00:00:00 --- log: started forth/02.03.06 00:03:21 --- join: rob_ert (~robert@kalys.unixtech.be) joined #forth 00:03:46 --- quit: rob_ert (Client Quit) 00:08:26 --- join: rob_ert (~robert@kalys.unixtech.be) joined #forth 00:56:24 --- quit: rob_ert ("CGI:IRC 0.4.3 [EOF]") 01:04:20 morning 01:21:07 --- join: Stepan (~stepan@Charybdis.suse.de) joined #forth 01:35:51 --- quit: onetom (carter.openprojects.net irc.openprojects.net) 01:39:28 --- join: onetom (tom@adsl52161.vnet.hu) joined #forth 02:04:17 --- join: rob_ert (~robert@kalys.unixtech.be) joined #forth 02:26:54 --- join: qless (~cerberus@clgr000977.hs.telusplanet.net) joined #forth 02:27:05 howdy folks 02:28:33 hmmm 02:28:57 --- quit: qless (Client Quit) 02:50:08 --- quit: rob_ert ("CGI:IRC 0.4.3 [EOF]") 04:32:44 --- join: aum (~david@l76-131.world-net.co.nz) joined #forth 04:39:45 --- part: aum left #forth 04:45:02 but i can play more easily w various multitasking/multiuser issues 07:09:32 --- join: Fare (~fare@samaris.tunes.org) joined #forth 07:49:49 pl mennyire hi 07:49:54 oops :) 07:49:56 hi 07:50:35 eForth v1.0f 07:50:36 Ok 07:51:19 s" architecture" environment? type 07:51:29 (or sg like that :) 07:51:54 :) 07:52:23 Linux 08:01:50 but eforth is usually used for embedded applications 08:02:32 do u have any 4th related problems or u r just here 4 fun? 08:03:14 4 fun 08:03:44 what r u working on nowdays? 08:03:45 unless by "Forth related problem" you include "has fundamental problems with the low-level nature of FORTH" 08:04:01 I'm working with LaTeX and HTML :( :( :( 08:04:10 :))) 08:04:11 programming with common lisp 08:04:14 poor u 08:04:17 or zsh 08:04:25 or (yuck) perl 08:04:48 i still use awk for such problems 08:04:57 but later i will use python 08:05:11 its clearer than perl, a lot 08:06:32 anyway, i also have problems w 4th low level nature 08:07:05 eg. i have to add logging facility to my program 08:07:54 and a printf like interface would b more convenient that the usual forth . ." TYPE EMIT <# # # #> ... 08:14:18 onetom: try common lisp one day 08:14:32 or maybe pop-11 08:14:42 pop has a high-level forth underneath 08:24:48 im working 4 the embedded market, so i have to 08:24:57 stick w forth 08:26:34 forth can b ported to microcontrollers, but i guess lisp is not :( 08:27:08 but what kind of problems do u solve in lisp? 08:28:22 [2002-03-06 17:26:41] -- infra: ON 15 5 4 08:28:22 [2002-03-06 17:26:41] -- infra: ON 14 5 2 08:28:22 [2002-03-06 17:26:42] -- infra: OFF 14 5 5 08:28:22 [2002-03-06 17:26:44] -- infra: OFF 15 5 4 08:28:22 [2002-03-06 17:26:45] -- infra: ON 15 5 4 08:28:44 oops - ive misrouted it 08:34:20 you solve all kinds of problems 08:34:34 about any problem except those where you need to directly program a microcontroller 08:34:59 but including problems where you metaprogram a microcontroller from LISP 08:38:41 --- quit: Stepan ("Do you think it is air you are breathing? Hmm?") 08:43:16 --- join: qless (~cerberus@clgr000977.hs.telusplanet.net) joined #forth 08:43:27 hi 08:43:33 heya onetom 08:44:03 Fare: "metaprogram a microcontroller" ??? u mean simulate it? or what? 08:44:14 qless: hows life? 08:44:21 friggin cold 08:44:45 -35C 08:44:58 hm... where r u (from)? :) 08:45:06 .ca 08:45:10 in a frige room? 08:45:18 :) no 08:45:30 ah, thats just the neigbourhood of .hu 08:45:33 I mean use a normal development workstation to develop programs that will run on the microcontroller 08:45:41 up in the mountains? 08:45:50 near to the mountains 08:45:56 which might include simulation, and might include interactive debugging 08:46:14 fare: that sounds cool... what kind of microcontroller? 08:46:20 (with most of the debugging facilities, i.e. symbolic disassembly, being done on the frontend rather than the host) 08:46:34 Fare: aham. and type of microcontroller have u ever programmed? 08:46:35 brb 08:46:46 qless: I know that some forth do that. Some people did that with LISP on the frontend, too 08:46:54 onetom: only 6502. 08:47:10 yep. forth is really good at doing cross-platform stuff 08:47:11 and it was in an Apple ][, not in a microcontroller. 08:47:57 you know those keypad the fedex guys carry? they were programmed in forth 08:52:53 qless: havent ever seen a fedex guy, but have u ever seen the film "cast away"? :) 08:53:59 qless: it starts w a russian fedex story. how does it contiune? well, dunno, havent seen yet 08:54:14 onetom: brb 08:54:30 Fare: interesting... most people met only 6502 :/ 08:54:58 --- nick: onetom -> onet-shower 09:02:24 --- join: futhin (thin@h24-64-175-123.cg.shawcable.net) joined #forth 09:06:32 hey futhin. you're up early. coding hard? 09:07:51 i'm up early? 09:08:04 i get up at 7am 09:08:10 yeah, its not even noon 09:08:27 --- join: rob_ert (~robert@h173n2fls33o898.telia.com) joined #forth 09:08:27 hah, i consider getting up after 9 am late 09:08:44 not coding hard 09:08:46 reading stuff 09:08:47 bleh 09:08:54 maybe i'll design the database thingie 09:09:00 for the forthmud.. figure out how it'll work together 09:09:10 either that, or get a job 09:09:13 i don't want it to be an ugly mess 09:09:14 ;-) 09:09:18 hi :-) 09:09:21 yeah, i was going to go get a job today 09:09:25 after i watch a movie :P 09:09:25 heya rob_ert 09:09:33 hi rob_ert 09:09:43 how's the coding going rob_ert? 09:09:49 fine, fine :-) 09:10:42 rob_ert: if you need something to code in forth.. there's an assignment for ya. create a program that displays a pyramid of asterisks 09:10:46 * 09:10:49 *** 09:10:53 ***** 09:10:58 hehe 09:11:00 ****** 09:11:03 ah 09:11:06 I get what you mean 09:11:08 you know what i mean :D 09:11:22 hehe 09:11:22 so.. 09:11:30 what's the word to deplay one char? 09:11:48 yeah.. make it so that it works like this: 10 pyramid and that prints 10 lines of asterisks 09:11:58 OMIT 09:12:00 emit 09:12:01 okay 09:12:02 lol 09:12:08 qless is pulling your legs 09:12:09 EMIT 09:12:19 and the pyramid thing is called PYRMIT 09:12:23 * rob_ert qlesses qless with a large qless. 09:12:41 (ouch) 09:12:59 * rob_ert starts his forth computer 09:17:23 we were just saying how the guy in the movie "Castaway" created the first fedex forth program using sea shells and crab legs 09:17:47 what?! sea shells and crab legs are too highlevel! 09:18:29 yes.. they didn't let him come back until he created the forth program 09:18:34 you and your high, mid, and lowlevel mania 09:19:05 naw, i get annoyed whenever some FRICKIN MORON says forth is low-level 09:19:22 forth is low-level. 09:19:26 (*hides*) 09:19:29 futhin, i think you should design a mud where people go bashing each other on the head with assembler opcodes 09:20:02 would that be low-level enough for you? 09:20:05 and then qless comes with a big forth word 09:20:06 :-) 09:20:25 when _I_ say it is low-level, i'm a) joking b) talking relatively (and everyone knows it's high-level) 09:21:34 forth isn't low-level or high-level. it is all of it. it can be extended to whatever level of abstraction you need. it is high-level too. 09:21:40 it is both. it is high level. it is low level. WASHER SPIN is about as high level as things get in programming, except maybe if you're talking to a fleet of washers 09:22:26 naw. you can take it up even higher to HUMANS EXIST 09:22:27 --- quit: Fare (Read error: 104 (Connection reset by peer)) 09:22:40 and then all the washer spining and stuff will occur because of HUMANS EXIST 09:23:24 oh no, an existentialist. and probably a capitalist one too 09:23:51 i know microeconomics :P 09:23:52 play "the international" in russian for him and check if he explodes... 09:24:21 talk to Fare if you want a hardcore libertarian 09:24:48 libertarians are all sheep in business suits 09:25:30 me, i think that everybody should get an equal _start_ .. like free education and health for the first 20 years of life. after that, all the people can go on to compete with each other. increased competition = good. 09:25:38 hm 09:26:07 how does cooperation fit in? 09:26:08 i could probably write up a paper about how mixing some aspects of socialism with capitalism can increase the competition and in fact increase the success of capitalism 09:26:26 --- join: Stepan (~stepan@pD951D7D8.dip.t-dialin.net) joined #forth 09:26:33 hi stepan :) 09:26:42 hi futhin :) 09:28:13 hi Stepan :) 09:32:21 --- join: MrReach (~mrreach@209.181.43.190) joined #forth 09:32:27 void main(void) { fork(); main(); } 09:33:01 qless: are you auto-greeting with a fork bomb? 09:33:19 hehe. me? naaaah. 09:33:58 you just caught me at a particluarly black moment 09:34:05 long time no see 09:34:20 good to see U@ 09:34:22 U2 09:34:23 HI MrReach :-) 09:34:30 greets, rob_ert 09:34:42 any forth chatter today? 09:35:04 * rob_ert points at futhin. 09:35:25 futhin: how's your MUD progressing? 09:37:48 it's not 09:37:49 heh 09:38:03 i need to install linux or somesuch 09:38:05 how come? lose interest? 09:38:18 naw, just lazy, and focusing on other stuff 09:38:26 oh, that's right @:^> 09:38:39 i'll try designing the database soon 09:38:44 and see if that gets me back into it 09:38:46 linux is not an absolute requirement, by the way 09:38:56 yeah, i have 2 shell accounts 09:38:59 i could use them 09:39:06 but the problem is 09:39:21 no, windows does sockets (albeit not as well) 09:39:25 it would help to have the man pages for sockets and friends 09:39:39 qless: don't you? 09:39:48 I can point you to the man pages online 09:39:50 i do, but i don't need them anymore 09:39:56 * MrReach laughs. 09:39:57 i have a router. and i'm using ssh to connect to the guy's port 23. it's not working, even when i portforward 23 .. 09:40:13 :-) 09:40:25 futhin: would that be a Cisco router using CBOS ??? 09:40:29 futhin: ssh uses 22 09:40:37 i know the default port is 22 09:40:44 but the guy made his ssh port 23 09:40:44 23 is telnet 09:40:59 mrreach: naw, sohoware 4 port router.. 09:41:05 ok 09:41:12 you might try nyx 09:41:30 qless: you don't need the manpages? show me some forth code opening a socket connection, etc :D 09:41:48 qless writes in C, I suspect 09:41:50 futhin: im also interested in database design, so id b happy if we could discuss it later 09:41:55 well first you need to create the socket with socket() 09:42:16 qless: uh? forth code i said.. 09:42:17 heh 09:42:38 and then it depends if you are connecting out or listening in 09:42:50 bind() or listen() 09:42:56 bravo! 09:43:07 um.. what's the forth code for interfacing with the library.. 09:43:10 etc 09:43:15 that's up to you 09:43:45 futhin: when using forth, one has to use the system libraries ... they are typically specified using C syntax 09:44:09 what does the code look like? 09:44:17 the forth code 09:44:22 I've seen a forth sockets package, from the card up, but it was proprietary 09:44:37 or am i doing something lame like passing strings s" socket(blahblah)" 09:44:53 also, socket stuff should *NOT* be implemented in user space 09:45:06 you could say 'domain type protcol socket_create' couldn't you? 09:45:08 hmmm ... 09:45:23 *sort* of like that, futhin 09:45:29 or is that too c'ian of me? 09:45:32 * futhin want's to see some dang forth code :P 09:45:46 that was forth code 09:45:47 qless was simplifying quite a bit 09:45:50 qless: yeah, nobody has implemented DOMAIN TYPE PROTOCOL SOCKET_CREATE ?? 09:46:14 the domain, type, and protocol are parameters pushed on the stack 09:46:22 futhin: heh, that is what it would look like 09:46:36 yeah, now gimme the code for SOCKET_CREATE :P 09:46:50 which forth? gforth? 09:46:57 does it matter? :P 09:47:00 bigforth or gforth 09:47:00 * MrReach goes to the man pages 09:47:03 oh jesus. that's why u need a unix box so you can peruse the glibc code 09:47:09 yes, it matters a great deal 09:47:22 if it's easier in bigforth, then i'll switch over to bigforth 09:47:31 if it's easier in gforth, i'll use gforth.. 09:47:39 I'm not sure which is _easier_ 09:47:44 mrreach: have you had experience with bigforth ? 09:47:55 I've installed and twiddled with it 09:47:57 qless: show me some code that interfaces with some library :) 09:48:01 with bigborth, you can call the libc.so directly 09:48:02 for bigforth 09:48:09 but I've never done anything substantial with bigforth 09:48:21 gforth does that, too 09:48:36 I generally use gforth on the linux box 09:48:45 why? why do you think gforth is better? 09:48:49 I don't trust my windows machine to keep working 09:49:14 I don't recall bigforth running on linux 09:49:15 me neither, but i've looked at the code that calls the libX11 library and it looks just like u would expect 09:49:38 oh, that's right, bf DOES run on linux 09:49:43 bigforth has its own X11 gui etc 09:50:06 bigforth runs slower than gforth right ? 09:50:17 the interface to system libs ought to be nearly identical between bigforth and gforth 09:50:38 ok, now i need some code to see so that i have a clue where to start :P 09:50:38 mrreach: i'm puzzled how anyone would call a .so lib from gforth? 09:50:42 futhin: depends, theoretically, bigforth should be substantially faster 09:50:50 eh? 09:51:15 bigforth compiles to optimised subroutine threaded code 09:51:17 bigforth is subroutine threaded native 09:51:25 :) 09:51:29 so it *should* be faster 09:51:45 but it often isn't, because gforth is pretty tight 09:52:24 believe me, when the machine displays "1000 BogoMIPS" ... it doesn't matter much 09:52:37 good point 09:53:06 qless: the lib calls in gforth are undocumented, but they're there 09:53:28 really? i thought that the library calls where in that compiled part like the other i/o calls 09:53:38 like putc() and friends 09:53:41 only for keyboard i/o 09:53:45 oh ok 09:53:56 the interface is awkward, though 09:54:12 the other thing is that a MUD is going to have to be threaded, no? 09:54:24 because it copies from forth stack to C stack, calls, then copies back to forth stack 09:54:32 ah 09:54:54 well, I don't think it HAS to be threaded 09:55:21 hmmm 09:55:23 getting threads started is a royal bitch in Forth ... but doable 09:55:24 gforth performed better than bigforth: http://www.bagley.org/~doug/shootout/ 09:56:11 futhin: you'll run out of sockets before you run out of compute cycles, so it doesn't matter 09:56:12 eh? gforth isn't threaded? :) 09:56:41 gforth is either direct or indirect threaded, depending on what processor you're compiling for 09:56:48 why don't i need to worry about threaded? 09:57:12 much more important is what utilities are provided by the forth you're using, and how comfortable you are with the forth 09:57:19 not 'threaded' as in implementation; 'threaded' as in multiple execution contexts 09:57:35 futhin: no, not about threaded, you don't need to worry about compute cycles 09:57:49 oh! I missed that part @:^> 09:57:52 well the mud is going to be multiuser.. and possibly multitasking??.. 09:58:18 futhin: in both linux and windows, there are processes which are divded into threads 09:59:07 a process has it's own, unique memory image, while all of a processes threads share the same memory 09:59:22 * qless wonders how bigforth would deal with calling the pthreads code in libc.so... 09:59:31 i imagine bigforth and gforth suck equally.. and have roughly the same utilities :P 09:59:42 qless: I've often wondered that myself 10:00:01 gforth can call pthreads code then ? 10:00:23 futhin: if they suck, it's because they both provide only minimal and incomplete documentation 10:00:27 futhin: dunno 10:00:32 yes it can 10:00:39 however 10:01:03 there is quite a bit involved in doing threads, not just calling the library 10:01:18 you have to set up a new set of stacks 10:01:39 then you have to initialize the forth interpreter to use them properly 10:01:44 but i don't need to worry about threads? 10:01:57 you might choose to 10:02:08 it might be hairy, but it might be hairyier to use fork() 10:02:11 well i could worry about it later, when there's more than 30 people right ? 10:02:24 my response was to your question "which forth is faster" 10:02:34 when more than 30 people start connecting, i can worry about implementing threads then? 10:02:40 yes 10:02:49 or you might never use threads 10:02:55 why not? 10:03:03 two reasons 10:03:33 brb 10:03:35 1. you want to do the items to be done sequentially, or you'll have probs with locking and race conditions 10:04:01 2. you can use select() to wait on many sockets at once 10:06:49 yeah.. probably use select. will that be good enough for me, even if 500 people are playing on the mud? 10:07:22 it should be 10:07:33 the advantage of using threads ... 10:07:49 is that you can pretend that there's only one socket in the world 10:08:11 no many sockets fighting for your attention 10:08:37 select sounds simpler to me.. 10:09:09 well, threads are substantially simpler ... until they start interacting with each other 10:09:23 erm ... simpler if you discount the startup headaches 10:09:47 it's important to me for the mud design to be very simple.. the simplest design is the best design :P 10:09:54 select() means that _nothing happens_ if everyone on every socket says nothing, right? 10:09:59 * MrReach envisions a poodle jumping through hoops for doggie treats. 10:10:11 what do you mean by threads interacting with each other? 10:10:19 qless: yes, unless it times out 10:10:19 oops, i forgot about the timeout 10:10:38 doh! 10:11:42 i wrote a little piece of nonsense to analyse gnutella traffic and select() worked just fine, but all the time i wished i used pthreads to be kool :-) 10:12:39 --- nick: onet-shower -> onet-away 10:12:44 futhin: pretend that your MUD has rooms, and two people come into a room simultaneously 10:13:17 futhin: furthor pretend that each room has a list of memory pointers to people that are in that room 10:13:58 now, two people join the room, each handled by a different thread that pretends it is handling the on socket in the world 10:14:56 one thread reads the count of pointers in prep for adding one to the end, then it places the new pointer, then updates the count, then stores the new count 10:15:21 *BUT*, while it's updating the count, the other thread does the same process 10:15:35 therefore, the list gets corrupted 10:16:18 futhin: do you follow that? 10:16:29 qless: did you write that code in forth? 10:16:33 hmmm, what if when a room was entered, a controlling thread for the room was created which handles this kind of race? then your monsters can have their own threads too. maybe i'm crazy 10:21:23 --- quit: futhin (Read error: 104 (Connection reset by peer)) 10:21:23 --- quit: MrReach (carter.openprojects.net irc.openprojects.net) 10:21:23 --- quit: qless (carter.openprojects.net irc.openprojects.net) 10:21:23 --- quit: onet-away (carter.openprojects.net irc.openprojects.net) 10:21:23 --- quit: oxygene (carter.openprojects.net irc.openprojects.net) 10:24:15 --- join: futhin (~thin@h24-64-175-123.cg.shawcable.net) joined #forth 10:24:37 ugh.. 10:24:41 what happened heh 10:26:12 --- quit: futhin (Client Quit) 10:26:15 --- join: futhin (thin@h24-64-175-123.cg.shawcable.net) joined #forth 10:27:59 --- join: onet-away (tom@adsl52161.vnet.hu) joined #forth 10:29:41 --- join: MrReach (~mrreach@209.181.43.190) joined #forth 10:29:41 --- join: qless (~cerberus@clgr000977.hs.telusplanet.net) joined #forth 10:29:45 wb 10:30:14 so i asked: what's the solution if two people enter the same room? 10:30:20 with the threading problem.. 10:30:40 --- quit: ChanServ (carter.openprojects.net irc.openprojects.net) 10:30:45 poof! 10:30:45 wb, folks 10:30:55 what was the last you saw from me? 10:32:11 futhin: do you follow that? 10:32:17 i said 10:32:18 yes 10:32:21 ok 10:32:23 and then "what's the solution 10:32:38 one solution is to create what is called a "semephore" 10:32:46 ooops, "semaphore" 10:32:55 semiphore? 10:32:58 heh 10:33:11 anyway, it is a single bit that is either set or clear 10:33:32 and you block your code until it is clear and you can set it 10:34:33 but this has to be done with a specific processor instruction that reads the byte, alters the bit, and writes it back to memory in an atomic operation. 10:34:58 semaphore 10:35:40 eh? wouldn't the thread just check if the bit is set itself? 10:36:15 I'm not sure about linux, but windows provides EXCELLENT semaphore facilities 10:37:08 yes, but the same thing can happen in reading and writing the bit as can happen when updating the memory list in the first place 10:37:15 i can make it up myself.. i can just have a bit in the room 10:37:17 that is 10:37:19 * qless shivers at the thought of developing under messysoft 10:38:07 you read the byte, then a context switch occurs and another thread sets the bit, but the curent thread doesn't know that and writes it back itself and proceeds 10:39:02 qless: does glibc have any semaphore calls? 10:39:08 afghanistan attacked usa disrupting some of the open project connections! :P 10:39:36 afghanistan couldn't attack an ant hill right now 10:39:38 not exactly. what you have i think is called SYSV ipc 10:39:56 which are semaphores, shared memory, etc at the os level 10:40:18 ok, is that like pipes and mailslots? 10:40:23 yep 10:40:28 well, a little lower 10:40:40 ok 10:40:44 --- join: oxygene (oxygene@linux-future.de) joined #forth 10:40:57 hello oxygene :) 10:41:16 --- nick: MrReach -> MrBRB 10:41:18 but the kernel code has spinlocks and semaphores all over the place using the native pentium-whatever opcodes 10:41:54 is it truely worth it for me to go the threading route ?? 10:42:43 dunno. it has advantages... it is a more modern approach than using select()... it could also be more complex... and u need to have a forth that can grok threads 10:43:26 maybe i should talk to i440r too, since i'll probably end up using isforth as a show of support :) 10:43:26 but i think u don't have to make the decision now 10:43:50 and isforth should run a lot faster 10:44:10 frankly i think you should be using Simula, but i'm old-fashioned :-) 10:44:23 simula ? 10:44:53 Simula did some really innovative things for 1968. probly more innovative than Java 10:45:36 bah, forth is my perfect language. it just needs to be developed and supported more. 10:45:55 have u given any thought about using object oriented methodologies for your mud in forth? 10:46:28 some.. for handling all the objects on the mud, players, equipment, mobs, etc.. 10:46:37 and the database 10:47:18 but i don't need full bullshit OOP to use some object oriented concepts :P 10:47:56 * futhin thinks OOP has way too much hype 10:48:21 you can do object oriented programming in ANY language 10:48:28 like Fortran or basic :P 10:48:41 yikes 10:49:05 OOP has nothing to do with the language :/ 10:49:26 it can though. and sometimes it can make your life alot easier 10:50:03 i was just wondering what thought you've given to oo 10:50:57 not necessarily.. OOP can obfuscate your code. since 80% of the coders in the world are not Real Coders (tm), they end up with obfuscated code 10:51:03 and more complicated code too 10:51:09 i know bigforth has a nifty class system 10:52:12 * futhin is away. has to go for a bit. 10:52:19 --- nick: MrBRB -> MrReach 10:52:24 hm 10:53:11 --- join: ChanServ (ChanServ@services.) joined #forth 10:53:11 --- mode: carter.openprojects.net set +o ChanServ 10:53:12 isforth will not run faster than gforth or bigforth 10:54:06 hm? 10:54:32 what part of what I said is unclear? 10:54:49 it it's simpler and there's less code in isforth than gforth/bigforth, i'd imagine it'd run faster overall 10:54:56 unbelievable? perhaps, but clear 10:55:19 it it's simpler and there's less code in isforth than gforth/bigforth, i'd imagine it'd run faster overall 10:55:34 except for the fact that it's your own words running 99% of the time 10:55:54 my dictionary is smaller than your dictionary! nyaah 10:55:59 but, like I said earlier, you'll run out of sockets before you run out of computing power 10:56:06 ;-) 10:56:19 i've no interest in wasting what little computing power i've got :P 10:56:22 so it doesn't matter spit which runs faster 10:56:49 even my pentium 200 linux bock can handle 500 users 10:56:55 linux box 10:58:06 all running fork() bomb's, mrreach? :-) 10:58:53 i see the pentium 75mhz box with 16 megs of ram is all the computing power anyone needs. we should be able to surf the net, write documents, be connected to irc, have many processes running at once, and running a mud with 1000 people on it. all simultaneously with no noticeable decrease in the responsiveness of the computer. :) 10:59:23 heh 10:59:38 except for 3d graphics, intensive math, decompressing .avi files 10:59:51 and a few other stuff 10:59:52 right on. i'm still looking forward to the day i can download the entire contents of the library of congress in 3 seconds flat 10:59:55 qless: actually, we tested the forking caps of linux with a fork bomb on our big servers 11:00:12 at the time, that was a K6/2-400 with 128MB 11:00:29 that's what i had, until it blew up 11:00:32