00:00:00 --- log: started forth/01.11.01 03:12:37 --- join: MrLunch (mrreach@209.181.43.190) joined #forth 03:13:28 --- quit: MrNap (Ping timeout for MrNap[209.181.43.190]) 03:21:41 --- quit: clog (Ping timeout) 03:21:41 --- log: stopped forth/01.11.01 03:22:50 --- log: started forth/01.11.01 03:22:50 --- join: clog (nef@bespin.org) joined #forth 03:22:50 --- topic: 'do drop >in ' 03:22:50 --- topic: set by I440r on [Mon Oct 29 10:56:53 2001] 03:22:50 --- names: list (clog MrLunch) 04:03:04 --- quit: clog (Ping timeout) 04:03:04 --- log: stopped forth/01.11.01 04:03:12 --- log: started forth/01.11.01 04:03:12 --- join: clog (nef@bespin.org) joined #forth 04:03:12 --- topic: 'do drop >in ' 04:03:12 --- topic: set by I440r on [Mon Oct 29 10:56:53 2001] 04:03:12 --- names: list (clog MrLunch) 04:15:04 --- quit: MrLunch (Read error to MrLunch[209.181.43.190]: Connection reset by peer) 04:28:12 --- join: MrNap (mrreach@209.181.43.190) joined #forth 07:31:35 --- nick: MrNap -> MrReach 07:47:46 --- join: Speuler (akhandel@orange.icafe.spacenet.de) joined #forth 07:47:52 g4day 07:48:03 good morning 07:48:16 afternon 4 me 07:48:18 noon 07:48:24 Why was it you disagreed with WinCE? Just that M$ made it? 07:48:38 so it is ... good afternoon, then 07:48:41 i was toying with a ce device a while ago 07:48:47 when i already had the psion 07:48:52 i found it a pain to use 07:49:03 tiny, undecipherable icons 07:49:11 crammed screen 07:49:21 that was ce 2.x 07:49:37 ok, I think I agree with you 07:49:47 but the fact that it is microsoft plays a role too 07:49:55 I bought the Jornada 07:50:11 they say that ce 3.x should be better 07:50:18 it's not bad, but I *KNOW* I can devise a better interface 07:50:25 i never considered giving it a try 07:50:58 just as much as i`m going to try xp 07:51:03 or win 2002 07:51:17 or what else is in mirosoft's cooker 07:51:33 may have to study a little bit soap though 07:52:24 jornada keyboard is said to be reasonable 07:52:37 but afaik you can't replace the os 07:53:18 is there an assembler for the journada ? 07:54:03 and linker, to make a running program ? 07:54:07 hi kintaro 07:54:31 the netbook looks nice. but psion went downhill 07:55:00 they say they'll support only corporate clients in the future 07:55:04 which is bad 07:55:23 very sleek 07:55:44 good keyboard 07:55:52 runs epoc 07:56:26 btw, i've uploaded some detail maps to the emap 07:56:49 still like it 07:56:59 * MrReach nods. 07:57:18 there is "PocketC" for the Jornada 07:57:29 I really *LIKE* the hardware 07:57:29 i would have bought the gps76 if it would have had exchangable flash 07:57:47 same size screen, better resolution 07:57:52 but it takes about 7 taps to start entering a new memo 07:57:56 slightly larger than the emap 07:58:06 when it should only take two or three 07:58:08 floats 07:58:31 no hotkey cabability ? 07:58:37 have you ever come across anything that might be the hardware spec for PDAs? 07:58:50 yes, there are four face-hotkeys 07:59:10 and two side hotkeys, currently used for recording audio notes 07:59:27 I'm peeved that they do not record into MP3 08:00:04 I very much like the fact that it comes with the spreadsheet standard 08:00:07 should be a ... 08:00:08 brb 08:00:45 nice processor, too ... Hitachi SH3/133MHz 08:06:16 back 08:06:24 ok 08:06:43 "you're looking like you would know ...' client inquiry 08:06:53 haha! 08:06:59 how did i give me away ? 08:07:07 you must have "Tech" written across your forehead 08:07:41 you're prob the only one in the room typing faster than 10wpm 08:07:49 amazingly few windows>linux problem questions here ... 08:08:02 that's surprising 08:08:38 i'm seeing messages from #assembler here 08:08:49 while i'm in #forth 08:08:57 does anyone in #assembler read me ? 08:09:04 I was wondering why you were greeting people not in this channel 08:09:29 i wasn't seeing spooks 08:09:50 no, I figured you were using a non-windowing client 08:09:59 I do that all the time with BitchX 08:10:15 currently running irc , not ircii 08:10:24 ircii doesn't do that 08:10:33 i tried before 08:10:51 should do /leave before /join i guess 08:11:13 mrreach, i hop over to assembler. will you type at me in a few seconds ? 08:11:39 (after i left, that is) 08:11:41 what me? type something? are your serious? 08:11:44 yes 08:11:51 oh, I see why you would ask 08:12:04 you're still in #Forth 08:12:10 you're still in #Forth 08:12:16 you're still in #Forth 08:12:23 you're still in #Forth 08:12:34 you're still in #Forth 08:12:45 well, you need to part channel 08:13:02 did you read me ? 08:13:08 nope 08:13:14 but i read you 08:13:18 but you never appeared to leave #Forth 08:13:35 it only seems to channel outgoing text 08:13:45 and mixes incoming text onto a single screen 08:14:16 yes, it has always done that 08:14:34 iirc, the displayed lines are prefixed by the channel name 08:15:49 how long does the journada runs on one battery charge ? 08:16:20 I can't say from experience, but supposed to have about 8 hrs of activity 08:17:21 it's amusing to note that my PDA has more physical memory than my desktop (but no swap) 08:17:39 (I bought a 128MB CompactFlash module to go with it) 08:21:13 that nice. will give you some slack 08:21:22 you might need it :) 08:21:40 the price was right, for a retail store ... only $80 08:22:01 pity the emap flash is pretty xpensive 08:22:20 garmin charges about 300 $ for 128 mb 08:22:22 doesn't use any of the standard ram modules available? 08:22:35 unluckily, no 08:22:36 pity indeed 08:22:38 smaller 08:22:44 perhaps they will change that 08:22:51 resembles the sony memory stick 08:23:13 heh, this compactFlash module is tiny, I was surprised 08:23:36 also surprised that they are able to stick modems and ethnet devices into the same form-factor 08:23:39 i hoped i could reuse the CF from the psion 08:24:01 there are pcmcia adapter for CF, to stick it into a pcmcia slot 08:24:13 how fast is the CF spec??? 08:24:31 that's great for exchangin big amounts of data 08:24:46 I'd rather have it the other way around ... a CF mod allowing use of pcmcia cards 08:24:55 about 2 mb per second 08:25:13 ok, that's is probably why it is not used more 08:25:33 8 bit bus width? 08:25:55 not sure 08:26:52 have you looked at the format of the emap database at all? 08:27:09 not at all. 08:27:32 I've looked at the SGML spec of the USGS topo data quite thouroughly, it would be tough to write a browser 08:27:48 so emap prob uses a custome format 08:28:03 because it would be both smaller and easier to parse 08:28:26 (but I'm quite sure the data originates from the same source) 08:30:10 i have roads+recreations from the gps162, and now a metroguide for the emap 08:30:27 emap can load R+R too, but gps162 can't load metroguide 08:30:44 brb 08:30:46 ok 08:31:48 back 08:33:08 * Speuler must have a look into the mirror about that "tech" think on my forehead 08:33:40 heh 08:34:40 how large a memory set do the mapping programs need? 08:35:03 (i'm thinking that might be incredibly useful when I'm doing real-estate inspections) 08:35:06 with R+R, munich takes about 1 MB 08:35:15 Seattle, is way bad about traffic 08:35:32 oh! that isn't spit 08:35:37 i can get the route and surrounding of munich - mainz (about 250 miles) in 3 mbyte 08:36:16 the metroguide is larger 08:36:43 it conntains a database with attractions, accomodation, restaurants, house numbers and the like 08:37:04 r+r can4t search for roads. metroguide can 08:37:30 I see ... how much room does metroguide need? 08:38:33 i got southern germany on it now, taking up 10 mb 08:39:13 i want a 128 mb module,to stick whole country on 08:39:19 still not bad ... surprisingly small, in fact 08:39:54 Id have a pretty tough time fitting that much data into that memory 08:41:35 i'm surprised myself. well, the street planner program for the psion uses maps comparable in size 08:41:40 but much cheaper :) 08:41:50 60 $ for western europe 08:42:05 r+r is 150 $ for one country 08:42:16 metroguide slightly more 08:42:29 * MrReach nods 08:42:50 i can connect gps to the psion maps too 08:43:15 but needing two machines for a job which can be done by one is bothersome 08:43:17 via the serial port? 08:43:21 yes 08:43:31 gps serial out, psion serial in 08:43:57 I still tend to think of "the computer with attached peripherals" ... even though the periphs may be smart 08:45:35 even if i got the psion with me ost of the time, and the gps too, that is not the case for the cable :) 08:45:47 and the psion will go out 08:45:57 be replaced against something quicker 08:46:28 yep, hopefully w/ a serial port and similar software 08:50:48 and cardbus slot 08:50:56 extension connector 08:51:10 want to stick a network card in it too 08:51:23 and a pcmcia scsi controller 08:51:26 * MrReach nods. 08:51:32 heh 09:02:07 brb 09:02:22 ok 09:10:34 back 09:10:46 another "tech" question? 09:10:56 no. personal maintenance 09:11:28 heh, ok 09:11:34 they call it the place where even the emperor walks to 09:17:14 having a nice cup of hot chocolate now 09:17:32 creamy mountain style 09:17:41 heh, was just refilling my coffee cup 09:17:51 lucky you, though, get your drinks delivered 09:18:15 oh well, I've got to have *SOME* exercise 09:18:49 i'm still pressing the mouse buttons myself 09:21:10 * MrReach laughs. 09:21:46 did I ever ask you about the timer interface I'm writing? would now be a good time? 09:22:45 "compaq offers free ipaq 3650 with armada notebooks" 09:23:08 i'm free till about 19:30 local time 09:23:10 ooooh! that might be a good deal, if you have $3,000 to spend 09:23:20 that's in about 1hour20minutes 09:23:44 well, you might be busy with something else 09:24:06 just hanging behind the machine, chatting here and there. nothng serious 09:24:14 reading some news 09:24:23 www.heise.de 09:24:28 (german) 09:24:50 the idea of the timer interface is that some bit of code would be called at some time in the future, either interupting the currently running code or running is a seperate thread 09:25:38 the simplest part is stolen from TCL's "after" procedure 09:26:03 call-back function ? 09:26:26 yes, a call-back 09:26:39 got a tasker ? 09:27:00 of course I've got a tasker, I'm on Windows at the moment 09:27:11 have it activate/revive a task maybe 09:27:17 most complicated tasker ever invented 09:28:00 "after" (in TCL) takes a delay and a script to execute, and returns a "timer ID" 09:29:08 at any time, "after cancel" can be used to prevent the script from being executed, if script has already executed, the after cancel is ignored 09:29:41 allowing periodic activation too ? 09:29:51 so, in Forth, that might look like "#delay ['] someword 1000 after ( hdl) 09:30:10 (i.e. unchanging timer id) 09:30:13 and, conversely, "( hdl) after-cancel" 09:30:28 yes, that is the next step, and much more complicated 09:31:13 it seems to me that, in Forth, that can't be done without a whole bunch of words, right now my proposed interface looks like ... 09:31:52 (grr... spec sheet in the car ... freehand, then) 09:32:22 TIMER ( udata xt delay -- hdl ) 09:32:36 TIMER-START ( hdl -- ) 09:32:50 TIMER-PAUSE ( hdl -- ) 09:33:06 TIMER-RESUME ( hdl -- ) 09:33:20 TIMER-STOP ( hdl -- ) 09:33:32 TIMER-FREE ( hdl -- ) 09:33:38 [end-list] 09:34:07 what difference between PAUSE and STOP ? 09:34:10 in all cases, xt is executed with one stack item, udata 09:34:50 stop shuts off the timer altogether, pause only disables the calling of xt, it doesn't lose it's syncronization 09:35:36 when the timer is restarted, it always calls xt immediately, then repeatedly every delay milliseconds 09:37:02 oh! wrong! xt starts with two item on stack, the u-data AND the hdl 09:37:16 hpw to preceed after STOP ? 09:37:26 because a timer is allowed to cancel itself from inside the callback 09:38:02 ok, when stop is issued, it shuts of the timer at the system level 09:38:46 when start is issued (perhaps after stop, perhaps after creation) the timer is (re)built and starts functioning 09:39:30 pause only stops the system timer from calling xt, it doesn't actually shut off the timer, so timing is maintained 09:39:37 so you could PAUSE, and START ? 09:39:57 yes, but not sure what to do with that 09:40:06 to avoid the need for STOP 09:40:17 I suppose that START should always restart the system timer 09:40:39 hmm no... 09:40:48 would still use up cpu cycles 09:40:50 PAUSE START might be a noop 09:41:00 a few, yes 09:41:08 you need STOP 09:41:19 quite a few, actually, because the stacks are built for the new thread 09:41:37 on modern systems, there isn't much issue 09:41:39 PAUSE START may undo the PAUSE 09:41:58 i don't like the idea that a process goes on after i told it to stop 09:42:19 yes, or PAUSE START might be equivelent to PAUSE STOP START ... which is the way I'm leaning now 09:43:15 START would imply STOP 09:43:22 before START, that is 09:43:30 makes sense 09:43:41 yes, and foul any repetitive timing, if the app depends on that 09:44:12 any provisions to adjust delay with a running timer ? 09:44:28 oh, good idea! 09:44:59 TIMER-DELAY ( delay hdl -- ) ??? 09:45:25 and TIMER-FUNCTION ( xt hdl -- ) ??? 09:45:52 possibly, for changing the xt it makes more sense to start another timer. 09:45:55 of course, if you're gonna chang it, you need a way to know the current value, so then ... 09:46:06 you could argue that this would hold true for changing the period too 09:46:18 TIMER-DELAY? ( hdl -- delay ) 09:46:29 that i feel that what makes a timer specific is the function it performs 09:46:30 TIMER-CUNTION? ( hdl -- xt ) 09:46:32 not the period 09:46:56 if you change the action, it is like having another timer elpase 09:46:57 elapse 09:47:04 that's true 09:47:14 but no reason it CAN'T be done 09:47:18 but changing period may be very useful, without restarting 09:47:31 the only thing I don't like about this is the plethora of words 09:48:40 actually, from within the call-back, one could start a new timer with the new delay, then destroy the old timer 09:48:55 not sure 09:49:16 because you'd like to control periodic timer by timer id 09:49:20 woud it be better to have the first callback performed AFTER the original delay? 09:49:30 starting a new timer oculd yield a different id 09:49:41 if the xt needs to be executed immediately, the starting word could call it 09:49:48 the controlling process would have to be informed about 09:50:03 yes, and the new id would go into the VALUE where the old id was stored 09:50:30 w/o changing ids, you may not need to store it in a value 09:50:50 which basically just remaps one id to another variable one 09:51:28 yes, and require special words to change delay 09:51:38 (or the xt, if one wants) 09:51:59 it is not like memory which you may want to reorganize, aand therefore maintain an indirectly accessed list of memory handles 09:52:36 could you rephrase that? 09:53:09 access to timer ids is less likely to require an additional level of indirection, than memory handles 09:53:28 oh, ok 09:53:39 the timer handle stays till you are finished using it 09:53:46 actually, under the hood, I intend to do just that 09:54:14 the hdl is opaque, t is meaningless to the application outside of giving to timer functions 09:54:18 it seems to me that it is just a control token 09:54:23 yep 09:54:29 could be anything 09:54:39 as long the timer is identified 09:54:46 I reserve the ability to change the implementation of timers at any time 09:55:06 but not while timers are runnning ... 09:55:25 I440r and I got into a quite heated debate, though 09:55:32 heh, not while running, no 09:55:57 about what should be passed to the xt when it starts running 09:55:57 as long the timer id is unique, it wouldn't matter how it is implemented 09:56:19 intuitive guess is: nothing 09:56:25 initially, I had thought of xt running as ( hdl xt -- ... ) 09:56:26 zilch in, zilch out 09:56:56 I4 was of the strong opinion that only hdl should be given to xt 09:57:17 and TIMER-DATA? be used to fetch the data token if it were neccessary 09:57:50 it would be useful to allow the timed process to indentify its timer indeed 09:58:14 it MUST ... the callbacks would be the one to cancel the timer, in most cases 09:58:50 the data token allows the same forth word to process multiple timers 09:58:59 want an example? 09:59:06 i get the idea 09:59:08 ok 09:59:34 yes, timer id passwd to xt makes sense 09:59:38 passed 09:59:49 what about the data token? 09:59:58 derive it from timer-id 10:00:26 ok, I'll change the spec, and add a word to the interface 10:01:28 brb 10:01:33 ok 10:02:43 another, orthogonal, interface would be the RAMP words 10:03:18 except that RAMP looks like RAMP ( data xt delay start stop -- hdl ) 10:03:24 any opinions? 10:05:22 back 10:06:00 lot's of arguments to pass 10:06:30 yes, I was looking at that 10:06:57 start stop and delay are absolutely mandatory 10:07:53 but xt user could be set after the data structure is created 10:08:08 well, I guess start stop and delay could be, also 10:08:13 how would you do it? 10:08:21 but would allow to implement words which require fewer parameters, by implying several of RMP 10:08:23 RAMP 10:09:05 yes 10:09:23 I would tend towards more parameters and fewer interface words 10:09:33 no need to always call ramp from the app directly. 10:10:08 well, RAMP might be as simple all allocating some memory and returning the address as the "handle" 10:10:12 as lowest-level interface to the timer word set, it is a good idea to have one word to funnel everything through 10:10:55 intermediate level word, as interface to app level timer word 10:11:38 may simplify replacing the implementation against another 10:11:45 so you think that there should be several words to setup the parameters before the ramp is started? 10:11:54 as the app-level words don't change 10:12:07 nor their implementation 10:12:08 hmmm ... good point 10:12:40 no, not lots of setup words. 10:13:02 myself, i try to limit the number of words which require a seperate setup 10:13:11 what would the interface look like if you were building it? 10:13:19 with DDA, i use a seperate setup word. 10:13:24 performance reason 10:13:28 * MrReach nods. 10:13:38 DDA itself is quite simple, but setup is complicated 10:13:48 as it produces lots of work variables once 10:14:14 heh, doesn't much of windows 10:14:22 I didn't know you did any windows work 10:14:37 i have written windowing routines 10:15:18 what would the ramp interface look like if you were building it? 10:15:50 together with the character generator stuff, which emulates text mode in graphics mode, text windows/menus etc in graphical programs are possible 10:16:21 i did the orbic program for the sirius/victor 10:16:24 that is certainly true ... hence the availability of the "console" function in Win32 10:16:28 which was written in forth :) 10:17:01 rumoured to be the first program, making use of graphics on that machine . 10:17:07 did you know that Win32 already implemented a console, complete with stdin, stdout, and stderr? 10:17:30 as graphical DOS shell, i'm not surprised 10:17:40 DOS had too 10:18:08 yes, but that might have been only part of the command interpreter program, and not made public 10:18:47 selecting output channel was within the scope of DOS funtion calls 10:18:55 no command line interpreter required 10:19:05 huh? 10:19:42 do you mean to say that the console was implied by dos shell? 10:19:45 i4ll leave now. somebody waiting ... 10:19:49 just called 10:19:52 ok, fare well 10:20:26 g'd bye 10:20:39 --- quit: Speuler (Speuler) 13:36:59 --- join: Speuler_ (l@d120073.upc-d.chello.nl) joined #forth 13:37:12 hi 13:37:43 --- join: Speuler__ (akhandel@orange.icafe.spacenet.de) joined #forth 13:37:49 --- quit: Speuler_ (Speuler_) 13:38:11 --- nick: Speuler__ -> Speuler_ 13:38:48 g'day 14:01:03 welcome back! 14:02:57 hi mrreach 14:03:22 i have to admit, that i didn't ponder the timer interface thing a lot since i left you 14:03:46 heh, you sound like that is a crime 14:04:05 do you think the topic important enough that you SHOULD have thought about it? 14:05:15 i was busy disassembling and integrating some mexican foodstuff 14:05:50 * MrReach laughs. 14:13:50 full set of courses 14:14:04 beg parden? 14:14:18 granted 14:14:36 * Speuler_ pardons mrreach 14:14:44 heh, what courses? 14:14:51 oh! the mexican food! 14:14:53 aperitive 14:14:55 starter 14:14:58 main course 14:15:02 dessert 14:15:07 digestive 14:15:11 coffee 14:15:23 cool 14:15:38 * Speuler_ stopped short of bursting 14:15:59 (by putting a cube of sugar less into the coffee) 14:16:05 heh 14:16:22 feel kind of heavy now 14:16:29 blood drain 14:17:47 I bet 14:28:57 you know, I'm thinking ... 14:29:29 that maybe AFTER TIMER and RAMP might all use the same datastructure, then ... 14:29:41 the PAUSE STOP etc, might apply to all of them 14:29:46 --- join: qless (qless@clgr000977.hs.telusplanet.net) joined #forth 14:29:51 greetings, qless 14:30:01 heya mrreach 14:30:45 was talking about timers, again 14:31:09 oh. what about them? 14:31:31 just wondering how to do a good, relatively portable interface 14:32:12 there was just a small forth talk on linpeople :) 14:32:12 hmmm 14:32:23 ha! 14:32:40 indiana jones land forth people of in 14:32:53 smallforthtalk? sounds like a good name for oo forth :-) 14:33:17 good grief! The WindowsCE 2002 SDK is nearly 48 megabytes, compressed 14:33:35 good you got yourself a 128 mb flash 14:33:56 ack! I'm not going to develop on a pda! 14:34:43 y not ? 14:34:52 too hard to type 14:35:04 OH! I bought a keyboard, lemme get it out of the box 14:35:13 keyboard not good ? 14:35:29 don't know, haven't tried it 14:50:23 later peoples 14:50:24 --- quit: qless ([x]chat) 15:11:08 --- quit: MrReach () 15:24:29