00:00:00 --- log: started forth/01.07.01 00:37:41 --- nick: aaronl_ -> aaronl 01:17:13 --- quit: MrReach (Read error to MrReach[209.181.43.184]: Connection reset by peer) 02:56:26 --- join: rob_ert (robost86@h3n1fls33o898.telia.com) joined #forth 03:34:26 --- nick: rob_ert -> rob_ert_aw_ay_from_he_re 04:08:20 * aaronl is away: signing off 05:24:52 --- join: klooie (kloo@213-84-79-23.adsl.xs4all.nl) joined #forth 05:24:55 hi. 05:25:05 Hi 06:08:04 --- quit: nate37 (zelazny.openprojects.net farmer.openprojects.net) 06:08:04 --- quit: rob_ert_aw_ay_from_he_re (zelazny.openprojects.net farmer.openprojects.net) 06:08:12 --- topic: set to 'http://isforth.sourceforge.net -- http://cvs.sourceforge.net/cgi-bin/cvsweb.cgi/isforth/?cvsroot=isforth' by ChanServ 06:08:22 --- mode: ChanServ set mode: -o clog 06:08:22 --- mode: ChanServ set mode: -o Trey 06:08:24 --- join: nate37 (nate@cx83983-d.irvn1.occa.home.com) joined #forth 06:08:24 --- join: rob_ert_aw_ay_from_he_re (robost86@h3n1fls33o898.telia.com) joined #forth 09:24:42 --- nick: rob_ert_aw_ay_from_he_re -> rob_ert_sub_urb_boy 09:25:13 --- nick: rob_ert_sub_urb_boy -> rob_ert_aw_ay 12:05:57 --- join: eddymaue (eddymaue@modemcable107.14-203-24.hull.mc.videotron.ca) joined #forth 12:07:28 --- part: eddymaue left #forth 12:18:33 --- join: MrReach (mrreach@209.181.43.184) joined #forth 12:18:44 hihi! 13:29:07 Hi 13:29:11 --- nick: rob_ert_aw_ay -> rob_ert 13:31:56 what structures packages have you used? what did you like and dislike about them? 13:32:20 I? 13:32:33 yes, or anyone who is watching 13:32:41 * rob_ert is wondering what a "structure package" is--- 13:32:44 * rob_ert is wondering what a "structure package" is... 13:32:45 =) 13:32:54 are you a forth newbie? 13:32:59 Hehe, yes 13:33:07 I met you just a few days ago 13:33:13 perhaps I should write this question on the Forth Wiki 13:33:21 you and I440r introduced me to forth =) 13:33:41 that may be so, but you might still have 20+ years of Forth experience 13:33:50 oh,ok 13:34:39 Oh, yes... 15 year old with 20 years experience... I started programming when I was -5. 13:34:52 bbiab, brunch is ready 13:35:00 oki 14:10:38 back 14:10:54 Welcome back from your bruch, Mr MrReach =) 14:10:59 "structures" refers to the layout of data in memory 14:11:01 *brunch* 14:11:25 similar to C structures, or Pascal records 14:11:33 yes 14:12:03 except that the programmer generally packs his own structures package with him from project to project 14:12:10 in Forth, that is 14:12:41 you have to use libraries to use structures? 14:12:47 nope 14:13:07 the structure functionality is usually not included in most forth systems 14:13:38 so the application writer brings his/her favorite flavor of structures 14:13:48 ok 14:14:09 gforth and win32f both have structures 14:14:17 most smaller systems don't 14:15:28 hmm.. how do you use structures in the small systems then? 14:16:00 I was actually looking at the Intel optimization manual, and warping my brain trying to figure out how to get forth to optimize for data aligment on cache rows, invisibly 14:16:33 rob_ert: the same way you do in big systems, except that the system's are usually shipped with structures 14:16:43 not shipped, rather 14:16:53 create ur own words 14:16:58 you port your favorite structures package 14:17:07 correct 14:17:34 also, with smaller systems, the lack of "alloc" is a stumbling block 14:18:09 alloc == allocate "external" (from forth environment) memory? 14:18:24 not necc external mem 14:18:28 well 14:18:33 "get me a block of mem this big" 14:18:41 but outside of what ur normally playing with 14:18:54 yes, not dictionary memory 14:20:28 MrReach: you have a ti85/86 calc? 14:20:38 but it looks like the optimization work (on memory data) cannot be abstracted from the application writer like the way code optimization can be 14:20:50 nate37: no, I don't 14:21:29 well 14:21:31 it could to a point 14:21:40 if the abstraction is enough 14:21:49 got an example? 14:21:52 as in there are words that create data and then access parts 14:22:05 and as long as the application writer only uses those words.... 14:22:14 the internal arragment could be however 14:22:19 everyone I know is going to determine the size of an array with a simple mutiplication 14:22:47 the simple mult will not align records on boundries 14:23:04 although the struct words could auto-align 14:23:41 within a particular record ... that would be totally defeated in allocating arrays of odd-sized records 14:26:29 MrReach: got any good ideas for uses of forth on a calculus/engineer calc (ti-85)? 14:26:47 yes, lots of ideas 14:26:56 is there already a forth ported? 14:26:56 besides embedded controller 14:27:01 i'm making a forth for it 14:27:09 what type of display? 14:27:26 128x 14:27:26 how much mem? 14:27:29 28KB 14:27:36 system will be aroudn 4KB 14:27:41 mem expandable? 14:27:42 prolly 5 or 6KB with everything 14:27:47 kinda not really tho 14:27:58 only black and white display 14:28:08 128x64 14:28:10 just remembered 14:28:13 ok, primitive PIM comes to mind ... appts, phone numbs, etc 14:28:35 also, spreadsheet 14:28:37 the little console i/o i have going makes it about 32x9 14:28:38 texxt 14:29:35 why the ti-85 in particular? 14:30:17 cause that's the one i have 14:30:18 :/ 14:30:29 * MrReach checks current price of TI 85 14:30:48 if anybody wants to send me rom dumps of 86 83 82 89 whatever i'll port at some point 14:32:15 * rob_ert wonders is there is any TI 85 emu... 14:32:22 (prize: 0) 14:32:25 yah there is 14:32:32 Virtual TI 14:32:34 is what i use 14:32:45 it can emulate all 14:32:46 ok, where can I get that? 14:32:50 you just need rom dumps 14:32:51 google 14:33:00 how much does a TI85 cost now? 14:33:17 i got mine for 100 but that was few years ago 14:33:20 its discontinued 14:33:31 so depends on source i guess 14:33:37 replaced by 86 14:33:41 which has more mem and some other things 14:33:45 but based on 85 14:34:15 ok, then how much does a TI86 cost? 14:34:32 100 prolly 14:34:44 maybe little less 14:34:45 80 90 14:36:02 i mean its getting to a point where its mostly working interactivly and fully 14:36:06 the palm m100 has MUCH more capability, and costs about the same 14:36:12 (about $110) 14:36:13 and you can link and unlink modules 14:36:24 s/modules/dictionarys-with-buzzword-name/ 14:36:32 I'd buy an m100 today if there were a Forth for it 14:36:37 m100? 14:36:39 hmm 14:36:44 what processor? 14:36:54 * nate37 hopes 86k 14:36:58 ermm 68k 14:37:03 * MrReach sighs, "Beats me. Lemme go find specs" 14:37:05 altho ti-85 is z80 14:37:06 k 14:37:08 heh 14:37:50 hmm that 14:37:53 i would love to get one 14:37:54 heh 14:38:08 i'd port forth to it 14:38:15 prolly already is one tho.. for palm os 14:40:08 can't find the processor 14:40:13 base 2mb memory 14:42:00 the palm m105 starts at $160 and has 8mb base mem 14:42:17 MC68EZ328 processor 14:42:23 just found it 14:42:38 hehe...MC68EZ328 i prefer short names like 8086 14:42:38 I'd like to discard palmos, and use a Forth OS 14:43:04 but MUST have drivers for all the hardware 14:43:21 Motorola processors are good 14:43:44 the reason I'd like to do this is that I'm fustrated w/ NOT being able to hot-sync with linux machine 14:43:46 * nate37 thinks of someway to get palm m10[0|5] 14:44:30 my SERVER should be the machine handling my e-mail and to-dos, not my pretty workstation 14:45:06 hmm 14:45:19 * nate37 gets motorola processor specs 14:45:29 MrReach: what do you code in? 14:45:39 for one, my workstation is not generally accessible from the outside net 14:45:51 assembler, Forth, TCL/Tk 14:45:55 Where can I get docs about TI calcs? How to program them 14:46:17 rob_ert: ticalc.org 14:46:25 those are my preferences, have dabbled in C and LISP and Smalltalk 14:46:26 get Virtual TI emulator (link broken there) 14:46:31 yea... 14:46:39 ok, thanks 14:46:52 i have done asm, C, forth, lisp (a bit).. researched smalltalk, never did anything it 14:46:54 et 14:46:54 c 14:47:02 MrReach: how much asm have you done> 14:47:41 oh, DUH! basic ... used to be REAL good w/ quickbasic-professional compiler 14:47:56 heh 14:48:00 i did kinda do basic 14:48:02 but i was... 14:48:05 younger 14:48:09 well, I metacompiled a native forth for the 8086 14:48:19 and used to write asm libs from QB 14:48:24 heh 14:48:27 * rob_ert did BASIC when he was 11... He's still playing with QB sometimes ;) 14:48:37 s/from/for 14:48:43 yah i'd say when i was 11 14:48:49 +- 1 or 2 years 14:49:03 MrReach: why not you make forth for m100? 14:49:13 don't have the time 14:49:30 and palm is a bit touchy about hardware specs, so far as I can tell 14:49:32 hmmmmm 14:50:21 christmas ways off and i don't know if i can scrounge up money hmmm 14:50:36 but forth is a perfect match for palm, limit hardware ... doesn't have to be fast 14:50:59 the only prob that I can see ... 14:51:39 is that by using a forth os, one would have to forego the nifty handwriting recognition that palmos does 14:52:13 unless you just integrate it 14:52:19 or establish a non-trivial programming team to duplicate it in freeware 14:52:43 yes, a forth-on-top of palmos already exists 14:53:00 have them side-by-side then 14:53:19 want a URL? 14:53:19 with a wrapper to call graffti 14:53:22 srue 14:55:07 http://www.quartus.net/products/forth/ 14:55:20 and ur looking at something to just replace Palm OS? 14:55:41 no, I want my palm to talk to my linux box 14:55:49 by whatever means 14:55:55 can't you just create software already to do that? 14:57:13 perhaps, but how hard is the protocol to interpret? 14:57:52 and the linux driver has to be perfect, or you scramble you're most valued data 14:58:53 doesn't it do checksums? 14:59:12 not sure, never seen the proto spec 14:59:14 even the little protocol in ti-85 checksums 14:59:19 prolly does 14:59:26 so just hope data is correct 14:59:27 hmm 14:59:35 that's not the point ... finding the protocol spec will prob be a real bitch 14:59:46 I'm sure the proto checksums 14:59:53 yea.. 15:00:05 but reverse-engineering the proto could be a hellish job 15:00:11 my dad has an orginal palm 15:00:15 he still uses it tho 15:00:25 moment he wants to upgrade i'm jumping on it heh 15:01:01 btw, Quartus Forth is only for the Pilot, the Palmpilot, Palm III, and the IBM Workpad 15:01:09 hmm orginal uses similar cpu 15:01:37 MrReach: it would be a cool project to hack palm tho 15:01:42 replace the OS 15:02:55 yes, and the character recog code, if one duplicated it's functionality ... would then be ported to every handheld in existence 15:03:13 probably taking Forth with it 15:03:14 "Automatic word-wrap and scrolling of text output" they listed that? 15:03:26 erm ... who? 15:03:35 quartus forth 15:03:40 on there litte list of features 15:03:43 the pam company in promotional materials? 15:03:51 yes, of course they did 15:03:56 i just mean 15:04:01 i would expect it to have that 15:04:06 mien does >:P 15:04:06 some of the earlier PIMs didn't do such things 15:04:16 ah 15:04:21 they were kinda crippled 15:04:26 damn 15:04:36 are those 16 bit or 32 bit processors?? 15:04:44 hmm i'm going to get a palm 15:04:45 32 15:04:55 68k are 32-bit 15:05:00 The z80 is 8-bit, right? 15:05:02 yes 15:05:04 7 General purpose registers 15:05:08 rob_ert: yes 15:05:09 7 Address registers 15:05:11 rob_ert: yah 15:05:19 rob_ert: Z8000 was 16 bit 15:05:28 * nate37 finds it amazing Mac SE is 32-bit :) 15:05:29 ok, thanks =) 15:05:34 Zilog has a 32 bit proc now, but not sure the number 15:05:56 i havn't done 68k programming.. but hoping to soon 15:06:07 ti-89 uses MC86000 15:06:27 hmmm ... nate37 ... do you think a portable forth kernel with matching drivers ported to all the hand-helds would be a feasible net-project? 15:06:27 ermm 15:06:29 68000 15:06:36 MrReach: sure 15:06:52 i'm not so sure 15:07:04 it would be nifty, for certain 15:07:09 hmm 15:07:25 but I'm not sure there would be enough demand to warrant the effort 15:07:59 there would be a LOT of work in leveraging the hardware specs out of the manufacturers 15:08:12 well 15:08:15 could be a forth kernel 15:08:23 but if we provide an interface ppl like 15:08:28 or programming environment 15:08:29 whatever 15:08:30 then 15:08:39 get more ppl easily 15:08:44 i think 15:09:15 yes, but how many would be willing to replace the PalmOS for a killer forth kernel? 15:09:28 You? 15:09:37 hmm 15:09:40 one would HAVE to start with the full functionality of the current Palm apps 15:10:05 yes it'll require work to get it to the same level 15:10:33 I consider the apps the easy part compared to obtaining harware specs 15:10:38 hardware specs 15:10:56 getting hardware specs may simply be impossible 15:11:11 hack palm os code :P 15:11:15 ugh i don't even wanna think of that now 15:11:24 * nate37 crys in pain 15:11:58 the neat part of all this is the ability to write a windowing OS similar to palmos in about 1/2 meg 15:12:17 that would be uniform across all handhelds 15:12:24 maybe yes 15:12:29 virtual desktops :P 15:12:42 either that or we go crazzy and push limits of calc.. 15:12:49 and make nifty divx player :P 15:12:59 HAHAHA! 15:13:31 "can hold and play 10 minutes of MP3 music!" 15:13:41 yup yup 15:13:46 through the crippled little speaker 15:13:57 heh that can only make that little click sound 15:14:17 so it sounds like somebody tapping there fingers like music 15:14:48 actually, I found a DOS program that played music through the PC speaker with fairly good fidelity 15:15:02 heh mp3? 15:15:08 will it work on 486/33? 15:15:10 don't recall what format 15:15:16 ah 15:15:24 yes, worked on my old 386dx/40 15:15:27 well there have been some pcspeaker sound drivers for linux 15:15:42 i havn't been able to get any to work (onlt spent a few minutes with it tho) 15:16:13 anyway, there would be many man-hours in such a product 15:16:13 just think tho if you somehow made a damn optimized system that 'll play divx on 486 15:16:23 what would the benefit be? 15:16:32 well i could watch movies easier while coding :) 15:17:08 well, an bovious one would be new apps being released for all platforms 15:17:24 hmm 15:17:37 and then platform specific "extensions" too maybe 15:17:40 also the ability to switch from one PIM to another effortlessly 15:17:45 no learning curve 15:17:57 but not sure how many people would be doing such things 15:18:24 www.palmos.com might have some hw specs 15:18:42 yes, some features are available everywhere, some features only on some systems, some features only on a particular brand/model 15:18:51 doesn't look like ti tho 15:18:54 some only 15:18:55 hmm 15:19:02 well, I'm not talking JUST palm 15:19:16 I'm also talking Casseopia 15:19:21 does handspring publish 15:19:24 ah 15:19:30 do you have any of these? 15:19:42 Visor 15:19:49 you have a visor? 15:19:49 not as yet 15:19:52 k 15:20:02 still using a paper dayplanner 15:20:16 heh 15:20:25 do code for living? 15:20:33 used to, just retired 15:20:42 retired? 15:20:55 burnt out after 15 years of coding 15:21:02 ah 15:21:10 retired coding or retired retired 15:21:10 not sure what I'm going to do next ... leaning towards RE investment 15:21:14 ah 15:21:27 heh, I'll always do SOMETHING 15:21:35 how old are ya? 15:21:38 34 15:21:45 ahh not retired retired then :P 15:22:03 * MrReach shrugs, "I could sit on my ass for a while ... but that's no fun." 15:22:15 yea.. 15:22:23 and, if you're going to do something, it's just as easy to do something that makes money." 15:22:46 maybe I'll start restoring old boats 15:23:34 hmm there is a palm os emulator 15:23:44 who knows? just about any hobby can make money 15:23:49 i wonder if it emulates hardware fully 15:24:00 yea.. 15:24:01 prob not 15:24:17 it prob emulates the palmOS for developing applications 15:24:40 kewl it does maybe 15:24:55 palmos, I believe, throws an exception if you try to access the hardware directly 15:25:21 "The Palm OS Emulator is software that emulates the hardware of the various models of Palm OS platofrm devices on Windows or Mac OS computers." 15:25:22 really? hmm 15:25:45 ROM images avaible for free too 15:27:06 oh, ok 15:27:12 I may be quite mistaken, then 15:27:17 maybe 15:27:26 i'm joining the alliance program 15:27:30 then i'll have to sign something 15:27:33 get rom images 15:27:35 and emulator 15:27:38 and see from there 15:27:45 in any case, I've got wash some carpets and change out the alternator in my motor home. 15:28:02 so will be back around dark 15:28:13 (9pm pacific) 15:28:15 ok 15:28:17 --- nick: MrReach -> MrGone 15:28:19 cya 15:28:20 see you 17:16:32 Hmmm.... how much does a z80 cost nowdays? 17:20:53 See you all 17:21:29 --- quit: rob_ert () 18:11:30 --- join: cleverdra (jfondren@1Cust206.tnt4.florence.sc.da.uu.net) joined #forth 18:11:31 --- mode: ChanServ set mode: +o cleverdra 18:11:42 --- mode: cleverdra set mode: +oooo aaronl clog klooie MrGone 18:11:42 --- mode: cleverdra set mode: +oo nate37 Trey 18:17:28 --- join: TheBlueWizard (tbw@ip-216-25-202-42.vienna.va.fcc.net) joined #forth 18:17:28 --- mode: ChanServ set mode: +o TheBlueWizard 18:17:35 hiya all 18:19:39 hey, blue 18:20:42 hiya cleverdra 18:20:52 curiously, what does your nick mean? 18:20:58 * cleverdra notices that dired is gone. 18:22:17 oh, just a description of me as a powerful wizard wearing a flowing blue robe 18:22:30 dired? 18:22:38 oh =) 18:22:51 dired is an infobot that used to hang in here. I wonder what happened to him. 18:24:13 oh...didn't know it's a bot :) 18:25:37 hey 18:26:05 hello, nate. 18:26:17 zup? 18:28:31 Today I tried to port Wil Baden's toolbelt and an ANS Forth game called 'sokoban' to upforth for cLIeNUX. I've decided that upforth (pforth) stinks. 18:28:41 hiya nate37 ... nothing much except I am fixing some conflicts in cvs :) 18:28:59 heh 18:29:11 heh...and I am stumped by the message I had on Amiga earlier today :) 18:29:25 cleverdra: sokoban == game where you move diamonds around and try to get them into the correct place? 18:29:31 TheBlueWizard: what was it? 18:30:43 nate - yeah. It's actually a very neat game, and the Forth implementation I have is very cool. 18:32:07 nate - I've improved it some. Changed the pages-long looped CASE and baggage to two tables (one for ESC-combinations, for the number pad) and some simple logic. Made some other improvements. I still have to factor things better. The heard of the program is a simple rule-based translator... very cool =) 18:32:15 ah...it said it can't write the file to MSH: --- object in use....I went huh? it is not being in use (MSH: is the device name for Messy-DOS devices...enables me to r/w MSDOS floppies 18:32:24 (as I said. especially interesting?) 18:33:15 so I rebooted Amiga and tried copying it to MSH: ---- same result...I think I will have to reformat the disk :P 18:33:26 cleverdra: kewl 18:33:58 rebooting Amiga takes only about 10 - 15 seconds...nice, huh? 18:34:07 any ideas for uses of forth on ti-85 calc/ 18:34:12 hah cool 18:34:19 almost as fast as dos itself :P 18:34:22 yeah, blue, that's cool =) Computers take too long to reboot today. 18:34:49 cleverdra: yeah, you say it! :) 18:35:39 * TheBlueWizard srenously advocates the idea of one-byte OS...blindingly fast bootup time! ;) 18:35:58 s/srenously/strenously/ 18:35:59 heh 18:36:05 loop: 18:36:08 jmp loop 18:36:20 (boot up code) 18:36:22 too long...RESET! 18:36:27 heh 18:36:42 or HLT! ;) 18:37:50 I could market this new OS as "Stop the Madness" OS (aka St. MOS), which consists of just one byte long code: HLT 18:37:54 blue - nah. 1k is small enough. 18:38:06 hehe 18:38:45 blue... that's a good idea! Have a subscript on that ad, "upgrade all your friends!" 18:41:17 ha ha ha! 18:41:39 then again, I'd eventually be crucified.... :( 18:47:42 dinner bbl 18:58:01 --- nick: MrGone -> MrReach 19:05:51 blah 19:05:55 * Trey has returned 19:06:07 blah? 19:06:18 TheBlueWizard: what is your favorite structures package? Why do you like it? What could be better? 19:06:35 sweating my ass off 19:06:44 * Trey did that all day. 19:07:41 what do you mean by structures package? 19:07:41 heh 19:07:50 C-type structures, Pascal-like records 19:08:17 the thing that you use to access various data within a contiguous region 19:09:02 ah...for Forth? I usually roll my own ;) 19:09:16 yes, of course ... what does yours look like? 19:09:29 I was thinking this afternoon of two things ... 19:09:43 I don't create a lot of Forth stuff though 19:09:56 1. the way the structs package usually ends up integrating into whatever object system is written 19:10:32 2. how it seems impossible to abstract the data memory optimizations from the application coder 19:11:12 MrReach - most of my structures are A) simple, and B) unique, and C) factored completely into a small set of words. There is a nice structs package by Anton though if you want it. It comes with gforth, and his objects package is based on it. See his site. 19:11:36 I've been through that with a fine-tooth comb 19:11:55 oh? I haven't really looked at it. What did you think? 19:11:58 just wondering if you liked another system 19:12:13 I've never used another structs package, if that's what you mean. 19:12:13 * Trey reads about Chuck's X18 chips. 19:12:24 there's quite a bit of overhead in alignment ... but other than that, it's excellent 19:12:32 Trey - oh, you're looking at his site too? Isn't that coo 19:12:40 cool?, even. 19:12:41 there isn't much support for odd sizes, also 19:13:44 MrReach - my conception is that it abstracts you from optimizing your code relating to data structures, which are usually central to a program that has much of them. 19:14:18 I'm not sure how I like that. As I've said, I haven't used the structs package much, or any other. 19:15:04 well, a structs package should either abstract or it shouldnt 19:15:15 that one only halfway does it 19:15:30 well... wait ... 19:15:49 it MIGHT allocate array elements on 32bytes boundries 19:15:59 I just don't recall that it does 19:16:10 how does it only halfway do it, MrReach? 19:16:23 cleverdra: most cool. 19:16:42 it aligns pointers for the next element of the record 19:17:02 but doesn't really make any effort to ensure that the record itself is aligned 19:19:02 Trey - his color code is different from what I've seen described and what Pringle has implemented. It's... interesting. Not difficult I think to put in Flux. 19:19:40 MrReach - so? stick a MAXALIGN if you want in whatever the struct-creating word is. 19:20:00 most Forthers ... 19:20:19 use something like "get-size 128 * ALLOT THROW" 19:20:24 I've never actually aligned my code, since I don't have a computer that demands it. What happens when you don't on such computers? 19:21:04 gotta go....bye all! 19:21:04 MrReach - 'ALLOT THROW'? Ick. 19:21:11 bye, blue! 19:21:15 --- part: TheBlueWizard left #forth 19:21:22 depends on the processor, on Pentium Pro, a normal mem fetch costs 3 cycles, but a misaligned fetch can cost 12 cycles 19:21:37 sorry! 19:21:42 ALLOCATE THROW 19:21:46 typo 19:21:47 oh. hm. 19:22:11 cleverdra: what on earth are you using for a PC that doesn't pipeline??? 19:22:50 MrReach - no, I think I'm just unaware of the penalty. Oops. 19:23:25 * cleverdra googles for alignment information. 19:23:36 I'll say this, though, looking at how Anton Ertl and Byrnd Payson tear into processor timings is actually intimidating 19:24:24 god forbid they should ever write an actual OS 19:25:27 their main goal re gforth is a *fast* portable ANS Forth, which is cool for various reasons. 19:25:59 I guess they got tired of slow forths that forced them to program in other languages :-) 19:26:04 yes, but I think that Forth, Inc. and LMI have the right approach, though 19:26:12 * cleverdra speaks theoretically. 19:26:17 yes 19:26:20 What approach do they have, MrReach? 19:26:43 they sell fully-optimised native Forths 19:27:26 and I'm discovering that optimising the assembled machine code is easier than optimising the application writer's memory usage 19:28:37 * cleverdra shrugs. 19:29:33 did I talk to you about writing a forth that was as capable/portable as TCL/Tk??? 19:29:37 There are some very simple optimizations you can do. One neat thing Flux does now is to forever eliminate the need for 1- and 2+ and such words with some logic in the + and - macros. 19:30:19 No you didn't, MrReach. 19:30:31 have you ever worked with TCL/tk? 19:32:01 No. Why? 19:32:18 well, it has some rather surprising qualities. 19:33:12 for one, one can write a graphical "hello world" program in about three lines of code 19:33:24 code development is incredibly quick 19:33:35 Yes, I've heard that. 19:33:46 need a button? ad two lines to your source 19:33:47 MrReach: yeah, three lines of code, hundreds of K of runtime. 19:33:58 Trey: that's the gotcha 19:34:25 I've done a lot of TCL, but very little Tk. 19:34:28 also, I write a windowing app on Linux/X, it works w/o change on Windows and Machintosh 19:34:50 * cleverdra hasn't even seen a Tcl "hello world". Never got around to it, though he knows many other languages. 19:35:05 the tk .dll is HUGE, though 19:35:06 MrReach - yes. 19:35:29 could Forth do this smaller and faster? 19:35:43 Of course it can. 19:36:08 there's was a LOT of manhours spent in designing a widget set that could operate identically on all platforms 19:36:40 Yes, but why do you want to do that anyway? 19:36:45 Well, let's assume that you have a good reason. 19:37:10 --- join: edrx (edrx@200.240.18.27) joined #forth 19:37:17 ok, assumed for now 19:37:22 greets, edrx 19:37:48 hi MrReach 19:37:58 OK. You've bound yourself to portability, and the quirks of several operating systems and their underlying machines, and the common denominator. You're already starting out pretty low on the efficiency scale :-) Why don't you just take the step many other languages have, then, and write an interface to tk.dll? Actually, bigForth has something too, as does iForth. 19:38:03 --- mode: cleverdra set mode: +o edrx 19:38:07 hey edrx. 19:38:22 hi cleverdra 19:38:36 bigForth and iForth both have their own portable widget sets, afaik. 19:38:41 s/afaik/iirc/ 19:38:42 because I think the tk.dll might be 1/10 its current size if it were written in forth 19:39:01 gforth can get at the shared libs, also 19:39:21 and run 10x faster (tk is a bit sluggish) 19:39:42 1/10 isn't that fantastic of an improvement, compared to other improvements made by switching to Forth. 19:39:57 * MrReach tips his hat to cleverdra. 19:40:05 what might those be? 19:40:12 So, you want a portable system like Tk for Forth? 19:40:24 that's exactly what I was thinking of 19:40:44 MrReach - I think Jeff Fox has expressed the gains fairly well at www.ultratechnology.com 19:40:49 OK. Cool. 19:41:32 For what Forth? Are you going to make it a shared library, C-style, or some source code that a Forth can generate an object file from, or just source code loaded at init time of a Forth GUI program? 19:41:59 Have you seen bigForth's thing? or iForth's? 19:42:18 I've seen bigforth's "thing" ... I can't make heads or tails of it, frankly 19:42:44 what do you mean, you can't? Of the code? 19:42:46 how do I get iforth? I thought he was charging for it? 19:42:52 yes, of the code 19:43:03 Yes, IIRC he charges 100 USD. 19:44:31 the model is not that important to me 19:44:49 but I lean towards optimised native code for sheer speed 19:45:10 (this is possible because one wouldn't want to rummage around the system-level code much) 19:47:56 Are you going to make it work like Tk? Is this why you're thinking about struct/object packages? 19:50:38 yes 19:50:47 not sure I'm going to do this project 19:50:55 just a consideration 19:51:19 an abstraction of the windowing routines is the only way to make it portable 19:51:33 and it's got a be a dmn good abstraction, at that 19:53:19 So, give an example of the vocabulary you'd use? What would a GUI 'hello world' in Forth look like? 19:53:34 I'd like to make it "look like" tk at the user interface level, not "work like" 19:53:51 because the tk byte-code compiler would be unnessary for forth 19:54:23 maybe I should just use tcl/tk and focus on more interesting things 19:55:40 Maybe. 19:56:30 --- nick: MrReach -> MrShower 19:57:12 I once seriously considered getting iForth, but US$100 plus another $50 for bank taxes for get something I couldn't use freely was too much 19:58:01 What do you mean, you can't use it freely? 19:58:02 tk byte-code compiler = tcl byte-code compiler? 19:58:56 cleverdra: can't give it to other people, can't comment parts of the source and upload to my page, for example 19:59:18 do you use iforth? 19:59:48 No, I've never used it. 20:00:03 But yeah, that's unacceptable :-/ 20:01:09 for a moment I thought you would say something on the lines of "forget that free software thing" :) 20:01:30 hah! rrriiiight. 20:02:22 * cleverdra ponders Chuck's Color Forth. 20:02:28 BIG FAT POINT: I don't have an AGP card. 20:03:20 can I ask you a question about English? 20:03:29 sure. 20:03:39 (sure you can) 20:04:53 I've noticed that recent American texts always use "she" as the third person pronoun when the author is referring to a person of any gender... 20:05:17 this is a recent tendency, isn't it? 20:05:30 This is a tendency so recent I haven't ever seen it. 20:05:37 example: http://people.debian.org/~jaqque/keysign.html 20:05:47 political correctness, I think 20:06:02 That's stupid. Female default is somehow more appropriate than Male default? actually, 'he' is not definitely 'male'. It's also nongenderal. 20:06:08 Definitely. 20:06:18 that's what I think, too 20:06:26 Actually, it's not, because that's DESCRIMINATION AGAINST MEN =) 20:06:45 rrriiiight. 20:08:09 worse yet, that suggests a difference (= war) between the two genders, and in completely unreal terms. 20:08:46 Only a war in the small minds of certain people =) 20:09:11 You may want to ask alt.english.usage what they think of the trend, though. 20:10:11 I've never been able to configure news correctly in my machine (33.6K modem and not much free time to spare) 20:10:46 groups.google.com/groups?group=alt.english.usage 20:11:00 alternatively, www.alt-english-usage.com (.org?) 20:12:20 the www.AEU.{com,net,org} addresses are not resolving here 20:12:45 grumbleth. 20:12:51 is it possible to post news via google? 20:13:49 Yes, I think so. deja had that, and google took deja's news archive and functionality, AFAICT. 20:18:14 * cleverdra blinks. 20:18:17 Oh, how odd. 20:18:23 cleverdra: I don't know what CF would do that would require AGP (except perhaps that he doesn't have anything but an AGP Rage card to work with). 20:18:28 www.alt-usage-english.org 20:18:48 Trey - well, he's doing programmed I/O. 20:19:05 Yeah, but AGP == PCI++ in this respect. 20:19:23 oh. hm. 20:19:58 I doubt he uses the GART on his motherboard, so the AGP requirement probably just means he only has tried the AGP Rage cards. 20:20:26 * Trey writes graphics drivers for a living, and knows the ATI family all too well. 20:20:42 Cool, cool. Thank you, Trey =) 20:21:11 Jeff Fox tells of trying chuck's Color Forth on his machine that had an AGP, and having it not boot, but this is evidently entirely unrelated to the video card. 20:24:22 gotta go - bye all 20:24:23 I want to know how the X18 accesses large memories. 2^18 only gets you 256K words. 20:25:38 bye edrx. 20:25:54 --- part: edrx left #forth 20:26:13 I'm still looking at X18. Why not email Chuck? 20:28:05 I want to make sure I understand what he has posted first. "Meddle not in the affairs of dragons..." 20:28:26 Good plan =) 20:30:05 back 20:43:10 Hmm, the X18 in the 5x5 configuration will require you to write code like you might have written for a Transputer array. 20:44:00 http://www.mindspring.com/~chipchuck/1percent.html 20:44:18 Trey - what characterizes such code? 20:46:38 bye for now. 20:46:38 Writing code for "massively" parallel machines without shared memories. 20:46:47 Goodbye, cleverdra. 20:46:49 --- quit: cleverdra (Leaving) 21:29:16 * Trey is idle: sleeping 21:47:38 --- join: I440r (mark4@purplecoder.com) joined #forth 22:26:01 --- quit: I440r () 23:29:51 --- quit: nate37 (varley.openprojects.net zelazny.openprojects.net) 23:29:51 --- quit: aaronl (varley.openprojects.net zelazny.openprojects.net) 23:29:51 --- quit: MrShower (varley.openprojects.net zelazny.openprojects.net) 23:29:53 --- quit: Trey (varley.openprojects.net zelazny.openprojects.net) 23:29:53 --- quit: klooie (varley.openprojects.net zelazny.openprojects.net) 23:30:06 --- join: nate37 (nate@cx83983-d.irvn1.occa.home.com) joined #forth 23:30:06 --- mode: farmer.openprojects.net set mode: +o nate37 23:30:09 --- join: Trey (bowser@ns.TreySoft.com) joined #forth 23:30:09 --- join: klooie (kloo@213-84-79-23.adsl.xs4all.nl) joined #forth 23:30:09 --- mode: forward.openprojects.net set mode: +oo Trey klooie 23:32:16 --- join: MrShower (mrreach@209.181.43.184) joined #forth 23:35:27 --- join: aaronl (aaronl@vitelus.com) joined #forth 23:44:55 --- topic: set to 'http://isforth.sourceforge.net -- http://cvs.sourceforge.net/cgi-bin/cvsweb.cgi/isforth/?cvsroot=isforth' by ChanServ 23:45:06 --- mode: ChanServ set mode: -o nate37 23:45:06 --- mode: ChanServ set mode: -o clog 23:45:06 --- mode: ChanServ set mode: -o klooie 23:45:06 --- mode: ChanServ set mode: -o Trey 23:59:59 --- log: ended forth/01.07.01