00:00:00 --- log: started forth/01.11.14 00:05:44 --- quit: MrGone (Read error: 54 (Connection reset by peer)) 00:06:05 --- join: MrZzzzz (~mrreach@209.181.43.190) joined #forth 05:34:32 * aaronl is away: nap 06:53:12 --- join: edrx (edrx@200.240.18.106) joined #forth 07:08:11 --- quit: edrx ("[x]chat") 09:41:58 --- join: edrx (edrx@200.240.18.73) joined #forth 10:53:35 --- join: Speuler (akhandel@preiselbeere.icafe.spacenet.de) joined #forth 10:53:47 hi 10:58:06 --- quit: Speuler ("ircII/tkirc") 11:32:44 --- quit: edrx ("[x]chat") 13:33:56 --- join: Speuler (akhandel@195.30.184.7) joined #forth 13:34:37 --- nick: Speuler -> Speuler_n 13:34:45 --- nick: Speuler_n -> Speuler 13:35:10 greetings! 13:35:17 --- nick: MrZzzzz -> MrReach 13:35:42 --- join: Speuler_ (akhandel@birne.icafe.spacenet.de) joined #forth 13:35:45 hi mrreach 13:35:52 mrZzz i mean 13:36:07 --- mode: ChanServ set mode: +o MrReach 13:36:15 how are you? 13:36:31 quite allright today 13:36:36 not so tired :) 13:36:37 --- mode: MrReach set mode: +oo Speuler Speuler_ 13:36:48 and yourself ? 13:36:48 heh, rough day yesterday? 13:37:04 pretty good, but a bit hung-over 13:37:44 you said you have obtained an assembler for the SH3 and intended to give it to me? 13:38:04 was out dining japanese yesterday 13:38:25 sounds like lots of fun (if you like raw fish @:^) 13:39:38 actually, i do, at least some kinds of 13:39:57 you see, dutch people like raw fish too 13:41:07 ok, that's fine 13:41:25 i don't have an sh3 asm yet but i think i will get one 13:41:40 I haven't eaten enough to know if I like it or not ... better to not tell me it's raw until after I've eaten some, though 13:41:55 it has been passed along, and the feedback was positive 13:41:57 oh! I thought you had gotten one already 13:42:19 why not write one in Forth? 13:42:30 next day after you stated your need, i asked the hitachi people 13:42:52 but they don't carry those assemblers in their pocket 13:43:03 erm ... why not? 13:43:03 you ask me ? 13:43:27 cause they have been sent over for some AEx / H8 job 13:43:29 well, yes I'm asking you, but I didn't mean Hitachi 13:43:43 i passed on your request to hitachi 13:43:55 i'm working with hitachi people at the moment 13:44:01 ah! ok 13:44:19 I should think Hitachi would write a freeware assembler for all their chips 13:44:22 same people i went for japanese restaurant yesterday 13:44:34 makes them MUCH easier to work with, IMO 13:44:50 i was asked the same question you did ... 13:45:04 why not in Forth? 13:45:09 "what kind of business model was i thinking about ..." 13:45:14 oh! 13:45:23 in regards to which project? 13:46:00 development system which runs on handheld/target itself instead of on cross dev platform 13:46:16 ok 13:46:23 i had to invent a need for the assembler :) 13:46:44 so i sticked with the truth :) 13:46:50 I was thinking more of a cross dev system ... I like my keyboard/monitor 13:47:19 but doesn't make sense not to have the full forth on such a capable handheld 13:47:27 why is everybody caring about making money so much ? 13:47:47 as if business models are the most important things on earth 13:47:58 because if you go broke, development comes to a grinding halt while you get a job 13:48:10 --- join: aum (~aum@210-86-60-200.jetstart.xtra.co.nz) joined #forth 13:48:28 not the most important, but certainly one of the most @:^> 13:48:29 will a business model keep you from getting broke ? 13:48:40 yes, if it's a good one 13:49:05 my business model is quite simple .. 13:49:17 the model is supposed to show if you're going to go broke *BEFORE* you start 13:49:31 write some code, charge the client, and go into holidays 13:49:38 haha! 13:49:48 ok, but who's the client? 13:50:01 how much do you charge? how do you decide? 13:50:01 think about what project to do next, and do something totally different then 13:50:44 last decision on how much i charge was based on "what i took before plus xx percent" 13:51:20 so you wouldn't consider this a long-term project? (maybe lasting a decade or more) 13:51:35 no. this is a ten-week project 13:51:44 before it was 12 weeks 13:51:57 next will be about 6 months or so 13:52:07 are we talking about the same thing? 13:52:10 if it is going to get off the ground at all ... 13:52:46 that's about my free-lance activities. that's what pays now 13:52:57 oh, ok 13:53:21 (does NOT pay the rent as this is done by my moonlighting activities in the place where i am right now) 13:53:33 I thought we were talking about multiplat forth for PCs/PDAs 13:54:14 i tend to get things working first, than decide how/whether to market them 13:54:30 ok 13:54:35 that's what i did with some previous forth systems 13:54:48 sort of no cure no pay 13:54:55 safe for the client 13:55:10 increases willingness to be pulled over the table 13:55:19 well, there wouldn't be a "client" ... per se ... in this project 13:55:34 more of a "user base" 13:55:46 i have not much experience with software, used my many clients 13:56:04 * MrReach nods. 13:56:15 instead, is marketed as part of the equipment by the client i wrote the thing for 13:56:26 (usually more than just the interpreter) 13:56:26 --- quit: aum (Read error: 104 (Connection reset by peer)) 13:56:33 firmware 13:56:49 diagnostic/test/verification software 13:56:54 yep 13:57:18 this would be a bit larger, I should think 13:57:32 what business model do you have in mind ? 13:57:43 I'm thinking in the 200-300 word range ... not including system calls 13:58:16 well, there's several ways to make money ... 13:58:32 but I want to keep the interpreter itself as low cost as possible 13:58:48 the hitachis also inquired whether i offer my services as individual, or through an agent ... 13:58:55 I don't $200 bill attached ... thus putting out of range of the hobbiest 13:59:08 i wonder why 13:59:24 they're obviously thinking of hiring you 13:59:44 could learn some more japanese then . 13:59:53 can charge like $20-40 to download the interpreter 14:00:10 or cripple the interpreter is some way, then charge for ful capability 14:00:20 bundle it with the handheld devices, charge the manufacturer 14:00:45 I like your idea of charging only for the applications written on the interpreter 14:00:46 you get it even if you don't want it 14:01:33 yes, we might talk Palm, HP, Compaq, etc into adding it to their ROMs 14:01:44 sometimes, compilers are given away for zilch, but you pay for the docs 14:01:59 sounds cool too 14:02:05 erm, we could do that, but I'm wary of it 14:02:24 lots of paper to send out ? 14:02:44 no paper at all ... electronic format 14:02:53 let the user print if needs hardcopy 14:03:15 with some paper. the client would have the feeling he receives something in return 14:03:30 electronic format is so .. insubstantial 14:03:57 yes, ummm 14:04:08 easier to appreciate value if you can hold it in your hands 14:04:29 (more difficult to copy too) 14:04:32 I'm thinking of a model where EVERYONE uses the system, it's too convenient not to 14:04:52 and then SOME people pay for it, or parts, or apps 14:05:01 some of the users can make better use of it because of the docs 14:05:33 germany just qualified for the world championship soccer 14:05:49 cool 14:05:51 last-minute effort 14:06:19 win-or-die situation 14:06:24 if we don't every ship a physical object, our distribution costs are zero 14:06:48 the turnover possibly too 14:07:17 it is difficult to make money with a forth system 14:07:19 turnover? 14:07:35 Quartus seems to be doing "ok" 14:07:40 but much easier to make money from what you do with a forth interpreter 14:07:44 not sure they're making money, though 14:07:54 agreed 14:08:14 Quartus is quite limited in its targets 14:08:17 that's where the non-zero-charge apps comes from 14:08:44 ha! nagware apps! 14:09:01 and nagware on a PDA is a serious annoyance @:^> 14:10:14 we have some advantages that most compilers/software writers don't have 14:10:54 but only we know about those 14:11:02 for example, if we write our APIs correctly, a single program can compile for just about any platform with little or no change 14:11:09 you'd have to convince the users first 14:11:34 a business model won't be enough to do that 14:11:41 if we wanted to get really fancy, we could make a binary that would run on all platforms 14:11:52 understood 14:11:53 multi-cpu ? 14:12:00 yes 14:12:13 byte-code interpreter 14:12:25 somehow precompiled, and finalized on the target ? 14:13:00 binary ships as byte-code ... platform dependant interpreter runs it 14:13:22 like java ? 14:13:40 also, we can do *MUCH* better at syncronizing than current software does 14:14:09 yes, something like Java ... but much faster ... I'm just brainstorming, don't take too seriously 14:14:12 probably, but it would mean diving into ms interfaces 14:14:36 we're gonna have to anyway, if we're going to support WinCE devices 14:14:38 which i don't consider to be a very pleasant thing to do 14:15:29 the advantage there is that the Forth on the desktop machine looks nearly identical to the Forth on the PDA 14:16:23 also, if we publish the file formats for stored data, users can bit-bang it themselves ... like they *USED* to do with mail files 14:16:25 almost like a command line interpreter :) 14:16:39 users will love that ... 14:17:31 but for devices with small screens, it makes efficient use of screen real estate 14:17:51 yes, exactly 14:18:16 desktop has larger API ... but for tools, not neccessarily apps 14:18:39 apps would look nearly identical on desktop and PDA 14:18:44 i can get a 106x30 console on the psion 14:19:30 that would be a 53x30 on the ipaq 14:19:39 pity, no 64 14:19:55 heh, so much for blocks? 14:20:28 i like comments on the same line with code 14:21:09 I'm expecting that users will develop apps and code of various flavors on their desktop machines 14:21:36 either with the built-in graphical IDE or with the editor of their choice 14:22:22 they CAN write directly on the PDA, but that's generally a nuisance 14:23:02 what is the keyboard of the pda good for, besides not entering text ? 14:23:22 most PDAs don't have a keyboard 14:23:31 bashing people on the head ? 14:24:10 i wouldn't buy a pda w/o keyboard . 14:24:18 god, this Stowaway keyboard of mine would explode into a bazzion little pieces if I ever hit anything with it 14:24:44 most PDAs do *NOT* have a keyboard 14:25:13 I spent $100 for this keyboard, I seriously doubt many people do 14:25:16 casting it into epoxy would solve that, and improve bashing performance as well 14:25:27 i would 14:25:31 yes, but then it would fuld up so neatly 14:25:47 yes, but then it wouldn't fold up so neatly 14:26:33 reduces aerodynamics too 14:26:51 in any case, can develop either on the desktop machine, or on the pda, whichever is more convenient 14:27:18 xactly that's why i'd need a keyboard 14:27:18 can make the tools roughly equivelent 14:28:17 or i'd be back to development on cross dev, and use the pda purely to execute those apps 14:28:34 in fact ... could even use a telnet-type application to sit at desktop and issue commands to the interpreter on the PDA 14:28:39 no need for such a pda for me 14:28:58 or type on the PDA and execute on the desktop 14:29:12 maybe one can stick the pda into a pci slot 14:29:24 there are pcmcia form factor pdas 14:29:30 usually USB or serial 14:29:41 no keyboard with those 14:29:49 but nice idea 14:30:14 yes, but nobody develops for ALL of them 14:30:17 very limited pdas, that is 14:30:57 but the idea is to NOT use those stand-alone so much 14:31:10 just for data retrieval, and the pim kind of apps 14:31:21 point and shoot 14:31:34 I really *LIKE* my 17" multimedia monitor and fancy keyboard ... but I can't take them with me in the car, hence the PDA 14:31:50 maybe one could use a pda to load such a pcmcia pda with data 14:32:11 i got two portables. 14:32:21 one is the psion (to be replaced against ipaq) 14:32:29 small, compact, light 14:32:35 and I find I don't really like the way the desktop machine and the PDA work together 14:32:42 the other is a somewhat heavier notebook 14:32:52 which i don't lug around all the time 14:33:01 but has a good screen 14:33:08 (1400x1050) 14:33:12 tft 14:33:16 yikes! 14:33:20 that's my mobile workstation 14:33:25 damn good screen, that's gonna COST you 14:33:32 against the psion, it looks huge 14:33:40 wasn't cheap indeed 14:33:58 god forbid it should be stolen 14:34:07 but i can't take my desktop machine with me all the time when i'm on the road 14:34:31 anyway, about the business model ... 14:34:35 well, it wasn't cheap but i do earn not bad 14:34:52 how many PDAs are there in the world, do you think? 14:34:58 cost me about a week, slightly more than that, of work 14:35:13 not as many as cellular phones 14:35:22 how many of them do you think might end up with our Forth on them? 14:35:36 one out of 200 14:35:49 that little ?? 14:36:08 pragmatic estimation 14:36:14 which apps can we charge for? which should we give away to promote the system? 14:36:17 better be wrong on the wrong end 14:36:50 anything ms specific should cost 14:36:56 punish those who use it 14:37:07 heh 14:37:21 besides, they are used to it 14:37:47 should the basic PIM stuff be freeware? nagware? 14:37:52 syncing software would be a seller 14:38:16 provided it works well 14:38:21 pay $20 to dump the MS sync library? 14:38:59 have ms pay us for not offering the sync software ? 14:39:11 ok, so maybe ... 14:39:23 not for free, but selling it instead ... 14:39:44 the PIM software is free and works flawlessly on both the PDA and the desktop, but pay $20 to get them to talk to each other? 14:39:47 threaten them to write a better wince :) 14:40:17 lince 14:40:32 fonce 14:40:34 4th 14:40:45 for-ce 14:40:54 there you go 14:41:21 now, can we write superior PIM software? 14:41:33 needs lots of user interface ... 14:41:41 i'm not good with that 14:41:47 i think 14:41:55 never really tried :) 14:42:03 it will prob go through several generations 14:42:21 in the process, the company goes bankrupt 14:42:42 ending up with a superior product but 0 cash-flow 14:43:04 incidently, we really do want to keep the originally OS ... whatever it is ... unless you want to write the i/o and handwriting recognition stuff. 14:43:24 epoc ? 14:43:39 EPOC, PalmOS, WinCE ... whichever 14:44:12 palmtop linux too 14:44:26 the challenge will be to get all of them to operate similarly and keep it small 14:45:11 maybe system could come on a cardbus card, as extension ... 14:45:33 I'm thinking, maybe ... PC version and IDE come free, but charge for ... 14:45:38 ram, flash, cpu in it 14:45:46 the platform specific interpreter 14:46:00 that way, we are encouraged to support more platforms 14:46:13 --- join: I440r (mark4@1Cust123.tnt2.bloomington.in.da.uu.net) joined #forth 14:46:20 greets, I440r 14:46:25 and the sys itself can be written in a rather platform specific way 14:46:35 build a pda for pdas 14:46:46 hi i440r 14:46:55 beg parden? how does that work? 14:46:56 put an SA110 in 14:47:13 it runs on its own cpu, in own ram, from own flash 14:47:23 just interfaces to the pda through cardbus slot 14:47:40 :) 14:47:48 has a client program which is pda specific 14:47:49 I440r: we are trying to find a good business model for a cross-platform Forth that runs on any PC or any PDA 14:48:08 blarg ... back to hardware again 14:48:19 that's what can sell 14:48:40 and there's still the need for software too 14:48:42 heh, outrageous development costs 14:48:56 covered by business plan :) 14:49:14 could sink $10 mill into something like that, easy 14:49:29 doesn't need its own screen 14:49:57 but an additional ethernet controller would be nice :) 14:50:11 I440r: would you be interested in such a project? 14:50:13 and scsi of course ... 14:50:21 firewire 14:50:31 bluetooth 14:50:54 heh, this is my PDA ... small box ... this is my 100Gig scsi drive for my PDA ... big box w/ cable 14:50:56 water pistol reservoir and trigger 14:51:24 and a slot to put pills in 14:51:39 ha! timed pill dispenser! 14:51:57 older people have a hella hard time remembering which meds to take when 14:52:05 pim integration 14:52:20 pill pim 14:52:44 that's where the water pistol comes in handy too 14:52:59 anyway, do you think the effort of bytecode compatiblity is worth it? 14:53:42 probably on a system being hosted by the pda. for a pda extension with own cpu, i'm not sure about 14:54:10 power is an issue 14:54:31 if it sucks the battery empty in half the time, that's not good 14:54:33 are you still talking hardware extension? 14:54:38 both 14:55:09 wouldn't know how to integrate the water pistol wothout extra hardware 14:55:12 in most cases, power conservation is handled by the underlying OS 14:55:34 I see no reason to interfere with that 14:55:51 did you read about electric generators, integrated into shoes ? 14:56:04 driven by walking energy 14:56:13 heh, no ... for charging PDAs? 14:56:17 yep 14:56:28 that's too damn funny 14:56:34 is it on the market? 14:56:34 good idea i think 14:56:39 not sure. 14:57:15 but methanol fuel cells have quite some potential too 14:57:26 don't like the idea of being forced to wear specific shoes 14:57:53 but would gladly clip something onto whatever shoes I happen to be wearing 14:58:30 should look it up with a search engine 14:58:50 I'll write an assembler for the SH3 ... Hitachi can go to hell 14:59:30 * MrReach sighs contemplating the Windows Portable Executable Format once again. 15:00:23 Speuler: do you have a syntax preference for labels or structured conditionals? 15:00:46 what other PDAs use the SH3? 15:00:49 i find labels like label: quite readable 15:01:11 yes, the webcam works again ... 15:01:14 I prefer a "label:" keyword 15:01:24 dark and small though 15:01:27 parses next token and makes an entry for it in a list 15:01:43 cool, what was wrong with it? 15:02:04 i actually don't know. i pulled out power, and plugged in again 15:02:13 "label: jump-here-later" 15:02:22 maybe it has crashed 15:02:36 time to replace with new tech? 15:02:42 and took 10 minutes to boot 15:03:11 it is an old image though 15:03:16 I440r: you awake? 15:03:17 still a week old 15:04:20 Speuler: your Psion uses the ARM processor? 15:04:43 that's RISC, right? 15:05:13 www.methanol.org/fuelcell 15:05:19 ARM7 15:05:38 in a CL7110 asic 15:05:47 risc 15:05:55 do yo think you can find the EPOC executable file format? 15:06:01 I looked and was unable to 15:06:59 is there any reason why the various assemblers should not operate identically, except for the nmemonics? 15:07:25 single-pass, 2-pass, multi-pass 15:07:58 single-pass, I should imagine 15:08:02 many assemblers are multi-pass 15:08:18 I can't think of a prob that can't be solved in single pass assembler 15:09:00 I'm thinking a Forth assembler, like any other Forth assembler 15:09:30 www.eren.doe.gov/RE/hydrogen_fuel_cells.html , scroll down, follow the methanol link 15:09:31 the only thing I can think of that might be unusual ... 15:10:03 is the .SECTION directives that are required to make more sophisicated executables 15:10:54 --- join: aum (~aum@210-86-60-200.jetstart.xtra.co.nz) joined #forth 15:11:05 hi - does anyone use oo in forth? 15:11:05 offsets changing because of different opcodes generated, as offsets allow, changing code size, recalculating offsets, changing opcodes .. until all refs are stable 15:11:09 greetings, aum 15:11:15 hi MrReach 15:11:18 aum: sort of 15:11:40 * aum suspects that OO in forth might be reeealllyyy ssllllooowww 15:11:54 Speuler_: do you think it's worth all the extra work? 15:11:57 pure forth is quicker 15:12:22 aum: some are, some are nearly as fast as raw forth ... depends on the lookups 15:12:26 mrreach: could be. single pass asm have to use some tricks. multi pass use brute force 15:12:33 and when the lookups are done 15:13:03 i could implement a super-fast oo, but its syntax would be ugly 15:13:22 heh, ugly Forth OO systems are not uncommon 15:13:35 instead of 'instname --> methodname', say,... 15:13:52 syntax would be 'instname classname --> methodname' 15:13:56 aum: if it is just syntax, you could convert it with a specialized line interpreter 15:14:18 except that i want to pass instance pointers as anonymous stack vars 15:14:24 aum: that under the hood, oo still costs 15:14:27 so line interpreters would be next to impossible 15:15:05 well, the scheme i have in mind would be hardcore early binding, which would create negligible speed cost 15:15:26 all the lookup work would be done at compile time 15:15:29 yes, I prefer that 15:15:43 aum: early binding applies only in some case. you still would need some late binding too 15:15:53 where early binding is the norm, and contortions must be done for late binding 15:16:15 i think i could eliminate all need for late binding 15:16:43 prob not, but it could be implemented using state variables and IF statements 15:17:08 aum: can you store a token of an object in a , say, var ? 15:17:17 a variable ? 15:17:36 and use it as object, after reading the token from variable ? 15:17:49 but syntax would be ugly - if a class contains an instance of another class, then the syntax for setting an attribute would be '5 instptr class1 --> inst2 class2 --> attrib1 ! ' 15:18:00 and yes, the whole class ref would then be a single pointer 15:18:11 (i.e. not known at compile time what object you are dealing with) 15:18:25 Speuler: that's what he's saying .... 15:18:39 that the type is explicitly told to the compiler 15:18:59 yes - by specifically mentioning them in colon defs, class names would work as 'casts' which would simply add an offset to a class ref ptr 15:19:39 aum: you could make compile-time lookups more sophisticated, if you want 15:19:41 so the cost would be a simple addition to a pointer for each level of class nesting - no table lookup - fast but very ugly 15:20:21 aum: at compile time, why not just compile the xt of the pecified method? 15:20:40 that's what i intend 15:21:00 or in the case of attributes, the xt of a word to push the attribute's address 15:21:11 then the lookup can be as complex as you want at compile time, with little or no runtime impact 15:21:17 yes! 15:21:38 ugly long-winded source code though 15:21:39 that is what SwiftForth does, it has an awesome OO package 15:21:53 I'm using a FICL derivative 15:22:13 the code inside a class def looks EXACTLY like regular forth 15:22:17 because an inherited method may have been overwritten 15:22:37 and has another xt than the compiler thinks it should have 15:22:42 but isn't swiftforth commercial? 15:23:42 its by forth inc so it prolly is :P~ 15:23:55 that's a problem for me 15:24:10 i'm using forth to create turnkeys that don't require any forth environment 15:24:19 --- part: I440r left #forth 15:24:19 i440r: i think i found a solution to our recent disagreement 15:24:23 --- join: I440r (mark4@1Cust123.tnt2.bloomington.in.da.uu.net) joined #forth 15:24:27 i440r: i think i found a solution to our recent disagreement 15:24:32 an example app i've just written is http://netforth.sf.net/psst 15:24:40 what was it again ? 15:24:42 i forget heh 15:24:44 its old age 15:24:48 im 38 now :P~ 15:24:49