00:00:00 --- log: started forth/02.06.05 00:05:05 --- join: davidw (~davidw@adsl-ull-66-108.42-151.net24.it) joined #forth 00:09:07 --- quit: onetom (carter.openprojects.net irc.openprojects.net) 00:09:07 --- quit: Etaoin (carter.openprojects.net irc.openprojects.net) 00:09:07 --- quit: davidw (carter.openprojects.net irc.openprojects.net) 00:09:07 --- quit: sbk_ (carter.openprojects.net irc.openprojects.net) 00:09:12 --- quit: ChanServ (carter.openprojects.net irc.openprojects.net) 00:09:18 --- quit: cdesousa (carter.openprojects.net irc.openprojects.net) 00:12:24 --- join: ChanServ (ChanServ@services.) joined #forth 00:12:24 --- join: davidw (~davidw@adsl-ull-66-108.42-151.net24.it) joined #forth 00:12:24 --- join: sbk_ (~500@dsl-65-184-98-221.telocity.com) joined #forth 00:12:24 --- join: onetom (tom@adsl52007.vnet.hu) joined #forth 00:12:24 --- join: Etaoin (~david@ljk3-7.sat.net) joined #forth 00:12:24 --- join: cdesousa (syf@faeldryn.demon.nl) joined #forth 00:12:24 --- mode: carter.openprojects.net set +o ChanServ 00:25:34 --- quit: sbk_ (carter.openprojects.net irc.openprojects.net) 00:25:34 --- quit: davidw (carter.openprojects.net irc.openprojects.net) 00:25:44 --- join: davidw (~davidw@adsl-ull-66-108.42-151.net24.it) joined #forth 00:25:44 --- join: sbk_ (~500@dsl-65-184-98-221.telocity.com) joined #forth 00:29:47 --- quit: sbk_ ("Leaving") 01:09:47 --- quit: davidw (Remote closed the connection) 01:09:49 --- join: davidw (~davidw@adsl-ull-66-108.42-151.net24.it) joined #forth 01:47:21 --- join: Serg_penguin (~snaga_NOI@nat-ch0.nat.comex.ru) joined #forth 01:47:43 --- quit: Serg_penguin (Client Quit) 01:57:14 --- join: Soap` (~flop@202-0-42-22.cable.paradise.net.nz) joined #forth 02:04:31 --- quit: onetom (Read error: 110 (Connection timed out)) 03:44:55 --- join: Serg_penguin (~snaga_NOI@nat-ch0.nat.comex.ru) joined #forth 03:46:30 Hi 03:48:30 hi 03:52:22 do u know any good source of START .. DIVE .. EMERGE ? 03:54:56 No. 03:55:58 no good... i tried to implement buttt failed 03:57:12 (but is spelled with one t, btw :D) 04:04:09 --- quit: Serg_penguin () 04:29:27 --- join: Serg_penguin (~snaga_NOI@nat-ch0.nat.comex.ru) joined #forth 05:05:35 --- join: sif (~siforth@ip68-9-58-81.ri.ri.cox.net) joined #forth 05:05:35 Type sif: (or /msg sif to play in private) 05:25:55 --- quit: Serg_penguin (Read error: 110 (Connection timed out)) 05:29:50 --- join: Serg_penguin (~snaga_NOI@nat-ch0.nat.comex.ru) joined #forth 05:48:41 --- quit: Serg_penguin () 08:51:05 --- join: Serg_penguin (~snaga_NOI@nat-ch0.nat.comex.ru) joined #forth 08:51:22 hi 09:29:30 --- join: Speuler (~l@195.30.184.4) joined #forth 09:29:34 g'day 09:29:53 Goddag, Speuler. 09:30:06 goddag, the_rob 09:49:30 --- quit: Serg_penguin (Read error: 110 (Connection timed out)) 10:07:52 what's a foth machine ? 11:06:57 --- join: kc5tja (~kc5tja@stampede.org) joined #forth 11:21:01 --- part: Speuler left #forth 11:32:14 --- join: I440r (~mark4@1Cust245.tnt2.bloomington.in.da.uu.net) joined #forth 11:55:54 --- join: miket2 (Mike@modem-181-4-60-62.vip.uk.com) joined #forth 11:58:59 hi miket 11:59:11 hello 12:03:15 --- join: I440r_ (~mark4@1Cust112.tnt2.bloomington.in.da.uu.net) joined #forth 12:03:36 --- quit: I440r (Remote closed the connection) 12:06:03 --- nick: I440r_ -> I440r 12:08:17 --- part: miket2 left #forth 12:45:26 --- join: gulliver (~gulliver@bi-node.teuto.de) joined #forth 12:55:45 * davidw listens to 'fast times at ridgemont high' wavs 12:57:13 --- part: gulliver left #forth 13:00:40 --- nick: kc5tja -> kc-food 13:10:17 you eat too much :) 13:11:17 --- join: jamc (~dne@as3-6-8.asp.s.bonet.se) joined #forth 13:11:43 hi jamc 13:11:58 hiya 13:12:21 im going thru the kernel sources adding comments where ther arent any :P 13:12:33 which would be faster... 13:12:33 sar ebx, 31 13:12:45 or 13:12:45 shl ebx, 1 13:12:45 sbb ebx, ebx 13:13:00 * jamc hopes you're not asking him :) 13:13:24 lol 13:14:22 Just read the manual, heh ;D 13:14:42 I'd use the upper one for size ;) 13:14:51 bleh 13:15:38 i know 13:15:38 its smaller 13:15:47 by 4 bytes :P 13:15:57 erm 8 ? 13:15:57 no 4 bytes 13:16:24 That much? 13:16:31 shl ebx,1 is 2 bytes, right? 13:16:38 And sbb ebx,ebx is too, isn't it? 13:16:45 What is sar ebx,31? 3 bytes? 13:16:56 ill go for the smaller of the 2 anyway because its ina word taht very very very very rarely gets called 13:17:05 Hehe 13:17:07 Good ;-) 13:17:33 yes shl ebx, 1 is 2 bytes 13:17:46 sbb ebx, ebx is prolly 2 heh 13:20:29 actually im getting the same size now 13:20:31 argh 13:21:19 heh this is weird 13:23:04 yes they are the same size 13:23:10 so which is faster ?:) 13:23:31 rtfm :) 13:24:04 bleh you suck :P 13:24:43 greatly 13:24:55 Hehe 13:25:04 På honom bara, slå honom med manualen, du... 14:22:31 80k pa each 14:22:39 not P0 14:22:44 erm wrong channel 14:25:21 ;D 14:27:48 --- join: CrowKiller (Vapo_Rulez@cnq5-233.cablevision.qc.ca) joined #forth 15:12:05 --- quit: jamc ("[x]chat") 15:12:09 --- join: jamc (~dne@mayonnaise.tsps1.freenet6.net) joined #forth 15:26:55 --- nick: kc-food -> kc5tja 15:28:21 wb :) 15:29:02 re 15:29:13 For two hours at least... 15:39:53 --- quit: jamc ("[x]chat") 15:43:21 * CrowKiller is trying to put some Forth code on top of Skiplists to make them understandable to him 15:44:42 * I440r has downloaded 97 megs of 650 in 4 days!!!! 15:45:22 one question i want to ask about something related: why do the FCC cap modems to 56k? 15:45:57 because they are assholes. nuff sed 15:45:57 :P 15:46:08 maybe they have a REAL reason 15:46:08 but i dont know what that is 15:46:31 --- quit: davidw (Read error: 113 (No route to host)) 15:46:43 i cant even get a 56k connection on my line 15:46:43 33.6 is the best i can do 15:47:39 im sure better modemds than 56k can be achieved 15:49:46 dsl uses fone lines 15:51:53 I know and that's frustrating 15:52:03 because modems could do the job 15:52:17 we could use them and pay less 15:52:48 i mean use the technology to transmitt and receive OUR data of any kind, not just internet traffic 15:53:03 --- join: mslicker (~mark@64.27.199.31) joined #forth 15:54:23 mslicker: :) 15:54:29 hello 15:56:01 does anyone have experience with pentium optimization of forth? 15:57:51 not me, my forth doesnt even have an assembler yet :( 15:58:25 the pentium is quite a strange chip. 15:59:19 what is your question? 15:59:27 yes 15:59:30 ive done some bits of assembly 15:59:55 maybe I can help, who knows 15:59:58 ;p 16:00:23 I don't really have a question, but it seems every "optimization" I try actually makes the thing go slower. 16:00:36 if not then go to irc.stealth.net and join #coders. those guys know everything about p4 optimization :) 16:02:09 I'll probably read the optimization manual, though it seems colorForth is quite optimized. 16:05:19 www.quantasm.com 16:05:19 o this site theres a lot of information about per instrcution op^timization 16:05:26 http://www.quantasm.com/optxref.html 16:05:32 thats the exact page, if this information is good then its very complete 16:05:46 maybe not about the newer p4 16:06:08 but chip variations up to the "pentium" generation are covered 16:07:40 CrowKiller: thanks 16:16:24 --- join: segher (~segher@a43195.upc-a.chello.nl) joined #forth 16:16:39 woohoo! 16:17:28 hi segher 16:17:28 you a forth coder ? 16:17:49 yeah 16:17:58 i'm writing an Open Firmware implementation 16:18:10 i just got its compiler working :) 16:18:18 cool! 16:18:31 nice. im writing a linux only x86 only forth 16:18:50 mine is a gcc-only runs-on-everything indirect-threaded forth 16:18:58 she compiles about 900k of source per second now too (im telling everyone that :) 16:19:22 is it open source? 16:19:26 :) 16:19:38 my kernel is assembled with nasm. i dont have an assembler extension yet 16:19:51 may never have :( 16:19:51 yes 16:19:58 my kernel is assembled by a little perl script + cpp :) 16:20:09 the source looks like: 16:20:13 its modified LGPL 16:20:13 isforth.clss.net has the whole source 16:20:13 on this box it compiles the extensions in .15 seconds 16:20:21 var(STATE 0) 16:20:21 col(STATE? STATE @) 16:20:21 imm([ STATE OFF) 16:20:21 col(] STATE ON) 16:20:21 imm(COMPILE, ,) 16:20:21 col(: PARSE-WORD HEADER DOTICK DOCOL COMPILE, ]) 16:20:23 imm(; DOTICK EXIT COMPILE, [) 16:20:29 which equates to VERY FAST heh 16:20:36 heh 16:20:55 so, my kernel source _almost_ looks like forth ;) 16:21:24 heh 16:23:20 segher: wher you from ? 16:23:27 the netherlands 16:23:50 oh yea. i noticed your /whois on entry lol 16:23:57 i just fergot:P 16:24:13 you recon you could code an x86 assembler in forth? 16:24:18 sure 16:24:22 one that DOESNT force "5 # eax mov" on you? 16:24:27 but why not just modify an existing one? 16:24:53 i realy realy need an assembler for my forth but i cant handle fucked upn backwards assembly syntax heh 16:25:03 because of lisencing. i have permission to use the assembler in bigforth 16:25:14 but its a bit weird 16:25:14 and tis backwards and i cant handle that 16:25:23 is the gforth assembler better? 16:25:27 OK, I need to hit the road here; aikido time. 16:25:38 in a perfect world i would get an assembler that can take the existing NASM sources with little or no midification 16:25:43 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK") 16:25:50 l8er kc 16:26:00 a if can make abstraction of the then easly in asm macros 16:26:14 you'll need _some_ modification to have the assembler code live like a good citizen in the forth world ;) 16:26:14 i dont think i can use that 16:26:14 or the win324th assembler 16:26:15 just make the if skip one call instruction ahead 16:26:19 anyway brb 16:26:21 nor any other GPL assembler 16:26:25 --- nick: CrowKiller -> CrowAway 16:26:47 why not? because you are lgpl? 16:26:51 seger my lisence is a modified LGPL 16:27:02 what's modified about it? 16:27:18 the modification being that ANY code you write can have ANY lisence you want it to 16:27:32 i don't follow 16:27:37 even tho your code is effectivly statically linked to MY kernel 16:27:43 i.e. part of YOUR executable is going to be my kernel 16:27:53 part of your executable is going to be YOUR code 16:28:07 --- nick: CrowAway -> CrowKiller 16:28:09 the part taht is my kernel is LGPL 16:28:11 ah, so you removed the "viral" part of the gpl 16:28:18 the part that is derived from your sources can have ANY fscking lisence you want it to have 16:28:28 exactly 16:28:43 erm, actually, it doesn't sound like you did 16:28:59 which means i CANNOT use any assembler that re-injects that viral property into my project 16:28:59 WILL not 16:29:22 you can ask the author of the assembler to re-license it under your license 16:29:25 how so? 16:29:49 if you write some application and distribute it. part of your application is going to be MY KERNEL and you must supply the sources to taht on demand 16:30:01 i could... maybe 16:30:04 if i link a, say, bsd-licensed thing into a gpl thing, the "big thing" will be gpl, but people can take the "small thing" and use it under bsd license 16:30:19 ah yes 16:30:44 why use gpl at all, if you don't like it? 16:31:05 my forth is also VERY non ans 16:31:11 heh 16:31:16 because i DO like gpl, i just dont like the viral property of it 16:31:20 mine is _almost_ ans :) 16:31:37 the whole point of the LGPL was to make what im doing possible 16:31:37 its a"lesser" gpl 16:31:47 Open Firmware is *supposedly* ANS. but the fscked up a few little things ;) 16:31:51 but it doubles back on itself with regard to staticaly linked things 16:32:07 if you use foo.lib and dynamically link to it you are not considerd a derivative work 16:32:15 stupid, eh? 16:32:22 but if yuo STATICALLY link to it you ARE a derivative work 16:32:31 i prefer to stay miles away from gpl 16:32:42 im stating that your application does nto NEED to be considered a derivative work of my compiler even though you are effectivly staticaly linked 16:33:04 yes. its a case of "lets see how far up our asses we can insert our heads" 16:33:30 isforth is part of a larger conspiracy tho 16:33:37 and it needs to be GPL of sorts for that to work 16:33:56 what conspiracy? :) 16:34:40 im trying to make forth accepted by the linux masses 16:34:49 if i achieve all the goals i have set for myself its a possability too 16:34:52 but the assembler is a HUGE stumbling block 16:35:13 to get the linux community to think of forth as something OTHER than a novelty 16:35:21 funny... paflof is going to do the same thing :) 16:35:47 i.e. if say.... isforth became part of the debian distro and debian modified all of its init scripts to forth 16:36:00 paflof ? 16:36:18 Portable And Flexible Linux Open Firmware 16:36:35 called paflof, 'cause it makes me drool ;) 16:37:01 my forht compiler already kicks some major ASS 16:37:13 ive almost totally replaced libc and nucrses in 30k of executable 16:37:25 not 100% tho 16:37:25 but close 16:37:29 hehe 16:37:37 -rwxr-xr-x 1 segher segher 18840 jun 6 01:37 paflof 16:37:47 ...and this is almost full ANS :) 16:37:52 portability is NOT one of my design goals 16:38:02 too bad 16:38:12 -rwxr-xr-x 1 mark4 mark4 16008 Jun 5 16:20 kernel.com 16:38:20 dos?!? 16:38:21 thats the unextended kernel 16:38:31 -rwxr-xr-x 1 mark4 mark4 30234 Jun 5 15:29 isforth 16:38:34 i'm on ppc, btw 16:38:42 thats the extended kernel 16:38:53 no. i called it kernel.com as a joke heh 16:39:10 -rwxr-xr-x 1 segher users 26248 Jun 6 01:37 paflof 16:39:18 that's the same, but on alpha 16:39:28 im currently downloading fbsd install iso so i can convert my kernel to fbsd 16:39:33 i also want ppc version of it 16:40:13 so you interested in helping me out with an assembler then ? 16:40:21 doyou have any x86 linux boxes ??? :) 16:40:35 i don't have time to help you, sorry 16:40:42 i have 6 or 7 x86 boxes, yes 16:40:47 and 4 ppc boxes 16:40:49 I440r: the linux community won't be interested in forth 16:40:51 ok :) 16:41:22 they will be if the forth has sockets, x, file and string handling etc 16:41:33 I440r: there is no hype, and it's not new. 16:41:40 hehe 16:42:05 within the year, most linux users will be using forth, i promise you :) 16:42:32 mslicker if i can get just ONE SMALL PIECE of the linux community intersted then ill have won :) 16:42:36 they probably won't know it's forth, though :) 16:42:57 im not holding my breath tho 16:43:37 later 16:43:41 --- part: segher left #forth 16:44:43 its all a pipe dream for sure if i dont get an assembler i can use 16:44:43 which i prolly never will get 16:44:59 I think the trend is toward less clumsy C's, more features, ect.. 16:45:42 c is a pile of shit 16:45:56 i hate it more and more every day 16:46:58 I agree, C is not very good. 16:47:17 such a verbose language 16:48:11 im doing a time make bzImage in /usr/src where all the obj's already exist 16:48:49 comparing the result to what i get with an extend with my forth 16:48:59 lol 16:49:35 real 0m51.354s 16:49:35 user 0m24.110s 16:49:35 sys 0m3.300s 16:49:49 thats make bzImage 16:49:49 0.12 user 16:49:49 0.03 system 16:49:49 0:00.14 elapsed 16:49:49 104% CPU (0 avgtext + 0 avgdata 0 maxresident)k 16:49:49 0 inputs 16:49:51 0 outputs (35major+20minor) pagefaults 16:50:01 0 swaps 16:50:01 thats my extend heh 16:50:13 what are the major and minor page faults ? 16:50:51 the minor page faults count seems to equal the number of source files compiled 16:51:04 prolly due to memory mapping 16:51:16 i.e. the file isnt actually mapped till you try access it 16:59:40 I440r: I was thinking, you might want to try chuck's approach 17:00:13 just use machine code macros, instead of an assembler. 17:02:16 mslicker the linux community would never accept tht 17:02:16 nor would they accespt 5 # ax mov 17:07:43 I440r: I once tried to implement a disasembler in colorForth, many many blocks were required. 17:08:13 It really makes me apreciate what Chuck is doing with the misc design and colorForth 17:08:23 yes 17:08:32 disassemblers are easy compared to assemblers 17:08:42 check out my 8051 disassembler aty isforth.clss.net 17:10:18 yes, this is simliar to what I was doing. 17:11:06 the disassembler is going to be part of a simulator 17:11:30 i dont think i could write an 8051 assembler yet i knnow the opcodes very well 17:12:19 i think a macro asembler is easy 17:12:23 if some things 17:12:25 are simplified 17:12:43 like relative "if"s only one call instruction long 17:13:12 crow make me a pentium assembler for isforth that has a SANE syntax :) 17:13:19 ok 17:13:22 mov eax,foo+[ebx+8*ecx] 17:13:23 seriously 17:13:37 ill try to make a vm for your forth 17:13:58 notice the spaces i liberally DIDNT distribute throughout that ???? :) 17:14:07 or maybe just the best vm i can and then use it with your forth 17:14:13 in other words 17:14:21 the most linear possible core dictionary 17:14:24 CrowKiller: most of the work is in the sheer number of instructions and diffirent forms of something like the pentium 17:14:41 i know 17:14:48 the x86 asm 17:14:48 yes 17:14:58 i have prinetd out the opcode machine codes 17:14:59 THATS whats got ME tied in knots over it heh 17:15:21 a precise intention means only one instruction sequence or block 17:16:28 thats only a maximum of 8 bytes 17:16:34 for any let say c18 macro 17:16:43 the 21 macros should not be very much 17:16:49 the f21 i meant 17:18:51 CrowKiller: I don't understand. 17:21:11 http://www.colorforth.com/forth.html 17:21:28 only the mulstep implementation get over 10 bytes, and its not even needed 17:22:14 yes, but this not an assembler. 17:22:25 in forth you dont need one 17:22:34 CrowKiller: yes 17:22:51 on the actual target?? 17:22:57 hmmm i doubt 17:22:59 CrowKiller: I440r wanted an assembler. 17:23:28 i know but for me there's no need to code into a forth system if there's a lit word 17:24:45 it can always be coded afterwards 17:24:51 once the forth works really good 17:24:53 I think the machine code macros are quite too. 17:25:07 I guess they are not for everyone. 17:25:27 quite elegant 17:25:33 these should be enough to write an assembler 17:26:03 it should be a trivial task using a good forth 17:26:35 if it cant do things like mov eax,foo+[ebx+8*ecx] <-- insert spaces anywhere is ok 17:26:38 no, really it quite a lot of work for a pentium. 17:26:48 it isnt much use to me 17:27:00 unless it can do the above 17:27:12 why? 17:27:12 i know :( 17:27:48 crow the linux community would NEVER accept having to code assembler doing 17:27:55 $5a c, $12 c, ..... 17:27:59 or 5 # eax mov 17:28:08 they would just REJECT both of those outright 17:28:15 CrowKiller: just the number of instructions, addressing forms, parameter orders. 17:28:17 as would i 17:28:22 if its kewl, the linux community would do anything 17:28:27 i have never liked the forth way of doing assembly 17:28:48 because linux persons think they are kewl 17:28:48 if fpc's assembler handnt allowed SANE assembly syntax i would never have learned forth 17:28:51 ;p 17:28:59 i too would have rejected it as a "trinket" 17:29:00 I440r: the forth way of doing assembly is misc :) 17:29:17 yep 17:29:23 im catering to THEM as much as i am to me 17:29:36 its pure insanity 17:29:45 look at the c18 instruction set 17:29:52 on a 2400mips machine 17:29:58 using the *+ instruction 17:30:01 mulstep 17:30:01 its NOT asm 17:30:01 i have always hated it 17:30:08 i learned to code assembler on MANY different processors before i learned forth 17:30:09 to do multiplications can be thought possible 17:30:24 but on the pentium you already have the shortcut mul 17:30:26 and i will never accept ass backwards assemblers 17:30:26 i cant THINIK that way 17:30:34 so you can take it instead 17:30:35 when i think assembler i think ASSEMBLER not forth 17:30:58 noneed to think in assembler when doing forth 17:31:09 unless you want absolute performance on a given architecture 17:31:33 good forth can resits call rets at each 3 or 4 words 17:31:52 I'm finding out, just coding in assembler can slow down code. 17:31:56 crow any forth that doesnt have an assembler built in is just pure crippled 17:31:56 it can resist intense call ret and use this caracteristic to "loose weight" 17:32:08 there are times when yiou NEED pure assembler 17:32:15 use nasm 17:32:19 do a list output 17:32:20 and you CANNOT wire a forth that does not at some point resort to assembly level code 17:32:33 until you find out thta you need a good editor 17:32:39 machine code, yes 17:32:39 to save assembly source into tokens 17:32:44 much like in the ahah system 17:32:56 but specific for the simplest possible assembler 17:32:57 fuck no 17:33:10 i want a BUILT IN assembler extension 17:33:10 if i could code it myself i would have already :( 17:33:22 ugh 17:33:24 the editor could convert existinga ssembly source into its more efficient format 17:33:38 CrowKiller: I don't think the aha method is practical for a pentium. 17:33:46 so it could rule and bring into submission any othe rmean of doing assembly 17:34:18 well, when you have finished rethinking my problem for me.... heh 17:34:27 it could by implementing tokens for each datatype 17:34:27 CrowKiller: I think it is ment for misc chips. 17:35:05 the stream would looklike (token of 1,2,3,4 bits)(data specified by token)token of 1,2,3,or 4 bit(data) 17:35:17 just implement the misc macros 17:35:34 and there's your forth system, it can always be enhanced 17:35:55 by discovering how a new implementation is made on the particular hardware 17:36:03 conevrting between versions of the source 17:36:07 CrowKiller: from my understanding it is a mixed object code/source code format. 17:36:12 would be made easy by coding converters 17:36:23 its the aha system 17:36:40 just plain old aha system 17:37:10 to me its the real object technology applied to computing and should be used since the beggining of computing, much like the forth language 17:37:13 pentium object code can not unambiguaously be converted to forth source. 17:37:14 crow distributing sources in BINARY format is not needed with isforth 17:37:39 partial compilation at EDIT time is just adding unneeded complexity to isforth 17:37:51 i recon isforth would already compile 10 * faster than aha ona pentium 17:38:13 no its unifying all the concpets into a SINGLE gui if look at the picture correctly, it still vague but its definetly there 17:38:16 and distributing sources in anything other than HUMAN readable format is not acceptable to me 17:38:28 not for isforth 17:38:59 l440r: so isforth will be the greatest lost of time and energy of your whoooooole life 17:39:15 I440r: human unreadable source, that is new to me ;) 17:39:20 l440r: i think its why i dont have a frenetic habit to code first, im afraid to lose time into coding 17:39:31 the unix community would not accept sources that were semi compiled already 17:39:44 mslicker stop looking at C sources then :P 17:39:47 fuck the community lets convert 'em to OUR community lol 17:40:02 they would be reprensented 17:40:08 into a special "xml" like language 17:40:16 I440r: good one :) 17:40:25 xml is truly the aha concept but not very efficeiently made 17:40:30 crow spend the next 10 years "designing" 5 days into the "coding" you suddenly realise you MISSED something and the whole fucking design is useless to you 17:40:41 there are things you CANNOT realise at the DESIGN stage 17:41:12 isforth is designed 17:41:20 i know, on architectures you dont know theres always that possibility 17:41:26