00:00:00 --- log: started forth/04.05.29 00:01:13 problem here is that I have no way of distinguishing a code word between a colon def after it has been compiled 00:01:17 unl.ess I wanna keep a flag 00:01:30 if they're both ASM how do they differ? 00:01:46 also will you have a SEE? 00:01:55 see comes much, much, much later :P 00:02:19 it was one of the first things i did :) 00:02:29 --- join: wmg (~weldon@bgp02689673bgs.flrdav01.dc.comcast.net) joined #forth 00:02:33 nowadays it does syntax highlighting 00:02:34 ^_ 00:03:27 slava: what I was thinking is have levels of optimization that you can choose 00:03:37 seems like needless complexity to me 00:03:44 why not always optimize to the max? 00:03:52 because when I optimize, I can't SEE 00:04:18 unless I really wish to keep a tab on the source or sich 00:04:30 i keep the colon definition around for ITC even if a word is STC compiled (which is optional). 00:04:46 .... 00:04:47 oh 00:04:56 so you're saying, compile both DTC _and_ STC? 00:05:11 my STC is unusual. 00:05:12 slava that makes me feel bad; I just decided not to keep the names in STC... I'm worried I'll regret it 00:05:20 it doesn't use a stack. 00:05:30 code with an unprovable stack effect is ITC 00:06:18 heh... 00:06:24 hrm 00:06:26 oh 00:06:28 coolness 00:06:33 wmg, will you feel worse if my 'see' shows comments too? :) 00:06:35 Hmm... I do code with provable dictionary constancy at compile-time as STC 00:06:48 no actually I love that 00:07:16 comments at the start of a colon definition are kept around; the colon definition has two pointers, oen to the start of comments, one to the start of actual code 00:07:18 My system is only interactive with difficulty though 00:07:19 its a bit weird but works 00:07:23 slava: what if I compile the code as DTC initially, and it gets copmiled to STC on demand? 00:07:36 you could do that I guess 00:07:44 well that's easier on some architectures than others 00:07:57 Intel makes you jump through hoops to write to executable memory 00:08:07 yeah, I cheated, my architecture is JVM bytecode haha 00:08:13 heh 00:08:23 have to jump through hoops to get good performance 00:08:42 I seem to recall SPARC makes it easier 00:08:56 good night 00:09:11 looks like i won't be able to fix my tail call optimization tonight... 00:09:24 all tail calls are jmp's.... 00:09:25 slava: good night, thanks for all the help. 00:09:37 (in the perfect world where socialism works, at least) 00:09:45 wmg, it also has to know if it can put a tail call here or not 00:09:56 slava: do what chuck does 00:09:59 slava: :) 00:10:03 slava look at how some scheme implementations did it some time 00:10:07 it's enlightening, or was to me 00:10:32 slava: another cool thing you can do is scratch the word "THEN" and replace it with ; 00:10:48 enforce factoring, _and_ make some tail calls nicer 00:11:00 That's the fun thing about STC is that THEN is just "ret" 00:11:05 its complicated by the fact that a) control structures are ordinary words b) i have to compile all at once, without changing previously output code, due to my compiler's design 00:11:14 hrm 00:11:16 oh 00:11:25 it just siad the thing about implied CFA in moving forth 00:11:29 I guess i'm blind 00:11:38 moving forth is interesting 00:11:56 --- part: wmg left #forth 00:12:18 also its hard writing unit tests for tail call optimization 00:12:27 since the code works either way, but without it, uses too much stack 00:12:38 only way to check is to look at disassembly 00:13:32 :) 00:14:08 tail-calling will be the only provided means of looping 00:14:20 to, once again, enforce factoring 00:20:35 slava: the tests really only need to make sure it works. 00:22:33 Herkamire: did you read what I said in c4th offtopic? 00:26:58 arke: yeah 00:31:03 --- join: arke_ (~Chris@wbar8.lax1-4-11-100-108.dsl-verizon.net) joined #forth 00:31:33 --- quit: arke (Remote closed the connection) 01:28:12 --- join: minus273 (~sandy@61.157.101.172) joined #forth 01:28:17 hi all 01:44:13 --- quit: minus273 ("That's it for today") 01:54:36 Robert: how do you exit the editor ... :) 02:06:19 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 03:48:58 --- join: crc_ (~Charles_C@0-1pool176-1.nas6.philadelphia1.pa.us.da.qwest.net) joined #forth 03:49:29 --- nick: crc_ -> crc 04:37:23 arke_: Which editor? 04:38:09 vim 04:38:13 hmm 04:38:16 what is question 04:38:43 10:54:39 < arke_> Robert: how do you exit the editor ... :) 04:39:01 If it's the F editor, I think it's the escape button 04:59:42 --- nick: SolarFire[ -> SolarFire 05:17:47 --- join: Serg (~root@193.201.231.126) joined #forth 05:20:05 hi ! 05:20:40 i'm again from byte-pay GPRS - unlim one is way too sucking ;( 05:21:05 :P 05:22:31 i was just kidding w/ ZX Spectrum emulator and CD of proggies ;)) 05:25:34 nice thing ;) 05:27:32 oops, i got so rotten console IRC proggie in my Knoppix 05:30:36 mur: do u have GSM mobile phone ? i wanna test SMS. i tested once w/ kc5tja - it never arrived 05:33:16 set novice off 05:33:24 --- part: Serg left #forth 05:33:53 --- join: Serg (~root@193.201.231.126) joined #forth 05:34:17 damn, i can be only in 1 chan at a time !! 05:34:45 --- quit: crc ("ChatZilla 0.9.61 [Mozilla rv:1.7b/20040316]") 05:39:35 :( 05:39:39 No fun.. 05:39:44 But hi anyway, Serg 05:43:36 hi ;) 05:44:10 i was about to do a little coding since i'm sick and can't go out - but i feel TOO sick even to code ;)) 05:46:18 plans are old 'noter' proggie and some game stuff - play w/ trees and realtime in Forth 05:47:47 btw, how do you sync realtime ? 05:47:47 my favorite method is retrace interrupt 05:55:51 brb 06:17:45 --- quit: Serg (Remote closed the connection) 06:33:14 --- join: I440r (proxyuser@adsl-67-65-218-133.dsl.snantx.swbell.net) joined #forth 06:33:27 --- quit: I440r (Client Quit) 06:37:45 --- quit: madwork_ (Read error: 110 (Connection timed out)) 07:11:16 --- join: I440r (~FooBlah@168-215-246-243.gen.twtelecom.net) joined #forth 07:11:28 Hey 07:12:16 hi 07:12:17 brb 07:12:19 :) 07:12:22 --- quit: I440r (Client Quit) 07:19:05 --- join: Serg (~guest@193.201.231.126) joined #forth 07:27:13 re 07:27:16 serg 07:27:25 wont /set novice off help? 07:28:00 pPRIVMSG #forth :i'm on plain telnet now ! 07:28:09 :) 07:28:21 07:28:33 Heh 07:28:43 you can still join if you dare :) 07:28:45 mur, i see u but dayta comes broken 07:29:21 what comes broken, there is inverted-H between day and ta (day*ta) 07:29:46 ;see 07:30:24 see me in users list ? 07:31:00 yes; 17:30 [freenode] -!- #forth Serg H 0 ~guest@193.201.231.126 [Serg] 07:32:10 dunno grok !!! i'm on chan or not ? 07:32:34 --- join: I440r (~mark4@168-215-246-243.gen.twtelecom.net) joined #forth 07:32:52 of course 07:33:22 i440r hi from plain telnet !! 07:34:06 i am riding raw connection 07:34:25 seeing all inner guts of IRC protocol 07:35:07 gonna code primitive client 07:35:25 Serg, want to make practical documentation 07:35:30 i supposed to do but then i got bored 07:35:53 hmm 07:36:00 first login 07:36:21 i did login a s u see 07:36:27 then insent messages, PRIVMSG <#channel>: asdfasdf or how did it go 07:36:51 yeah i'm doing this now 07:36:59 by hand ;) 07:37:15 well RFC is good help when you want to look the all codes 07:37:36 but the practical documentation to understand it in practise woudl boost developement a lot 07:37:40 RFC is scary mess in this case 07:38:07 heh 07:38:22 i was seeking it madly how to MSG, after login 07:38:23 there are few parts that require too much thinking 07:38:36 PRIVMSG 07:38:52 how'd the whole clause went? 07:40:01 i wrote PASS NICK USER PRIVMSG 07:40:12 JOIN of course 07:41:00 nm 07:41:17 ??? 07:42:28 i gonna need threads... 07:43:54 /me is thinking how 2 implement ze client 07:44:18 what was my last message ? 07:44:29 (it had /me at start ) 07:44:45 thats an "ACTION:" 07:44:51 not a"PRIVMESG:" 07:45:35 ACTION uncnown cmd 07:45:44 no 07:45:52 i mean in the raw irc protocol 07:46:16 if you want to do a message you- use privmesg: blah blah 07:46:33 if youw antt o do a ctcp you do a privmesg: \001 blah blah blah \001 07:46:45 to do a /me blah blah you do action: blah blah blah 07:47:47 i wrote ACTION #forth acrting ;)) and got'unknown cmd' 07:50:09 in my RFC ACTION is even not on contents table 07:52:15 im kinda sure its an ACTION: message 07:52:30 hang on ill tell you 07:52:32 how can i di o it w/o threads ? non-blk read of STDOUT ? 07:53:08 how can i do client w/o threads ? 07:53:12 use a pollfd 07:53:22 non-blk read of 07:53:26 yes 07:53:37 you can have poll NOT block. the pollfd says how long to block for 07:53:42 look at the isforth irc bot 07:53:52 which im waiting on right now lol 07:54:00 --- join: [Forth] (~Forth@168-215-246-243.gen.twtelecom.net) joined #forth 07:54:11 * I440r tests the format of this message 07:54:31 :I440r!~mark4@168-215-246-243.gen.twtelecom.net PRIVMSG #forth :ACTION tests the format of this message 07:54:33 i was close 07:54:59 scary scramble comes 07:55:30 it is a privmesg but the first token in the message body is an "ACTION " 07:55:37 protocol is waste ! 07:55:44 yup 07:55:48 no 07:55:51 proto is 07:56:01 the first token is a special character and THEN "ACTION " 07:56:16 the message must end with the same special character 07:56:18 i bug 07:56:35 i bug out to sketch a client 07:56:40 \001 07:56:42 ? 07:56:50 same as a CTCP ? 07:57:05 --- quit: Serg () 07:57:23 right! 07:57:43 viewing the raw messages with my bot doesnt show the \001 characters at all 07:58:05 yeah 07:58:09 same as CTCP 07:58:11 btw serg poll deprecates select 07:59:25 i don't understand how poll select etc. work 08:00:13 no there is no action in rfc 08:00:20 hmm 08:00:49 different rfc 08:00:59 its in the CTCP and colors rfc 08:01:19 nm 08:01:37 --- join: blockhead (default@dialin-743-tnt.nyc.bestweb.net) joined #forth 08:07:46 i found my commodore vic 20 forth cartridge :) 08:11:08 question is 08:11:15 do you have a vic 20 to plug it into 08:11:22 no :/ 08:12:14 I have 1.8G in my ~/doc directory 08:12:18 need to prune 08:20:58 --- join: kc5tja (~kc5tja@66-74-218-202.san.rr.com) joined #forth 08:21:09 --- mode: ChanServ set +o kc5tja 08:34:54 sup kc5tja 08:39:36 I am, now. 08:42:24 Planning on doing some Kestrel work this weekend, including reviewing the peripheral interconnect bus technology again. 08:42:42 --- join: MrReach (~spam@209.181.43.190) joined #forth 08:42:42 --- mode: ChanServ set +o MrReach 08:42:53 hihi 08:42:56 hey 08:43:36 I'm thinking about interface syntax ... looking for opinions 08:43:47 interface tow hat ? 08:43:52 to what 08:44:02 good 'ol lib loading again 08:44:03 hopefully its an interface to a generic programmable audio tool 08:44:16 warpzero: which platform?? 08:44:38 which do you think is the better syntax ... 08:45:02 MrReach: x86/PPC please 08:45:10 LIB: /lib/somelib.so SomeLib 08:45:13 or ... 08:45:30 s" /lib/somelib.so" LIB: SomeLib 08:45:45 and, for proc addresses ... 08:46:16 PROC: Someproc ProcName 08:46:18 or ... 08:46:25 I would personally make it S" somelib.so" S" SomeLib" make-lib, which then lets the programmer write whatever he wants. 08:46:32 s" Someproc" PROC: ProcName 08:46:48 s/make-lib/bind-lib/ 08:46:53 * MrReach nods to kc5tja 08:47:10 You need to specify a library to indicate what domain the procedure address can be found in. :) 08:47:12 warpzero: does the system have /dev/dsp ??? 08:47:21 oh! damn! thanks, forgot 08:47:24 MrReach: yeah sure why not 08:47:26 duh! 08:47:36 :) 08:47:50 then you can open an audio channel by opening /dev/dsp like any other file 08:48:03 Yeah I know that. 08:48:13 * kc5tja is thinking once again of making the Kestrel's interconnect bus an interconnect *loop*. 08:48:19 what functionality were you looking for, then? 08:48:32 I'm just saying I would like a generic programmable postfix expression live-performance audio tool with good timesources and stuff. 08:48:42 Thats all. 08:48:56 oh, you're looking for multimedia functionality? 08:49:03 syncing and whatnot? 08:49:08 No just audio. 08:49:22 i hate s" 08:49:35 * MrReach giggles, "I know, I440r" 08:49:37 i hate you I440r... no i don't sup dude 08:49:42 lol 08:49:51 * I440r thwaps warpzero with the isforth manual 08:49:57 I440r: yeah, nobody asked you.. sheesh. :) 08:49:57 still a bit light but effective:) 08:50:05 i want your isforth make me a threadsafe version please 08:50:24 warpzero, just do a clone syscall :P 08:50:36 ... 08:50:45 also port it to PPC please 08:50:56 whats a ppc ????? <*snicker*> 08:51:03 you are so dead 08:51:13 I440r: Download PearPC and find out. :D 08:51:29 better than x86 is what it is 08:52:47 MIPS is still my all-time favorite RISC because of its sheer simplicity, but good lord, the PowerPC is so close a second that I consider them about equal. 08:52:49 back in 40 mins ... gonna go watch New Yankee Workshop 08:53:03 Moo. 08:53:03 --- nick: MrReach -> MrGone 08:53:05 --- nick: arke_ -> arke 08:53:13 * arke yawns 08:53:19 MIPS has better MMU architecture 08:53:19 * kc5tja is now known as kc5tja 08:53:55 warpzero: What MMU architecture? :) It's all software controlled. 08:54:04 Some PPC models are also software controlled, but not many. 08:54:09 I emean you know what I mean. 08:54:23 --- quit: Robert ("brb") 08:54:23 I was being facetious. 08:54:38 Software-controlled TLBs rock. 08:54:43 :D 08:55:01 Not having a necessity for MMU is best. 08:55:05 The future MMU for the Kestrel's CPU (and I do mean FUTURE) is going to be entirely software managed. 08:55:06 kc5tja: do you just have to learn enough emacs to use it, or enough emacs to gr0k it 08:55:34 in that case do you know lisp lol 08:55:39 arke: I don't want to grok it in its entirety, but I do need to learn enough of elisp to be able to configure it to my preferences. 08:55:47 warpzero: Yes, I do. 08:55:56 But knowledge of Lisp does not connote knowledge of emacs. 08:56:03 It helps. 08:56:37 I find that my knowledge of Lisp is utterly useless when trying to figure out how .emacsrc file works. 08:56:51 oh :( 08:57:08 The problem isn't the language, it's the API that Emacs exposes. I have no clue what anything means. 08:57:17 * kc5tja is thinking of getting the Emacs manual from the local tech book store. 08:57:20 kc5tja: why emacs? 08:57:29 work computers don't have Vim? 08:57:36 uh oh 08:57:39 editor war time 08:57:48 * arke fears the flame war about to ensue 08:57:49 * blockhead hides. another vi vs emacs war 08:57:53 Herkamire: Because of all the e-mail clients we have at work, that's the best one they have, the most configurable, and the most powerful for my needs. 08:58:20 Evolution sucks oil-infested car exhaust from a 16-cylinder 1930's engine with cracked oil seals. 08:58:26 kc5tja: you're using emacs build-in email client? 08:58:35 Herkamire: That's all I'd use it for. 08:58:41 I see 08:58:46 kc5tja: yeah i fucking hate Evolution 08:58:46 blockhead: VIM wins. But VIM doesn't have an e-mail client in it. :) 08:59:00 kc5tja: if you say so :D 08:59:11 I honestly have to admit that Outlook, the program Evolution is modeled on, is actually *BETTER* an e-mail client than Evo is. 08:59:23 blockhead: Care to enlighten me then? 08:59:26 Yeah fuck Evo. 08:59:27 arke: I didn't think it'd start a war because I know kc5tja prefers vim :) 08:59:28 kc5tja: (1) Pine? :) (2) What do you think of a join DTC/STC Forht system? Words are compiled to STC and cached as STC as needed. 08:59:36 Herkamire: :) 08:59:40 arke: Pine fucking sucks ass too. 08:59:40 Herkamire: so do I/ 08:59:52 er 08:59:55 I've gotten to like mutt well enough (it "sucks less") 08:59:55 kmail? 08:59:57 :P 09:00:15 kmail isn't installed on the systems. 09:01:07 arke: I've thought of that plenty of times before. No time to research it in depth though. A good place to start is researching how Transmeta's stuff works under the hood. Unfortunately, there isn't much detail online. It's still proprietary information. 09:01:31 VLIW is an interesting concept 09:01:38 kc5tja: so should I just do good old STC then? 09:01:58 arke: I don't know. 09:02:06 I have never had any explicit need for speed in a Forth system. 09:02:20 hrm 09:02:24 I use STC/native code only out of convenience -- I didn't want to bother with the hassle of establishing two levels of interpretation. 09:02:42 ok, then I will just go for the regular STC for simplicity. 09:03:49 you guys and your forth implementations have it easy :( 09:04:04 :) 09:04:15 warpzero: not really. :) x86 isteh suck 09:04:29 im going to have to deal with that too 09:05:01 * arke fires up VMWare 09:05:10 Damn, these people JUST don't GET IT. 09:05:18 what people 09:05:25 When you have a loop configured network, you don't need collision detection!! 09:05:27 GAHH! 09:05:58 --- join: arke_ (~chris@wbar8.lax1-4-11-100-108.dsl-verizon.net) joined #forth 09:06:06 Firewire proves that a loop architecture network is doable without collisions, and has been demonstrating this for years and years. 09:06:20 --- nick: arke_ -> teehee 09:06:36 define loop architecture network please 09:06:48 Let us assume a network with four nodes, A, B, C, and D. 09:06:57 Each node's network interface has two ports, IN and OUT. 09:07:02 (no pun intended.) 09:07:11 okay 09:07:18 The OUT of node A is attached to the IN of node B. 09:07:24 lol 09:07:25 OUT of node B is connected to IN of node C, etc. 09:07:38 okay 09:07:39 OUT of node C is connected to IN of node D. 09:07:45 And OUT of node D is attached to IN of node A. 09:07:48 thus completing the loop. 09:07:58 Now, this architecture is damn nice for several reasons. 09:08:06 First, there's no need for collision detection. 09:08:08 kc5tja: can data go two way? 09:08:23 blockhead: In more advanced implementations, yes -- SONET for example uses two, counter-rotating rings. 09:08:33 Firewire also supports two counter-rotating rings. 09:08:41 kc5 do you have gr-core 253 ? 09:08:46 But that's a refinement to the general architecture, and not a strict requirement. 09:08:58 so without counter-totation, for B to send to A it would have to go 'round the long way? 09:09:32 blockhead counter rotation means the in's go one way and the outs th eother 09:09:34 blockhead: Yes. But this isn't as bad as it seems. Firewire introduces only one bit (yes, only one bit) of network delay per node in the network. Most other networks, like SONET, introduce four bits to eight bits, and that's it. 09:09:38 I440r: What? 09:09:47 kc5tja: yeah i dont see how this would need collision dectection 09:09:48 isnt that right ? 09:10:05 the in port only recieves data, the out port only transmits it 09:10:15 kc5tja: 'k. Even I can see how ther would be no need for collision detection ... and I'm a blockhead. :D 09:10:18 I440r: No. Counter-rotation means you have two distinctly separate rings, which means each device has two inputs, and two outputs. 09:10:20 you cant send data out your in port as a shortcut 09:10:25 The direction of data travel is reversed, however. 09:10:26 oh! 09:10:48 Right! 09:11:03 The neat thing is this: the early days of loop-based networks, you might call "token ring." 09:11:30 The idea was that there was a single token that traveled around the loop. When you received the token, that was necessary and sufficient to grant you the permission to write to the loop. 09:11:51 When your packet came back to you, you re-transmitted the token, so someone else could transmit. 09:12:07 A few years later, a refinement to this approach came along -- token ring with early token release. 09:12:35 It was realized, "Gee, there's no real reason why we have to hold onto the token after we're done transmitting -- we can just let it go after we're done." 09:13:02 The nice feature of this is that other nodes downstream can *append* their packet data to the already existing packet train. This is how SONET networks work, for example. 09:13:11 The next refinement is this: forget the token entirely. 09:13:46 Each packet has a start/stop marker (which for brevity, I'll call it a "flag" because that's the informal name for it in general. Note that this "flag" is really a sentinel market, but I digress). 09:14:05 When a node receives a flag, therefore, either a packet is just starting, or just ending (or both). 09:14:32 In this case, if you have sufficient buffer space, you can insert your packet immediately on the OUT port, while buffering the incoming packet on the IN port. 09:14:45 When you're done with your packet transmission, just re-transmit what you've buffered. 09:15:01 This is called Packet Insertion Multiple Access, and this is precisely what Firewire does. 09:15:11 oooh the packet IS the token! 09:15:16 Yes. 09:15:32 so if your relaysing data anyway you might as well relay the shit you wnat to send too :) 09:15:56 The nice thing about this is that every node has guaranteed minimal latency when accessing the network. The disadvantage is that isochronous data streams MAY fall out of specs if everyone tried transmitting at the same time. 09:16:04 Exactly. 09:16:07 That's precisely the idea. 09:16:21 This still seems better than ether for small loops. 09:16:36 It's actually vastly superior to Ethernet, even for large loops. 09:16:49 --- quit: [Forth] ("abort" Reality Strikes Again"") 09:16:50 In fact, the benefits grow as the loop gets larger. 09:17:19 Hmm... the maximum latency is loop_size-1 bits... 09:17:25 yeah... thats not bad at all... 09:17:45 well unless there is other traffic i mean 09:17:46 For example, using SONET as the I/O model (since it's already a large, well-understood networking model), each network node introduces only an octet of delay in the network stream. 09:17:50 kc5tja: what if one computer in the loop is shut down. does that break the loop? 09:18:11 A network with 10,000 nodes on it thus introduces a delay of 80,000 bits from the time you transmit to the time you receive your packet again. 09:18:28 At *gigabits* per second rates, 80,000 bits is so small a fraction of time that it's insignificant. 09:18:48 BUT, I openly challenge you to even consider a 1-gig-E network with 10,000 nodes on it. It won't happen. 09:19:16 blockhead: For a single ring, yes. For counter-rotating rings, no. You need two nodes to shut down for twin-ring systems. 09:19:31 And you can still use switching if you have counter-rotating rings right. 09:20:18 I think this is why ring systems are ideal for either very small (e.g., desktop-area networks) and for very large (e.g., metropolaton(sp??) area networks), because in both cases, the nodes are either all on, or at least, highly unlikely to be shut off to a degree capable of shutting down the network) 09:20:39 Well, if you have intelligent network devices, you can get by with a single ring too. 09:21:03 The idea being, you don't insert yourself into the ring until you know there is no packet being transmitted, AND the device's power is on. 09:21:09 IBM's token ring worked that way quite successfully. 09:21:30 Well in the FW model... I'm slightly confused on one point. 09:21:41 Firewire does something similar, come to think of it. That's how it supports hot-swapping. 09:22:49 What's that? 09:23:22 can a FW device just transmit if the line is dead? 09:23:26 Yes 09:23:29 all it has to do is make a tag right? 09:23:53 A Firewire's packet flag is, by design, chosen to mimick a dead bus, yes. 09:24:32 And with a counter-rotating ring type of system where we say our wires are directly connected [instead of on one shared rail] 09:24:47 we can still use a packet-switcher to join loops right? 09:25:09 Yes 09:25:22 why the fuck doesnt the internet do this 09:25:31 Each packet transmitted still has a source and destination address, with destination address transmitted first to speed up a switch's job. 09:25:44 The Internet is medium independent. 09:25:52 :) 09:25:55 Internet doesn't care if you're using Firewire, Ethernet, or carrier pigeons. 09:25:58 start your own internet! 09:25:58 Well I mean why aren't my computers doing this right now. 09:26:04 (and yes, there is an RFC that documents IP over carrier pigeons!) 09:26:38 warpzero: Inertia, for the most part. Ethernet is said to be cheaper to implement, and it is because of huge production volumes. But in terms of the cost of materials, it's about the same. 09:27:10 Packet insertion multiple access wasn't invented until sometime in the mid-80s to early 90s (I forget precisely, but it's circa that time period), and so, it was a little too late -- Ethernet already owned the market by that time. 09:27:27 I applied for a cool job today 09:28:01 field tech for specialised medical equipment 09:28:25 Ahh. 09:28:28 Good luck. :) 09:28:33 I enjoy helping people out and getting their situation fixed, but I'm really over the PC mess 09:28:40 :) 09:28:51 * teehee is installing bochs 09:29:02 after that, I will start development of F2 09:29:03 --- nick: MrGone -> MrReach 09:29:12 I'm back ... anybody miss me terribly? 09:29:17