Dilemma.htm | Forth: using bits without limits | x86 Linux Forth coded in asm - http://isforth.clss.net | home of forth - http://www.ultratechnology.com' 09:19:40 --- topic: set by futhin on [Thu Aug 08 18:57:06 2002] 09:19:40 --- names: list (clog Speuler kc5tja I440r cleverdra @ChanServ Robert onetom goshawk` sif) 09:25:10 wb clog :) 09:26:12 onetom - doomsday thing? 09:26:34 onetom - That's Speuler's. 09:29:54 true. hi 1tom. that's one way to determine day of week 09:30:20 was just searching for unusual ways to do so 09:30:51 methods which you can use to do the calculation mentally 09:31:25 my previous formula only worked till 1999 09:37:19 hmmm strange things r these 2 me 09:37:41 why was it just good till 1999? 09:37:55 what has happened after it? 09:38:04 --- part: kc5tja left #forth 09:38:09 or IN that year? 09:45:36 worked until dec31,1999, then calc'd the wrong dow 09:45:45 was a simple formula 09:49:05 --- join: proteusguy (~irc@24-197-147-197.charterga.net) joined #forth 09:50:33 oh-bringer-of-the-light-whose-liver-was-mutilated-so-brutally 09:51:29 fire steal if liver eagle feed then 09:54:55 : feed ( food feedie -- ) begin >r grow r> over reasonable u> if 2dup ! then again ; 09:55:54 : grow 1+ ; 09:56:03 :) 09:56:48 * Speuler forgot to shrink liver after ! 10:01:26 --- join: proteus (~irc@24-197-147-197.charterga.net) joined #forth 10:09:07 --- quit: proteusguy (Connection timed out) 10:23:30 --- join: kc5tja (~kc5tja@user-24-214-86-42.knology.net) joined #forth 10:27:54 --- quit: cleverdra ("Leaving") 10:36:01 --- quit: Robert (Read error: 104 (Connection reset by peer)) 10:36:27 --- join: Robert (~Robert@robost86.tsps1.freenet6.net) joined #forth 10:59:41 --- join: proteusguy (~irc@24-197-147-197.charterga.net) joined #forth 11:06:42 * kc5tja sighs 11:08:48 --- quit: proteus (Read error: 110 (Connection timed out)) 11:21:31 Hmm... 11:21:46 * kc5tja ponders how to handle the case where, in a paned GUI environment, you run completely out of screen space. 11:57:17 --- join: futhin (thin@h24-64-175-61.cg.shawcable.net) joined #forth 12:08:53 --- quit: proteusguy (Read error: 110 (Connection timed out)) 12:12:44 Besides, I don't think I could reach you anyway. 12:12:46 grr 12:19:08 hm?? 12:19:29 kc: what did you want to talk about last nite? 12:20:48 Not right now. 12:20:53 k 12:20:57 I'm still a bit hot-headed from a drag-out fight in #hamradio 12:25:30 ah :) 12:30:27 --- join: proteusguy (~irc@24-197-147-197.charterga.net) joined #forth 12:44:07 futhin: hi 12:45:04 kc5tja: use tabs! or scrollable panel 12:47:16 onetom: ah? you said something interesting today? :P 12:48:23 onetom: i looked in the log, didn't see anything ?? 12:49:17 He was responding to my rhetorical question regarding paned user interfaces. 12:49:39 er 12:49:40 I have four console windows open in Win2K right now, each doing a different task, but I need to see all of them concurrently. 12:49:41 doh 12:49:55 "kc5tja:" != "futhin:" 12:49:58 heh 12:50:11 I've arranged them to cover the screen in the proportions I need then, emulating a paned UI. 12:50:29 yeah, i want a 40" monitor too 12:50:31 But I ran out of screen space for my SSH window and e-mail program. 12:50:42 As I went searching high and low for P6 opcodes, I kept coming across people from Intel saying that there were only two new opcodes in the P6. So I tried and I tried to count up the opcodes that I know about, to see if 1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 = 2. No matter how I tried to accumulate it, CMOV + FCMOVB + FCMOVE + FCMOVBE + FCMOVU + FCMOVNB + FCMOVNE + FCMOVNBE + FCMOVNU + FCOMI + FCOMIP + FUCOMI + FUCOMIP + RDPMC + 12:51:14 i440r: it cuts off at: RDPMC + 12:51:14 CMOV was introduced with the Pentium. 12:51:28 + INT01 + SALC + UD2 didn't add up to 2. Now I see the FDIV bug in a completely different way. 12:51:33 Ditto for RDPMC and its friends. 12:51:46 SALC has been in the instruction set since the 386. 12:51:55 UD2 has always existed, just never documented. 12:51:55 yes but not documented heheh 12:52:17 UD stands for undocumented :) 12:52:22 Actually, SALC might even date all the way back to 8086, now that I think about it. 12:52:32 What I want to know is, where's UD1? :) 12:53:34 hah, PTSC sells $1,000 development board for use with the ShBoom processor they sell 12:53:53 kc5 ya good point hehe 12:54:06 it must still be a BIG secret :) 12:54:57 Yeah. Nobody ever really uses. 12:54:59 it. 12:55:16 must suck pretty bad then? 12:55:27 how many instructions are there on the P5 anyways? 12:55:29 100? 12:55:45 More than that. 12:56:14 does anybody look at the pentium chip and just laugh their ass off when mentally comparing it to forthchips or mips/68k processor or whatever 12:56:41 futhin: i got me a psc1000 (shboom core) board+cpu for about 200 $ ... 12:58:28 speuler: i emailed PTSC and they sell the IGNITE/shboom cpu for $15 12:58:37 + shipping and handling 12:59:00 speuler: have you coded for your psc1000? do anything useful with it? :) 12:59:29 not really, just toyed around with 12:59:59 thought about putting it into a small case for portable use 13:00:13 but can't remember where i've put it :( 13:00:48 speuler: kc5tja and I talked about a forthchip desktop computer awhile ago. we were talking about 25X forthchips, but it would cost about $100 bucks or more per chip.. however, since i discovered that ShBoom chips are 15 bucks, it might be possible to build a desktop computer with some 13:01:13 you lost your pcs1000? :( 13:01:40 ShBoom & F21 seem to be the only forthchips/stack-based machines available for purchase on the market 13:01:41 i should be somewhere ... t'is not a huge board ... 13:02:01 cpu clocked @ 80 MHz (internal) 13:02:11 performance is not outrageous 13:02:18 80 mhz = 80 mips i think 13:02:25 mem clock 20 MHz 13:02:28 1 mips for each mhz 13:02:36 for the shboom chip 13:02:37 yes, about 13:02:56 and, it gets hot 13:02:57 That's true if you execute one instruction per clock cycle. I thought the ShBoom had some instructions which could effectively execute in parallel? 13:03:39 well, not really extremely hot, but noticable. theres a passive cooler on the chip 13:03:45 kc5tja: i dunno, a short pdf detailing the features of ShBoom is available here: http://64.186.225.66/download/IGNITE_Processor_ASIC__Data_Sheet.pdf 13:03:58 (very short, only 2 pages pdf) 13:04:14 nope. none that i've seen. reads 32 bit (4 instructions) per memory cycle, and 13:04:59 because internal clock is 4x, it executes those 4 instructions until it is ready for the next instructions from next memory cycle 13:05:16 one by one 13:05:17 speuler: the pdf says it operates between 225 mhz (worst case) and 360 mhz (typical) 13:05:25 how come it's only 80 mhz for yours? 13:05:34 (maybe they updated it i guess) 13:05:40 psc1000 ? 13:05:45 or another ? 13:06:19 futhin: ive just browsed ~guest/www/test/netcat/scripts 13:06:25 that's 225 mhz internal or external clock ? 13:06:32 it that part of the netcat distribution? 13:06:35 pdf says: Power/performance factor is 4545 MIPS per Watt :) 13:06:41 onetom: yes 13:06:51 hmm... thats cool 13:07:00 speuler: go to the pdf.. :P 13:07:08 psc1000 core is ShBoom-II core 13:07:15 iirc 13:07:22 how come i (and any of my more experienced friends) wasnt able 2 mention it b4 13:07:39 u can learn an awfully lot from those lil scripts 13:07:59 spueler: IGNITE isn't shboom anymore.. it's based on ShBoom.. probably ShBoom-III renamed to IGNITE 13:08:10 ah 13:08:16 onetom: the scripts aren't forth .. 13:08:43 IGNITE seems to be used widely in embedded applications 13:08:43 * Speuler purchased psc1000 because i couldn't manage to get me a f21 sent over 13:08:50 tried thrice 13:08:57 no reaction at all 13:09:03 gave up then 13:09:03 you emailed jeff fox? 13:09:10 --- quit: proteusguy (Connection timed out) 13:09:10 called tings dad 13:09:21 oh 13:09:40 or was it the i21 13:09:42 not sure 13:11:22 speuler: if you still want to buy f21 go here: http://www.ultratechnology.com/store.htm scroll down to F21d prototype chips i think 13:11:35 they are about 100 bucks or 150, i forget 13:11:46 I thought they had them for $27 13:11:57 (which is still very expensive for a CPU that small) 13:12:07 futhin: jeff told me he still got a couple for seriously interested people 13:12:20 i'm not sure whether i qualify :) 13:12:42 i have no real application for it at the moment 13:12:46 kc5tja: eh?? i thought you told me the f21 chips were 100 bucks 13:13:06 jeff fox only has a couple f21 chips left ?? :( 13:13:11 That's the price Jeff told me they were, but if you look on his site, they're only like $27. I really don't know which figure to believe. 13:13:35 It may also have to do with who is fabbing the chips too. 13:13:46 At any rate, I need to get some food. 13:13:48 I'll be back. 13:13:49 --- nick: kc5tja -> kc-food 13:14:06 kc-food: give up, your friend has eaten all already 13:15:01 kc-food: i don't see any price listed for f21 on the store.html part.. 13:19:58 --- join: jamc (~dne@as3-6-8.asp.s.bonet.se) joined #forth 13:21:52 --- nick: Speuler -> CaffeineJunkie 13:50:20 --- join: tcn (tcn@tc2-login10.megatrondata.com) joined #forth 13:51:39 --- nick: kc-food -> kc5tja 14:06:57 --- join: geakazoid (JB@adsl-63-202-179-17.dsl.snfc21.pacbell.net) joined #forth 14:07:08 Hi there :) 14:07:53 --- quit: CaffeineJunkie ("using sirc version 2.211+KSIRC/1.1") 14:08:49 hi geak 14:08:54 * kc5tja sighs 14:09:07 kc5tja: What's up? 14:09:11 Why so sad? 14:09:29 Because I want to beat the living piss out of someone. 14:09:30 hello onetom and all 14:09:39 geakazoid: http://hermantom.homeip.net/~tom/forth/tile-forth-2.1.tar.gz 14:09:58 geakazoid: get it now! while the connection is working :) 14:10:01 I've just been a mean old man these past few days, and nothing I do vents it! 14:10:55 it is dl'g now, what is it? 14:13:31 geakazoid: ? the TILE forth we r recommending beside gforth & bigforth 14:13:59 I'm also getting very upset at these tests that I'm working on, because they keep failing for bizarre reasons. Apply one bandaid and another fault appears. 14:15:11 geakazoid: w lots of very didactic examples and useful libraries 14:15:56 are you sure that was the correct url, I am getting nothing 14:16:13 wait a sec 14:17:04 and that looks like a Debian server, does that mean that TILE only runs on Debian; I have mandrake 14:19:13 sure, it should work on any *nix 14:19:26 onetom ok, what about the dol'g of the file? 14:19:30 cant understand, i also get permisson denied... 14:19:41 dol'g? 14:20:02 OH MY! onetom asks about an abbreviation! :-) 14:20:10 nevemind, ftp it! 14:20:12 kc5tja: :p 14:20:24 lol 14:20:45 kc5tja: uknow, i use mines. that doesnt mean i know all. it wouldn b ez 14:20:51 * kc5tja now enters the food psychosis. 14:21:44 geakazoid: ftp://guest:x@hermantom.homeip.net/home/tom/www/forth/tile-forth-2.1.tar.gz 14:23:19 * kc5tja is too tired even to get a cup of water. :( 14:27:57 onetom maybe because I am on windose I am not able to ftp to it. my linux is not on the net 14:28:19 shouldn't matter if you are on windows 14:28:39 works for me.. did you click on the link and let i.e download it ? 14:28:44 geakazoid: Are you behind a firewall? 14:28:45 futhin I have tried ncftp, ftp, and the MSIE 14:32:09 http://hermantom.homeip.net/~tom/forth/tile-forth-2.1.tar.gz 14:34:04 works for me.. so it's weird if it doesn't work for you 14:34:52 hmm... great :) 14:34:56 ftp://ftp.taygeta.com/pub/Forth/Compilers/native/unix/tile-forth-2.1.tar.z 14:35:03 or did I miss something? :) 14:35:12 heh 14:35:17 that works too 14:35:29 jamc: no, but that has some "bugs" 14:36:04 jamc ok, that works 14:36:07 u have 2 add/adjust an include and delete 1 signal (a line) from a table 14:36:12 ok, then why does the fixed one have the same name? 14:36:23 :) 14:36:31 mine is gz not z ... 14:36:42 otherwise im not sure thats the corrected 1 :) 14:37:41 well well... last time I looked at tile (ages ago) I didn't like it much 14:38:00 why? 14:38:25 don't remember... must've been in '94 or '95 or something :) 14:38:40 amen 14:39:03 onetom: Are you the author of TILE Forth? 14:40:02 ? 14:40:25 kc5tja: certainly not. im just a fan of it, coz ive learnt a lot from it 14:40:26 kc5tja: no 14:40:53 kc5tja: so why do u ask? 14:47:14 * geakazoid is gone, autoaway/10m (l!on) 14:50:54 --- join: proteusguy (~irc@24-197-147-197.charterga.net) joined #forth 15:03:03 --- join: Fractal (dubqg@24.77.171.228) joined #forth 15:08:30 --- quit: tcn ("Leaving") 15:44:38 --- quit: I440r () 15:57:28 --- quit: jamc ("[x]chat") 16:07:02 --- join: proteus (~irc@24-197-147-197.charterga.net) joined #forth 16:10:02 --- quit: proteusguy (Read error: 110 (Connection timed out)) 16:12:33 --- join: I440r (~mark4@1Cust88.tnt2.bloomington.in.da.uu.net) joined #forth 16:12:53 what are "gdi resources" in windows? 16:13:13 Anything you can see on the screen is a GDI resource. 16:13:27 well ive got almost NONE left 16:13:32 Icons, colors, fonts, etc. 16:13:35 and ive got 256 megs of ram in this box 16:13:46 What version of Windows? 16:14:13 98 se 16:14:36 The amount of RAM doesn't matter for GDI resources -- they're allocated in the kernel's address space. 16:14:37 1024x768 truecolor 16:14:47 However, you have a program which is leaking resources profusely. 16:14:53 The only way to recover them is to reboot, really. 16:15:39 probably opera 16:15:45 i440r: naww 16:16:06 i440r: hit ctrl-alt-del. and shutdown EVERYTHING except: explorer, systray, trillian, mirc, opera 16:16:21 What is trillian? 16:16:24 er trillian is a pile of crap. i deleted it 16:16:39 its a multi protocol chat program 16:16:45 I440r I use memokit, don't know what kc5tja uses, but I keep track of all the memory, and if an app is too resource consumptive, I uninstall it: I no longer have to reboot a few times a day 16:17:03 * kc5tja doesn't use anything. 16:17:10 When I see my resources going out the window, I reboot. 16:17:27 kc5tja my machine is our LAN server, that is not an option 16:17:28 This is true even in Linux. 16:17:29 i see it as soon as i reboot 16:17:35 could it be a virus ? 16:17:42 Could be. 16:17:48 Or it could be a false alarm too. 16:17:58 Is the system complaining about low resources? 16:17:58 I440r yes and yes 16:18:02 bah, i still use trillain, its smaller than icq & msn & aim & yim, so i might as well use it 16:19:48 i installed norman anti-virus but its not flagged anything as being infected 16:19:59 norman anti-virus is TBAV baught out 16:20:11 and tbav is the absolute BEST anit virus ever 16:20:17 You mean Norton? :) 16:20:40 no 16:20:41 norman 16:20:41 ummm 16:20:52 look up thunderbyte anti viurs 16:21:00 norman baught them out 16:21:48 I440r did you clean out your cookies? 16:21:57 theres a group of ppl who rate anit virus programs and they REFUSED to rate TBAV because it never missed a single virus they fed it 16:22:01 you can have an adware eating you up 16:22:05 im always cleaning out my cookies 16:22:07 i haven't used virus software for 5 years now 16:22:17 however, i doubt TBAV is the best 16:22:22 i forget my favorite.. 16:22:33 tbav can decompress/decrypt code by single stepping through it and then do huristic analysys of the code to determin if its a virus or not 16:22:40 futhin wow, I use AVG the free edition 16:22:42 so can norman av now 16:22:43 my favorite virus scanner is obviously the best ;P 16:22:59 i forget the name.. fscan? 16:23:03 grr.. 16:23:28 futhin so without a virus checker, how is your pc running online? 16:24:08 viruses are myths 16:24:10 i never run anti virus, only reason i installed this is cuz i think i might have one 16:24:11 i never get them 16:24:20 and i run win 98 16:24:26 futhin for 5 years i activly sought out viruses 16:24:32 i TRIED to be infected 16:24:41 i never once seen an in the wild virus 16:24:45 the only way ppl get viruses is thru email or thru disks 16:24:49 i440r: heh :) 16:25:02 futhin you do not get email? 16:25:07 yeah i do 16:25:13 nobody sends me viruses tho 16:25:16 a good way to not get viruses is to not use microsoft lookout express 16:25:38 I440r although you may be right, it is the easiest to use 16:25:52 geek hell no 16:25:55 its horrible 16:25:56 geakazoid: bullshit 16:26:06 ive seen very very few microsoft apps that i liked 16:26:06 geakazoid: you only say that cause it's the "easiest" to install 16:26:21 it's not hard to download a free email program that is easy to use 16:26:51 futhin yes, it is already there, I have tried PINE and I use JBMail, but other people who share this machine are not programmers and cannot get used to what they see as archaic 16:27:05 there are HUNDREDS of email programs 16:27:09 opera just performed an illegal operation and now i have 78% gdi free 16:27:27 there is definitely one email program out there suitable for your needs 16:27:29 I440r I am no MS fan, believe me if I could switch all the machines to a custom self-efficient OS I would have done it long ago 16:27:50 geek - linux :) 16:28:02 futhin yeah but the email programs are either quirky if they are free or shareware 16:28:04 geakazoid: have you tried eudora mail? (there's an older free version) perhaps try pegasus 16:28:25 ah hah, that applies to outlook "quirky if free" 16:28:31 futhin yes I tried eudora and immediately I had some problems withh it 16:28:34 I440r: I seem to recall that 50% to 75% is normal for a modern Windows environment. 16:28:50 If it gets above 80 to 85%, that's when I'd worry. 16:28:57 geakazoid: what kind of problems? which version? the older version was better i thought 16:29:01 netscapes mail client is the best ive used, its real good if you can stand the fact that its netscape 16:29:19 opera's email client is useable 16:29:24 their news client ISNT 16:29:40 futhin it was about a year ago, so I cannot recall. This machine is about a year and a half old and the prupose was that it would be on dsl and that caused a ton of issues 16:30:10 geakazoid: it takes awhile to find a good email client.. because there are so many. avoid getting lazy and you'll find a good one 16:30:14 i never really found a good one 16:30:26 i use yahoo mail 16:30:29 I440r yes, there are issues with netscape because they are a painful group of programmers 16:30:32 so i just end up getting on the website all the time 16:30:59 geekazoid netscape USED to be realy good. that was before AOL got their hands on it 16:31:11 now its buggy pile of crap spy-ware 16:31:12 futhin now I have been working on setting up a Eserv, a forth server that has an email server 16:32:17 geakazoid: does it run ontop of an os or just a forth system? 16:32:22 like enth or something ? 16:36:19 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK") 17:08:54 --- quit: proteus (Connection timed out) 17:25:04 futhin Eserv? runs on Windows 17:26:17 futhin this machine has dsl internally and I needed a server to run on it. I chose Eserv because it is written in forth and I can write forth plug-ins 17:26:54 --- join: TheBlueWizard (TheBlueWiz@ip-216-25-205-132.vienna.va.fcc.net) joined #forth 17:26:58 hiya all 17:31:05 hi 17:32:47 futhin (et al...) here is our server http://azedia.sytes.net/ 17:33:16 hiya geakazoid 17:33:48 TheBlueWizard how's it 17:34:57 I just checked http://azedia.sytes.net/ and got a URL not found message 17:35:33 (or server not found, for that matter) 17:35:41 TheBlueWizard ok, I have been having that problem and have been trying to fix it 17:36:11 then why advertise the nonexistent website? 17:37:00 TheBlueWizard this is a new server that I have been working on for a few months and I am trying to test it 17:39:22 TheBlueWizard ok, try it again 17:39:30 hmm...are you getting some help already, or are you trying to fix it on your own? 17:40:10 TheBlueWizard I am the admin, I have been setting it up for a few months. It works on our LAN but I needed to have it work over the net 17:40:40 I got an "Action Cancelled" message (I think it refused the request) 17:42:42 TheBlueWizard I received some issues about my password. I need to login with the right psswd I guess 17:43:06 what setup are you running? 17:43:19 TheBlueWizard that must have been one of the main issues I had not noticed 17:43:41 TheBlueWizard an Eserv HTTP/FTP/Email server 17:44:08 hmm...never heard of it 17:44:38 TheBlueWizard http://www.eserv.ru/eserv/ 17:46:53 I see...no wonder I never have heard of it 17:47:04 sorry I can't help you there 17:48:04 TheBlueWizard the only thing I need is to make sure that someone on the net can view it on http, after that it is set up 17:49:37 TheBlueWizard apparenlty I cannot recall the passwd 17:52:11 TheBlueWizard well, I reset the passwd anyway, so can you try it again? 17:52:29 you said you have been working on for a few months...sounds like you have forgotten some of what you've done n months ago and when that question shows up that calls for something you've done back then...ouch! 17:52:57 TheBlueWizard well, that is true; that always happens when one is rotating projects 17:53:11 TheBlueWizard but I reset it so I do not have to worry about it now 17:53:34 it reports a 404 error 17:53:37 here is the url again http://azedia.sytes.net/ 17:55:02 TheBlueWizard hrm... for a month, I thought it was the firewall not port forwarding and I tried a few different firewalls but that was not it, I think 17:55:26 now it was only the passwd and that is not it, so I am not sure that I have the port set right 17:56:16 looks like you have a lot of troubleshooting to do... 17:57:55 well, gotta go....bye all! 17:58:13 TheBlueWizard see ya 17:58:27 bye geakazoid 17:58:31 --- part: TheBlueWizard left #forth 19:00:07 I have to sign off 19:00:27 --- quit: geakazoid ("Leaving... ") 20:32:12 --- quit: I440r () 20:44:39 sad sad sad asdfdfadfakdf;adfjdaksfjasd;fkjdsfkljsdklfssadasad 20:51:39 --- join: JosephSpiros (~Snak@1Cust136.tnt3.oberlin.oh.da.uu.net) joined #forth 20:51:50 --- part: JosephSpiros left #forth 22:19:41 --- join: kc5tja (~kc5tja@ip68-8-206-226.sd.sd.cox.net) joined #forth 22:21:26 hello kc5tja 22:21:37 rere 22:22:01 are you free to talk about ShBoom forthchip desktop computers? :( 22:22:20 Sure, for the time being. 22:22:48 have you gotten a chance to check the specs? 22:22:55 apparently they are about 350 mips 22:23:00 for 15 bucks 22:23:13 and lets assume that it's 1 instruction per cycle 22:23:17 No, I didn't -- I couldn't even find them except for the document link you sent me. 22:23:19 but it fetches 4 instructions at a time 22:23:35 the ASIC pdf had the specs.. 22:23:51 Right, so it's really 1.25 cycles per instruction. Still not at all bad considering there is no pipeline in it. 22:24:28 want me to provide the ASIC pdf link again? 22:25:02 heh, too late: http://64.186.225.66/download/IGNITE_Processor_ASIC__Data_Sheet.pdf 22:25:17 hrm.. 22:25:40 so the ShBoom might not be that useful for video card stuff.. 22:25:56 but for an initial desktop computer.. ? 22:29:12 Not for video, but as the central CPU... 22:29:50 what part of the specs make it not good for video? no analog i/o? 22:30:54 Lack of dedicated video support. 22:31:20 i'm not sure what that means? 22:31:35 does 25X have dedicated video support or only F21 ? 22:31:45 Only the F21 and P21 does. 22:32:03 ah 22:32:19 what kind of dedicated video support? 22:32:52 It has a dedicated video processor where each instruction either spits out a pixel value or a (horizontal/vertical) sync. 22:33:15 hmm 22:33:33 do you know if P21's are available for purchase? :) 22:33:55 I don't think so. But the F21s are apparently still in limited supply. 22:35:02 so if it is unfeasible to acquire p21s and f21s for video card stuff, what to do? how many shbooms would we want for the desktop computer ? 22:35:33 I'd be happy implementing a video processor/device using an FPGA of some sort. 22:36:03 you said previously something about NTSC video.. would that be done with the fpc? 22:36:13 fpc? 22:36:14 s/fpc/fc 22:36:17 forth computer 22:36:17 :P 22:36:22 forth PC 22:36:24 heh 22:36:49 um.. NTSC video is basically outputing to TV right ? 22:36:59 Well, if we use one dedicated CPU just for video, we can have "software generated video." This is, actually, how the Atari 2600 worked, except that it used only one processor for both the games and the video display. 22:37:24 a processor for the games ? 22:37:39 or do you just mean the central processor running the games also handled the video display ? 22:37:52 No; the Atari 2600 had only one 6502 CPU, which did both video refresh and game logic. 22:41:55 what kind of FPGA for the video? are they cheap enough? 22:42:23 is it better to do an forthchip ontop of an FPGA instead of a dedicated chip like 6502 or whatever is good ? 22:42:28 I really can't answer any of those questions. I'm not at all familiar with FPGAs, how to program them, or their prices. I know only the most rudimentary stuff about them. 22:43:01 It always will be; 6502 will steal a lot of cycles because you're emulating a stack architecture on an accumulator architecture. 22:44:10 i'm discussing what we were discussing previously, creating forthchip desktop computers that would be sold to third-world countries and other places.. 22:44:22 but because 25X would be too costly to invest in initially 22:44:39 it would be better to start of with ShBoom (15 bucks) 22:44:41 what do you think? 22:45:21 for the video card, it doesn't have to be a forthchip.. 22:45:47 No, it doesn't. I mean, it's nice if it is, but it's really not necessary. 22:45:50 then later on, the computers can use 25X and F21s and stuff.. 22:46:33 do you think there is a good chip for the video card that doesn't consume too much power ? 22:47:05 brb 22:47:24 ok 22:52:15 back 22:52:36 wb 22:52:58 i'm reading an article about internet appliances heh 22:53:22 Heh 22:54:52 kc5tja: what do you think of the whole thing.. starting up a company to create forthchip desktop computers starting off with ShBoom chips and later upgrading to 25X and F21s when the money is available? 22:55:46 If we have the money to upgrade to 25xs or F21s, we'll be able to afford our own chip designs. 22:57:08 sure, they would be similar to 25X or whatever.. we'd hire chuck moore ;P 22:57:25 wouldn't we? aren't his chips going to be the best? ;P 22:57:39 in his email to me about forthchips 22:57:46 Generally speaking, I suppose. :) 22:57:57 he said he would customize the 25X for whatever features that would actually cause people to buy them 22:58:07 like analog i/o, or whatever 23:00:55 I envision one 25X for video, one for audio, one for generic peripheral support (read, USB, legacy RS-232, legacy IEEE-1248-compatible parallel ports, timers, etc) 23:02:32 cut'n'paste of chuck moore's response to me: 23:02:33 The trick is to get a large number of Forth coders to agree on anything. Any real product must appeal to more than the Forth market. 23:02:33 But to answer your question. Using a conservative process, I could get 25 chips for maybe $5,000. That would cost out at $200/chip. Without paying for my time, which I've about given up charging for. 23:02:33 The next step is a wafer run (5 wafers) that would produce 20,000 chips for maybe $60,000. Or $3/chip. This is much more attractive, but requires $60,000 and a customer for 20,000 chips. 23:02:33 I'm pursuing several paths that might lead to chips. The question then becomes, exactly what capabilities does the chip have? Several computers? Analog I/O? 23:02:36 One path would be to ask Forthers to prepurchase a specific chip and development board to finance their development. What do you think is possible? 23:04:56 He hit the nail on the head. If we do make such a device, we must make sure to not publicize Forth in the initial press copy. It has too much of a negative stigma. 23:07:16 agreed 23:08:00 but that's not how i interpreted that sentence :P 23:08:14 i interpreted it as "it must be useful to more than the forth market" 23:08:27 to more than forth hobbyists :P 23:09:12 Exactly. Forth is just a means to an end, not to be promoted as a fad. 23:11:24 The biggest problem I find with forth is it's complexity. Most forthers seem to pride themselves on the language's simplicity... ANS Forth has hundreds of words in the core set, C has maybe 20 keywords... 23:11:33 well i emailed him just now, asking if that $60,000 for 20,000 chips includes packaging or not.. i just want to make _absolutely_ sure 23:11:55 20 keywords, but a near infinite array of possible expression forms. 23:12:17 As does forth. 23:12:22 Remember a C program describes a highly abstract, near mathematical form of a program. Forth is much lower level. And if you think ANSI Forth is bad, try ANSI Lisp sometime. 23:12:22 As does any language./ 23:13:01 No, Forth is quite different. For starters, it's not applicative -- it's not mathematically founded. However, it's interesting that it is actually mathematically provably correct language. 23:13:04 Common LISP isn't that bad actually. 23:13:09 fractal: forth comes across more complex than it is.. annoying.. i think that forth should be learned from the ground up.. start off with understanding the virtual machine, the data stack and return stack, and the coreset vocabulary (27 words or so..) 23:13:26 Second, Forth is very imperative, while C is, like Fortran, a model of an arithmetic expression system. 23:13:52 No, but it definitely has a lot more than 200 functions in its core functionality. 23:14:20 My point is that the size of the language's core vocabulary/library doesn't affect how complex a language is to learn, only to implement. 23:15:18 I disagree. In languages like forth and lisp, where words are built upon eachother, knowing the base words is essential. 23:15:44 The problem is, ANSI Forth's core words aren't guaranteed to be based on other core words. 23:16:09 is knowing the asm for the primitives essential? ;) 23:16:18 For example, WORD and PARSE rely on string manipulation functionality that is simply non-existant in the CORE wordset. 23:16:52 futhin : No, primitives ought to be implicitly obvious. 23:17:44 Fractal: I think what you wrote is critically true, and is the major complaint that Chuck Moore has against ANSI Forth. 80% of the words that sit in the CORE wordset are superfulous and represent nothing else except common usage. 23:18:14 Agreed. Forth could definitley do with a redesign. 23:18:23 Housecleaning anyways. 23:18:57 agreed 23:19:09 Well, FS/Forth is decidedly not ANSI compliant. :) But it's more compliant than, say, isForth. I remain ANSI compatibility only where it makes sense to do so. Otherwise, I have no complaints with ripping out the fluff and replacing it with something that works better. 23:19:48 i've been thinking that forth needs to be learned from the ground up.. learn the basic words, keep the whole thing simple 23:19:59 3 looping constructs??? 23:20:30 I guess that isn't too bad, but it's more than's needed. 23:20:35 --- quit: kc5tja (bear.openprojects.net irc.openprojects.net) 23:20:35 --- quit: goshawk` (bear.openprojects.net irc.openprojects.net) 23:20:35 --- quit: Fractal (bear.openprojects.net irc.openprojects.net) 23:20:53 --- join: kc5tja (~kc5tja@ip68-8-206-226.sd.sd.cox.net) joined #forth 23:20:53 --- join: Fractal (dubqg@24.77.171.228) joined #forth 23:20:53 --- join: goshawk` (goshawk@panix1.panix.com) joined #forth 23:21:21 heh 23:21:32 --- quit: goshawk` (bear.openprojects.net irc.openprojects.net) 23:21:32 --- quit: kc5tja (bear.openprojects.net irc.openprojects.net) 23:21:32 --- quit: Fractal (bear.openprojects.net irc.openprojects.net) 23:21:41 --- join: kc5tja (~kc5tja@ip68-8-206-226.sd.sd.cox.net) joined #forth 23:21:41 --- join: Fractal (dubqg@24.77.171.228) joined #forth 23:21:41 --- join: goshawk` (goshawk@panix1.panix.com) joined #forth 23:22:21 I'm not sure. I don't think assembly language is essential. Definitley the basic ideas behind assembly should be learnt, but it's largely machine dependant. 23:23:01 I think CS should concentrate more on algorithms and data structures early on. The coding will seem obvious if these concepts are understood. 23:24:18 Yeah, but I mean as a first actual *language*, assembly language is the thing to start off with. 23:24:48 WIthout assembly level knowledge of what's going on inside the computer, using a language like C is harder than it should be. Consider how hard it is to teach someone the concept of pointers to a new C student! 23:25:33 Maybe... But there's simply too much crap to deal with in assembly. Stuff that doesn't really make one inkling of difference in understanding the computer. 23:25:43 "Assembly Language Step-By-Step" by Jeff Duntemann (edition 2) was an EXCELLENT book.. covered the assembly background/theory thoroughly, and really did not get into assembler specifics.. it was beautiful and easy to read.. i read the whole book without needing to touch the computer (did all the testing in my head heh) 23:26:00 --- quit: goshawk` (bear.openprojects.net irc.openprojects.net) 23:26:00 --- quit: kc5tja (bear.openprojects.net irc.openprojects.net) 23:26:00 --- quit: Fractal (bear.openprojects.net irc.openprojects.net) 23:26:13 --- join: kc5tja (~kc5tja@ip68-8-206-226.sd.sd.cox.net) joined #forth 23:26:13 --- join: Fractal (dubqg@24.77.171.228) joined #forth 23:26:13 --- join: goshawk` (goshawk@panix1.panix.com) joined #forth 23:26:35 --- quit: goshawk` (bear.openprojects.net irc.openprojects.net) 23:26:35 --- quit: kc5tja (bear.openprojects.net irc.openprojects.net) 23:26:35 --- quit: Fractal (bear.openprojects.net irc.openprojects.net) 23:26:44 --- join: kc5tja (~kc5tja@ip68-8-206-226.sd.sd.cox.net) joined #forth 23:26:44 --- join: Fractal (dubqg@24.77.171.228) joined #forth 23:26:44 --- join: goshawk` (goshawk@panix1.panix.com) joined #forth 23:26:53 ewww 23:26:56 I actually think C is a fairly decent first language (although I know I'll get flamed on #forth for saying that). It's sufficiently low level to illustrate memory managment, etc. 23:29:33 when C coders try to learn Forth, they find it forth to be very complex.. 23:29:44 but i think that C coders forgot the learning curve they went thru to learn C 23:30:44 C isn't really necessary to learn.. why not just get people to learn asm and lisp? :) 23:31:12 ASM, then Forth, then Lisp. That's all you need to teach. Everything else falls in place from there. 23:31:33 I look at C as ASM without all the crap. 23:31:44 Portable ASM without all the crap. 23:31:56 I still think C hides too much from the newbie programmer though. 23:32:14 /me has the whole "I learned it the hard way" attitude. :) 23:32:16 forth is much more like a high-level asm than C is 23:32:45 forth hides a lot less.. the stack is a reflection of how asm is like 23:33:28 C isn't actually portable. it's only "portable" because C compilers have been laborously coded for each architecture 23:33:54 forth can be coded for an architecture quite rapidly.. it could be as small as 1 or 2k 23:33:56 Well, obviously. Forth is no different there. 23:34:11 well i would argue forth is the _most_ portable language out there 23:34:20 especially since it is a natural virtual machine 23:34:29 It's been ported from everything from 4-bit CPUs to 64-bit. :) 23:34:36 Forth has the potential to be portable, we'll just leave it at that. 23:34:54 hehe 23:34:54 Forth code has the potential to be portable, I mean. 23:35:03 I would say that C hasn't ever been used with a 4-bit CPU before, but I'd be lying: there exists a crippled C compiler that supports HP calculators, whose processor is only 4-bits. :) 23:36:03 kc5tja: any point in learning lisp? 23:36:43 Definitely. 23:36:45 Definitley. Lisp is the only language that is more simplisticly elegant than forth that I know of. 23:37:03 Although kc5tja is right. Common Lisp is ridiculously bloated. 23:37:24 more elegant than forth? 23:37:35 I'd say so. 23:38:04 As forth is built around 2 stacks, lisp is built around lists. 23:38:14 car and cdr pff :P 23:38:17 jk 23:38:19 Your program is a list of instructions, which each can be a list of instructions, etc. 23:38:25 * futhin is just about to go to bed ;P 23:39:15 lisp code makes me think of forth except its sort of reversed 23:39:28 my friends were saying you have to think recursively to code in lisp 23:39:31 but somehow i don't believe them 23:39:33 Lisp places its emphasis on correctness, while Forth places its emphasis on pragmatics. 23:39:53 That's really the only real difference between the two philosopheis. They're otherwise very similar languages. 23:40:04 --- quit: clog (^C) 23:40:04 --- log: stopped forth/02.08.14 23:40:18 --- log: started forth/02.08.14 23:40:18 --- join: clog (nef@bespin.org) joined #forth 23:40:18 --- topic: 'http://www.msmisp.com/futuretest/Forth's_Dilemma.htm | Forth: using bits without limits | x86 Linux Forth coded in asm - http://isforth.clss.net | home of forth - http://www.ultratechnology.com' 23:40:18 --- topic: set by futhin on [Thu Aug 08 18:57:06 2002] 23:40:18 --- names: list (clog sif onetom Robert futhin @ChanServ kc5tja Fractal goshawk`) 23:40:19 Lisp definitley encourages recurisve code, if for no other reason than the clunky looping constructs. :) 23:40:21 i think that i only have to think Forthish to code in lisp, rather than thinking recursively 23:40:35 (repeated for clog's benefit heh) 23:40:48 well ok, good night all 23:40:50 ttyl 23:40:59 Night. 23:41:28 --- quit: futhin ("sleep") 23:59:59 --- log: ended forth/02.08.14