00:00:01 --- log: started forth/04.03.02 00:00:01 If the Rabbit 2000 is pin compatible with the Z-80, then it needs to maintain the same bus interface, and that means, 4 clocks *per memory cycle.* 00:00:12 no 00:00:12 The 6502, OTOH, has precisely *one* clock per memory cycle. 00:00:22 its not pin to pin compatible 00:00:23 Which is quite largely why it's so fast. 00:00:37 OK, let me ask this then: does it have /RD and /WR signals? 00:00:47 The 6502 has A0-A15, D0-D7, R/W, and CLK. That's it. 00:00:48 the memory access scheme is the same 00:00:56 Then it still needs 4 clocks per memory cycle. 00:01:02 at 25mhz ;) 00:01:22 for 3 bucks in 1000 order 00:01:22 Run that chip at 16MHz, and pit it against a 6502 at 16MHz. 00:01:27 And then we'll see who wins. 00:01:48 Remember: the higher the clock speed, the more CMOS transistors suck power. 00:02:14 So the more work you can get done with a lower clock, the more power efficient it is. 00:02:23 ok, but it runs on 2,8v 00:02:32 The 65816 runs at 1.5V. Your point? 00:02:39 1.5v ??? 00:02:42 Yep. 00:02:47 1.5V to 5.0V. 00:02:55 Pure static CMOS logic. 00:03:01 4000 gates. 00:03:03 Have a nice day. :) 00:03:07 http://lists.trolltech.com/qt-interest/1996-10/thread00134-0.html 00:03:17 i´ll research this... 00:03:56 http://www.rabbitsemiconductor.com/ <- please, look 00:04:48 rabbit 3000 runs on 1.5 to 5.0b 00:04:50 ops 00:04:52 5.0v 00:05:31 kc5tja: 1.5V at what clock speed? 00:05:51 they abandoned their low cost parts. shame :/ 00:06:29 As far as I recall, the 1.5V clock speed can be as high as 14MHz -- it's what the chip is rated for. 00:07:09 Higher voltages can go faster, but not by much. At 5V, the 65816 is still rated for only 14MHz, but some have taken it as high as 20MHz (see Commodore 64 attachment SuperCPU for example). 00:07:53 Another great reason to use the 65816: $6 in quantities of 1. :) 00:08:03 (roughly the same price as a 68000) 00:08:24 the problem (back in the 1980) to get faster cpus was on the memory speed 00:08:31 * kc5tja nods 00:08:42 my father owned a Apple II and got a speed upgrade for it 00:08:56 It still is an issue with the 65816, which is why it's speed limited to 14MHz to 16MHz "officially." 00:09:24 Frankly, it was a *very* hard decision for me to make: 68000 or 65816. 00:09:34 The 68000 is ==such== a nice microprocessor. 00:09:41 Bar none, it's my all-time favorite CISC processor in the world. 00:09:55 And it's easily available, relatively cheap, and can be driven to respectable speeds. 00:10:01 What computers is it used in? 00:10:05 mac 00:10:06 The deficiency with it is the fact that it's a 64-pin DIP package. 00:10:41 Wasn't early Suns running 68K? 00:10:45 Robert: Macintosh, Atari ST, Amiga, and almost *every* Unix workstation (well, 68010 and later at least) before RISC technology became fully developed. 00:10:57 Oh. What's so great about it? 00:11:19 68000 is (to the programmer viewpoint at least) much better than 8086/80286 00:11:38 For what reasons? 00:11:47 16 registers (8 data registers, 8 address registers), powerful addressing modes, very fast instruction execution rates, etc. 00:11:53 is 68000 free of damn "segments" ? 00:12:04 For the love of God, there is NOTHING WRONG with segments. 00:12:05 and much better protected memory model 00:12:26 (better than 80286 wich is crappy) 00:12:27 * kc5tja is so freakin' pissed off at everyone who condemns the concept. 00:12:47 That being said, the 68000 has a flat address space. 00:12:54 And its registers are true 32-bit registers. 00:13:04 i wonder why PC, not Amiga became hit ^( 00:13:22 kc5tja: Does it have 16 32-bit registers? 00:13:22 i wonder why VHS not Betamax :( 00:13:26 Robert: Yes. 00:13:55 8 32-bit data registers (D0-D7) and 8 32-bit address registers (A0-A7, A7 used as hardware stack pointer, but A0-A6 can all be stack pointers too). 00:14:05 That's really neat... 00:14:17 nice ;)) 00:14:28 all 8 have equal rights ? 00:14:33 The big problem I have with using the 68000 in my design is that there are basically 16+24 pins to have to worry about. :/ 00:14:36 Serg: Yes. 00:14:49 kc5tja: kool ! 00:14:55 The 68000 was heralded as a mainframe on a chip for many years. 00:15:04 I guess it doesn't support floating point numbers, does it? 00:15:10 The 68000 is about as powerful as an 80386 running in flat, 32-bit protected mode. 00:15:11 no 00:15:23 Robert: Built-in FPU support came with the 68040. 00:15:26 there was a FP unit for it 00:15:29 x86 regs are braintwist 00:15:38 kc5tja: And no memory protection? 00:15:50 Robert: Support for external MMUs came with the 68010. 00:16:01 However, the software supported the concept of memory protection since the 68000. 00:16:35 and remmeber he was to compete with 8086 :P 00:16:40 Hehe. 00:16:42 The 68000, for example, supports two modes of operation: User mode (where applications run) and supervisor mode (where an OS kernel runs). 00:16:58 i just wonder why 8086 won ;((( 00:17:03 $$$ 00:17:06 I B M 00:17:07 :D 00:17:17 Serg: Because Intel and IBM were bed-buddies back then (not today), and they cross-invested in each other heavily. 00:17:42 In fact, IBM deliberately chose inferior technology to power their IBM PC because they were afraid their PC would eat into mainframe sales (which, actually, to this day, has not happened!). 00:18:14 IBM did a lot of stupid decisions 00:18:23 Haha! 00:18:24 money spoil everything :((( 00:18:34 who really ended making lot of money from PC was microsoft and intel 00:18:55 And anti-virus companies. 00:19:00 hehe 00:19:05 and peter norton :P 00:19:17 * Serg really wants 64 forth CPU's on a chip clocked by 3GHz 00:19:48 * kc5tja doesn't want clocks on any of his chips, ideally. 00:20:03 That's the one thing that the 68000 has that I absolutely *LOVE* -- an asynchronous memory bus. 00:20:15 i want lots of 50 mhz cpus 00:20:20 kc5tja: so how things gonna be synchronized ? 00:20:24 in one computer :D 00:20:26 clustered 00:20:40 Serg: The same way two people synchronize against each other. 00:20:43 A: "You ready?" 00:20:47 B: "Yep." 00:20:48 cluster-on-die 00:20:55 A: "The answer is 42." 00:20:59 B: "Ok." 00:21:09 kc5tja: it sucks long time 00:21:16 Serg: You are wrong wrong wrong. 00:21:39 how hard would be to design a cluster of stack based microprocessors ? 00:21:40 The 68000 does a complete asynchronous bus transaction in four clocks (same as Z-80, same as 80x86, same as PowerPC, same as SPARC, same as . . .) 00:22:01 But, the 68000 can handle slow devices without batting an eye. 00:22:35 The Intel and Zilog parts, however, (and 65816 for that matter), need obtuse external logic explicitly to handle wait-state generation for slow devices. 00:22:35 kc5tja: ok 00:22:43 Makes interfacing arbitrary pieces of logic very difficult. 00:23:06 This is why my MISC processor designs will all have async buses. 00:23:17 16MHz is really too fast for a synchronous bus design to keep up. 00:23:37 Not without resorting to bizarre techniques like packet switching and routing, like Hypertransport does. 00:24:13 For the longest time, PCs had a 33MHz bus and that was *it*. :) 00:24:32 It wasn't until regular adoption of Hypertransport that we started to see 800MHz FSBs. 00:24:42 a cluster of stack based processores would be feasible ? 00:24:45 And the latency of such buses still leaves a lot to be desired. 00:24:56 to be used like a beowulf cluster 00:25:02 ? 00:25:07 AldoBR: A cluster can get aggregate processing bandwidth, but wouldn't address latency. 00:26:06 no, i´m just asking if its possible (and if we gain something) in designing a cluster of small stack based processors, in place of designing a system around just a big fast processor 00:26:30 Well, that depends on what kind of parallelization you're expecting to achieve. 00:26:39 Some programs parallelize better than others. 00:27:06 i think 3d games would benefict very much... 00:27:08 Or, if you're running a multitasking OS, you can run multiple programs across multiple CPUs on an as-needed basis. 00:27:17 and other kinds of simulation 00:27:18 Oh, well, yeah. That's all vector processing. :) 00:27:46 Vector processing can make huge performance gains without even multiple CPUs -- a simple dumb vector unit like that used in the Cray can be used for that. 00:27:54 But yes, a cluster would be great for that. 00:28:00 DSP is another great application. 00:28:16 It's possible to arrange the CPUs as a systolic array, which is nearly ideal for implementing software FIR filters. 00:28:20 just wondered about the day when i will be sitting in front of 1000 forth running microprocessors :P 00:29:01 Well, when the ForthBox II is released, we'll at least have a start... :) 00:31:06 See, now I'm leaning back towards the 68000 again. :) 00:31:26 Oh, but I know the CPU timing for the 65816, which means I can build the video hardware around it. 00:31:34 altivec... 00:31:49 But I'd have to verify the 68000's instruction timings so as to ensure the CPU can keep up with the video shift register speed. 00:31:59 jdrake: Altivec is quite limited. 00:32:13 but this dont eats too much processing power ? 00:32:33 Altivec can process vectors of length 2 (two 64-bit floating point values) while the Cray's vector processing unit can manage *64*. :) 00:32:38 how much power does a 65816 consume or a 68k? 00:32:56 I think the 65816 is MUCH less power draw than a 68000, but I'd have to verify. 00:32:59 kc5tja, one is a super computer, the other is not... 00:33:10 that would seem to indicate that 65816 might be a better choice 00:33:13 Plus the 65816 has an 8-bit data bus, so it's easier to wire into a circuit. 00:33:17 being that is one of your goal 00:33:46 however your indication of 68k higher speedness may out way that 00:33:48 One ROM chip, one RAM chip, versus two of each. 00:33:57 Yeah. 00:34:08 See, to stuff the video shift register, I'd need 20 instances of the following code: 00:34:12 you seem to know your stuff 00:34:14 LDA (LINEPTR),Y 00:34:17 STA VSR 00:34:19 INY 00:34:20 INY 00:34:25 NOP 00:34:39 while in the 68000, I can use 20 instances of this: 00:34:53 ERR, sorry, 10 instances of this: 00:35:07 MOVE.L (A0)+,D0 ; load D0 with next 32 bits of graphic data 00:35:15 MOVE.L D0,VSR ; Load VSR 00:35:21 and that's it. :) 00:35:56 They both take roughly the same amount of memory. The question is, how many clocks will it take? As long as the CPU completes one move faster than the shift register can keep up with, then I'm OK. 00:38:26 Basically, what it boils down to is I'm going to need to design two circuits on paper, add up their respective costs (both in terms of software overhead and in physical hardware component selection) and see what turns out to be cheaper overall. 00:38:51 kc5tja: you're purple 00:40:34 warpzero: That's what happens when you're as cool as I am. 00:40:39 Anyway, I have school tomorrow. 00:40:43 * kc5tja will get some rest now. 00:40:51 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK") 00:42:12 this quit msg is a forth program ? :P 00:42:46 i dont know 00:42:54 ive often wondered about it myself 00:52:57 Its radio abbrev. 00:53:42 KC5TJA is his call sign, THX is for "thanks" etc. 00:54:42 l33tspeak for morse code. 00:55:12 Try to telegraph his message without using the short forms, and you'll be wasting valuable hours. ;) 00:55:21 Anyway, time for school. 00:56:57 not for morse and short only, but for i18n and voice code too 00:57:07 --- quit: clog (^C) 00:57:07 --- log: stopped forth/04.03.02 00:57:13 --- log: started forth/04.03.02 00:57:13 --- join: clog (nef@bespin.org) joined #forth 00:57:13 --- topic: 'A channel dedicated to the Forth programming language, its implementation, its application, and its philosophy. kc5tja: http://216.239.57.104/search?q=cache:cWNnp2m5yzAJ:dec.bournemouth.ac.uk/forth/euro/ef99/chapyzhenka99.pdf+Transport+Triggered+Architecture&hl=en&ie=UTF-8' 00:57:13 --- topic: set by warpzero on [Tue Feb 10 23:23:11 2004] 00:57:13 --- names: list (clog chandler_ warpzero Robert hovil Serg jdrake AldoBR madgarden @ChanServ ASau slava scope_ skylan ianp cmeme mur fridge) 00:57:33 kilo-charly-five-tango-juliet-alfa ;) 00:58:02 good night/day 00:58:08 now i have to sleep 00:58:21 the word "location" is different in every language, but "QTH" is all the same ;) 00:58:26 g'nite 00:58:28 --- nick: AldoBR -> AldoBR[Away] 01:24:01 --- quit: Serg () 01:42:08 --- join: warp0b00 (~warpzero@dsl.142.mt.onewest.net) joined #forth 01:54:28 --- quit: warpzero (Read error: 110 (Connection timed out)) 02:12:20 --- quit: ASau () 02:22:16 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 03:23:15 --- quit: warp0b00 (Read error: 110 (Connection timed out)) 03:23:54 --- join: warpzero (~warpzero@dsl.142.mt.onewest.net) joined #forth 03:35:45 Serg: QSO = 'signing off' ? 04:37:24 --- nick: chandler_ -> chandler 05:03:36 --- quit: hovil ("Leaving") 06:12:17 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 07:01:37 --- quit: tathi ("leaving") 07:17:42 --- join: arke (~Chris@wbar8.lax1-4-11-099-104.dsl-verizon.net) joined #forth 07:22:11 --- quit: warpzero (Read error: 60 (Operation timed out)) 07:32:25 --- join: warpzero (~warpzero@dsl.142.mt.onewest.net) joined #forth 07:33:57 --- quit: AldoBR[Away] (Read error: 60 (Operation timed out)) 07:55:26 --- join: TreyB (~trey@cpe-66-87-192-27.tx.sprintbbd.net) joined #forth 08:08:49 --- join: Herkamire (stjohns@c-24-62-119-5.ne.client2.attbi.com) joined #forth 08:11:31 --- quit: warpzero (Read error: 110 (Connection timed out)) 08:13:25 --- join: warpzero (~warpzero@dsl.142.mt.onewest.net) joined #forth 09:02:46 --- join: networm_ (~networm@L0650P11.dipool.highway.telekom.at) joined #forth 09:33:49 --- join: ASau (~asau@158.250.48.196) joined #forth 09:33:57 Dobryjj vecher! 09:36:48 heh, a google search for dobryjj brings up logs of this channel :) 09:37:00 :D 09:39:01 --- join: Aldo[Away] (Aldo@200.195.110.69) joined #forth 09:39:16 I noticed that too. ;) 09:41:30 --- nick: Aldo[Away] -> AldoBR 09:44:54 --- join: niklaus (~niklaus@210-210-75-119.lan.sify.net) joined #forth 09:45:15 why isn't SP0 and SP@ same for stacks of size 1 and size 0 09:47:01 anyone there ? 09:47:09 Yes, we're just slow. 09:47:54 madgarden: SP0 and SP@ should be same for stacks of size 1 or 0 ? isn't it so 09:48:27 Which forth? 09:49:00 madgarden: do you have forth in your pc ? can you just see if sp0 and sp@ are same for stack of size 0 or 1 09:49:05 i am using gforth 09:49:29 what he means is that its probably implementation dependent 09:50:02 (i dont know what sp@/sp0 is supposed to do ;) 09:50:12 32 ok. 09:50:12 sp@ ok.. 09:50:12 . 1224572 ok. 09:50:12 . 32 ok 09:50:12 32 ok. 09:50:13 sp0 ok.. 09:50:14 me either. :) 09:50:15 . 1228716 ok. 09:50:17 . 32 ok 09:50:32 thats on winforth 09:50:35 sp0 and sp@ don't exist in FICL 09:50:58 what was that show cmd again? 09:51:00 to see a def 09:51:09 :) 09:51:21 see 09:51:22 right 09:51:57 well, for winforth, sp@ is code, and sp0 is forth 09:52:02 see sp0 09:52:03 4240884 -4 USER SP0 Value = 1224576 ok 09:52:08 see sp@ 09:52:09 SP@ IS CODE 09:52:09 4013B0 53 push ebx 09:52:09 4013B1 8BDC mov ebx, esp 09:52:09 4013B3 2BDF sub ebx, edi 09:55:09 qFox: i need some help 09:55:27 well i have my doubts about being the right person, but i'll try :) 09:55:32 qFox: don't put anything on stack. Run SP0 at startup and print the output by . 09:55:49 what is SP0 on an empty stack 09:55:51 sp@ . 1224576 ok 09:55:51 sp0 . 1228716 ok 09:56:05 qFox: how about with an element 09:56:07 say 2 09:56:28 2 3 ok.. 09:56:29 sp@ . 1224568 ok.. 09:56:29 sp0 . 1228716 ok.. 09:57:38 qFox: not 2 only 1 element like 2 or 1 or 3 not multiple elements 09:58:00 oh 09:58:00 heh 09:58:27 2 ok. 09:58:27 sp@ . 1224572 ok. 09:58:27 sp0 . 1228716 ok. 09:58:47 qFox: So they aren't same 09:58:53 obviously.. 09:59:01 qFox: They should be same 09:59:25 qFox: because SP0 is memory location of bottom of stack and SP@ is memory location of top of stack 09:59:26 they are defined differently to begin with.. 09:59:58 networm_, you should search for: "GOST 7.79-2000 Pravila transliteracii kirillovskogo pis'ma latinskim alfavitom" 10:00:06 no idea on that 10:00:10 networm_, http://www.orwell.ru/info/tliter 10:00:26 ASau: what are you talking 10:00:29 i suggest you see what works best to you 10:00:34 see = use 10:00:39 niklaus, they should no be the same. 10:01:09 ASau: why shouldn't they be same ? In an empty stack and top and bottom are same i guess 10:01:12 See nesting: COLD -> ( WARM -> ) ABORT -> QUIT 10:01:26 ASau:or a stack with one element 10:01:30 INTERPRET calls your SP@ 10:01:51 ASau: can you be a bit more elaborate ? 10:02:15 If you refer to your anti-standard you may see that noone guarantee anything. 10:02:47 As far as I know ANS 10:03:02 ASau: Can you please explain why shouldn't the memory location of top and bottom namely SP@ and SP0 not be the same when it has no elements or 1 element 10:03:35 it states nothing about stack usage by interpreter. 10:04:18 ASau: i see but then why can't they be same ? We understand that as a convention ? Don't we ? That stack top and bottom are same if it is empty 10:04:23 It may happen that you are not guaranteed of SP@ and SP0 presence at all. 10:04:54 is sp@ and sp0 ansi standard? 10:05:00 Ah! 10:05:00 i mean are they standard words? 10:05:09 I see your mistake. 10:05:19 You should use: S0 ? 10:05:21 not S0 . 10:05:28 sorry S0 10:05:40 that was a typo 10:06:02 oh right, to fetch the data, who both point to the same location? 10:06:03 qFox: sorry can you please try S0 10:06:14 nop 10:06:17 s0 10:06:17 Error: s0 is undefined 10:06:48 Here're my results: 10:06:50 S0 @ . 14096 OK 10:06:51 SP@ . 14096 OK 10:06:51 ASau: it doesn't work here with S0 and SP@ 10:07:01 ASau: which forth are you using ? 10:07:22 Consider it is FIG-Forth. 10:07:59 I started from it. 10:08:13 :) 10:08:16 ASau: thanks it worked. 10:08:21 well s0 is undefined here 10:08:34 qFox: it is actually S0 @ 10:08:49 um 10:08:52 still nothing :\ 10:09:04 and s0 @ are two commands.. 10:09:10 WORDS! i mean words! i really did :p 10:09:36 s0@ isnt a word either incase you meant that 10:09:53 s0 @ you have to use two wrods 10:10:05 aye, but my dictionairy doesnt know s0 10:10:14 so s0 and s0 @ wont change anything to the result 10:10:21 then try sp0 @ 10:10:33 qFox, uppercase. 10:10:46 i believe winforth isnt case sensitive, but i'll try 10:10:50 nope 10:10:55 sp0 works 10:10:58 qFox: did you try sp0 @ 10:11:12 i already pasted that above here, that sp0 works i mean 10:13:24 ASau: how long have you been working with forth 10:19:48 The earliest source I've found now is dated 25 May 2001. 10:20:26 But I can't state it is really the earliest. 10:42:43 --- quit: niklaus ("Client exiting") 10:45:25 --- join: imaginator (~George@georgeps.dsl.xmission.com) joined #forth 10:51:35 --- quit: Robert (Read error: 104 (Connection reset by peer)) 10:51:43 --- join: Robert (~snofs@c-8f5a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 11:14:34 --- quit: Robert (Remote closed the connection) 11:14:44 --- join: Robert (~snofs@c-8f5a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 11:24:39 --- join: Robert___ (~snofs@c-ef5a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 11:36:14 --- quit: Robert (Read error: 113 (No route to host)) 11:39:17 --- nick: Robert___ -> Robert 11:46:39 --- join: Robert___ (~snofs@c-775a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 11:52:43 --- join: zardon (~zardon@h24-68-59-145.gv.shawcable.net) joined #forth 11:54:52 --- quit: Robert___ (Remote closed the connection) 11:55:05 --- join: Robert___ (~snofs@c-775a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 11:57:47 --- join: scope__ (~junk@njd.paradise.net.nz) joined #forth 11:58:32 --- quit: scope_ (Read error: 104 (Connection reset by peer)) 11:59:42 --- quit: Robert (Read error: 110 (Connection timed out)) 12:02:27 --- nick: Robert___ -> Robert 12:03:26 --- join: scope_ (~junk@njd.paradise.net.nz) joined #forth 12:08:58 * warpzero is away: Sunny day, sweeping the clouds away, on my way to where the air is sweet! Can you tell me how to get, how to get to Sesame Street? 12:09:17 --- quit: scope__ (Read error: 60 (Operation timed out)) 12:09:46 --- join: Nutssh (~Foo@gh-1081.gh.rice.edu) joined #forth 12:16:04 --- part: imaginator left #forth 12:21:13 --- quit: Robert (Read error: 60 (Operation timed out)) 12:22:34 --- join: Robert (~snofs@c-775a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 12:25:58 good afternoon 12:28:01 Dobryjj vecher! 12:30:43 Dough-bridge? 12:31:09 What's "Dough-bridge"? 12:31:25 That's how "Dobryjj" sounds in my head when I read it. 12:31:29 Dough-bridge. 12:31:37 здравствулте! 12:31:38 Like, a bridge made of dough. 12:32:06 (that's what my translation bot spits out for !trans en|ru hello) :) 12:32:27 Eh! networm_, use plain Latin. 12:32:31 Please. 12:32:33 Hmm. Doesn't like my font. 12:32:42 I don't send Cyrillic here. 12:32:48 heh, you should all upgrade to unicode 12:33:05 Unicode is anti-Forth. 12:33:22 how so? 12:33:23 It is also unuseful in its bloatness. 12:33:33 UTF8 seems very integrable to Forth to me.. 12:33:49 it already has that number=first char often.. so no problem parsing UTF8 12:33:51 It requires two times more memory to process. 12:34:17 Also variable character size. 12:34:29 I still don't know what a dough-bridge is! :P 12:34:41 yes, of course.. but it's the only sane way to support internationalization I think 12:34:49 It causes even more troubles than having 3 different encodings. 12:35:14 and some encodings need multi-byte chars anyway AFAIK 12:35:36 talking about encoding... 12:35:37 The only sane way to support internationalization is forcing everyone to use KOI-8. 12:35:45 can somebody explain me how mime encoding works? 12:35:48 i need to script it for something 12:36:02 hm, don't know KOI-8 12:36:16 qFox, what do you mean under "MIME encoding"? 12:36:19 Base-64? 12:36:31 networm_, are you listening to SLAY still? 12:36:31 Or RFC 822 headers? 12:36:45 no.. preparing to log out already 12:37:06 gaim and xchat are next now.. so, good night! :) 12:37:07 networm_, so learn it. 12:37:21 later! 12:37:33 --- quit: networm_ ("Client exiting") 12:38:06 asau> yea 12:38:08 Why it doesn't cause so many troubles to me to use Latin enc? 12:38:54 no actually just how the encoding works 12:38:55 :P 12:39:05 i expect its not that difficult 12:40:20 Well, these are my headers in e-mail. 12:40:27 MIME-Version: 1.0 12:40:32 yes 12:40:33 well 12:40:35 msn uses it too 12:40:43 to encode his msnobj 12:40:57 unfortunatly you get these objects supplied in url encoding 12:41:23 Content-Type: text/plain; charset=koi8-r 12:41:37 Content-Transfer-Encoding: 8bit 12:41:44 well... 12:41:46 this 12:41:56 12:42:08 needs to be sent encoded 12:42:34 searching online for it is like searching for a needle :\ 12:42:57 Maybe you need "URL-encoding"? 12:43:03 no 12:43:08 url encoding is how it is supplied 12:43:10 Or "base-64" encoding? 12:43:26 %3Cmsnobj%20Creator%3D%22maill%40hotmail.com%22%20Size%3D%2220223%22%20Type%3D%223%22%20Location%3D%22TFR2E.dat%22%20Friendly%3D%22AAA%3D%22%20SHA1D%3D%22JM%2BAhe6uPPxv6YLtFNYAnaRX7M8%3D%22%20SHA1C%3D%22c8sI8zpPm17HVTIGPy4Bk7pxpWU%3D%22%2F%3E 12:43:38 base64 12:43:51 Why not read RFC? 12:44:04 I don't remember its number. 12:44:15 It's easy to find. 12:44:28 not? :p well rfc's have the tendency to explain something in 50 pages, what can be explained in 1 12:44:39 Keywords: Multipurpose Mail Extensions 12:45:16 I don't learn by heart. 12:45:28 I had no need in base-64 encoding. 12:45:40 It's somewhat similar to UU. 12:46:01 You join 3 sequent octets. 12:46:04 --- join: wossname (wossname@HSE-Sherbrooke-ppp79747.qc.sympatico.ca) joined #forth 12:46:16 Break them into 6 bit parts. 12:46:36 URLs are not encoded with base64 12:47:07 And translate them into octets again using al-num characters 12:47:08 url encoding is a utf8 variation 12:47:23 and some special characters 12:47:27 hm 12:47:36 as you have too little of alphas. 12:47:47 as you have too little alphas. 12:48:20 Better you learned Cyrillic. 12:48:50 In that case you never bothered with digraphs and trigraphs. 12:49:15 And of course you had octet-based main encoding. 12:49:52 Since the very beginning of Internet. 12:50:59 All these troubles are caused by septet-oriented ASCII. 12:52:53 :))) 12:53:34 what are you talking about? 12:53:52 this encoding: %3Cmsnobj%20Creator%3D%22ma in URLs is nothing to do with utf or base64 12:54:21 no 12:54:23 that is how i GET it 12:54:36 Yes, it's called "URL encoding" 12:54:41 aye 12:54:44 what are you talking about thin? 12:54:46 but its not how i send it 12:54:48 s/thin/then 12:55:01 send what? 12:55:05 PG1zbm9iaiBDcmVhdG9yPSJxZm94eEBob3RtYWlsLmNvbSIgU2l6ZT0iNDU3MiIgVHlwZT0iMyIgTG9jYXRpb249IlRGUjM2LmRhdCIgRnJpZW5kbHk9IkFBQT0iIFNIQTFEPSJXMnNObi9SYmlCUWJLazFORXhLWEtHRDRmWHM9IiBTSEExQz0iMGJ5czRkNkdLY1VDYTRrLzNnYS9BODYvbXpnPSIvPg== 12:55:14 thats an example of what i should send 12:55:16 That's base-64 12:55:27 qFox: what are you sending that to? 12:55:34 msn server 12:55:35 Read RFC. 12:55:47 to get a p2p file transfer 12:55:53 and, this example works 12:55:56 but its hardcoded.. 12:56:06 obviously i want other display pictures too (thats what its for) 12:56:07 what protocol? 12:56:12 msn p2p 12:56:20 qFox, read RFC it explains base-64 encoding very well. 12:56:32 alright i'll try 12:56:34 msn chat has an RFC? 12:56:37 no 12:56:38 IIRC, it even shows C code. 12:56:41 but this encoding will :p 12:56:47 ye c code is useless to me tho 12:57:01 i already have a vb example that didnt make things clear :\ 12:57:09 base64 is pretty simple 12:57:13 There's no need to use it. 12:57:18 Herkamire> i only need to know the encoding method, so i can script it 12:57:26 You get a description. 12:57:49 Just read RFC. 12:57:57 yes yes 13:00:12 qFox: what language are you writing in? 13:00:51 BTW, IIRC, there should be a Forth code for base-64 13:01:17 Possibly, "Ugly page" 13:01:36 And v.d. Ven's 4ePost. 13:01:39 Herkamire> mircscripting :p 13:02:11 qFox: :) hehe :) you're writing a mircscript to do msn chat? 13:02:16 actually the whole thing is done. i now just need to finetune it so i can request any picture/emoticon. add some file transfer messages, and then go with games 13:02:42 my script connects and communicates with the msn server and it acknowledges me as being online 13:02:53 in fact i have a aibot (random chatterbot) running on a script 13:03:11 artificial_intelligent_bot@yahoo.com 13:03:13 :p 13:03:21 geepers. I just use bitlbee 13:03:25 now just working on display pictures 13:03:35 and file transfer and the whole p2p protocol 13:03:38 and later games 13:03:50 see if i can write a nice aibot to play minesweeper against :p 13:04:26 oh and btw, god that msn client is so wasted when you look at its behaviour behind the scenes 13:04:35 ack's? we need those? why? 13:04:37 :) 13:04:58 --- join: kc5tja (~kc5tja@66-91-231-74.san.rr.com) joined #forth 13:05:09 --- mode: ChanServ set +o kc5tja 13:05:33 Well, I think ForthBox Kestrel will definitely remain a 65816-based project. 13:05:38 Dobryjj vecher, kc5tja! 13:05:42 The 68000 is attractive, but it's largely unavailable anymore. 13:05:43 kc5tja> my friend has some questions 13:05:51 he should be joining any second now... i think 13:06:01 lets say... his intrest is peaked :p 13:06:10 That being said, however, the next generation ForthBox *WILL* use FPGAs, and *WILL* run with a MISC processor core. 13:06:22 And *WILL* run asynchronously. 13:06:55 I found some *really* neat online resources relating to FPGAs, including links to _free_ software that I can use to program Xilinx and Altera FPGA chips. 13:07:09 Free as in $0, not free as in open. 13:07:14 how much for a generic FPGA ? 13:07:25 AldoBR: About $15 or so in quantities of 1. 13:07:33 hmm... 13:08:00 So if I get a generic FPGA and program it with a MISC core, (and there'll be room for other things too, like timers, I/O devices, etc!), that'll definitely be overwhelmingly more cost effective than an all discrete component solution. 13:08:20 --- join: BoomBoX (BoomBoX@cd5114735.cable.wanadoo.nl) joined #forth 13:08:26 there 13:08:30 kc5tja, there're open cores already. 13:08:32 kc5tja > BoomBoX > friend > talk 13:08:32 :p 13:08:48 enough introduction, i need to check on my cookie 13:08:52 Good afternoon, BoomBoX. 13:08:56 hello all 13:09:02 ASau: I'm not sure how that at all relates. 13:09:10 re BoomBoX 13:09:19 well qfox informed of an interesting forth board, i was curious about it 13:09:31 like what processor? 13:09:51 65816 @ 12.6MHz. 13:09:58 kc5tja, there're ready open FPGA cores for Forth CPUs. 13:10:12 ah okay, and you use an small bootloader to make an forth processor around it 13:10:14 ASau: Yeah, but they're synchronous in design. I'm looking for asynchronous logic. 13:10:18 yeah i tought about that too 13:10:26 BoomBoX: boot loader? 13:10:43 the 65816 cant excecute forth directly. right? 13:10:44 kc5tja, I've not looked at them yet. 13:10:49 The only booting process this board will partake in is copying the contents of ROM to RAM, then switching out the ROM so the CPU can run at full speed. :) 13:10:56 BoomBoX: Correct. 13:11:12 BoomBoX: However, I have another design in the works that involves a custom Forth CPU embedded in an FPGA. 13:11:22 yeah thats what i wanted to suggest :) 13:11:33 any idea what fpga? 13:11:35 But since FPGAs aren't exactly "homebrew friendly" when it comes to soldering them onto boards, I'm concentrating on something that I can turn into a kit very easily. 13:11:41 or how dense the design will be? 13:11:46 BoomBoX: Probably Xilinx Spartan II. 13:11:49 yeah, most of them are in TQFP 13:11:53 or similar 13:12:14 BoomBoX: Yep. The ForthBox Kestrel design is largely similar to the design of a Sinclair ZX-80/81. 13:12:32 kc5tja: maybe if the design isnt too dense you can use an smaller cheaper PGA from altera 13:12:38 The CPU would even be responsible for video refresh. 13:12:45 kc5tja: 10000gates, 13:12:54 kc5tja, Maksim Makarov designed, IIRC, Forth CPU for Spartan II. 13:12:57 Gate count isn't so big an issue as routability is. 13:13:07 kc5tja: video refresh could be handled in an seperate PLD? 13:13:16 true true 13:13:25 BoomBoX: If I were to use FPGAs, then yes, video refresh would certainly be handled by dedicated logic. 13:13:26 kc5tja, though I'm not sure if it is really Forth CPU. 13:13:27 how big was that design? 13:13:40 The point of the Kestrel is minimum cost, minimum parts count, and minimum skill required on the part of the builder. 13:14:37 kc5tja, I've heard of several such projects. Not only that. 13:14:50 See, most of the people who expressed interest in this thing wanted to purchase one, but since I'm *not* a manufacturer for electronics at all, I lack the resources to have pre-fabricated boards (besides, tooling up for that kind of stuff is expensive). Hence, I'm restricted to kits. 13:14:56 yeah, thats how i want to build an small FPGA kit now 13:15:26 But then came the issue of people not having any kit-building experience. :D 13:15:48 So that left me with no choice: minimize skill required, also at all costs. 13:15:53 kc5tja: well thats why qfox asked me to possibly assemble something 13:16:11 qfox hates hot metal :\ 13:16:32 Well, I don't have any FPGA programming equipment at all at the moment. This would be a big step for me. 13:16:40 qFox: try molten solder :P 13:16:42 Plus I'd have to design PC boards, and pray they worked first time. 13:16:59 I'm on a *very* tight budget. 13:17:06 kc5tja: well you could get something easy like this : http://www.fpga4fun.com/board_pluto.html 13:17:23 I'd need several. 13:17:40 One for microprocessor, one for I/O chipset, one for memory controller (thinking of going with SDRAM), etc. 13:17:41 kc5tja: same here same here, i was planning to build the programming cable from the altera site, this way you can program quite some dense devices 13:18:12 kc5tja: well if you want to select several chips, you need to know what their capabilities need to be 13:18:14 Yeah, fpga4fun.com is what convinced me that FPGAs are well within my reach for this kind of stuff. 13:18:41 Well, I don't think there is any need for different families of chips for different devices. 13:18:49 kc5tja: me too, i just ordered several smaller "Tryout" PLD's too see if i could program them and how useful they are 13:19:01 E.g., the same FPGA chip can function as the CPU, or as the video display controller. It's just that I might need two of them. 13:19:46 Question: is there any kind of indication from an FPGA chip that indicates that it's done reading its EEPROM program, and is operating in I/O mode? 13:19:47 kc5tja: well it could but would depend on the rescources spent in that chip, 13:20:02 I'm not sure I'm making myself clear. 13:20:42 What I'm saying is that I can have *two* Spartan IIe FPGAs, both 144 pins (I forget the exact part number). ONE of those is the CPU. The OTHER is the video chip. 13:20:43 kc5tja: usually the external programming roms have an pin to signify when they are ready, when PGA's are not programmed, all the pins float 13:21:00 kc5tja: sure that would be possible 13:21:13 That's why I said I would need multiple Pluto's for the purposes of development. 13:21:36 kc5tja: ah i see 13:22:03 pluto ? the mickey´s dog ? 13:22:06 hehe sorry... 13:22:33 AldoBR: Check the link posted above. It's an FPGA module that homebrewers can hack around with and play with the chip without having to custom etch their own PC boards. 13:22:34 well the pluto's dont have enough capacity for such an complex device like an forth processor 13:23:21 Well, if the Steamer 16 can be coded into a moderate sized CPLD, then I'm sure *something* can be made to fit in it. :) 13:23:43 After all, CPLDs aren't supposed to have enough logic in them for making custom CPUs either. :D 13:23:44 well an 8bit 6508 style micro sure :) 13:23:53 Umm...no. 13:23:54 :) 13:23:58 Steamer 16 is a 16-bit CPU. 13:24:09 However, it's internal resources are limited. 13:24:17 But it is performance competitive with an 80386 on average. 13:24:26 oww quite nice then 13:24:36 MISC, gotta love it. :) 13:24:42 8 instructions too. 13:24:48 8??? 13:25:01 well thats probably the "limited" aspect i think 13:25:02 NOP # @ ! + AND XOR and ZGO. 13:25:17 ah its an forth cpu then 13:25:19 # loads a literal onto the 3-deep data stack. 13:25:26 @ and ! fetch and store whole 16-bit words to memory. 13:25:28 + adds 13:25:33 AND and XOR do as you'd expect. 13:25:34 am i the only one who thinks that its a good idea these two live a few thousand miles away from eachother? :p 13:25:38 And ZGO is a conditional branch. 13:26:31 So, now, I'm confused. 13:26:33 what kind of device was it implemented on? densities? 13:26:56 Do people want the minimalist, easy-to-build solution, or do they want the MISC-based system that offers more power for the buck? 13:27:24 kc5tja: first easy to build, to check it out 13:27:29 then more complex devices 13:27:34 http://www.stringtuner.com/myron.plichota/steamer1.htm 13:27:45 thats' what i did when i started 8051 development 13:28:18 So, start with an all discrete component-based system like Kestrel, and offer that as a kit. 13:28:23 This would give my customers some kit-building experience. 13:28:42 Then release a more advanced computer design based on MISC later, after they've gotten their feet wet with building kits. 13:28:51 indeed, and then when you get enough input from your customers and more experience you can get more complex 13:28:55 exactly 13:29:09 How about a Forth stamp? 13:29:29 madgarden: Easily done. :) Not with Kestrel's design though. But I can do it once I move towards the FPGA-based designs. 13:29:40 Seems like they could sell well. 13:29:45 hmm, didnt knew that cypress built PGA's 13:29:56 If you look at all the BASIC stamp projects out there. 13:30:42 Yeah, PICs need to die. Their processor core sucks ass. Sure it's fast for straight-line code, but their interrupt system is, shall we say, sorely suboptimal. 13:31:05 kc5tja: agreed, i once touched them, but they are quite "weird" 13:31:27 BoomBoX: Actually, they concentrate more on CPLDs than on (F)PGAs per se. 13:31:53 kc5tja: well that doesnt matter, maybe some free samples could be ordered 13:32:11 * kc5tja nods 13:32:37 It seems like the right thing to do now is to move forward with the Kestrel design. 13:33:13 65816 * 12.6MHz, 512KB of static RAM, some ROM, two 65C22 VIA chips to handle IDE and other I/O interfaces, and some PIC chips to handle VGA timing and audio generation. 13:33:45 65C22, where are you going to find those? 13:33:54 Bwaahahaha! I have my sources... :) 13:33:56 they are quite rare (at least here) 13:34:22 http://www.westerndesigncenter.com 13:35:37 mmm, i usually used 8255 for the IO stuff, but the 6522 are good too 13:36:15 no distributor for the netherlands... 13:36:45 You might want to contact Mike Nazemberry (I think that's how it's spelled); he runs http://www.6502.org -- he's now an official distributor. 13:37:14 --- quit: jdrake ("Oops. This machine just fell asleep") 13:37:59 thanks 13:38:22 But that is OK; if I were to finish the Kestrel design, and one were to order it from me, I'd supply all the parts. 13:38:34 Including the pre-programmed PIC chips required for VGA and audio signaling. 13:39:19 BoomBoX> so when's my cpu ready? :D 13:39:35 heh well if its trough the hole parts only then its no problem for me to build 13:39:40 just a buttload of wireing 13:39:50 so... next week? 13:39:54 Yes, the Kestrel is intended to be through-hole. 13:40:13 qFox: well let me first get an new regulator tube for my scope then we are in business :) 13:40:21 alright alright, 2 weeks 13:40:22 :) 13:40:22 :p 13:40:33 qFox: You forgot something. 13:40:40 It's going to take me time to design and build this thing. :) 13:40:51 kc5tja: i was already finding an excuse to buy an SMT soldering kit 13:41:11 That'll be for ForthBox II (whatever it's final name will be). :) 13:41:24 kc5tja: aah yes designing... if most of the parts are trough the hole breadboarding would be easely done 13:41:28 well yea, but you were gonna speedfix that in like 2 days, 3 days for mail, 2 days for assembly, a week for that regulator tube, 2 weeks :p 13:41:29 kc5tja, ForthBox ][ 13:41:32 :))) 13:41:41 2 ForthBox 13:41:41 That'll definitely be MISC core, and have a bit more advanced graphics (like 256 colors on a 640x480 display, MAYBE, if you're lucky, even some sprites!) 13:41:59 madgarden: Hahah! I like it. 13:42:05 ;) 13:42:12 kc5tja: so when will it have PCI bus??? :P 13:42:14 ASau: ForthBox //e+ pro Alpha. :D 13:42:40 BoomBoX: It probably won't. Ever. Seriously, if anything, it'll have a Zorro-II bus or a Zorro-III bus. :) 13:43:07 kc5tja: is that an offical spec? 13:43:32 BoomBoX: It was when Commodore was still in business. It's the expansion backplane for the Amiga 2000 and 3000/4000 series of machines, respectively. :D 13:44:15 My MISC design will have an asynchronous bus interface (since I'm going to make the MISC asynchronous as much as possible, it only makes sense), just like the 68000's. 13:44:35 kc5tja: asynchronous... asin clockless? 13:44:43 Yes, that's my goal. 13:45:16 kc5tja: wow, designing it would be quite an complex affair then, 13:45:40 kc5tja: but you already offered to do so :D, i am curious about the results 13:45:46 BoomBoX: The idea behind these machines is to escape the "Upgrade twice a year" cycle that seem pervasive in the PC industry. 13:46:08 I've got more important things to upgrade than the microprocessor. :) 13:46:25 The biggest reason for going async with the CPU is to minimize RF emissions. 13:46:30 <-- is a ham radio operator. 13:46:35 <-- is my callsign. :) 13:46:46 Subsequently, I want to enjoy my radio hobby while using the computer. 13:46:50 kc5tja: aaaah, okay i tought it was maybe an design fetish of yours or something 13:46:58 My current PC hinders this a great deal, largely due to the clocks that are in it. 13:47:15 kc5tja: offcourse, but with such highfreqency signals are you going to have intereference anyway? 13:47:20 Well, it's also well documented that, besides having much lower EMI, it's also VASTLY lower in power. 13:47:35 kc5tja: shielding is not an option then? 13:47:36 BoomBoX: Want me to send you a recording of the emissions from my computer? :) 13:47:41 kc5tja: true true 13:48:04 kc5tja: sure :) 13:48:13 BoomBoX: You can't really shield a computer that has expansion slots. 13:48:33 kc5tja: becouse of the long bus wires? 13:48:43 * kc5tja notes his Amiga 500 is next to dead silent compared to his PC, despite both being shrouded in sheet metal. 13:48:54 that's got a lot to do with it, yes. 13:49:17 The other thing is that shields work better (due to inverse square law for power) when placed close to the source of interference. 13:49:23 this is why a ground plane on a PC board is so nice to have. 13:49:29 kc5tja: well what is the fastest clock in an amiga? 13:49:41 28.363636MHz for NTSC version. 13:49:48 kc5tja: forgive my ignorance but ham radio, what frequencies are we talking about 13:49:51 But the most prevalent clock is 7.15909MHz, squarely in the 40m amateur radio band. 13:50:08 kc5tja: do you have the same problem with older computers (like 80386 machines and slower) 13:50:15 BoomBoX: Yes. 13:50:42 When you take apart a Commodore computer, you better wear cut-gloves, because the shielding is thick, strong, and very sharp, sheet metal aluminum. 13:50:49 You take apart a PC, and it's all open space. 13:50:53 kc5tja: Not in this region :( 13:50:54 Not a *lick* of shielding anywhere. 13:51:03 kc5tja: hmmm. strange, well the amiga 500 doesnt have a lot of long connections 13:51:07 kc5tja: that too 13:51:23 Robert: ?? 13:51:42 Er, nevermind. 13:51:47 I misread that frequency. 13:52:18 Thought it was 100kHz higher. 13:52:20 Robert: Well, it's outside ITU Region 3's allocations, but for us Americans, 7.15909 is smack in the middle (our 40m goes from 7.0 to 7.3MHz). 13:52:29 Er, damn. 13:52:33 And now I forgot the band limits. 13:52:38 It's not as useful as you might think. 13:52:49 Hehe. :) 13:52:56 Because 7.1 to 7.3 happens to be *loaded* with international shortwave broadcasters. 13:53:13 Yeah, I've heard some of those. 13:53:27 * kc5tja has privileges only from 7.100 to 7.150MHz, which is prime territory it seems for SWBCs. :( 13:54:00 Ouch. 13:54:27 SWBC's? 13:54:30 Must feel terrible that morons like me have access to more than that... 13:54:36 Short-Wave Broadcast Companies. 13:54:51 ahaa 13:55:45 So, the only problem I'm having now is how to get VIA chips at 3.15MHz to talk to the 65816 at 12.6MHz. :) 13:55:57 * kc5tja has to make some kind of RDY-signal logic. 13:59:15 you could use something configurable waitstate generator :) 13:59:41 No, I can't. The bus timings of the 65816 and the VIA chips are quite rigid. 13:59:57 kc5tja: i the type of 8051 i use has an special register where you can regulate how may cyles will be used to acess an external device 14:00:02 I have to wait for the VIA clock to go low, then high, then low again before I can let the 65816 go to its next cycle. 14:01:43 kc5tja: or you could make it an user responsibility? like wait before writing again 14:02:01 The problem is the bus on the 65816 is 4x faster than the VIA chips can go. 14:02:15 i see 14:02:19 And the VIA chips have registers that I need to *read* from too. 14:02:28 So I can't just do the "chip select stretch" technique. 14:03:27 Otherwise, I definitely would. :) 14:04:11 well i am thinking right now "how would you do that?!" 14:04:50 its not possible to latch on the edges when the I/O range is acessed? 14:04:52 The first thing to realize is that the 6502/65816 places address information on the bus during clock low. 14:05:10 This is why I need to wait until the VIA's clock goes low first time around. 14:05:24 When that happens, I can then gate the 65816's address and data information onto the 3MHz bus. 14:05:46 exactly 14:06:00 Then I need to wait for the clock to go high, because that's when the data is actually transferred. 14:06:05 but you would need to dedicate an special amount of space just for I/O 14:06:08 But I can't relinguish the bus yet because the VIA is still driving it. 14:06:20 Well, you have to do that anyway. 14:07:09 The 65816 or 6502 CPUs do not have a distinct I/O space. 14:07:19 In fact, **ONLY** Intel CPUs ahve distinct I/O spaces. 14:07:24 well the databus to the IO devices can be seperated from the "speedy" bus with 74244? when the processor has seen the value they would just disable the 244 outputs 14:07:31 indeed, i was confused 14:07:38 (and yes, the Z-80 is an 8080 clone, and thus falls under the category of an Intel CPU for this definition) 14:08:06 BoomBoX: Yes, that's what I'm planning, precisely. 14:08:21 i found memory mapped more elegant then special I/O spaces 14:08:27 First high-to-low clock transition: gate address and/or data onto slow bus. 14:08:30 easy to protect by OS too 14:08:38 Low-to-high transition: wait -- data is being transferred. 14:08:54 Second high-to-low transition: Transaction complete. Relinguish the slow bus. 14:09:26 Since ForthBox is going to be designed for imminent hackability, and will run Forth as its OS, protection isn't an issue. :D 14:09:28 problem is... when you do 2 fast actions to I/O in sucession 14:09:46 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 14:09:47 then you need to stop the CPU 14:09:49 Yep, you have to wait for the extra wait state. 14:10:07 Nothing you can do can stop this from happening. 14:10:09 the 65816 does have an wait pin or something? 14:10:13 RDY 14:10:46 how wide is the bus of the 65816? 14:11:09 When RDY is low, the CPU is forced to repeat its previous bus cycle -- basically halting the CPU dead in its tracks. 14:11:18 BoomBoX: 8-bit data bus, 24-bit wide address bus. 14:11:22 this is so way above my head... i cant believe you and i differ only a week BoomBoX :\ 14:11:33 (16 bits directly accessible; the upper 8-bits are accessible by demultiplexing the data bus). 14:11:37 qFox: different minds man :) 14:11:47 i cant even pretend to know and talk along :p 14:12:09 kc5tja: well that would be quite some latches 14:12:19 maybe PAL's are an option? 14:12:41 BoomBoX: For address decoding and driving the bus drivers, maybe. But I still need those bus drivers. 14:13:06 * qFox goes about scripting base64 14:13:07 But that being the case, I would need a source for PAL chips. I don't have a programmer for those either. 14:13:19 okay, so 2 74F573 devices then and an 244 14:13:32 just a sec,,, i had an programmer right here 14:13:51 http://www.geocities.com/mwinterhoff/galblast.htm 14:13:54 there 14:14:36 Yep. Just 2. Even though it's a 24-bit address bus, I will have much less than 64K of combined I/O and ROM space. 14:14:47 I expect only 8K of ROM and 4K of I/O space, *max*. 14:15:13 well 4k can be quite cramped imo 14:15:19 4K is monsterous. 14:15:27 what if i wanted to add some SID chips or something 14:15:51 Audio will be handled via one of the two PIC chips in the system, which communicates to the 65816 via one fo the 6522's built-in serial ports. 14:15:53 i am never going to utilise exactly 34 registers (or something in that range) 14:16:09 Here are the devices I plan on supporting: 14:16:12 kc5tja: well just as an example :) 14:16:14 IDE interface: 16 bytes at most. 14:16:17 --- quit: Nutssh ("Client exiting") 14:16:30 Video Shift Register: 16 bytes, all of which reference the same register. 14:16:36 (makes address decoding easier). 14:16:37 IDE is 4 bytes i know for sure, i made such an interface myself 14:16:41 VIA #1 (16 bytes) 14:16:46 and VIA #2 (16 bytes) 14:17:29 No, it's technically 9 bytes -- 7 bytes and 1 16-bit read/write port for the actual data transfer. 14:17:34 okay so no provision of adding some extra hardware 14:17:38 The 7 bytes configure CHS 14:17:52 BoomBoX: I'm not sure where you get that idea from. 14:17:59 I've detailed only 64 bytes out of a total of 4096. 14:17:59 well... 14:18:06 There is a *ton* of room in that 4K space. 14:19:15 yeah i see now btw, the 4 bytes i gotten from here, but i used an 8255 for the controlling the whole setup 14:19:16 http://www.pjrc.com/tech/8051/ide/ 14:19:27 * kc5tja nods 14:19:28 Right. 14:19:34 --- quit: slava (Read error: 104 (Connection reset by peer)) 14:19:46 I considered controlling the IDE interface via the VIAs, but then I decided against it for performance reasons. 14:20:18 well the performance is going to be bad, unless you use an PLD 14:20:36 Define "Bad." 14:20:37 becouse i maxed out at 30kb when the 8051 was running at 11.0592 mhz 14:20:50 Remember the CPU runs at most at 12.6MHz. 14:21:02 That's way underneath the maximum IDE bus bandwidth. :) 14:21:15 the controller was memory mapped in 8051 bus 14:21:21 this was also an high speed 8051 14:21:28 so no 12 cycles per machine cycle stuff 14:21:31 Well, the 8051 also requires 12 clocks per memory cycle too. :) 14:21:39 it was an risc 8051 14:21:47 so 1 cycle is one machine cycle 14:21:48 not risc 14:21:51 i mean pipelined 14:21:55 (doh) 14:22:06 Yeah, but did it access *memory* at that speed? 14:22:40 (BTW, just what the *hell* was Intel thinking when it decided 12 cycles per memory cycle? Never mind. That's a topic for another day. I digress.) 14:22:48 as far as i could see, my scope isnt the newest thing on the block but counting the ALE pules it was acessing it at fullspeed 14:23:03 OK. 14:23:15 Well, then, that settles it. 14:23:26 I don't see how a PLD will in any way make it go faster. 14:23:56 it could read the 512 sectors in one go, instead of querying the IDE interface for each byte 14:24:07 i mean 512 bytes in one sector 14:24:14 But that's exactly *what* the PLD will be doing. 14:24:24 The bus is only 16-bits wide. 14:24:40 yes, but it would do that at maximum possible speed 14:24:54 * kc5tja shrugs 14:25:18 well. i am thinking too, what would make it so slow 14:25:21 Polled I/O is plenty sufficient for nearly anything I can think of for the Kestrel. 14:25:41 The IDE drive delivers the next set of bytes as soon as it sees a read from its data port. 14:25:42 well 16kbps was my goal for that project 14:25:59 So if the IDE interface can poll one word per cycle, it'll at least match DMA in speed. 14:26:04 yeah, i have no actual reason why it would be so much slower 14:26:36 I would think it would have to be the drive's electronics then. 14:26:49 but i tried many different drives 14:27:00 from modern devices to the not so modern 14:27:04 But anyway, for the Kestrel, which will have a real-world performance close to 3.15MHz thanks to CPU-enforced video refresh, full-speed IDE isn't a priority. 14:27:12 all around 29 till 31kbps 14:27:33 Also are you sure it's kbps or is it kBps? lowercase b is bit, uppercase B is byte. :) 14:27:39 it is byte :) 14:27:40 sorry 14:27:44 OK. 14:27:46 :) 14:27:58 Well, we'll see how it goes. 14:28:18 The Kestrel is designed to be low cost, low maintenance, so uber-performance isn't a requirement. 14:28:25 exactly 14:28:31 It's sole purpose is to just get the job done, and be as frugal as possible about it. 14:28:52 Actually, come to think of it, I may just consider using another PIC chip exclusively to control the IDE interface. 14:28:59 And just drive that off the VIA's serial port. 14:29:33 with a little bit of intelligence you could add some more features at the controller level 14:29:35 That'll free up some of the ports for other uses. 14:29:40 that too 14:30:01 And i want as many bits available for hacking as possible. I miss the Commodore 64 user port! :( 14:30:35 never been an cbm64 user (too young) i have to make do with an LPT :) 14:31:11 There isn't much difference, except for the fact that you have more control over the pins (each pin is individually controllable to be input or output, for example), plus interrupts are handled more logically. 14:31:41 i see 14:31:57 But it's that level of control that made everything so possible. 14:32:13 how many pins are on such a port? 14:32:24 Oh, and the user port exported the 6526's serial port too, so you can use the parallel and serial port in conjunction to get virtually limitless I/O expansion off that thing. 14:33:06 The User port had 24 pins IIRC. 8 of which are data pins, 2 are control pins (equivalent to _STB and _ACK), a serial port, some power pins, and I forget what else. There were also a couple of no-connects too. 14:33:58 that serial port, quite useful indeed 14:34:15 Yes. Synchronous serial port, that can safely operate up to 1/4 the clock speed of the VIA chip. 14:34:39 Which in my case, would be somewhere in the vicinity of 787500kbps. (yes bits. :D) 14:34:40 heh 14:34:46 hey, where's the point? 14:34:53 7.87500kbps. 14:35:09 Geez, learn to type. 14:35:15 787.500kbps <-- that's right! 14:35:25 heh but dont you put points ar... that what i wanted to remark :) 14:36:19 Well, before you arrived here, I *just* got back from college, and we use scientific notation there all the time. So I tend to "instinctively" write 7.875 even when I mean 787.5. :) 14:36:33 Here's a better way. 14:36:37 787k5 bps. :) 14:36:42 :) 14:36:59 Still though, 787kbps is rather nice data rate. 14:37:15 well in holland we use , to seperate groups of 3 and an . to signify the radix 14:37:20 For the sake of convenience, I'd probably drop the data rate down to 500kbps, or as close to it as possible. 14:37:22 quite "nice" 14:37:22 we do? 14:37:29 :p 14:37:38 yeah, didnt they teach you that in primary school? :P 14:37:40 BoomBoX: So do we. 14:37:49 I used the decimal point to set the decimal point. :) 14:37:57 i have _no_ idea what it was about :) 14:37:58 I generally don't use , to separate groups of three. 14:38:04 eehmm... qfox 14:38:06 you are right 14:38:09 ignore me 14:38:10 :) 14:38:11 , we do it other way around 14:38:18 lol 14:38:23 ye 14:38:25 Both are technically correct, as per Arabic numeric conventions. 14:38:32 thats why i had problem with maths when i started programming :) 14:38:35 Which makes it all the more confusing. 14:38:40 132.34.253.14 14:38:45 hm 14:38:48 132.347.253.14 14:38:49 :p 14:38:50 you missed some :) 14:39:26 becouse the , is in a lot of languages an seperator and . is well the decimal dot for floating point numbers 14:39:44 --- join: slava (~slava@CPE0080ad77a020-CM.cpe.net.cable.rogers.com) joined #forth 14:39:57 Dobryjj vecher, slava! 14:40:03 privet! 14:40:13 Kak dela? 14:40:16 Well, anyway, I think this is going to be a rather interesting project. 14:40:29 Zhizh' b'jot kljuchom? 14:40:30 Whether anyone actually sends money to get one of these things remains to be seen. :) 14:40:50 i am indeed quite interested where it will be going 14:41:05 slava, zhizn' ... 14:41:12 Cool. 14:41:22 BoomBoX: Are you familiar with the Sinclair ZX-80 at all? 14:41:34 That's the inspiration as far as minimalist hardware construction is concerned. 14:41:36 i have some processors lying here somewhere in my junk box 14:41:43 but the arch i dont know 14:41:55 The CPU feeds the video shift register directly. 14:42:03 jikes 14:42:05 That's why the machine seems so unnaturally slow. 14:42:31 is the zx80... where you piggyback those video memory chips to get uppercase letters? 14:42:36 Well, that's the same thing I'm doing with the Kestrel, but since I have a 12.6MHz clock, the machine itself will feel like it's running only at 3MHz. Which is about 4x faster than the ZX-81. :) 14:42:56 BoomBoX: Huh? No. The ZX-80/81 do not have text display. It's only graphics. 14:43:01 BoomBoX, you're mistaken. 14:43:17 ah okay, what computer was that then?.... 14:43:22 read: wrong. 14:43:44 Graphics only display. 14:43:46 no i am just thinking what computer that was, but nevermind 14:43:51 BoomBoX: I'm not familiar with any that required this. The only thing I can think of that came close as a TRS-80 Model 1, but that was never sent overseas. 14:44:04 (AFAIR) 14:44:35 well guys, its been a pleasure to chat with you, but i am getting sleepy 14:44:39 * kc5tja nods 14:44:58 Anyway, that's the approach I'm taking with the Kestrel, so as you can see, it's a very dirt-cheap system to build. :D 14:45:01 That's my goal at least. 14:45:07 The main difference is direct video. 14:45:14 Real ZX used ULA. 14:45:15 well dirt cheap is always been my primary goal :) 14:45:23 good night all 14:45:27 ASau: Real ZX? 14:45:31 Good night. 14:45:36 --- quit: BoomBoX () 14:45:46 Clones also. 14:46:07 The ULA is a programmable logic chip as far as I know -- PROM-based (e.g., like most PLA chips). 14:46:08 Even the most cut of resources emulated ULA. 14:46:13 Not sure why they called it a ULA though. 14:46:37 kc5tja: you're calling your forthbox concept Kestrel? 14:47:07 Herkamire: Yeah, for the cheapie cheapie kit model. Someone named it here on the channel, and I just kept it ever since. 14:47:42 Kestrel was my favorite ship it Escape Velocity 14:47:45 Herkamire, I wanted to ask you. Aren't you native English-speaker? 14:47:51 ASau: yes 14:48:16 Wow. I have officially caused a major floating point issue within Python. 14:48:20 >>> 12.6*2 14:48:20 25.199999999999999 14:48:25 :) 14:48:35 Gotta love binary-based floating point numbers. 14:49:03 computers suck ;) 14:49:27 Herkamire, I always thought that such an error as in "you're calling your forthbox concept Kestrel?" is priviledge of non-native speakers. 14:50:08 Herkamire, or is it called "mistake"? I'm lost in languages. 14:50:29 ASau: Mistake or error is acceptable; they are synonymous. 14:50:59 --- quit: AldoBR () 14:51:04 kc5tja, yes. It's the hell of English. 14:51:10 what's wrong with my sentence? 14:51:12 Though personally, I don't see the error. 14:51:18 *TONS* of synonyms. 14:51:48 --- join: AldoBR (Aldo@200.195.110.69) joined #forth 14:52:21 I do often take shortcuts on irc that forigners often make, but in that particular case, I see nothing incorrect about my english. 14:52:28 I don't know folk language. High style approach in foreign languages is always taught here. 14:53:23 ASau: there's nothing slang about it except maybe the apostrophy 14:53:47 OK. I'm completely lost. 14:53:56 I'll think it over later. 14:53:57 ASau: How would you have written the sentence? 14:54:21 You might have wanted to write: you're calling your forthbox concept "Kestrel?" 14:54:52 Hm. I've been taught to write: "Are you..." 14:55:07 Of course, I don't write this way. 14:55:35 Because it's more natural to me to write: "You're ... ?" 14:55:42 Yeah, technically, it should be "Are you," so that is kind of "slangy" in that sense. 14:56:10 But you all are wrong. :) forthbox is an improper noun, while ForthBox is proper. >:D 14:56:49 It's actually not so much slangy, now that I think about it. 14:56:51 It's just informal. 14:57:10 I've even seen Britons do it, so it can't be that bad (and we all know how proud they are of their language). 14:57:19 ForthBox is proper? Forth-box mayhap... 14:57:45 madgarden: Proper nouns are always capitalized. 14:57:58 That's why we don't say, "i went to the store," but rather, "I went to the store." 14:58:08 :) 14:58:09 Ja. 14:58:21 Sure... but Forth*B*ox... 14:58:37 CamelCase isn't part of proper English yet. ;) 14:59:21 ImSureICanFindSomeCases...:) 14:59:40 lp777777777777777777254 14:59:47 Ahh, CAT! 14:59:55 * kc5tja loves cats. :) 15:00:03 Thanks, Lloyd. 15:00:06 They go good with ketchup and mustard and relish... :) 15:00:14 hmmm. I suppose it's a bit informal to make a statement and put a question mark at the end. 15:00:14 (j/k) 15:00:21 Yes? 15:00:25 not to bad eh? 15:00:26 I love that one. 15:00:44 Yea, cats are pretty neat? 15:00:56 everything I say is a question? 15:01:01 We did some major spring cleaning yesterday? 15:01:04 The next time someone says, "Yes?" when I call their name, I swear I'm oging to reply with, "Thank you, that's all I needed to know." 15:01:14 Hm. Never I've heard cats being tasty. 15:01:40 ASau: It was more a play on words than anything else. 15:01:44 My cat would make good noodles. 15:01:46 I do like cats though. 15:01:51 He's quite floppy. 15:02:01 I don't like cats. 15:02:10 Poor ol' Lloyd. 15:02:28 Heheh :D 15:02:34 Well, here's my plan of attack thus far: 15:02:38 I've got a negative experience in teaching them. 15:02:43 1. Cook some freakin' food. I'm freakin' hungry. 15:02:58 I just had a hamburger patty with cheese on it, and some hot peppers. :D 15:03:04 2. Try to rework my loop antenna to use 300 ohm twinlead cable, to see if that makes any beneficial difference in performance. 15:03:05 (cooked) 15:03:05 It seems only Kuklachjow succeded in it. 15:03:12 3. Get to programming on my client's code. 15:03:39 kc5tja, what is your other job? The one that's not INO? 15:03:42 BTW, I'm on the second BLOCKs wave. 15:03:46 madgarden: Ahh, have TWO patties with TWO slices of cheese, no bun (not even lettuce wrapped), and you got yourself a Flying Dutchman. :D 15:04:13 I'll do that next time. And my bowels will surely love me for it! 15:04:21 madgarden: Programming an embedded Linux client on the desktop PC, and hoping it'll port fairly easily. 15:04:21 A year ago I was working on blocks-only system. 15:04:41 kc5tja, so two contract jobs and a burger joint? 15:04:49 madgarden: What's the other contract job? 15:04:52 I only have one. :) 15:05:05 I thought you said you worked 3 jobs. 15:05:17 The other will be forthcoming; working at an Internet cafe. 15:05:23 Ah... 15:05:42 And, hey, if my kit business takes off, I can ditch that one, and concentrate on kit sales instead. 15:05:49 Or keep it as part-time, 4th job. :D 15:05:55 (until I get rid of In-N-Out completely) 15:05:59 how much would you be charing for a kit ;D 15:06:00 madgarden, I'm working two jobs and going to Univ. the same time. 15:06:04 I may be ending up with less than software development for my next job. 15:06:07 That's the hell. 15:06:13 I got laid off in November. 15:06:25 I bet. 15:06:28 I'm dropping one job this week. 15:06:37 I worked part time delivering pizza during college... that was quite fun actually. 15:06:48 Hmm, I might end up doing that again. :O 15:07:33 I need to be making at least $12 an hour before my severence runs out if I expect to maintain my current standard of living. The wife just got a full-time job, thankfully. Eases the strain of responsibilities a bit. 15:07:51 But... blarg. 15:07:54 It's the only thing I'm thinking recent time of. To give one job up. 15:08:18 ASau, don't overtax yourself. 15:08:23 wossname: That remains to be seen, but I suspect it will be less than $100 when all is said and done. I suspect somewhere in the $70 to $80 range -- THIS IS JUST A GUESS AT THIS TIME. 15:10:02 madgarden, thanks the month is over. 15:10:35 i'm going to selectively quote those prices later when it turns out to cost much more ;D 15:11:01 madgarden, the only thing is to get salary. 15:11:02 And I'm going to selectively quote the big, bold, flashing sign that reads, THIS IS JUST A GUESS AT THIS TIME. >:) 15:11:49 Seriously, the CPU is $6, the two VIA chips themselves cost in the $5 range each (I'm being conservative here), so that so far is $16, plus some of the glue logic to hold everything together -- $20. 15:12:13 The PC board will be another $20 or so in small quantities, so we're looking at $40. Plus teh ROM and RAM chips, which might add another $20 to the circuit -- $60. 15:12:25 Then comes the instruction booklet for putting the thing together. 15:12:58 $40 instruction booklet? :P 15:27:06 --- quit: ChanServ (Shutting Down) 15:27:43 --- join: ChanServ (ChanServ@services.) joined #forth 15:27:43 --- mode: irc.freenode.net set +o ChanServ 15:32:41 --- join: Nutssh (~Foo@gh-1177.gh.rice.edu) joined #forth 15:33:58 Dobryjj vecher, Nutssh! 15:34:09 Hi. 15:34:52 --- quit: wossname (":(~") 15:38:00 --- nick: arke -> chris-xp 15:38:34 --- quit: qFox (Read error: 60 (Operation timed out)) 15:40:14 madgarden: Yeah, easily. 15:40:59 hi kc5tja' 15:41:52 re 15:43:22 whats up? 15:49:49 we are the priests...of the temple ... of syrynx .... *sing* 15:51:12 syrinx :P 15:51:59 whatever :P 15:52:03 what a kickass song 15:52:12 Yea, long-ass too! 15:52:46 Attention all planets of the Solar Federation... We have assumed control... 15:53:16 * madgarden is listening to C64 SIDs right now, though. 15:56:27 heh 15:56:30 listening to live rush 15:56:43 Live, recorded Rush? :) 15:57:32 And I'm listening "Dire Straits" now. 15:58:11 * kc5tja has just returned from eating. 15:58:46 kc5tja, no you haven't. 15:58:56 I saw you return earlier, for a second. 15:58:57 So there. 15:59:07 "just"... pff! 15:59:08 I was done cooking then. 15:59:28 Did you eat *anything* during or immediately after cooking? 15:59:47 Only what was stuck to the spatula. :) 16:00:15 Food is food, my friend... try to pull one over on us... *grumble* Attica! 16:00:48 chris-xp: The decision has been made. The ForthBox Kestrel will NOT have Ethernet support. Just too doggone expensive. I'm sorry. 16:01:12 USB, though? 16:01:15 No 16:01:27 Not enough CPU horsepower or memory bandwidth to support it. 16:01:28 kc5tja: :( 16:01:47 So I guess I just lost a few customers, eh? :/ Sorry. I tried. 16:02:31 RS-232? 16:02:47 What I can do is expose a large number of I/O pins (including the serial port used by the PIC support chips), and let someone with better development resources than I make an add-on for it. 16:03:04 ASau: Yes, it'll have RS-232, but it will be bit-banged by the CPU. E.g., no UART chip. 16:03:28 Uff 16:03:39 It should still handle fairly high baud rates though, seeing as how the CPU is running at 12MHz and the VIA chips are running at 3MHz. 16:03:59 C'mon -- the point is minimum cost. :) 16:05:04 The alternative is to let another PIC chip handle the RS-232 port, and access that via the same serial bus I use to control sound and VGA timing parameters with. 16:05:08 That's a possibility. 16:05:30 That VIA's serial port will run close to 500kbps so there should be plenty of bandwidth to go around for that kind of thing. 16:06:49 In all honesty, I have to wonder whether the PICs can keep up with the VIA's serial port speed. :) 16:08:10 It seems logical for me to develop that serial bus as much as I can. It seems like I'll be depending on that quite a bit. 16:09:23 If a PIC chip is handling the RS-232 port, then data is sent or received from the PIC unit via packets. 16:09:36 Kind of like IEEE-488 packets. 16:09:48 (in fact, I'll make it as close to IEEE-488 in design as I can) 16:10:21 hrm. 16:10:30 kc5tja: could you make prebuilt boxen for an extra fee? 16:10:40 kc5tja: residing in a pizza box :P 16:10:47 I can make a pre-built PC board, sure. 16:10:59 So you want me to ship you a pizza box with this thing in it? :) 16:11:02 cool, cool. 16:11:11 show box would be better, actually. 16:11:15 shoe* 16:11:33 just whatever, I'm flexible :) 16:11:47 How about I just pre-build the PC board, then you can fit it in whatever box you want? :) 16:12:03 I need a way to somehow hook it up to eth eventually.... maybe you can find somebody who makes RS232-ethernet thingies for you... 16:12:12 kc5tja: better even :) 16:12:21 kc5tja: just have to ship it in something, right? :) 16:12:27 kc5tja: how much extra would you charge for that? 16:12:31 Yes. Anti-static bubble wrap. 16:12:35 chris-xp: I don't know. 16:12:48 Ok. 16:13:00 Depends on how long it takes for me to build it. 16:13:08 And hwo much it impacts my regular schedule. 16:13:13 Ok. 16:14:51 What I like about my serial bus concept is that I can hang pretty much any device I want off of it. 16:15:07 So if I do the RS-232 thing, I'll have three PIC chips hanging off of it. 16:15:41 I can release different devices for the platform easily. 16:16:04 The bus will probably consist of less than 8 pins total: SD, SC, ATN, SRQ, CONFIGIN, CONFIGOUT, and some power pins. 16:16:23 cool 16:16:47 how much memory bandwidth/CPU speed does eth need? 16:16:53 Oh, scratch those CONFIGIN and CONFIGOUT pins. But do add a RESET pin. 16:17:17 :) 16:17:30 chris-xp: For full speed, a 10Mbps Ethernet connection needs about 1MBps throughput on the bus, assuming direct memory access. 16:17:45 100Mbps requires 10Mbps, etc. 16:17:53 10MBps, rather. 16:17:58 heh, oops.. 16:18:11 it can't run slower? 16:18:16 This is, again, assuming a simple streaming DMA channel. 16:18:17 No. 16:18:19 How can it? 16:18:23 :( 16:18:26 It's 10/100 megabits per second. 16:18:31 It needs to supply data that fast. 16:18:38 Otherwise, it won't be 10/100 megabits per second. :) 16:18:48 well, it doesn't have to transmit that fast. 16:18:55 Actually, it does. 16:18:59 oh :( 16:19:11 so it has to send NOPs or whatever when theres nothing to send? 16:19:16 Since other devices on the network may possibly be contending for the bus, any given device MUST transmit for as little a time as possible. 16:19:22 No. 16:19:28 When there is nothing to send, it releases the line. 16:19:44 :( 16:20:16 so the eth RS232 "modem" would have to do alot to be able to feed the CPU and network correctly, eh? 16:20:21 To gain access to the line, the Ethernet card has to transmit a preamble which is designed to do two things: *guarantee* a collision with any other node on the net, and synchronize any receiver who is currently listening. 16:20:32 Yes. 16:20:53 The RS232/Ethernet bridge would need to queue up packets of data from the computer in bulk, to then *blast* it onto the network. 16:20:54 how much throughput would the RS-232 achieve? 16:21:08 heh :( 16:21:16 massive ping times.. 16:21:17 RS-232 itself achieves at most 115200bps at half-way reliable speeds. That's about 11KBps. :) 16:21:21 Yes. 16:21:44 11kbps, vs. 10Mbps ... :( 16:22:02 Now, the nice thing about the VIA's serial port is that, when the VIA is clocked at the 3.15MHz rate, the serial port can move data at 787.5kbps -- 0.7875Mbps. 16:22:09 E.g., it's still nowhere near close to Ethernet speeds. 16:22:15 But it's a HECK of a lot better than RS-232. 16:22:28 good stuff.. 16:22:55 If I clocked the VIAs at 5.0MHz even, I could pull off 1.25Mbps throughput on them reliably. 16:23:04 cool! 16:23:21 That's almost T1 data rate (T1 is 1.54Mbps throughput). 16:23:51 * kc5tja should put some ideas on the serial bus and on Kestrel on my Wiki. 16:25:02 Hm. It seems RS-232 can go twice as faster. On short lines. 16:26:09 ASau: The hardware specifications for RS-232 can support arbitrary bit rates. Amiga's RS-232 port could handle 1.54Mbps throughput, for example (MOST impressive for a machine running 7.15909MHz!!), but as you say, the cables needed to be so short that it wasn't ever exploited. 16:26:39 However, to be compatible with the ISO RS-232 standard, I *think* (I may be wrong on this) that 115200 is the max. Beyond that is outside of specification. 16:26:49 I remember reading about it somewhere. 16:26:55 But in any case, it's not really important. 16:27:22 I don't have UART spec. handy. 16:27:40 It's not the UART spec that'd tell you; it's the ITU-T recommendation on RS-232C. :) 16:28:04 I think of what could be used to connect to PC. 16:28:05 kc5tja: anyway, do you know somebody who could make a serial->eth bridge? 16:28:17 Why I remember this is because I was reading on the RS-485 standard, which is a balanced version of the RS-232 spec that IS legal to go into the megabit range (possible because it is balanced) 16:28:32 chris-xp: No. Not off hand. 16:29:37 kc5tja, I don't remember throughput. RS-485 is definitely faster due to balance. 16:29:52 Yep. 16:30:03 whats different between balanved and unbalaned..? 16:30:18 chris-xp: That requires some knowledge of electronics to explain fully. 16:30:31 An unbalanced signal is a signal whose voltage is referenced to 0V -- e.g., to ground. 16:30:32 well, explain away, I might get it. 16:30:57 A balanced signal is a signal whose voltage is referenced to a mirror image of the signal in another wire. 16:31:07 Thus, it does not depend on there being a solid ground present. 16:32:10 Because of this, a balanced signal emits less radio-frequency interference. Remember that it takes energy to radiate radio waves; hence, balanced circuits conserve power. But more importantly, and reciprocally, they are *less influenced* by external sources of power too. 16:32:30 Therefore, the signal of a balanced circuit is cleaner, and immune to external sources of RF interference. 16:32:58 Now consider that everything inside a computer operates at radio frequencies, and you can see how having balanced communications links can be important to sustaining high throughputs. 16:33:02 :) 16:35:54 Quick -- I need a name for my other ForthBox line. :) 16:36:13 RAVEN! 16:40:44 I like it. 16:40:50 ForthBox Kestrel and ForthBox Raven. :D 16:49:27 heh? 16:49:34 whats the difference between kestrel and raven? 16:52:06 The first one is "ForthBox I", the second one is "ForthBox ][". 16:52:20 That's how I understand. 16:52:26 :P 16:54:26 Basically, yes. 16:55:05 ForthBoxKestrel is the beginner's computer kit, suitable for construction by anyone with a Radio Shack-brand soldering iron and maybe a digital voltmeter. 16:56:03 ForthBox Raven is a fundamentally different architecture, based on much higher performance parts (but harder to solder yourself; still doable, but you need the experience first!), based on my own MISC implementation, with larger memory capacities, and vastly superior graphics. 16:56:10 (and vastly superior audio for that matter) 16:56:20 The difference can be summarized like this: 16:56:43 Sinclair ZX-80 is to Kestrel, as ZX-Spectrum 3+ is to Raven. :) 16:57:29 er... 16:57:52 Kestrel: Minimum cost, Minimum Parts, Minimum functionality, minimum performance. 16:57:53 :) 16:58:17 Raven: Moderate cost, Moderate parts, Maximum functionality, maximum performance, huge memory capacities. 16:58:36 so your own CPU design? 16:58:40 Yes. 16:58:45 Integrated into an FPGA chip. 16:58:49 whats it based on? 16:59:05 You have already seen my MISC processor specs. It's on my Wiki site, remember? 16:59:17 :) 16:59:34 oh, that weird one? :) 16:59:55 Weird?! That is perfectly sane by comparison to the other MISC concepts in existance. 17:00:07 No 20-bit buses for me. 16-bit or 32-bit, but no 20-bit. 17:00:12 heh.... 17:00:19 MISC means minimal instruction SC, right? 17:00:24 (actually, the Raven processor [HA! Now that I have a name for it!] will be a 32-bit processor) 17:00:33 Yes, minimal instruction set computer. 17:00:52 --- nick: AldoBR -> Aldo[Working] 17:01:40 what about my 16-instruction RISC? would that be considered a MISC? 17:05:14 Marmalade. 17:07:19 I don't recall. 17:07:54 ForthBox Marmalade 17:13:39 hi everybody 17:14:02 hi slave 17:14:04 a 17:14:21 tuck drop 17:15:32 tuck drop == swap, my friend. 17:22:55 SWAP DROP 17:23:31 What, you don't like "slavae?" :P 17:23:53 SWAP DROP == NIP 17:24:09 NIP == SWAP DROP 17:26:05 --- join: Sonarman (~matt@adsl-64-160-166-159.dsl.snfc21.pacbell.net) joined #forth 17:27:06 --- join: qFox (~C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 17:27:24 madgarden: The machines are named after birds. :) 17:27:35 Not jellies, jams, preserves, or other canning technies. 17:27:45 techniques even. 17:27:59 couldnt you at least make it a bird we all know? hummingbird or something :p 17:28:10 i believe firebird is free now ;) 17:28:14 --- quit: zardon (Read error: 54 (Connection reset by peer)) 17:28:23 Aww! But marmalade is so tangy... 17:28:35 qFox: The top-of-the-line box is named Raven. 17:28:48 hey i know that one 17:28:50 Ravens are annoying birds that mess up your garbage. 17:28:52 black ugly bird :p 17:29:03 that go for shiny things 17:29:14 qFox: Ugly is a matter of opinion. 17:29:19 true 17:29:23 madgarden: Then put it in cans, like everyone else does. :) 17:29:26 Why not "ForthBox Shithawk" if you're going to choose squawky annoying birds? 17:29:51 madgarden: If you don't like the naming convention, you can always build and market your own computers. :) 17:30:03 Put the Raven in cans? Preserves! Raven Marmalade... see? ForthBox Marmalade, and yout logic follow. ;P 17:30:11 your 17:30:15 follows 17:31:20 I recall the "Chjornyjj voron" song. 17:31:51 Good name for muP. 17:32:50 Raven is believed as death messenger here. 17:33:51 ASau: Ravens are characteristically associated with prankster or delivery with bad news amongst many native tribes. 17:34:00 s/prankster/pranksters/ 17:34:15 I see. That's the case. 17:34:16 This is because the birds are quite intelligent, and work in teams to get at food. 17:34:38 An interesting fact. 17:34:40 One group of birds will often distract an animal while the other birds sneak in from behind and take the food. 17:35:25 Raven is used on Radio- Chemical and Biological Defense Forces emblem. 17:35:26 Crows do not exhibit this behavior, for example, despite being very close bretheren to ravens. 17:36:02 go for owl 17:36:05 night animal 17:36:09 like us.. 17:36:10 :p 17:36:13 me anyways 17:36:43 qFox, morning is not far in future. 17:36:50 Wait for 2 ours. 17:36:52 Well, if Raven is a bad choice for marketing purposes in Russia, I'm open to alternative suggestions. 17:36:56 MARMALADE IS RIGHT OUT. 17:36:58 its 2:30 am 17:37:08 owl. 17:37:11 kc5tja, better Raven. 17:37:33 ostridge (uh spelling correct?) 17:37:46 Owl doesn't work for me. 17:37:52 Ostrich doesn't either. :) 17:37:59 those birds that cant really fly and hide by sticking their head... right 17:38:13 Actually, they don't actually stick their head in the ground. 17:38:20 They do lay it flat on the ground though. 17:38:32 would be a bit tough on hard soil 17:38:36 Exactly what it gets them, I'll never know. 17:38:56 It's the classic "I can't see you, so you can't see me" line of thought, I guess. 17:39:15 must've worked at some point 17:39:16 :) 17:39:20 for them 17:39:56 If I meet ornitologist I'll ask. 17:40:06 In case I won't forget. 17:40:43 isnt that a bird guy? 17:40:45 :) 17:41:15 I know some of them. 17:41:59 I've got acquainted on the Black Sea shore. 17:42:04 heh :) 17:42:29 anyways 17:42:32 * qFox goes bugtracking 17:42:54 They were wandering around doing their research works. 17:44:23 Ah. No. They were actively spending their vacation. 17:45:43 Though there were enthomologists that were collecting insects for their research. 17:49:32 Wow. 17:49:36 The sky is red outside. 17:50:18 Well, it's blue, but the sunset is so vivid that *any* cloud what-so-ever is just luminous in this super brilliant red shade. 17:50:21 I wish I had a camera.' 18:02:53 Blazin 18:03:02 Blazin' Forth on a C64 emulator is fun :) 18:03:37 ForthBox: Chicadee 18:04:02 Er chickadee 18:04:47 --- quit: Nutssh ("Client exiting") 18:04:57 Chickadee sounds like the name of a northern-californian hamradio kit 18:05:39 --- quit: Aldo[Working] (Read error: 104 (Connection reset by peer)) 18:05:44 How would one implement a REDEFINE word, like on the Jupiter Ace? 18:05:56 --- join: AldoBR (Aldo@200.195.110.69) joined #forth 18:06:51 i'm not exactly clear on what REDEFINE is supposed to do 18:06:55 if you have: 18:07:01 : foo ." Hello" cr ; 18:07:12 : bar foo ; 18:07:19 : foo ." Blah" cr ; 18:07:57 Not familiar with Blazin' Forth for Commodore 64. 18:07:59 redefine bar ." Marmalade!" cr ; 18:08:00 then before REDEFINE, BAR prints "Hello", and after REFEFINE, BAR prints "Blah"? 18:08:29 In a way, I wish I were familiar with Forth for the 64; I might not have had a need to get the Amiga. Though the Amiga was purely an impulse thing anyway. Ohhh...it was the machine to have back then. :D 18:08:41 Yep, if the second FOO is your redefinition. 18:09:06 hmm 18:09:07 I saw International Karate + and R-Type on the Amiga, and had to get one. 18:09:20 REDEFINE does just what you expect... redefines an existing word. 18:09:33 oh, then i have it wrong 18:10:47 you had a Jupiter Ace? 18:11:05 Sonarman, the example you gave would not change the behaviour of BAR with the second FOO definition. 18:11:10 I'm back... 18:11:27 Doughbridge vecher! 18:11:29 madgarden, why can't you do: : same-name ... ; ? 18:11:55 ASau, well that creates a new word. 18:12:01 REDEFINE changes an existing word 18:12:10 : same-name POSTPONE ... ; IMMEDIATE ? 18:12:12 Jupiter Ace: http://hem.passagen.se/tiletech/forth.htm 18:12:36 Right now, I'd just plan ahead and use a DEFERed word... but, I am wondering how the JA did it. 18:12:51 --- quit: qFox (Read error: 104 (Connection reset by peer)) 18:12:59 It uses REDEFINE because it has no editor. 18:14:13 Hm. If you have enough parameters reserved then you can do: : new-name new-behaviour ; ' new-name CFA ' same-name ! ' ;S CFA ' same-name CELL+ ! 18:16:37 I mean memory in PF. 18:17:40 If you have modern-style DOES> then you're able to replace old CF and one PF cells. 18:17:48 Yea, I was thinking of something along those lines... basically a retro-defer on the old word to a new noname entry. Seems a tad unsafe though. 18:18:00 Right. 18:18:12 There's better way: FORGET 18:19:25 Sure, but that's not the question. ;) 18:19:42 Anyway, your conclusion is the same as mine. 18:20:08 Unless they shift and relink the dictionary after the redefined word, which seems a bit excessive. 18:24:03 Hey cool... Blazin' Forth for the C64 was "OK TO COPY, BUT NOT TO SELL!" 18:24:17 Hmmm. 18:24:46 Any automatic binary editing of dictionary is unsafe. 18:25:02 Most time it crashes environment. 18:25:20 Define: automatic binary editing 18:25:52 I don't mean adding new words or FORGET-ting. 18:26:23 Heh :) 18:26:33 I mean editing PF-s of coloned words. 18:27:48 Well, if it can be guaranteed that a coloned word has one parameter, then the redefinition method we discussed should be almost OK. I just see a problem if words after the redefined word are forgotten. 18:27:51 kc5tja, do you have Blazin' Forth? 18:29:36 When you come editing coloned words, you should raise FENCE 18:30:52 I've had good experience of crashing after ' NW CFA ' W ! FORGET X when X was defined earlier than NW 18:31:30 "Good" means it was not critical. 18:35:02 i finally discovered why the : NAME FACTORS ; syntax is the way it is. 18:36:08 Whaddya mean? 18:36:19 How : parses ahead? 18:36:36 it maps nicely to the way a 'real' assembly forth stores the dictionary. 18:46:25 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 18:46:55 Dobroe utro! 18:47:14 madgarden: No. 18:48:55 kc5tja, want it? 18:51:28 Sure. I don't have VICE set up anymore, but I can play with it later on. 18:51:37 Please send it to me via e-mail 18:51:53 * kc5tja may not be able to DCC RECV 18:52:08 you could just download it 18:52:14 it's on the taygeta ftp site 18:52:18 That'd be even better. 18:53:40 ftp://ftp.taygeta.com/pub/Forth/Compilers/native/misc/commodore64/blazin64.zip 18:53:47 VICE is so nice :) 18:54:12 I'm slow, sorry. Well yea, there you go. 18:54:26 * madgarden lubbors the VICE. 18:55:11 If you get yourself into an infinite loop in Blazin' Forth, press run/stop-restore. :) 18:55:18 Yeah, VICE is bar none the best C64 emulator I've ever used. 18:55:41 : bdrfun 53280 c@ begin 18:55:41 Yes, because that triggers an NMI. :) 18:55:57 16 0 do i 5380 c! loop 18:56:05 53280 c!, i mean 18:56:08 And in the C64, the NMI cannot be masked at all (like it can be in the PC, via a special I/O port). 18:56:15 CCS64 is my alternative C64 emulator. It requires DX9 now though, so I haven't updated it lately. 18:56:17 ?terminal if 18:56:30 53280 c! ." Goodbye!" exit 18:56:32 then again ; 18:56:35 whee :) 18:57:28 You know, if you're careful enough with that, you can bit-bang the VIC-II's color registers to communicate with other devices (assuming that they had some optical means of communicating). :) 18:57:56 Those Casio watches that got appointments and current time information from the PC comes to mind. 18:59:02 hacking c64's? hehe 18:59:13 did any personal computers from the 70s and 80s come with built in forth instead of basic? 18:59:28 slava: Jupiter Ace 18:59:31 --- join: Nutssh (~Foo@gh-1177.gh.rice.edu) joined #forth 18:59:35 Canon Cat 19:00:02 slava: Jupiter Ace 19:00:06 Sun SPARCstations. :D 19:00:22 chandler: From the 70s and 80s? I doubt that. 19:00:34 heh 19:00:38 sparcs are fun 19:00:41 I thought by the end of the 80s Open Firmware was around in initial form. 19:01:04 sun 3's had something similar, with a different name. 19:01:07 i think that's late 80's. 19:02:04 Blazin' Forth runs much faster than C64 BASIC, that's for sure. 19:02:22 chris-xp, i started a factor tutorial, but its still really lame: http://slava.kicks-ass.org/slava/ftut.html 19:04:13 madgarden: Does it compile to native 6502 machine language, or is it a traditionally threaded Forth implementation? 19:05:38 kc5tja, I really couldn't tell you. Just did a quick visual test with flashing the screen border. 19:06:09 it's traditionally threaded, i'm pretty sure 19:06:20 slava: I'll look through it later. 19:06:34 Argh, where's the damn ' key... 19:07:00 anyone know how to get to the secondary set of foreground colors in VICE? :-) 19:07:04 chris-xp, nothing in it is new. it only convers basic postfix notation so bar. 19:07:14 Ctrl-# accesses the primary set 19:07:28 I did this: 19:07:28 : test 1000 0 do 16 0 do i 53280 c! loop loop ; 19:07:43 madgarden: MINE IS BETTAR 19:07:45 And the equivalent under BASIC. 19:07:58 Sonarman, :P 19:08:11 madgarden: set the keyboard mapping type to symbolic 19:08:20 then the keys are easy to find :) 19:08:35 CTRL-#? 19:08:35 (under the right-click menu in linux) 19:08:56 sorry, by # i meant one of the numbers 19:08:57 I'm under WIndows... I'll check the menu. 19:09:43 btw, when you're loading something from a "disk" in VICE, switch to warp mode (M-w) 19:09:51 Anyway, the Forth version of that test took a few (< 10) seconds to complete, and the BASIC one was several minutes. 19:10:14 Sonarman, yea I always use that warp trick. :) 19:10:54 :) 19:11:12 Plus, you can visually see the difference in the retrace on the border... the FOrth version has quite thin colour-cycling bars compared to the randomly blocky BASIC version. 19:11:52 Wish I had found Forth as a kid. :( My games would have been so much faster! :D 19:12:09 ok, i've determined that Blazin' is traditionally-threaded 19:13:14 RTFM'ed? 19:13:33 --- join: dasbear (das@adsl-64-219-100-35.dsl.lgvwtx.swbell.net) joined #forth 19:13:38 no, just did a little test 19:13:41 : foo ; 19:13:45 : bar foo ; 19:13:57 ' foo . 19:14:09 ' bar >body . 19:14:54 Yea, if I coulda found the damn ' key... ;-/ 19:15:20 --- join: jdrake (~uniq1@CPE00045afdd0e8-CM014410113717.cpe.net.cable.rogers.com) joined #forth 19:15:32 hi jdrake 19:15:41 hi jdrake 19:15:43 EH 19:15:49 hi jdrake 19:16:06 hello 19:16:10 thank you for the warm welcome :-) 19:16:14 :) 19:16:19 We're playing with C64 Forths. Come join us! :D 19:16:25 :) 19:16:34 theres this C64 Unix ... it makes me want to play with it 19:16:45 i have to go have a shower :p 19:17:05 Is there a d64 for that unix? 19:17:13 lng.sf.net 19:17:25 LUnix Next Genarayshun! 19:17:39 and there's also WINGS 19:17:50 which is a unix-like GUI OS, I think 19:17:59 wait, WINGS requires a SuperCPU 19:18:16 Q: has anyone ever made a (portable, but possibly in host os) forth environment similar to how basic was when it booted up on systems in ancient times? (so it was already in graphics mode for example, and you could interactively do things without 'running a program', like you do now. 19:19:24 jdrake: i think the Jupiter Ace was like that, if I understand you correctly 19:19:37 --- part: dasbear left #forth 19:19:45 i just recall the power that used to have 19:20:46 things like python have claimed they are to advantage in that you skip the compile cycle. those old basics were of advantage that you didn't have to compile or do much of anything that you could run it right there interactively more so than python does 19:20:48 btw, ' bar >body . show have an @ after >body 19:21:09 and forth could be made to do the same sort of thing (it does already work that way in console of course) 19:21:11 /show/should 19:21:46 jdrake, you could say that Blazin' Forth for the C64 is like that as well. 19:22:16 C64 is that way natively so it would make sense 19:22:53 I was also considering making a graphical Forth console using Allegro, to do something like you describe. 19:23:03 Using my Forthy system. 19:23:32 Sonarman, argh, can't seem to resolve this damn C64 keyboard mapping issue. 19:23:58 i would probably stick to ficl as i know it works 19:24:27 you could easily 'sell' (not $$) such a system on almost complete portability while running at native speeds 19:24:50 Sure, FICL could easily be hooked into something like Allegro. 19:25:35 madgarden: well, ' should be shift-6 or somewhere around there, then 19:26:16 Getting latest VICE... 19:26:41 are you talking about that allegro graphics library? 19:27:58 Yep. 19:29:43 --- quit: chris-xp (Read error: 104 (Connection reset by peer)) 19:29:44 --- join: arke (~Chris@wbar8.lax1-4-11-099-104.dsl-verizon.net) joined #forth 19:29:47 i would use sdl, because i don't think allegro is osx compatible 19:30:12 hmm, they say it is now 19:30:15 might check that out 19:30:24 sdl is limited by comparison when i looked at it prior 19:31:13 i remember about 15, 16 months ago i was trying to get allegro working on OS X 19:31:13 Allegro is definitely OSX compatible. 19:31:30 by "trying" i meant trying to compile it, not coding on it 19:31:32 We've got an OSX build of our Allegro-based Sword of Fargoal remake. 19:31:56 i finally got the demo prog to run in X11 hehe :) 19:32:08 but that was a long time ago 19:33:54 have you guys ever tried the 'shareware-equivelent' form of free software -> Free (GPL) with a donate button? 19:34:04 jdrake, i do this for jedit 19:34:04 Allegro is nice and high-level, and you can bang stuff together with it pretty quickly. 19:34:14 slava, does it work? 19:34:26 jedit i believe was a popular editor 19:34:27 jdrake, yes. not huge money, but still ok 19:34:39 slava: are you Slava Pestov? 19:35:09 my figuring is - i like the idea of free software, but I also like the idea of shareware. If it is free your paying customer base could be higher than with shareware in some cases 19:37:42 Sonarman, yes 19:39:09 wow :) 19:43:29 --- quit: qFox ("if at first you dont succeed, quit again") 19:46:00 Sonarman, why? 19:46:44 why what? 19:48:15 * kc5tja loves the idea of shareware. 19:49:18 I need a really exotic text editor 19:49:22 non-X11 19:49:31 I'm considering ed.... lol 19:49:55 mcedit 19:50:07 link? 19:51:02 Don't have a link. Used it under Redhat, came with the distro. 19:51:19 anyone remember what editor I440r uses? is it joe? 19:51:36 It's a lot like DOS edit. 19:51:37 http://www.falvotech.com/cgi/fsforth/ForthBoxKestrel -- feel free to let me know what you think. 19:51:59 arke: you have mcedit 19:52:49 well, if its like DOS edit, then its not exotic... 19:53:49 So you want like... PornEdit? 19:54:23 Oh wait... that's erotic. 19:54:28 LOL 19:54:42 :) 19:56:39 arke, what about vi or emacs? 19:56:50 Not sure what you mean by exotic, or why... 19:57:10 DON'T SAY EMACS AROUND ARKE 19:57:49 madgarden: EMACS IS @#^#$^@$^@^$%^#@%#@#$^#@^$ DONT EVER MENTION THAT EVIL NAME AROUND ME EVER AGAIN 19:58:03 Oh, I didn't... I said... e... mush... yea. It's like a MUD editor. 19:58:09 arke, emacs! 19:58:11 actually @#^#$^@$^@^$%^#@%#@#$^#@^$ is TECO 19:58:23 slava: @ &$@%$^@!^^^#!$^$#$!@%&!@$@!&!@&#@%&#@%&@%#&%#@!&@!&!@&@!&&#@%!@!#%&@!#&@%&@%#&@@^&@!&@!&@!&@#!%& 19:58:35 C-m-X ^- actuate-frobincator C-c X-k is emacs 19:58:36 slava: do you have cramps yet? 19:58:43 arke, i don't use emacs 19:58:44 fetch-store-fetch-fetch-store-fetch!! 19:58:54 slava: esc ctrl+b m+aeqw shift+f1 19:59:00 madgarden: lol 19:59:09 arke, i use jedit 19:59:11 Sonarman: lol 19:59:13 slava: good boy 20:01:05 i added a cool recoil effect in my game when you fire big guns 20:01:47 You folks should stop developing for the PC. Start coding for ForthBox today! :D 20:02:28 kc5tja, OK... can I get a free sample? 20:02:55 madgarden: :) 20:03:19 Sure, I'll send you a free, complimentary ForthBox Kestrel once I receive your free, complimentary payment of $95. :) 20:03:37 that $95'll $95 shipping and handling 20:03:46 ummm...yeah. :) 20:03:47 That's it. 20:04:15 Sounds just like eBay. 20:04:34 no, MY free, compilmentary $95 to YOU will require $95 in shipping&h from YOU 20:05:26 Hehe 20:05:33 If only the world worked that way. :D 20:06:58 Only one RS-232 port? 20:08:59 Quite likely. 20:09:04 Depends on cost, again. 20:09:26 If it turns out cheaper to use a PIC chip to drive the serial ports, then I will probably go with 2. 20:12:27 Any chance of a parallel port? 20:14:57 i just fuxored my BForth session :) 20:15:07 BForth? 20:15:13 a forth written in Brainfuck? 20:15:23 why, Blazin' Forth for the Commodore 64, of course! 20:16:01 ooooooooooooooooooooooooooooooooooooooooooh 20:16:02 ok 20:16:02 lol 20:16:26 i wrote past the framebuffer 20:17:28 allegro compiled nice 20:17:32 it works on osx 20:17:39 Sonarman: :) 20:17:43 one potential problem is it seems to studder on the display 20:20:18 jdrake, how so? What are you doing, software double buffered drawing? 20:20:37 --- quit: tathi ("leaving") 20:20:49 i didn't do much , just chose quartz window mode in the demo 20:21:03 Hmm, I have no idea what that is. :) 20:21:04 i do have other things running that are taking avg. 30% of the cpu, so that could be it 20:21:18 quartz is the display technology that osx uses 20:22:18 Sonarman: Does Blazin' Forth use character-mode display or a bitmapped display? 20:22:37 Well, if you have any problems, go to www.allegro.cc and post in the forums. Or there are mailing lists at alleg.sourceforge.net. But, it's got some pretty flexible options for graphics modes and such, so you should be able to optimize it in various ways. 20:24:04 kc5tja: chararcter-mode, at least by default 20:26:47 Ahh 20:27:21 was framebuffer the wrong word to use? 20:27:40 No; but it has a slight bitmapped connotation. 20:27:48 But it's not incorrect usage. 20:29:19 * kc5tja thinks he is going to implement a PIM package for the Kestrel in Forth, just to demonstrate what is possible with the machine. 20:30:58 Sonarman, do you have any other C64 forths? I seem to recall a few others... one's on a cartridge. 20:31:01 And implement a new variation of LOAD called +LOAD, to load blocks using *relative* block numbers. 20:32:30 madgarden: there's an interview with Ballantyne in an issue of C=Hacking where he says that all other C64 forth's suck :) 20:33:04 kc5tja: a lot of forths have +LOAD 20:33:26 well, maybe not a lot 20:33:42 i don't know 20:34:09 how would you like this baby... http://www.weaponmasters.com/index.html?ID=5fa70ff82006b8b3a598b307228032eb&SORT=&ITEM=UC-1250 20:34:15 Sonarman: I'm not aware of any. 20:34:20 But it can't be that hard to implement. 20:34:30 Sonarman, that's probably why I don't have them then... I know I tried some other ones, but can't find them. Must have deleted 'em. :) 20:34:45 : +LOAD BLK @ + LOAD ; :-) 20:35:27 kc5tja: bigFORTH does 20:35:41 lol haha i guess you're right it isn't hard to implement :) 20:36:22 Hey, gforth has it too. Implemented exactly as I've done it too. :) 20:37:23 BoxOS is going to be more like a ColorForth though. 20:37:41 I'm thinking of using proportional-width fonts (due to 320 pixel screen width being so small). 20:37:50 And monochrome screen leaves zero room for antialiased fonts. 20:38:09 kc5tja, aren't you going to have to use a shit load of code to actually get this thing running? 20:38:30 jdrake: Not really. 20:38:31 kc5tja, What's wrong with 40 columns? 20:38:45 madgarden: Code density. 20:38:56 320x480 though? 20:39:11 --- quit: Sonarman ("Reconnecting") 20:39:15 I was thinking of booting the system in 320x240 "mode." 20:39:19 can't you get it any larger than 320? 320 truely is limiting in the extreme 20:39:27 Keeps the framebuffer (sorry!) size reasonable. 20:39:55 jdrake: http://www.falvotech.com/cgi/fsforth/CpuDoesVideoRefresh 20:40:19 --- join: Sonarman (~matt@adsl-64-160-166-159.dsl.snfc21.pacbell.net) joined #forth 20:40:47 Forth code density is already pretty good... let the user switch to 320x480 to get more. 20:41:36 Well, I can get 64 characters across if I use a 5-bit wide font. 20:41:47 But that involves using the same code as a true proportional-width font anyway. 20:42:27 40x30 is still 1200 characters per screen. 20:42:58 --- join: Sonarman_ (~matt@adsl-64-160-166-159.dsl.snfc21.pacbell.net) joined #forth 20:43:01 --- quit: Sonarman (Client Quit) 20:43:24 True. 20:43:25 Seems plenty to me, especially for the beginner hacker. 20:43:41 one thing I have not been really understanding is how you actually expect to do the video refresh. Is it something that the user program will have to do or are you going to have another process run doing that 20:44:16 The video PIC will periodically interrupt the CPU via its NMI (non-maskable interrupt) pin. 20:44:42 The NMI routine will then dump 40 bytes of data into the video shift register. 20:44:58 When the NMI routine returns, it'll return to whatever program was executing at the time (even another interrupt routine). 20:45:53 basically its just like an interrupt would work on an XT 20:46:00 Now, if the application program WANTS to subjugate the OS to do its own video refresh tasks, it certainly can. 20:46:30 jdrake: ?? Last I recalled, interrupts worked the same way on any computer. The IBM XT, however, did not use the CPU to refresh the display. It was woefully too slow to do it. 20:47:21 The nice thing about the 6502/65816 is that it has *excellent* interrupt response times. This is entirely doable. 20:47:38 If a Z-80 could do it back in 1979 at 3MHz, a 12.6MHz 65816 can *definitely* do it. :) 20:48:31 * kc5tja adds discussion on how the 65816 knows when to fill video shift register to Wiki... 20:50:54 is it possible for the kit constructor to replace this 65816 chip with a faster chip? 20:50:59 * warpzero is back (gone 08:42:03) 20:55:43 BLAZIN' FORTH TIP OF THE MOMENT: don't forget to include a ?TERMINAL conditional in words that contain an infinite loop 20:56:09 (brought to you by Phillip Morris USA - visit our website for help on quitting smoking) 20:56:20 SEE ?TERMINAL 20:56:21 ?TERMINAL IS UNKNOWN OK 20:57:07 I love how Blazin' Forth patronizes me with every command... 20:57:07 Like, DUP is CODE OK!?! 20:57:35 haha :) 20:57:38 i'm working on my SEE implementation right now. 20:58:01 madgarden is lose 20:58:11 I still have to write some sort of SEE. 20:58:54 madgarden: ?terminal exists, you just can't SEE it; i don't know why 20:59:06 i need to make SEE print the definition in some kind of usable format. 20:59:19 Sonarman_, maybe you can't SEE immediate words in that forth? 20:59:21 Problem is, I have to do it in a language and output independent fashion. I may just throw a stack of XTs and strings onto the stack for the user to peruse. 20:59:56 slava: maybe 21:00:10 btw, how do you do recursion? 21:01:02 jdrake: in ANS forth, i think you use RECURSE 21:02:39 is RECURSE a GoodThing? 21:02:43 should i add it to my language? 21:02:56 (i don't have the restriction that a word can only be referenced after its been defined) 21:03:42 slava, in lisp - recursion is the ONLY loops available 21:03:54 i highly recommend them :-) 21:04:33 jdrake: he's talking about using the word RECURSE instead of the actual name of the word currently being defined 21:04:59 he KNOWS lisp :) 21:05:09 for some definition of 'knows' 21:05:24 never written anything > 400 lines or so in it. 21:06:03 jdrake, in fact my language is lisp + forth :) 21:08:23 slava: I recomend using recursion 21:08:53 Herkamire, no i use recursion a lot 21:08:57 but shoudl I add the RECURSE word 21:13:14 is needing a 3DUP indicative of bad style? 21:13:35 needing 3dup is a bad sign 21:13:43 use variables or something 21:13:59 my language doesn't have locals. 21:14:24 slava: I don't see why you'd want recurse if you can do the same thing without it 21:14:43 slava: I don't mean locals 21:15:00 I mean variables 21:15:44 slava: can you recurse like this in Factor? : jason ... jason ... ; 21:16:24 Herkamire, yes 21:18:01 Done. 21:19:13 jdrake: The bus interface is 65816-specific. If you design your processor module to look like a 65816 to the rest of the computer, I don't see why not. 21:19:29 Fortunately, the 65816's bus interface is pretty trivial. 21:22:19 The CpuDoesVideoRefresh page now has information on how it basically works (err, will work, more accurately, but you knew that). 21:23:31 you're on a roll, kc5tja! 21:24:12 Heh :) 21:27:16 madgarden: know how to get a random number on the C64? 21:28:33 Sonarman_: Read the random number (aka White noise) register of the SID chip. 21:30:53 thanks kc5tja 21:32:28 --- quit: Nutssh ("Client exiting") 21:36:34 no problem. 21:39:38 i don't know anything about the SID (or the rest of the C64) though, but address $dc04 seems to work 21:53:29 why are you guys obsessed wtih the c64 all of a sudden? :) 21:54:02 kc5tja: where can i find out what a video shift register is? 21:54:36 slava: i don't know. it's just fun :) 21:58:15 --- join: hovil (~hovil@CommSecureAustPtyLtd.sb1.optus.net.au) joined #forth 21:59:50 Sonarman_: It's just a normal shift register. 22:00:11 You load it with a byte, and each clock cycle, it shifts to the left or to the right, one bit at a time. 22:00:29 They're also called Parallel to Serial converters. 22:00:55 (also, a shift register can be used to convert serial back into parallel, but I'm ignoring that application for the Kestrel) 22:04:29 --- join: Serg (~z@212.34.52.140) joined #forth 22:19:17 --- quit: Sonarman_ ("leaving") 22:26:25 Dobroe utro, Serg! 22:29:21 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK") 22:32:55 hi 22:33:36 ASau: take a look at my new perl scripts. http://cryptomancer.narod.ru 22:34:12 Serg, I never have been interested in Perl. 22:34:35 And I do not plan to learn it in future. 22:34:56 It's a kind of boycott. 22:35:09 why ? 22:35:25 i was not interested too (writing in Forth only long time) 22:35:36 I looked at it once. 22:35:44 It is sufficient. 22:36:17 but once i needed a command-line news client, but failed to cope w/ strings and sockets in SP Forth 22:36:53 in Perl the task was trivial, despite zero knoweledge and looking up a book for writing every line 22:36:56 I plan to write news client in Forth. 22:37:13 dialog or dumper type ? 22:37:37 I think I'll use netcat utility. 22:37:53 It should be half-off-line. 22:38:18 RUnet really lacks a good news client ! 22:38:37 Two passes: 1) headers only; 2) messages. 22:38:41 Outlook is buggy, ForteAgent - codepage handicapped 22:39:07 Serg, there's an option to tune Emacs. 22:39:15 --- join: MarkT (mark@tom.nss.udel.edu) joined #forth 22:39:27 Dobroe utro, MarkT! 22:39:34 Hi! 22:39:35 :) 22:39:43 i found emacs overkill complex 22:40:07 and Lisp is hard to grok, after C and Forth 22:40:10 question, are there any docs in english for SP-Forth? 22:40:19 source ;)))) 22:40:31 MarkT, use ANS standard and source. 22:40:47 ok, thanks :) 22:41:09 There is no Russian docs, only English version of ANS. 22:41:19 Authors are too lazy to write docs. 22:41:24 lol 22:41:28 MarkT: is SP popular outside RU ? 22:41:37 I dont much care to either, so I understand 22:41:54 Serg: I don't know, I just got it myself. 22:42:12 (On recommendation from ASau) 22:42:32 Serg, I don't know another non-minimalist environment for Windows. 22:42:32 i know web server is written in it ;) it runs http://forth.org.ru 22:42:47 Serg, I run this server. 22:42:53 It's impressive. 22:43:03 you run forth.org.ru ??? 22:43:17 No, I mean acWEB. 22:43:41 ;) 22:43:55 You can sit at SPF console with HTTP running in background. 22:44:14 i run Apache on WinXP to test my site 22:44:24 is acWEB free software ? 22:44:35 I don't know. 22:44:48 You may contact authors. 22:44:57 Author. 22:45:23 so, anyone can add extra features to server, even while it running ? 22:45:26 http://sourceforge.net/projects/acweb/ 22:45:28 cool ! 22:46:20 I remember DZ offered his Palm DS-Forth for all RuFIG connected folk. 22:46:56 Now it seems they made it GPL or smth like. 22:48:18 Or it was not DZ... Read: "authors". 22:49:18 palm ds-forth? 22:49:43 I am not to remember everything. 22:49:45 I only remeber finding 2 palm forths 22:49:52 quartus and one other 22:49:54 Consult RuFIG main page. 22:50:00 * hovil ponders 22:52:09 Yes, RuFIG news: "20.01.2004 day: Dragon Forth for Palm goes GNU PL" 22:52:31 excellent! 22:52:53 russia seems to have a lot of forthers =) 22:53:28 Try to dig in RuFIG site. 22:54:03 It may happen you'll find useful stuff there. 22:54:47 http://www.forth.org.ru/ 23:05:18 --- quit: jdrake ("Oops. This machine just fell asleep") 23:20:17 http://russian.foundries.sourceforge.net/repository.pl?section=russian&op=list 23:20:23 odd that it returns no list 23:55:25 --- quit: Serg () 23:59:59 --- log: ended forth/04.03.02