00:00:00 --- log: started forth/04.05.23 00:08:09 i read 2/3 of the /. interview with chuck moore! 00:08:13 but thats about it 00:18:50 --- join: gl (~foo@opcode.org) joined #forth 00:25:32 --- join: mur (~mur@smtp.uiah.fi) joined #forth 00:32:56 opcode.org - thats an awesome dns :) 00:33:21 but couldnt you thunk of a better username ? 00:33:21 lol 00:46:00 :) 00:48:13 * arke yaawn 00:48:23 I should code, hrm? 00:50:59 bleh 00:51:01 or do homework? 00:51:07 or, maybe, do my finances? 00:51:15 which will surely be depleted VERY VERY VERY soon 00:54:11 --- quit: Herkamire ("bed") 01:10:35 --- quit: Sonarman ("leaving") 01:21:47 dammit 01:21:50 Sonarman left 01:21:51 ack 01:21:52 :( 01:23:54 --- quit: kc5tja ("Lost terminal") 01:25:22 --- quit: I440r (Read error: 110 (Connection timed out)) 01:41:01 --- join: Pierpa1 (bernardp@ppp-99-220.98-62.inwind.it) joined #forth 01:54:34 --- quit: Pierpa1 (Operation timed out) 02:41:27 --- join: Serg (~knoppix@193.201.231.126) joined #forth 02:42:45 hi 02:42:58 my unlim GPRS turns to be mega suxx 02:43:53 claimed 4 KBit/sec (500bytes/sec), really down to _3_ b/s someimes 03:10:29 ;) 03:16:11 * arke is away: sl33[ 03:44:10 --- join: killaweb (~killaweb@pD9E64267.dip.t-dialin.net) joined #forth 03:44:35 hello 03:45:55 may i build a machine to compile forth code for pic micro PIC18 controller on windows? 03:46:10 is there any solution in doing this? 03:46:58 i want to write programms for PIC18fxxx series and develop them on a linux or windows host system 03:48:16 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 04:06:13 --- quit: Serg (Read error: 110 (Connection timed out)) 05:07:10 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 05:07:55 killaweb: there is a PIC forth, runs under gforth. 05:08:21 Think you can get a windows binary for gforth... 05:09:55 http://www.rfc1149.net/devel/picforth 05:11:36 http://www.complang.tuwien.ac.at/forth/gforth/ 05:13:14 --- quit: tathi ("leaving") 07:42:20 --- join: I440r (proxyuser@adsl-67-65-218-133.dsl.snantx.swbell.net) joined #forth 07:56:08 --- quit: I440r ("Leaving") 08:36:39 --- quit: qFox (Read error: 54 (Connection reset by peer)) 08:36:41 --- join: qF0x (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 08:38:48 C00K13S 08:38:48 C00K13S 08:38:48 C00K13S 08:38:48 C00K13S 08:41:44 --- nick: qF0x -> qFox 08:46:38 C00K13S 08:46:40 C00K13S 08:46:40 C00K13S 08:46:40 C00K13S 08:46:41 :) 08:46:46 whats up? 08:54:14 --- join: kc5tja (~kc5tja@66-74-218-202.san.rr.com) joined #forth 08:54:24 --- mode: ChanServ set +o kc5tja 08:55:34 kc5tja: :) 08:55:37 re 08:56:09 was looking for a minimal kernel I could build upon 08:56:12 didn't find anything 08:56:21 so I'll have to write that myself 08:56:34 OS kernel you mean? 08:57:49 --- quit: killaweb (Read error: 110 (Connection timed out)) 08:57:58 yeah 08:57:58 bare essentials 08:58:00 * kc5tja looks into re-organizing his website files on his local machine to ease administration... 08:58:19 you got anything you know of that might help me? 08:58:19 For what are you looking for this for? 08:58:33 I'm thinking just flat out writing a minimal OS kernel in asm for every platform, and have forth on it. 08:58:42 For my OS :) 08:59:06 You might want to research the concept of a nanokernel, exokernel, and cache kernel. 08:59:21 * kc5tja is personally going to use a cache kernel design for Dolphin. 08:59:54 nanokernel? thats a new term 08:59:56 is it like a mini-exokernel? :) 09:01:03 arke> anything wrong with my ident? :p 09:01:18 qFox: I like it :) 09:01:28 A nanokernel is closer to a HAL than any real kernel. 09:01:50 It's job is to provide the "ideal interface" to the CPU's exception logic, memory management logic, etc. 09:02:05 Otherwise, it's nothing spectacular. 09:02:05 But it might be just what a Forth environment would want. 09:02:20 i just read that my max speed is not the 470kBs i get continously for months, but in fact 320, which will be upgraded to 500kBs... 09:02:50 kc5tja: yeah, that's _EXACTLY_ what I want 09:03:00 * qFox wonders either his isp even knows how to setup bandwithlimits 09:03:07 qFox: lol... 09:03:12 --- quit: I440r_ ("Leaving") 09:03:41 arke: I sent you e-mail containing an attachment with some *ancient* Dolphin 0.5 code in it. It won't compile, but it might be educational all the same. Since it's been something like 4 years or more since I last looked at it, don't ask me to support it. I don't even know what's in it anymore. 09:04:41 ok, thanks 09:08:18 hrm, cool. 09:08:20 * arke tries to find loader code 09:08:24 * arke will likely steal 09:08:40 For the new Dolphin, I'm just going to use GRUB. 09:08:56 I might not even have to write a single line of code for the nanokernel (or at least initially). 09:09:17 Got Dolphin, ree's Pre-Linux DNOS, and some of Robert's stuff. 09:09:20 :) 09:09:58 :) 09:10:30 Well, you got an "Early Design Concept" for Dolphin 0.5 09:10:41 It isn't going to be what the final Dolphin is. 09:11:37 call ss1 09:11:37 db "Booting Dolphin...",0 09:11:37 ss1: pop si 09:11:37 call PutString 09:11:38 heh, that is cool :) 09:11:42 of course. 09:13:38 Hi 09:16:30 Damn, my site is a mess. 09:16:53 hi 09:17:09 Robert: F is real mode, correct/ 09:17:12 ? 09:18:50 No 09:19:00 oh, pmode? 09:19:01 cool 09:19:07 Yep.. 09:19:19 * arke wonders why he thought real mode 09:19:19 oh well 09:19:31 Robert: I'm gonna be stealing lots of code from it. 09:21:05 eeeek 09:21:08 * arke just looked at disasm.c 09:22:04 Robert: what about Primula? 09:22:43 err. pmode 09:22:44 :) 09:25:19 Yep.. 09:25:30 What are you writing? 09:26:08 A cool Forth 09:26:09 :) 09:26:23 somewhat CF-ish 09:26:30 * Robert is thinking about writing a Forth compiler for the Microchip PIC family, and then make a small computer system to replace his sister. 09:26:44 o.O 09:27:05 It'll be a challenge, since the PICs are VERY simple. 09:27:16 o.O 09:27:21 to replace your sister!? 09:27:28 just get a girlfriend 09:27:28 sheesh. 09:27:41 :) 09:27:51 Girlfriends actually require sensitivity and emotions and all that crap. PICs are much easier to get along with. 09:28:10 And they're a heck of a lot cheaper. 09:28:21 True... 09:28:36 Robert: are you open-sourceing your sister? 09:28:46 kc5tja: Yes, they sent me free samples. 09:29:19 Replacing my sister as a telegraphist, that is. When testing radios... 09:29:19 * fridge wonders why Robert's sister needs replacing 09:29:34 fridge: because she's only like 12? :P 09:30:33 if thats the case it would be a good 80 years or more before she needs replacing 09:30:56 fridge: The problem is she's no real radio freak. 09:31:18 fridge: I have to give her candy and stuff before she agrees to take care of the radio :( 09:31:29 =D 09:31:30 bribes huh 09:32:05 Yep.. 09:32:19 Such a wh... Oh, well, better get on with the coding. 09:33:38 * arke just submitted part of this convo to bash :) 09:33:52 Bah 09:34:02 They rejected a quote of mine yesterday. A _good_ one. 09:34:13 And the only one I've submitted so far, I'm not a basher. 09:34:36 heh 09:34:41 * arke is a bash junkie 09:34:51 they reject quotes all teh time 09:34:51 funny ones too 09:34:53 for no reason 09:35:01 They suck. 09:35:12 i would relate it to PMS, but there's no such thing as a 2-year PMS 09:38:55 but you know what, the quotes they do accept are funny 75% of the time 09:39:32 Maybe the first 10 you read. 09:39:36 Then it gets boring. 09:39:42 heh 09:39:42 :) 09:43:48 terve mir 09:43:48 er 09:43:48 mur 09:43:48 :) 09:44:01 terve 09:44:23 mir (russian) : peace, world 09:44:42 mir (german): me 09:45:27 Robert: :) 09:45:36 mur: :) 09:45:36 kc5tja: :) 09:45:36 futhin: :) 09:45:46 EVERYBODY: :) 09:45:47 oh 09:45:58 slava, qFox and Herk get special mention 09:45:59 :) 09:46:34 Heh 09:46:36 Yay, flooding. 09:47:31 bleh 09:47:31 talk about flooding 09:47:43 special mention? 09:47:45 we have a typical P.O.S. american toilet in our bathroom 09:48:06 that thing is so b0rked that the output valve repeatedly gets stuck 09:48:11 causing the toilet bowl to overflow 09:48:53 Robert: whats "wie geht's, mein freund" in sewedish? 09:49:01 arke: That was the very last thing I had to deal with working for In-N-Out on Friday. Fortunately, I don't have to deal with that anymore. :D 09:49:21 ^_^ 09:49:29 kc5tja: and then you got to race 09:49:29 :) 09:49:31 Yep. 09:49:32 And I won. 09:49:34 Morons. 09:49:40 :) 09:49:41 how fast did you end up going? 09:49:46 bout 100? 09:49:53 arke: 115. 09:49:53 :) 09:49:56 With plenty of room for more. 09:49:59 115 isn't fast either. 09:50:06 oh, i know 09:50:10 arke: "Hur går det, min vän?" 09:50:16 A modern sports car can easily swamp my car in a heartbeat. 09:50:16 its not uncommon to see people cruising around at 160 in germanyt 09:50:39 Mustangs do it all the time, but then again, my car is heavy for its engine power, and well, the engine is just plain underpowered and choked off. 09:50:53 arke: 160 mph? :) 09:50:54 man 09:51:03 Robert: yup 09:51:03 Heh. 09:51:05 which is about 280km/h 09:51:09 Insanity. 09:51:14 If I had the cash and the assurance that I can get the car smog-legal, I would strongly consider swapping the engine for a 13B-REW. 09:51:15 wait wait, can we go back to that special mention thing? what was that about? :) 09:51:27 That's like 70m/s. 09:51:41 No, 80m/s! 09:51:50 qFox: arke was flooding the channel with `:)' lines, and certain people got special mention. Nothing more. 09:52:08 :) 09:52:08 Robert: let's just say a helluvalot 09:52:13 oh 09:52:13 oki :) 09:52:48 Robert: Two things. First, you're crazy to push a Mercedes Benz 300-series to those kinds of speeds. So don't try. 09:53:03 77.777777777777777777777777777778m/s 09:53:03 Robert: But if you have a car that CAN go that fast, why not allow it? 09:53:07 what does DET GÅR BRA, uij, kanske mean? 09:53:12 Robert: Second, the Autobahn is *BUILT* for cars traveling at those kinds of speeds. 09:53:27 There are no sharp bends in the Autobahn from what I understand, and if there are, drivers have PLENTY of advance warning. 09:53:28 kc5tja: The human body isn't. 09:53:39 kc5tja: yep yep yep. exactly. 09:53:48 Robert: heh 09:53:48 kc5tja: That's currect.. I've been on those roads. 09:53:59 Robert: there's not even HALF the amount of car accidents there than here 09:54:02 Robert: drivers there get educated MUCH better 09:54:14 Yes. Learn to drive your car, and you'll be safe. 09:54:27 Learn to rely on God, or rely on some other form of blind faith, and you'll die. 09:54:29 It's as simple as that. 09:54:37 * arke can drive his car well 09:54:37 kc5tja: yep :) 09:54:47 My RX-7 can go up to 150MPH. 09:54:47 Stock. 09:54:47 80 m/s on a public road - no way. 09:54:47 * arke avoids big SUV's with Jesus fish stickers on them 09:54:50 It takes a while to GET there, but it can do it. 09:54:58 Robert: :) 09:55:03 Sure, for mad drivers on a racing track it could be OK. 09:55:11 Robert: The *WORST* thing you can do is go SLOW on a FAST road. 09:55:33 kc5tja: What does that have to do with anything? 09:55:33 kc5tja: indeed. 09:55:44 Look at how a pebble in a stream causes a major pileup of water in its vicinity. That's precisely what happens on a freeway. 09:55:50 Robert: It has EVERYTHING to do with safety on a freeway. 09:56:15 kc5tja: Going in 80 m/s ? 09:56:24 Robert: in the fast lane, indeed :) 09:56:32 Literally, *THE* biggest cause of accidents on California freeways are people who are going more than 20MPH *slower* than regular traffic, who then cut someone else off in another lane. 09:56:56 Regular traffic is far from those speeds. 09:57:07 Robert: You're confusing the issues. 09:57:08 Big-time. 09:57:12 kc5tja: ...usually people in V8 SUVs who are either asian or christian 09:57:20 If, by regular traffic, you mean intra-city driving, then of course you're going to go slower. 09:57:23 kc5tja: don't ya hate it? :) 09:57:23 Duhh. 09:57:31 I'm not a driver, I hade roads, so I'll just shup up and leave you alone. 09:57:35 hate* 09:57:39 But on the *FREEWAY*, you don't. You, and everyone else, is driving in the same direction. 09:57:47 You, and everyone else, are more often than not traveling intER-city. 09:58:21 Laws of physics: velocity differentials -- if A is moving 80MPH in one direction, and B is moving 75MPH in the SAME direction, then should a collision ever occur, only a 5MPH difference is there. 09:59:02 So if EVERYONE moved the same speed in the same relative direction, that is every bit as safe (I argue safer) than two people moving at 25MPH but in OPPOSITE directions (like on inner-city roads), where a collision guarantees a 50MPH (!!!) speed difference. 10:00:01 wwhich is why the freeway would be safer IF (!!!!!!!!) people in CA knew how to fucking drive 10:00:13 Everyone are NOT moving in 80 m/s, those freaks have a 50 m/s speed difference to normal people. 10:00:30 Yes. They need to stop being so god-damned afraid of their cars, and to respect the vehicle they're in. SUVs ARE NOT SPORTS CARS!! DON'T DRIVE THEM LIKE THEY ARE! 10:00:36 Robert: that's why there's several lanes 10:00:39 And everyone going in 80m/s would waste tons of fuel. 10:00:51 Robert: depends on the gear ratio :) 10:00:52 Robert: I think you need to go back and hit the Physics books again, dude. 10:00:53 kc5tja: indeed. 10:01:00 kc5tja: which is why I love my small POS 10:01:06 Robert: In particular, re-read the section on velocity vectors and how to do vector math. 10:01:17 To bystanders on the side of the road, each car is only moving 25MPH. 10:02:16 arke: What car are/were you driving now? 10:02:28 kc5tja: 1986 Dodge Colt Vista 10:02:34 kc5tja: 45 or 50 HP 10:02:37 kc5tja: 4 cylinder 10:03:09 kc5tja: the best driver won't be able to go faster than 40 up the 101 by thousand oaks :) 10:03:14 I'll just take the train and use the gas I save to start fires. 10:04:19 OK, so basically a Mitsubishi Mirage hatchback. I know those cars. My ex-roommate had one. They can move and handle quite well. 10:04:32 Though still not a true sports car, they're exceptionally agile. 10:04:48 kc5tja: !?!!?!??!?!?! 10:04:49 kc5tja: dude 10:04:54 kc5tja: no 10:04:54 kc5tja: lol 10:04:59 arke: Dodge Colt = Mitsubishi Mirage 10:05:11 arke: Note that the Dodge Stealth = Eagle Talon = Mitsubishi Eclipse 10:05:11 kc5tja: yes, but I have a Colt _VISTA_ 10:05:21 Notice a pattern here? 10:05:21 kc5tja: MUUUCH different 10:05:21 heh 10:05:24 not vista though 10:05:33 OK, so you have a weaker engine. 10:05:33 But the car's chassis is the same. 10:05:33 its like a colt as a wannabe minivan 10:05:36 heh 10:05:43 lemme find+show you a pic 10:06:05 Yes, please, because Google is only showing me Colts that look like every other Colt/Mirage I've ever seen. 10:09:25 http://auto.consumerguide.com/auto/used/reviews/photos/index.cfm/id/2135/img/91103291990705 10:09:48 only mine does not have the optional 4WD and is older 10:10:00 it does _not_ take curves too well :) 10:10:49 Well, like I said, they're not sports cars. 10:10:58 I never said my roommate's Mirage ever took a curve well. 10:11:04 But for freeway driving, they're amazingly agile. 10:11:33 * kc5tja notes the new Mazda 3 series looks amazingly like a Colt Vista, and they're still considered sound, agile cars. 10:16:04 they're great for cruising around town 10:16:04 but thats about it 10:16:04 :) 10:16:07 yet, I love it 10:19:58 Botta-bing, botta boom. My website has had its phase 1 clean up completed. 10:20:13 Still the same size, but it's organized better, and now I can start pulling out the weeds. 10:20:22 (over the course of the week or so) 10:22:31 20MB of space consumed by my website, 14MB of which is from the darcs repository meta-data. 10:22:37 Time to write a few scripts for updating the website. 10:25:45 :)P 10:31:54 Well, I'm just optimizing things a little bit to make my coding efforts easier. 10:32:06 col 10:32:06 cool 10:32:29 If I move my software development repositories into the HTML infrastructure directly, then I'm more likely to maintain my website documentation in accordance with my coding efforts. 10:32:53 :) 10:33:55 Likewise, it enables me to make a complete backup of all my projects, simply by rsync-ing the *whole* directory tree with that on the webserver. 10:35:13 After the website cleanup is completed, the total site size will drop back to around 4MB or so, which is seems to be about ideal. 10:35:43 btw, I have released FTS/Forth's cross compiler on my site, for anyone who is interested. 10:53:35 c00l 11:01:23 --- quit: rsync (zelazny.freenode.net irc.freenode.net) 11:01:35 --- join: rsync (~pavel@CPE000c41aac435-CM000a73a75acb.cpe.net.cable.rogers.com) joined #forth 11:08:29 --- quit: gl (zelazny.freenode.net irc.freenode.net) 11:08:29 --- quit: warpzero (zelazny.freenode.net irc.freenode.net) 11:08:29 --- quit: skylan (zelazny.freenode.net irc.freenode.net) 11:08:29 --- quit: futhin (zelazny.freenode.net irc.freenode.net) 11:08:29 --- quit: cmeme (zelazny.freenode.net irc.freenode.net) 11:09:40 --- join: futhin (thin@bespin.org) joined #forth 11:09:40 --- join: gl (~foo@opcode.org) joined #forth 11:09:40 --- join: warpzero (~warpzero@dsl.142.mt.onewest.net) joined #forth 11:09:40 --- join: skylan (~sjh@vickesh01-4798.tbaytel.net) joined #forth 11:09:40 --- join: cmeme (~cmeme@216.184.11.30.unused.swcp.com) joined #forth 11:09:40 --- mode: irc.freenode.net set +o futhin 11:18:16 --- quit: rsync (zelazny.freenode.net irc.freenode.net) 11:18:53 --- join: rsync (~pavel@CPE000c41aac435-CM000a73a75acb.cpe.net.cable.rogers.com) joined #forth 11:19:58 --- quit: skylan (zelazny.freenode.net irc.freenode.net) 11:19:58 --- quit: cmeme (zelazny.freenode.net irc.freenode.net) 11:19:58 --- quit: gl (zelazny.freenode.net irc.freenode.net) 11:19:58 --- quit: warpzero (zelazny.freenode.net irc.freenode.net) 11:19:58 --- quit: futhin (zelazny.freenode.net irc.freenode.net) 11:19:58 --- quit: tgunr (zelazny.freenode.net irc.freenode.net) 11:19:58 --- quit: Callidus (zelazny.freenode.net irc.freenode.net) 11:19:59 --- quit: Robert (zelazny.freenode.net irc.freenode.net) 11:19:59 --- quit: rsync (zelazny.freenode.net irc.freenode.net) 11:19:59 --- quit: slava (zelazny.freenode.net irc.freenode.net) 11:20:00 --- quit: onetom (zelazny.freenode.net irc.freenode.net) 11:20:26 --- join: futhin (thin@bespin.org) joined #forth 11:20:26 --- join: rsync (~pavel@CPE000c41aac435-CM000a73a75acb.cpe.net.cable.rogers.com) joined #forth 11:20:26 --- join: cmeme (~cmeme@216.184.11.30.unused.swcp.com) joined #forth 11:20:26 --- join: skylan (~sjh@vickesh01-4798.tbaytel.net) joined #forth 11:20:26 --- join: warpzero (~warpzero@dsl.142.mt.onewest.net) joined #forth 11:20:26 --- join: gl (~foo@opcode.org) joined #forth 11:20:26 --- join: tgunr (~davec@vsat-148-63-4-108.c001.g4.mrt.starband.net) joined #forth 11:20:26 --- join: Callidus (austin@Callidus.wikipedia) joined #forth 11:20:26 --- join: Robert (~snofs@c-bf5a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 11:20:26 --- join: onetom (~tom@novtan.bio.u-szeged.hu) joined #forth 11:20:26 --- join: slava (~slava@CPE00096ba44261-CM000e5cdfda14.cpe.net.cable.rogers.com) joined #forth 11:20:26 --- mode: irc.freenode.net set +o futhin 11:23:19 freenode is going roxor 11:23:40 r0x 11:23:52 --- quit: gl (zelazny.freenode.net irc.freenode.net) 11:23:52 --- quit: warpzero (zelazny.freenode.net irc.freenode.net) 11:23:52 --- quit: skylan (zelazny.freenode.net irc.freenode.net) 11:23:52 --- quit: cmeme (zelazny.freenode.net irc.freenode.net) 11:23:52 --- quit: futhin (zelazny.freenode.net irc.freenode.net) 11:23:52 --- quit: tgunr (zelazny.freenode.net irc.freenode.net) 11:23:52 --- quit: Callidus (zelazny.freenode.net irc.freenode.net) 11:23:52 --- quit: Robert (zelazny.freenode.net irc.freenode.net) 11:23:52 --- quit: slava (zelazny.freenode.net irc.freenode.net) 11:23:52 --- quit: onetom (zelazny.freenode.net irc.freenode.net) 11:23:58 and theeeere they go 11:24:07 --- join: futhin (thin@bespin.org) joined #forth 11:24:07 --- join: cmeme (~cmeme@216.184.11.30.unused.swcp.com) joined #forth 11:24:07 --- join: skylan (~sjh@vickesh01-4798.tbaytel.net) joined #forth 11:24:07 --- join: warpzero (~warpzero@dsl.142.mt.onewest.net) joined #forth 11:24:07 --- join: gl (~foo@opcode.org) joined #forth 11:24:07 --- join: tgunr (~davec@vsat-148-63-4-108.c001.g4.mrt.starband.net) joined #forth 11:24:07 --- join: Callidus (austin@Callidus.wikipedia) joined #forth 11:24:07 --- join: Robert (~snofs@c-bf5a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 11:24:07 --- join: onetom (~tom@novtan.bio.u-szeged.hu) joined #forth 11:24:07 --- join: slava (~slava@CPE00096ba44261-CM000e5cdfda14.cpe.net.cable.rogers.com) joined #forth 11:24:07 --- mode: irc.freenode.net set +o futhin 11:24:09 * kc5tja is getting sea-sick. 11:24:39 * kc5tja is currently researching using texinfo to write some documentation files for GCOM, and eventually, other projects like CUT, FTS/Forth, and Kestrel. 11:25:08 Although I'm not so sure it'd work for the Kestrel, since Kestrel might require graphic figures to be included in the text. 11:27:59 --- quit: onetom (zelazny.freenode.net irc.freenode.net) 11:29:33 --- join: onetom (~tom@novtan.bio.u-szeged.hu) joined #forth 11:37:03 --- quit: onetom (zelazny.freenode.net irc.freenode.net) 11:37:04 --- quit: gl (zelazny.freenode.net irc.freenode.net) 11:37:04 --- quit: warpzero (zelazny.freenode.net irc.freenode.net) 11:37:04 --- quit: skylan (zelazny.freenode.net irc.freenode.net) 11:37:04 --- quit: cmeme (zelazny.freenode.net irc.freenode.net) 11:37:04 --- quit: futhin (zelazny.freenode.net irc.freenode.net) 11:37:04 --- quit: tgunr (zelazny.freenode.net irc.freenode.net) 11:37:04 --- quit: Callidus (zelazny.freenode.net irc.freenode.net) 11:37:04 --- quit: Robert (zelazny.freenode.net irc.freenode.net) 11:37:04 --- quit: slava (zelazny.freenode.net irc.freenode.net) 11:37:24 --- quit: kc5tja ("leaving") 11:37:29 --- join: futhin (thin@bespin.org) joined #forth 11:37:29 --- join: onetom (~tom@novtan.bio.u-szeged.hu) joined #forth 11:37:29 --- join: cmeme (~cmeme@216.184.11.30.unused.swcp.com) joined #forth 11:37:29 --- join: skylan (~sjh@vickesh01-4798.tbaytel.net) joined #forth 11:37:29 --- join: warpzero (~warpzero@dsl.142.mt.onewest.net) joined #forth 11:37:29 --- join: gl (~foo@opcode.org) joined #forth 11:37:29 --- join: tgunr (~davec@vsat-148-63-4-108.c001.g4.mrt.starband.net) joined #forth 11:37:29 --- join: Callidus (austin@Callidus.wikipedia) joined #forth 11:37:29 --- join: Robert (~snofs@c-bf5a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 11:37:29 --- join: slava (~slava@CPE00096ba44261-CM000e5cdfda14.cpe.net.cable.rogers.com) joined #forth 11:37:29 --- mode: irc.freenode.net set +o futhin 11:57:56 --- quit: arke ("Leaving") 11:58:44 --- join: arke (~Chris@wbar8.lax1-4-11-100-108.dsl-verizon.net) joined #forth 11:59:52 FUCK 11:59:58 i seem to have some sort of spyware shit 12:00:17 FUCK 12:00:19 i hate windows 12:02:29 i started working on a wiki 12:03:35 why the fuck will windows not utilize more than 1,3% of my network bandwidth!!?!? 12:05:53 because it sucks 12:05:53 oh 12:05:53 --- join: crc (~Charles_C@0-1pool88-101.nas48.philadelphia1.pa.us.da.qwest.net) joined #forth 12:06:02 it actuazlly used 2.3% jsut now!! :0 12:10:20 --- quit: slava ("Leaving") 12:11:20 well, ad-aware instantly found one 12:11:25 and its looking around 12:12:17 its got 44 12:12:17 ugh 12:27:50 nice going 12:30:19 --- quit: qFox ("this is mirc's last attempt of communication...") 12:31:42 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 12:51:49 --- quit: crc ("ChatZilla 0.9.61 [Mozilla rv:1.7b/20040316]") 12:55:45 --- quit: arke (Read error: 104 (Connection reset by peer)) 12:57:41 --- join: kc5tja (~kc5tja@66-74-218-202.san.rr.com) joined #forth 12:57:45 --- mode: ChanServ set +o kc5tja 13:05:02 --- quit: tgunr (Excess Flood) 13:07:00 --- join: tgunr (~davec@vsat-148-63-4-108.c001.g4.mrt.starband.net) joined #forth 13:30:01 --- quit: warpzero (Remote closed the connection) 14:02:04 --- join: rO| (rO|@pD9E593B2.dip.t-dialin.net) joined #forth 14:02:13 * rO| greets 14:05:12 re rO| 14:05:21 Hi rO| 14:05:57 :-) howdy 14:06:37 can't find any newsfeed rss atom related chan on freenode 14:07:23 Hmm 14:07:33 I never bothered to use such things. 14:07:41 Although my weblog has an RSS and atom feed on it. 14:08:03 It's there just because it came standard with the weblog software. 14:08:46 but maybe somebody knows of some forth newsfeed code? 14:09:47 Not here. 14:10:00 --- quit: tgunr (Read error: 54 (Connection reset by peer)) 14:15:11 has several thousands of newsitems per day now, time to code some intelligent filters 14:15:11 kc5tja: useful for seeing live how mass media is manipulated, collecting evidence 14:15:31 Yes, I am sure. 14:19:04 once has outed an US ambassador as an liar with this technique/resources. on a live webchat. so, it's good for putting overmighty smart*sses were they belong. 14:19:48 computer must be good for something, after all ;-) 14:19:59 right now, all I'm hoping for is that, come 2005, I don't get drafted into the military. 14:21:52 --- join: TheBlueWizard (TheBlueWiz@pc67dn1d.ppp.FCC.NET) joined #forth 14:21:52 --- mode: ChanServ set +o TheBlueWizard 14:23:08 hiya all 14:24:15 re TheBlueWizard 14:24:15 --- join: lalalim (~lalalim@p508ABA05.dip.t-dialin.net) joined #forth 14:24:15 --- join: Herkamire (stjohns@h000094d30ba2.ne.client2.attbi.com) joined #forth 14:24:15 kc5tja: yep, may this come true. 14:24:17 hiya kc5tja...I had some interesting ideas re: the rearchitecturing the network protocol thingie (by the way, I don't see any critique of BEEP on your website...presumably it isn't finished) 14:25:11 No, it's not finished. I'm still collecting data. 14:25:35 I'm also learning how to use GNU Texinfo, as well, which is what I'm going to be using to write it, probably, since it'll be a non-trivial document. 14:26:26 * TheBlueWizard nods, and notes that it is a good idea to collect data on various aspects and current practices and trends...not easy 14:30:23 The thesis of my essay will be that BEEP does in ASCII what ATM does in binary, and therefore, BEEP is a grotesque waste of communications bandwidth as a result. 14:30:59 * TheBlueWizard chuckles 14:31:07 This includes things like routing packets!!! 14:31:19 debugging BEEP presumably would be easy with Telnet though 14:31:30 (e.g., it's part of the BEEP protocol specification that 'relays' can exist to route specific channel data, independently of IP routing!) 14:32:44 Of course, BEEP does solve a pretty big problem, and that is, you can multiplex many independent streams of data between different parts of two programs. 14:33:00 You can do that with TCP/IP or UDP/IP too, but it sucks up ports like a siv sometimes. 14:34:33 And this is where the second part of the paper comes in: the architecture of the Internet, as it exists today, is outmoded. It needs to be changed and modernized. And this means, in no uncertain terms, a new Internet built on the idea that it's not *computers* that serve as end-points of communications, but the programs that run *on* those computers. 14:34:38 It's hard to express. 14:34:41 This is why I'm taking a while to collect my thoughts. 14:36:36 anyway, one interesting train of thought I have is that, by discarding the ports, one would naturally have to extend the IP-like enumeration *into* the parts inside the computer. So, a FTP daemon would get its own number, a web server has one, and so on...and I then realize it would be possible to set up a hierarchy (something akin to nested directories) on them, and can selectively jail those...and I could expand the addressing to include va 14:37:04 --- quit: lalalim_ (Read error: 110 (Connection timed out)) 14:37:10 Not sure how all that would work out though...I decide that it would be a good idea to write up a simulator of a new architecture and let it evolve 14:37:36 TheBlueWizard: You truncated. "...and I could expand the addressing to include va" 14:38:19 TheBlueWizard: What you describe is precisely how ATM network addresses work. They have *20 octet* long addresses (that's 160 bits), the first N octets are treated as a network number, and the remainder is left for the rest of the sub-hierarchy to use. 14:38:32 And that's pretty much the model I was shooting for, personally. 14:39:02 The only real difference is that in my scheme, the addresses would ideally be variable length. 14:39:35 "... include various "objects"...a la LDAP or Active Directory." 14:39:58 I don't know how LDAP works, or Active Directory. 14:40:19 I don't even see how they would apply. I am familiar with DNS though. Can you explain how things would work relative to that? 14:41:01 I don't know a lot about them, but I plan to investigate them regarding my line of thoughts 14:47:07 What is LDAP? http://www.gracion.com/server/whatldap.html 14:49:02 --- join: warpzero (~warpzero@dsl.142.mt.onewest.net) joined #forth 14:50:34 --- join: slava (~slava@CPE00096ba44261-CM000e5cdfda14.cpe.net.cable.rogers.com) joined #forth 14:50:36 hi all 14:50:37 * kc5tja reads 14:50:40 hi kc5tja! 14:50:43 Hi 14:51:18 hiya slava 14:54:38 re slava 14:55:08 TheBlueWizard: I should be more specific: I could never technically follow LDAP -- the protocol is ***SO*** hard to grok, and its database format **SO** hard to set up, that I just gave up on it years ago. 14:55:43 I distinctly remember attempting to use LDAP when I worked for Armored Internet, trying to use LDAP as a means of storing our user account information in a convenient, central location for our other servers to authenticate against. 14:56:03 No freakin' way. OpenLDAP would have *none* of it. No applications would connect to the server, and those that did, got only error messages back. 14:56:31 We spent about a month trying to get it working before I just gave up and switched everything over to a distributed account management system built on CORBA. 14:56:49 CORBA is quite complex too :) 14:57:17 CORBA is a piece of cake to grok. 14:57:31 LDAP is . . . retarded. 14:57:33 going to release fts/forth any time soon? 14:57:43 slava: Check my website. 14:58:20 Also, PLEASE do not confuse FTS/Forth with its cross compiler. They really are two different programs. 14:58:25 FTS/Forth doesn't even exist yet. 14:59:08 ah 14:59:20 you're making the straps to bootstrap with :) 15:00:05 Right now, I'm working on GCOM -- I haven't worked on this since I last worked for Amiga. 15:00:14 kc5tja: ok re: LDAP...I am only focusing on ideas, and I just pick various potential examples...we can always redesign things anyway :) 15:00:17 (speaking of CORBA... :D) 15:00:28 TheBlueWizard: *nods* 15:06:03 --- join: I440r (~mark4@168-215-246-243.gen.twtelecom.net) joined #forth 15:06:17 Hey I440r 15:06:25 hiya I440r 15:06:30 hi 15:06:35 tbw!! hi :)( 15:06:40 hi I440r 15:07:01 :) 15:08:44 :) 15:08:56 life going well? 15:09:11 yea. im at work lol 15:09:28 im burning 6 dvd's of data so a guy can take it home with him when he leaves 15:10:01 I440r: My site has an update -- FTS/Forth's cross compiler is online. 15:10:13 cool 15:10:26 I440r: What kind of data is that? A movie trilogy? ;) 15:10:53 no 15:10:56 work related 15:11:02 schematics, sources, documentation 15:11:04 ALOT of it 15:12:50 kc5tja: That links on your page gives me a 404. 15:12:53 link* 15:13:18 big change in my language -- referencing a word before its defined raises an error at parse time, instead of adding the word to the dictionary with no definition 15:13:34 Robert: What link? 15:13:48 kc5tja: To FTS/Forth. The "dowload" one. 15:14:07 http://www.falvotech.com/downloads/ftsforth-0.1.0.tar.gz 15:14:42 WTF? 15:14:45 It somehow got deleted. :( 15:14:48 > fetch http://www.falvotech.com/downloads/ftsforth-0.1.0.tar.gz 15:14:49 fetch: http://www.falvotech.com/downloads/ftsforth-0.1.0.tar.gz: Not Found 15:18:08 fixed 15:19:01 thanks 15:19:07 And don't you folks do anything to delete it again! :) 15:19:20 * slava 31337 h4x0rz kc5tja's b0x 15:21:36 vibe - is that your block editor? 15:21:52 --- part: rO| left #forth 15:22:37 yes 15:24:43 OK. It's nice that the files are stored in a human-readable format. Don't like it when I get some block of Forth code that can only be read in some obscure editor... 15:25:05 Robert, forth blocks are just 60x15 ASCII! 15:25:11 Robert: 64x16 15:25:20 slava: 64x16 actually 15:25:26 oops 15:25:42 I think Robert was meaning something like ColorForth blocks, which are binary packed with a custom character set and all that. 15:25:43 i pulled two feasable sounding numbers from thin air that could multiply to 1024 15:25:52 but 60*15 = 960 15:26:05 kc5tja, or Herkforth blocks 15:26:38 This is why, for as much as I love the concept of ColorForth, I feel that FTS/Forth is best kept a plain-ASCII, punctuated Forth system. 15:27:01 fuck 15:27:12 fuckiing shitheads who made "usermod" are fucking morons 15:27:24 --- quit: I440r ("Leaving") 15:27:32 Heh. 15:27:49 OK, thanks. 15:30:59 --- join: I440r (~mark4@168-215-246-243.gen.twtelecom.net) joined #forth 15:46:21 --- join: rO| (rO|@pD9E593B2.dip.t-dialin.net) joined #forth 16:09:24 Man, writing a manual is really, really, really hard work. :/ 16:12:58 yeap 16:22:52 if its for newbys you cant possibly be clear enough, if its for techies you cant possibly be correct enough :p 16:23:31 kc5 yea lol 16:24:13 hm 16:24:25 i never thought i'd ever be intrested in a fps kinda game for a console 16:24:33 * kc5tja is trying to re-write the documentation for GCOM so that it's actually half-way usable, letting me put it up on the site. 16:24:36 but i'm really starting to like metroid prime, after playing a bit 16:24:40 (gamecube) 16:25:27 oh and i never played one of the earlier versions so i'm not familiar with the whole "who's the bad guy and whats the basic mission again" thing 16:26:18 its nice how the gamecube, at least in metroid, "hides" loading places 16:26:27 you dont get this pause screen stating "loading..." 16:26:42 the door to the adjacent area just wont open untill he loaded it 16:26:59 i can hear the disc spinning like crazy every time tunill the door opens :p 16:28:37 its actually a pretty good and friendly technique if you ask me, rather then having to wait every time, even for a second or two 16:29:14 qFox: yeah. However, it's also slightly unrealistic in that sense too. 16:29:38 But it's clear that a multithreaded game design is superior to one that isn't. :) 16:29:53 kc5tja> so how many times in your life do you see "loading... please wait" in your life? ;) 16:30:32 well, i cant really tell if this technique is used in other games i've played since i cant really sense whats happening in the background 16:30:51 i just noticed here because the disc spins when i reach a door, the door opens sometimes with a delay, sometimes instantly 16:31:18 but in most games you usually just enter a hotspot where he loads the other side of the area 16:31:34 (if the game allows backtracing, otherwise just loading the next area period :) 16:32:28 i'm kinda puzzled atm, i am CERTAIN i used a grappling hook earlier in the game, when i was unfamiliar with the game, and now i really really need the hook, and i seem to have... misplaced it.. :p 16:32:55 on the bright side, i'm searching all visited area's and i'm finding a shitload of bonusses that i've missed earlier on :) 16:33:02 "secret places" and so 16:33:33 qFox: You start the game with a full compliment of tools, but you then lose them. Part of your mission is to find them again. 16:33:40 partially because i have more items/weapons now then i had previous, some which are required to get to those places, but nevertheless 16:34:07 kc5tja> is that the basic story of metroid? or just in prime? 16:35:14 qFox: I don't know -- I've never played the earlier Metroids. 16:35:24 k :) 16:35:30 The basic story is the same as one earlier game, but I think the precise mechanics of the game are different. 16:36:18 do you get the weapons/items in order of the pause screen? 16:36:31 it seems a bit weird because i have weapon 1 and 3 16:37:05 so i'm not quite sure either i was just sloppy because i wasnt familiarized with gameplay earlier (and so missed area's and the weapon), or its just not that order... 16:47:39 --- quit: I440r (Read error: 60 (Operation timed out)) 17:04:26 qFox: I don't know in what order you get them. 17:04:38 I haven't played the game long enough to know. 17:04:49 I mostly watched my roommate play, who is a metroid fiend. 17:04:54 ah ok 17:04:59 i'll figure it out 17:05:00 :) 17:06:45 hard part i find in these nonlinear games is finding a route that will allow me to have been everywhere, without redundant paths, some times you have multiple ways to go... which is the best 17:06:58 and which does not lead to a point of no return 17:07:12 (ok not really an issue in metroid, but other games...) 17:10:27 * kc5tja nods 17:15:01 hey kc5tja howsit 17:20:17 Wow. I just managed to make a very strong case pro-COM versus CORBA. 17:34:40 I take it that you were pro-CORBA, right? 17:35:01 I love CORBA 17:35:05 But I also love COM too. 17:35:22 DCOM seems pretty nice, but it's noticably lower performance than CORBA. 17:35:43 DCOM == Distributed COM (invoking COM objects between physical computers; COM is just local machine only) 17:36:32 * TheBlueWizard knows what those acronyms stand for...and never have actually worked with them 17:38:20 Writing a CORBA or (D)COM server isn't terribly all that hard. In fact, it's quite easy. 17:38:47 It's *exactly* the same model as writing an event-driven program in any other environment. 17:39:28 I can make a direct one-to-one mapping of the operations that occurs in AmigaOS's IDCMP event distribution system to CORBA and COM, for example. 17:39:40 Likewise with SMTP servers. 17:40:25 * TheBlueWizard knows IDCMP fairly well....and thinks he has written a few apps mucking with it in the past...boy, how his memories has faded! 17:40:36 Heheh :D 17:40:40 SMTP is pretty simple yeah 18:19:40 --- join: MrReach (~spam@209.181.43.190) joined #forth 18:19:40 --- mode: ChanServ set +o MrReach 18:25:48 --- quit: qFox ("this is mirc's last attempt of communication...") 18:27:22 re MrReach 18:28:35 finally im adding wordlists to my language. 18:28:45 its a very simple implementation, minimal code changes 18:30:24 hiya 18:40:18 --- join: crc (~Charles_C@0-1pool88-28.nas48.philadelphia1.pa.us.da.qwest.net) joined #forth 18:40:29 --- join: arke (~Chris@wbar8.lax1-4-11-100-108.dsl-verizon.net) joined #forth 18:40:52 hi guys 18:41:00 hey man 18:41:17 hi futhin, kc5tja, TheBlueWizard, MrReach, Robert, Herkamire, madgarden_, slava, warpzero 18:41:34 may I ask the ops a question? 18:42:34 Hooters- "Chicken tastes better with a boner." 18:44:12 Hello 18:44:36 I can't believe I spent six hours on the introduction to GCOM's documentation. 18:44:44 Almost 7 hours: it's 6:45. 18:45:00 heh 18:45:19 kc5tja: I need an IRC birthday party. 18:45:26 kc5tja: trying to figure out a place 18:45:48 An IRC birthday party? 18:45:57 yeah :P 18:46:07 I'm gonna invite all my IRC buddies 18:46:17 I know it'll have to be freenode 18:46:27 because IRCNet is a biotch 18:46:32 unless you can think of a better one. 18:46:37 then, I need a channel 18:46:40 new one, or existing one? 18:46:57 I would just assume use this network, in your own channel. 18:47:08 mmkay 18:47:12 hrm. 18:47:25 00:00 Pacific Time in #frapiar? 18:47:27 sound good? 18:47:37 Although, if it were my birthday, and I actually had the day off, I'd be about as far away from the computer as mankind can possibly get. :) 18:47:43 since I got that channel regged, I can do neat things. 18:47:45 No. 18:47:48 I'm asleep at that time 18:47:51 (except in rare cases) 18:47:52 Noooo :( 18:48:05 arke: You forget: I have a full-time job. :) 18:48:06 kc5tja: you mean, like yesterday, and the day before, and the day before, and the day before? :) 18:48:36 I'd have to check the logs, but I'm pretty sure I was, if not IN bed by that time, I was GOING to bed. 18:48:52 ^_^ 18:48:53 I know yesterday was a definite exception. 18:49:02 But I seriously couldn't see the keyboard. 18:49:09 And, I *really* don't remember what I was doing. 18:49:09 heh :P 18:49:12 :) 18:49:16 Ergo, I was probably asleep at the keyboard. 18:59:09 hiya arke 19:16:15 hi TheBlueWizard 19:16:16 --- quit: crc (Read error: 104 (Connection reset by peer)) 19:16:45 --- join: crc (~Charles_C@0-1pool176-7.nas6.philadelphia1.pa.us.da.qwest.net) joined #forth 19:31:02 --- join: Sonarman (~matt@adsl-64-169-95-220.dsl.snfc21.pacbell.net) joined #forth 19:41:16 --- join: I440r (proxyuser@adsl-67-65-218-134.dsl.snantx.swbell.net) joined #forth 19:44:59 hiya I440r...again 19:45:48 --- join: snowrichard (~richard@adsl-068-209-159-248.sip.shv.bellsouth.net) joined #forth 19:46:30 hiya snowrichard 19:46:36 hello 19:47:42 --- quit: snowrichard (Client Quit) 19:47:49 --- quit: Sonarman ("leaving") 19:49:10 hi :) 19:57:11 I440r: will you join my b-day party tonight? 20:06:15 eh ? 20:08:02 I440r: I'm having an IRC birtday party in 4 hours 20:08:08 lol 20:08:17 how do i get cake 20:09:27 buy it, then put it on a plate, then look into a mirror, and hand it to yourself 20:12:35 .oO( ) 20:12:36 I440r: anyway, join #frapiar in 4 hours :) 20:12:40 kc5tja: and you too :) 20:12:53 kc5tja: heh, interesting 20:12:54 Sorry, I am definitely going to be sleeping during your party. 20:13:01 * kc5tja just woke up. 20:13:01 :( 20:13:03 I have a headache. 20:13:07 I'll have a second one tomorrow, mmkay? 20:13:14 And in general just plain am not feeling too good. 20:13:28 kc5tja: :( 20:14:30 I think I'll stop working on GCOM's documentation for the time being. 20:14:38 I got a lot done, but it's only one section of its documentation. 20:14:47 Which is not nearly as far along as I would have hoped. 20:23:37 * arke has the OS design figured out 20:23:57 but one more thing 20:24:23 kc5tja: could you explain cache kernels very simply to me? 20:24:35 the website is confusing me (then again, I'm tired...) 20:24:59 --- join: wmg (~weldon@bgp02689673bgs.flrdav01.dc.comcast.net) joined #forth 20:26:56 well, gotta go...bye all 20:27:08 --- part: TheBlueWizard left #forth 20:27:24 NOOOO! 20:27:24 :) 20:40:28 --- quit: crc ("ChatZilla 0.9.61 [Mozilla rv:1.7b/20040316]") 20:43:49 oh man 20:43:52 audacity kicks ass 20:53:04 arke: Cache kernels work on the principle that you can switch to something (be it a process, a thread, an address space, whatever) unless its state is loaded into the kernel's cache of objects. 20:53:14 For example, consider an application's page table. 20:53:24 Page tables are freakin' huge monstrosities. 20:53:37 --- join: Sonarman (~matt@adsl-64-160-167-110.dsl.snfc21.pacbell.net) joined #forth 20:53:44 Who wants to maintain them? I know I certainly don't. I'd much rather compute them on an as-needed basis. 20:54:06 But that would ordinarily require dynamic memory management, and dynamic memory management in kernel space is hard to get right. 20:54:14 So cache kernels just avoids the whole issue all-together. 20:54:22 Everything in a cache kernel is *fixed*. 20:54:24 especially since on modern systems you actually have to deal with page table table table tables 20:54:47 Sonarman: join my IRC bday party in 3 hours, #frapiar. and also, tell me what new mp3's you got :) 20:54:53 So if you only have 16 processes that you can switch to, and you have 80 processes that are running in the OS, how do you reconcile this? Simple: you treat the 16 thinsg as a cache. 20:55:05 aah.. 20:55:24 wmg: You're over-exaggerating. 20:55:34 only slightly 20:55:39 page table table tables 20:55:59 Umm...no. 20:56:03 A page table is a page table. 20:56:08 yes 20:56:08 and the things that are cached and those that are flushed are based on (1) priority and (2) some sort of counter or list to determine which process must go next 20:56:10 correct? 20:56:11 A page table is also a page. 20:56:16 and a page table table is a table of page tables 20:56:37 Ergo, we get a natural degree of recursion that lets us naturally break a huge, monolithic "page table" into a tree. 20:56:44 But it's still considered, as a whole, a page "table." 20:57:32 arke: As with all caches, a least-recently-used (LRU) algorithm is used to determine what gets flushed. 20:57:36 from a user's standpoint, yeah, it's a page table system 20:57:48 wmg: No, from a kernel programmer's point of view too. 20:57:52 kc5tja: thats what I thought :) 20:57:54 wmg: Trust me. I've written several OSes. :) 20:57:55 ("user" in this case meaning a developer of the top-half of a kernel) 20:58:02 kc5tja: well, cool. Then I have decided for me as well. 20:58:18 kc5tja then you know three layers of indirection on page tables can be a bitch 20:58:20 which is what I was saying 20:58:32 wmg: I don't see it as being any more a bitch than 2 layers. 20:58:44 I probably did it wrong, then 20:59:04 The problem is that these layers are used to handle sparse populations of pages. 20:59:30 And it's widely known that the tree-style of page table is just plain NOT suitable for anything above a 32-bit address space with 4KB-sized pages. 20:59:50 This is why IBM uses (and why I *really* love the idea of) inverted page tables, even for 32-bit systems. 21:00:14 kc5tja: I'm using a cache kernel design, with 16 caches (it seems like a good number) 21:00:17 although, frankly, the only way to go is with CPU-managed TLB caches. 21:00:26 arke: 16 *caches*? 21:00:28 of course, that'll be trial-and-error, and maybe even configurable at runtime 21:00:30 er 21:00:41 did i get something wrong? :) 21:00:56 --- part: wmg left #forth 21:01:00 oh 21:01:00 yeah 21:01:06 a cache of size 16 21:01:07 :) 21:01:13 2 of which are static OS processes 21:01:23 6 highest-priority 21:01:26 and 8 flushables 21:01:26 :) 21:01:35 * arke is making this up as he goes along 21:01:36 arke: A cache kernel has typically only 4. You have a "kernel" object cache (usually caches up to 16 kernel objects), the "address space" cache (usually in the vicinity of 64 address spaces), the "thread" cache (around 256 threads in this cache), and a "process" cache (also 64 usually). 21:01:49 er.. 21:02:11 Dolphin does away with the concept of "kernel" objects, because I'm adopting the exokernel idea of a library-OS that is linked with a process. 21:02:21 So my cache system will only support three object types. 21:04:36 hrm.. 21:04:41 * arke has an idea 21:05:00 a kinda modified cache system. 21:05:10 basically, the same thing as I said before. 21:05:34 except that processes are, basically, caches. 21:06:03 which contain a relatively low number of address spaces and threads. 21:06:05 :) 21:06:19 A process, by definition, has precisely one address space, and zero or more threads. 21:06:22 16, 16, and 16 seems like good numbers to start with 21:06:35 I think you need to re-read the V++ paper. :) 21:06:42 it was confusing :) 21:06:50 remember, i'm not a native :) 21:07:03 vocab implementation done 21:07:10 Trust me; you're making things more complicated than they need to be, and with arbitrary limitations to boot. 21:07:48 there is the notion of the IN: vocabulary where new words are added 21:07:54 By separating the idea of address spaces from processes and threads, you gain a lot of very cool (and considerably 'advanced') capabilities. 21:07:58 and a list of USE: vocabularies that are searched when words appear 21:08:22 kc5tja: er.. 21:08:23 oh 21:08:24 ok 21:08:25 each source file must fully declare all USE: vocabularies, but at the outer interpreter, there are a few available automatically like the arithmetic vocab etc. 21:08:30 I had the wrong idea of address space 21:08:40 after startup is done, the current vocabulary is set to the user vocabulary 21:08:49 s/address space/allowed memory ranges+permissions 21:10:51 don't know quite how it works in EROS etc. 21:14:07 arke: You can read the EROS documentation on the EROS website. 21:14:15 efr 21:14:17 er 21:14:18 sorry 21:14:24 that was supposed to be continued 21:14:25 but 21:14:30 They claim they use a nanokernel, and maybe that's true as far as it goes. But what's built on top of the nanokernel is in fact a cache kernel. 21:14:48 here is what I'm thinking for my nano/microkernel 21:14:55 or cache, almost :) 21:15:26 flat 36 bit address space (for intel at least). 21:16:23 32/64 for others, as applicable 21:17:05 each process has 1 or more threads and 1 or more ranges of addresses with permissions 21:17:09 (RWX) 21:19:31 Described through abstract data structures no doubt. 21:20:17 feh. 21:20:39 struct { byte perms; void *begin; void *end; void* next; } 21:20:46 actually 21:20:48 not even the next 21:21:11 its cached :) 21:21:50 Sonarman: I like the rotting christ song :) 21:21:57 great :) 21:22:06 Sonarman: for some weird reason... 21:22:25 No, you need the link. 21:22:26 Sonarman: you're type of music isn't really mine, but every song I've gotten from you I've liked at least somewhat 21:22:31 Well, you don't NEED it... 21:22:34 kc5tja: :) 21:22:47 But what I've read in the cache kernel paper is this: 21:23:03 An address space is just a (in our example) a linked list of structures, as you describe above. 21:23:45 The CPU *cannot* directly interpret this structure, so when it's installed into the kernel's address space cache, the kernel interprets the structure (when necessary!!) to reconstruct the relavent CPU page table structure. 21:24:04 Note that applications do not ever directly see the CPU's page tables. 21:24:14 they can only manipulate the address space structures. 21:24:14 of course not --- wouldn't want them to 21:24:38 arke: Don't be so hasty; exokernels reveal *everything* to the application, including page table structures (on a read-only basis, but still) 21:25:30 well, mine won't be quite an exokernel 21:25:37 more like a weird-ified microkernel 21:25:40 :) 21:25:58 man 21:26:02 ideas racing through my mind 21:26:10 this microkernel is gonna kick ASS 21:28:12 Man, for as much as I love CORBA, I have to admit -- I think it would be in my best interest if I used COM as the standard object model for Dolphin. 21:29:37 COM being MS's COM? 21:29:43 just not for MS? 21:30:49 Yes. 21:31:07 Microsoft COM is the *only* piece of solid innovation they've done, and it shows -- it really is a great piece of technology. 21:31:27 And I have an open source implementation of it (in-process objects only right now) called GCOM. 21:32:02 It implements the core of COM up to the 0.9 spec, and then departs from there, since Microsoft's documentation gets rather hazy beyond that. 21:32:10 (and decidedly Windows specific) 21:32:43 aha 21:33:03 Wasn't COM deprecated in favor of .NET? 21:33:19 .NET uses COM :) 21:33:20 kc5tja: could I have a link to the cache kernel reading material? 21:33:25 arke: not quite. 21:33:32 kc5tja: and how applicable is it to PPC? 21:33:41 typical Herk :) 21:33:57 Herkamire: #frapiar, bday party 21:33:58 Callidus: .Net relies heavily on COM. 21:34:00 kc5tja: same 21:34:02 it starts now 21:34:05 a few hours early :) 21:34:17 .Net is NOT COM, but under the hood, it relies on it very, very, very heavily. 21:34:25 No, it relies on a model that supersedes it, even if there's a compatibility layer in place. 21:34:33 arke: I saw. I'm hoping to go to bed before 3am though. 21:34:38 It is true that it's a direct descendent. 21:34:46 (I live on the east coast) 21:34:47 Callidus: I'm not sure I agree with that at all. 21:35:12 Herkamire: its starting early 21:35:25 everybody --- bday party starting early -- right now --- #frapiar 21:35:31 Herkamire: http://www-dsg.stanford.edu/papers/cachekernel/main.html 21:36:49 Callidus: .Net is basically a java-like environment. It's object model depends on the host processor executing a virtual machine (either via JIT or by direct bytecode interpretation). The object model, therefore, can be "better" than COM in taht it's more abstract, supports things like garbage collection, et. al. 21:37:32 wow 21:37:38 4 people in #frapiar already 21:37:41 kc5tja? 21:37:44 kc5tja: gonna come? :) 21:38:15 arke: At best I can lurk. But I really don't see how one can hold a birthday party on IRC. 21:38:18 I really don't. 21:38:23 * kc5tja despises parties in general. 21:38:38 kc5tja: go lurk then :) 21:39:14 Callidus: But to say that it "replaces" COM . . . I'm still utterly unconvinced of that ever happening. 21:39:48 kc5ta: that's Microsoft's position, and the implementation is clearly different. 21:39:56 Explain runtime-callable wrappers otherwise. :) 21:40:10 I don't even know what the phrase infers, so I can't. 21:40:13 -1s/ta/tja/ 21:40:28 It's how .NET implements COM+ backwards compatibility. 21:41:53 OK. 21:42:01 I don't see why this is somehow something "new." 21:42:32 AmigaBASIC had similar capabilities (the ability to directly invoke system libraries, whose interfaces are structured very similarly to COM interfaces) directly from within BASIC itself. 21:42:58 And I know that VB has a similar facility (though it really does prefer using IDispatch if it can get away with it) 21:44:32 I am just not convinced of .net's so-called technical superiority. 21:44:39 Neither am I. :) 21:44:52 my dad works with .NET daily 21:44:52 But why COM? 21:45:08 he says its good because there;s jsut sOOOOOO much shit which makes a lazy coder like him happy 21:45:08 CORBA is nice, but it has some issues. 21:45:30 Such as? 21:45:31 arke: Tell him he should try Java. There is even MORE shit to make him even MORE happy. 21:45:56 Callidus: CORBA's POA is its achillies heel. 21:45:59 kc5tja: heh 21:46:04 kc5tja: he doesn't like Java 21:46:18 kc5tja: and i think there's actualy more in .NET than java 21:46:19 i mean 21:46:22 for christ sakes 21:46:29 1.5GB worth of DOCUMENTATION ONLY 21:46:50 It prevents binary-level interoperability, because to get any kind of binary interoperability, an object implementation needs to dispatch through two levels of indirection every time, through some really bizarre method tables. And that's *assuming* a C binding. C++ is different still. 21:47:11 True. 21:47:52 Without a binary-level interoperability standard, it's not possible to optimize for the most common case: in-process objects. 21:48:00 In fact, it's not even possible to support the concept at all. 21:48:35 This leads to other issues as a direct consequence: the massive proliferation of ORB implementations, all of which basically do the same job, but none of which share any code. 21:49:30 Lack of object life cycle management as an integrated facility is another issue. 21:49:31 I don't think I've seen a perfect (or even "good") implementation yet. 21:50:00 OmniORB is one of the best ORB implementations I've ever seen in my life. ORBit is excellent too, but it cheats to get higher throughputs. 21:50:07 But it's still not as fast as COM for in-process objects. 21:50:41 The other major problem that I have is that CORBA very tightly couples the local object implementation with its proxy and stub implementations. 21:51:16 Consequently, it's not possible, for example, to write an in-process object implementation, then auto-generate proxies for it after the fact. 21:51:22 With COM, all the interfaces are perfectly clean. 21:52:24 Life cycle management too -- reference counting creates problems, sure, but it solves way more problems than it creates. This is evidenced at Gnome's ready acceptance of reference counting as the basis for Bonobo objects. 21:54:15 CORBA Component Model represents a critical failure for CORBA -- instead of fixing the underlying architecture to support multi-interface reflection on an object (which involved changing *one* ORB method's semantics), they created a 200+ page document detailing a WHOLE NEW way of interacting with CORBA objects that isn't quite compatible with the old, and certainly only marginally compatible with EJBs. 21:54:36 I guess what I'm trying to say is, OMG just made a series of wrong decisions since CORBA 2.4 came out. :( 21:55:06 I mean, how many ORBs are you aware of that actually implements the security service? NONE. Why? Because IT DOESN'T WORK. 21:55:34 It's too ahrd to put into practice, and the security mechanism is so fundamentally abstract that it's next to impossible to map to Unix or Windows NT security policies. 21:56:08 Not only that, but when you examine how object capability security works, COM is a *perfect fit* for it, and not *one* operational semantic of COM needs to be changed. 21:56:19 It's almost like it was custom made for it. 21:58:17 * Callidus nods. 21:58:19 COM is easier to implement, faster in the common case, smaller in resource consumption size, permits custom marshalling objects if performance requires demands it, is more highly factored, is system-wide, and supports object capability-based security out of the box, fully transparently. CORBA supports few, if any, of these things. 22:24:21 --- quit: Sonarman ("leaving") 22:25:15 kc5tja, you're coding with COM or CORBA? 22:25:30 COm 22:25:31 :) 22:32:52 --- quit: Herkamire ("goodnight everybody") 22:36:02 slava: I do both coding. 22:36:09 But I've written my own COM implementation. 22:37:24 how well does COM go with FORTH? 22:41:16 COM is much easier to implement via Forth than CORBA is, that's for sure. 22:41:41 But COM uses C calling conventions as a rule, so C-language bindings are required to support invoking methods and receiving method calls. 22:42:00 But it is doable. Many Forth systems for Windows do this with amazingly little code. 22:43:09 win32forth even lets you write full GUI apps using the win32 api, no? 22:43:18 Yes. 22:44:04 does anything similar exist for x11? 22:44:35 Not to my knowledge. 22:45:05 i tried bigforth but i didn't like it very much 22:45:14 FTS/Forth will provide low-level bindings to Xlib as I need them (e.g., I don't intend on thunking every single Xlib function for the sake of completeness; I just intend on implementing those functions that I need, when I need them) 22:46:00 have you looked at xcb? 22:46:10 its a new binding eventually intnded to replace xlib 22:47:26 Never heard of it. 22:47:44 What language is xcb designed for? 22:47:44 the main change is that its async i think; less round-trips than xlib. 22:47:48 C. 22:48:06 xlib is already very asynchronous. 22:48:19 It accumulates X requests until its buffer is full, then transmits it. 22:48:26 http://www.freedesktop.org/Software/xcb 22:51:41 I'll have to look at it tomorrow. 22:51:53 I'm reading up on some COM-related software development techniques. 23:35:29 --- join: Serg (~z@212.34.52.140) joined #forth 23:59:59 --- log: ended forth/04.05.23