00:00:00 --- log: started forth/05.06.08
00:00:13 <arke> seah
00:00:15 <arke> yeah
00:00:26 <arke> usually foo? returns teh boolean and ?foo does something based on a condition
00:00:40 <arke> so, like : foo? bar @ 0 ;
00:00:48 <asymptote> I think of ?foo as having side effects,
00:00:49 <arke> : ?foo foo? if bar then ;
00:00:50 <arke> or something
00:00:56 <asymptote> and foo? just having stack effects
00:01:26 <asymptote> but that may just be me
00:02:40 <arke> :)
00:03:20 <sproingie> cute.  i'm used to creating a directory for projects in a language, then checking out into that directory
00:03:28 <sproingie> so for factor, i end up with a path like: 
00:03:33 <sproingie> chuck@tetsuo:~/proj/factor/Factor/Factor/factor$
00:03:38 --- join: Topaz (~top@sown-88.ecs.soton.ac.uk) joined #forth
00:03:54 <sproingie> speaking of factoring ...
00:04:22 <alexander_> hmm
00:05:02 <slava> i wish people would stop using the CVS sources unless they are contributors
00:05:08 <slava> your next question will be "why doesn't this work?"
00:05:23 <slava> what's so hard about clicking a download link?
00:05:30 <sproingie> i usually go from cvs sources first and if it doesn't work, work backward from there
00:05:35 <slava> why?
00:05:38 <asymptote> why?
00:05:45 <slava> the latest release is packaged, tested, and known to work
00:05:48 <asymptote> (sorry, I've just always wondered why people do that)
00:05:58 <slava> CVS has various problems right now because of a major compiler redesign
00:06:01 <sproingie> because cvs update is much faster than download, move, untar, etc 
00:06:15 <sproingie> but if you say head is borked, i'll grab the release
00:07:06 <slava> do you use the latest cvs version of the kernel and libc and x11 too? :)
00:07:48 <asymptote> if so, switch to #gentoo
00:08:20 <slava> actually, i have a pretty effective strategy to prevent people from using the CVS version unless they have a resaon
00:08:37 <slava> I don't document the location of the CVS bootstrap images ;)
00:09:02 <sproingie> not in cvs, eh? 
00:09:05 <slava> no
00:09:14 <sproingie> guess that'll do it
00:11:23 <sproingie> oh how god damned wonderful
00:11:40 <sproingie> my 32 bit gcc isn't.  or it's messed
00:11:52 <slava> you're using gcc 3.3 and there's a internal compiler error?
00:11:54 <sproingie> got the most UGLY error output i've seen from a compiler
00:12:01 <sproingie> yep
00:12:04 <slava> yes, you need gcc 3.4 or 4.0
00:12:10 <sproingie> spilled a screen full of RTL at me
00:13:02 <sproingie> 4.0 it is
00:13:51 <sproingie> argh.  the makefile does not respect the value of $CC
00:16:41 <sproingie> getting compile errors in the bootstrap
00:16:51 <slava> ignore them
00:17:20 <slava> i should change it to print 'compile warning' instead of 'compile error' to stop people from reporting this
00:18:01 <sproingie> oh heh there it is at the end
00:18:20 <sproingie> >>> Unless you're working on the compiler, ignore the errors above.
00:30:46 --- join: qFox (~C00K13S@92pc222.sshunet.nl) joined #forth
00:42:17 --- part: asymptote left #forth
00:47:22 --- quit: sproingie (Remote closed the connection)
01:14:12 --- quit: Topaz ("Leaving")
01:47:36 --- quit: slava ("USE: factor")
04:51:48 --- join: onetom_ (~tom@ns.dunasoft.com) joined #forth
04:51:48 --- quit: onetom (Read error: 104 (Connection reset by peer))
05:39:06 --- join: PoppaVic (~pete@0-1pool67-236.nas22.chicago4.il.us.da.qwest.net) joined #forth
05:39:23 <PoppaVic> Mornin'
05:52:28 <ASau> Evenin'
05:52:42 <PoppaVic> lo
07:08:46 --- join: Herkamire (~jason@c-24-218-95-147.hsd1.ma.comcast.net) joined #forth
07:08:46 --- mode: ChanServ set +o Herkamire
07:28:11 <I440r> were not allowed to call this channel #forth
07:28:20 <I440r> because we dont own the forth programming language
07:28:31 <PoppaVic> no own owns it
07:28:36 <I440r> we have to call it ##forth according to the freenode channel naming polacy
07:28:42 <I440r> hmmm
07:28:45 <PoppaVic> even Moore made it pd until forth, inc
07:28:47 <I440r> maybe thats true
07:36:57 <I440r> well i wasnt planning on changing the channel name anyway
07:37:01 <I440r> ill move it before i change it
07:37:12 <I440r> err tho its not MINE any more lol
07:37:20 <I440r> its more like "ours"
07:37:36 <PoppaVic> check the ##C ops - we autojoin those that are lost
07:37:51 <I440r> ?
07:38:21 <PoppaVic> we are set up to redirect #C joiners to ##C
07:38:44 <PoppaVic> I now join ##C direct, but the former was useful for a long time - and still is, it appears
07:40:02 <I440r> same as #freebsd
07:40:12 <PoppaVic> I dunno why there WOULD be a #forth channel under the new rules, since only forth,inc is legit
07:40:29 <madwork> #froth
07:40:48 <PoppaVic> I said the same in #C, but... *sigh* It's easier to just set it up and drop it
07:41:26 <PoppaVic> it seemed - to me - that #gcc or #msc or #ansiC and such made more sense.
07:42:17 <PoppaVic> I440r: maybe you and herk and whatever should all NAIL the official names for your releases and route everyone to ##forth?
07:43:05 <I440r> noooooo
07:43:22 <I440r> this is #forth if fweenode (camp voice) tries to force a change to ## im leaving
07:44:20 <PoppaVic> well, that's fine. But lilo & friends have that 'right'. otoh, I am positive one can make a case for the fact there IS no damned std
07:44:36 <PoppaVic> ansi never made me all that excited
07:45:06 <PoppaVic> ..and I can never envision those folks bothering to come to freenode and #forth
07:45:29 <PoppaVic> STILL, makin' lilo happy can't hurt <shrug>
08:01:00 <I440r> chuck did a few times, jeff fox was also semi regular
08:01:11 <I440r> but E.R. never came tho she was invited :)
08:02:08 <madwork> I don't see a problem with ##forth. It's just a naming convention for channels.
08:02:24 <I440r> its anal retention
08:02:46 <PoppaVic> yes, it IS anal-retentive - it's also painless
08:02:48 <madwork> It's a sharp... press it twice, and you're good to go.
08:03:16 <I440r> well i never ever understood the need for the FIRST # on channel names
08:03:21 <PoppaVic> painless, no-cost "make lilo happy" solutions do not bug me
08:03:29 <I440r> im not here to make lilo happy
08:03:33 <I440r> im here to make ME happy :P
08:03:35 <PoppaVic> I440r: welcome to a shitty protocol
08:03:39 <I440r> lol
08:04:05 <PoppaVic> no, yer here to jabber, and lilo makes it possible... It's like living at home: make them happy or move the hell out.
08:04:28 <I440r> no. i can always find another irc server or even start my own
08:05:34 <PoppaVic> that's fine
08:05:43 <I440r> yup
08:05:58 <I440r> i doubt ##forth is going to be a requirement anyway
08:06:05 <PoppaVic> seems pretty stupid to me, but go for it
08:06:18 <PoppaVic> I440r: ask him, lilo is not unreasonable
08:06:44 <PoppaVic> I'd wager my left nut they expect #forth for ANS
08:06:58 <I440r> ans does not own the name forth
08:07:03 <I440r> and ans is not forth anyway
08:07:05 <PoppaVic> I agree
08:07:10 <I440r> its just a language of the same name
08:07:13 <PoppaVic> to both - and so what?
08:07:27 <I440r> so why ask him and make him decide against :P
08:07:37 <PoppaVic> you will not know until you ask <shrug>
08:07:41 <I440r> just be quiet and maybe he wont even THINK we need to change
08:08:29 <PoppaVic> hehe... ##forth is open for takeover ;->
08:10:20 <I440r> lol
08:10:27 <I440r> all yours
08:10:30 <I440r> :P
08:10:50 <PoppaVic> I've no interest in adding channel-owner/admin to my chores
08:11:03 <I440r> which is why i no longer admin this channel :P
08:11:04 <I440r> futhin does :P
08:15:32 --- quit: PoppaVic ("brb - beat the isp")
08:16:56 --- join: PoppaVic (~pete@0-1pool65-214.nas22.chicago4.il.us.da.qwest.net) joined #forth
08:17:40 <PoppaVic> beat the isp
08:18:31 <I440r> lol
09:02:43 --- quit: cmeme (Connection reset by peer)
09:03:03 --- join: cmeme (~cmeme@216.184.11.2) joined #forth
09:11:40 <PoppaVic> MAn, I need a few days of "stare at the printout" *sigh*
09:26:20 --- join: sproingie (~chuck@64-121-15-14.c3-0.sfrn-ubr8.sfrn.ca.cable.rcn.com) joined #forth
09:31:35 --- join: PoppaVic_ (~pete@0-1pool66-169.nas22.chicago4.il.us.da.qwest.net) joined #forth
09:31:48 --- quit: PoppaVic (Nick collision from services.)
09:31:51 --- nick: PoppaVic_ -> PoppaVic
09:54:50 <I440r> a friend of mine who sells radio active stuff on ebay just make coffee from heavy water lol
10:11:05 --- quit: PoppaVic ("Laters, inflex")
10:29:48 <sproingie> how's it taste?
10:50:30 --- quit: cmeme (Read error: 131 (Connection reset by peer))
10:52:32 --- join: cmeme (~cmeme@216.184.11.2) joined #forth
10:52:47 --- quit: cmeme (Remote closed the connection)
10:53:34 --- join: cmeme (~cmeme@216.184.11.2) joined #forth
10:55:39 --- quit: cmeme (Remote closed the connection)
10:56:52 --- join: cmeme (~cmeme@216.184.11.2) joined #forth
10:56:53 --- quit: cmeme (Remote closed the connection)
10:57:38 --- join: cmeme (~cmeme@216.184.11.2) joined #forth
11:03:43 --- quit: ASau (Remote closed the connection)
11:44:21 --- quit: qFox ("this quit is sponsored by somebody!")
12:35:46 <arke> teh hies
12:46:06 --- quit: Al2O3 (Remote closed the connection)
14:48:26 <crc> hi all
14:49:24 <alexander_> hey
14:53:38 --- quit: KB1FYR (Read error: 104 (Connection reset by peer))
14:55:00 --- join: KB1FYR (~Alex@196-220.suscom-maine.net) joined #forth
14:56:28 <Frek> http://www-128.ibm.com/developerworks/power/library/pa-dw-slof.html
15:01:42 --- join: tathi (~josh@pcp02123821pcs.milfrd01.pa.comcast.net) joined #forth
15:20:36 --- join: Al2O3 (~Al2O3@c-24-1-126-202.hsd1.tx.comcast.net) joined #forth
15:48:11 --- quit: sproingie (Remote closed the connection)
16:23:37 --- quit: I440r ("Leaving")
16:23:57 --- quit: tathi ("leaving")
16:27:02 --- join: OrngeTide (~orange@rm-f.net) joined #forth
16:27:08 <OrngeTide> hio
16:27:40 <OrngeTide> x86 macs?! what will we run powerpc forth on?
16:28:02 <crc> on Gumstix :)
16:28:10 <OrngeTide> yea... ARM is good
16:28:22 <OrngeTide> I have a stack of sparc classics :)
16:29:10 <alexander_> the ARMs are looking great
16:29:15 <alexander_> they're coming down in price, and going up in functionality
16:29:15 <OrngeTide> you could make a cool processor for a XESS board and run forth on that. it has VGA and keyboard jacks even. and almost the same price as a gumstix
16:29:31 <OrngeTide> yea. someone pointed out to me that they have 600Mhz+ Xscales.
16:30:07 <OrngeTide> all the game consoles are going powerpc. 
16:30:16 <OrngeTide> i want an ARM game console :)
16:30:43 * crc wants some ARM hardware that's not tied to PalmOS or PocketPC
16:31:00 <OrngeTide> it can be one of those direct-to-tv joysticks and just put an MMC slot in it. that would be a cool game system :)
16:31:20 <alexander_> crc: gumstix, GameBoy Advance, NetWinder, etc.
16:32:08 <OrngeTide> crc, i know. i was going to hack my sony clie but it was pretty tough to find enough info to hack it. i'd have to reverse engineer it with hexdumps and possibly wire jtag straight to the processor pins
16:33:12 <OrngeTide> gba is pretty louse at 32-bit arm. it likes arm thumb (16-bit bus for most of the ram and rom)
16:33:24 <OrngeTide> s/louse/lousey
16:33:32 <OrngeTide> is that how you spell it? I don't know
16:34:19 <alexander_> ARM thumb isn't that bad
16:34:40 <alexander_> I guess it depends on what you want.
16:35:02 <OrngeTide> i wonder why there are all these game consoles with ethernet and 56k modems but none of them have a web browser. yet people buy video cards with tv-output and hook their pcs up and browse the net, watch movies, play games 
16:35:19 <OrngeTide> alexander_, i want a blazing fast 32-bit architecture :)
16:35:41 <OrngeTide> although thumb instructions do use 32-bit registers.. but the instruction set is kind of crippled.
16:35:46 <alexander_> OrngeTide: yah, these are cheap microcontrollers and video game consoles. :)
16:35:51 <alexander_> and you can use 32-bit ARM if you need to.
16:36:17 <OrngeTide> although targeting a minimalist forth for thumb would be easy. since it has more than enough instructions to implement everything in forth.
16:36:35 <OrngeTide> alexander_, yea. 32-bit arm is just a bit slow unless you have 32-bit memory
16:42:35 <alexander_> Atmel's AT91s even have protected memory.. for a $15 a microcontroller 
16:43:33 <OrngeTide> really? is it a full on mmu?
16:44:11 <alexander_> no, you couldn't implement VM with it, but it'll let you do protected processes
16:44:40 <alexander_> it's like a MMU, with only the permission aspect
16:51:43 --- quit: OrngeTide (brown.freenode.net irc.freenode.net)
16:51:43 --- quit: cmeme (brown.freenode.net irc.freenode.net)
16:51:43 --- quit: vitaminmoo (brown.freenode.net irc.freenode.net)
16:51:43 --- quit: Raystm2 (brown.freenode.net irc.freenode.net)
16:51:57 --- join: cmeme (~cmeme@216.184.11.2) joined #forth
16:52:12 <alexander_> go split
16:59:58 --- join: OrngeTid1 (~orange@rm-f.net) joined #forth
16:59:58 --- join: vitaminmoo (~vitaminmo@dsl-95-113.peak.org) joined #forth
16:59:58 --- join: Raystm2 (~vircuser@adsl-69-149-62-224.dsl.rcsntx.swbell.net) joined #forth
17:00:20 --- quit: OrngeTid1 (Client Quit)
17:01:24 --- join: OrngeTide (~orange@rm-f.net) joined #forth
17:01:28 <OrngeTide> grr
17:05:16 --- join: OrngeTid1 (~orange@rm-f.net) joined #forth
17:05:56 <alexander_> yoh
17:06:05 <alexander_> did you get what I said about the ARM's memory controller?
17:13:34 <OrngeTide> yea
17:13:40 <OrngeTide> just protected. can't do virtual memory
17:14:14 <OrngeTide> so it's good enough to run uclinux, but not regular linux. (I think linux on a $15 device is novel)
17:20:30 <alexander_> you could run regular Linux
17:20:35 <alexander_> but you don't have the RAM
17:20:58 <alexander_> and obviously since you can't manage the memory (e.g. virtual/physical) VM is out of the question
17:31:44 --- join: sproingie (~chuck@64-121-15-14.c3-0.sfrn-ubr8.sfrn.ca.cable.rcn.com) joined #forth
17:48:18 --- quit: KB1FYR (Read error: 104 (Connection reset by peer))
17:48:30 --- join: KB1FYR (~Alex@196-220.suscom-maine.net) joined #forth
17:48:50 <OrngeTide> well uclinux is linux without virtual memory. the stock linux from Linus can't run without virtual memory.
17:49:16 <OrngeTide> alexander_, i think some of the AT91s have a bus for doing external memory.
18:00:45 <alexander_> I know what ucLinux is.. I'm just saying you could make the stock Linux work without a full MMU.
18:01:02 <alexander_> it'd be a bitch, but it's not out of the question.
18:01:11 <alexander_> ucLinux is good for when it's completely out of the question.
18:01:48 <alexander_> OrngeTide: and yah, I forgot that most of the older Atmel ARMs expose the memory bus, because they don't come with internal RAM
18:10:55 <OrngeTide> well isn't uclinux just linux without mmu support? the tree is usually 10 steps behind though.
18:11:36 <alexander_> yah, but I mean, you almost have an MMU on these ARMs. :)
18:11:58 <alexander_> the bitch is that binaries can't rely on virtual addressing.
18:12:04 <alexander_> so you have to relocate everything
18:12:18 <alexander_> which I think uCLinux has to do 
18:12:46 <alexander_> so maybe it'd be easier to add protection to uCLinux, than take the virtual aspect away from Linux
18:13:43 <OrngeTide> well uclinux does support memory protection on things that can do it. (like some of the 68k quiccs can)
18:14:06 <OrngeTide> at least as far as I know. I haven't actually touched uclinux in ages.
18:14:20 <OrngeTide> I actually worked on ELKS more than uclinux. now that was a weirdo linux port
18:14:40 --- part: OrngeTid1 left #forth
18:16:40 <OrngeTide> well time to go home. seeya.
18:17:30 <sproingie> uclibc-based micro-distribution would be nice
18:18:04 <sproingie> might actually save enough space to statically link most apps
18:19:03 <crc> or dietlibc
18:19:16 * crc has used dietlibc, but not uclibc
18:22:07 <OrngeTide> dietlibc is good but used to have problems on mips. newlib is okay. uclibc is quite nice. I usually end up writing a libc from scratch or just copying libc from netbsd whenever I start a new embedded platform
18:22:51 * crc likes NetBSD :)
18:23:05 <OrngeTide> sproingie, there are quite a few dietlibc based distros.
18:23:33 <OrngeTide> I originally wanted to do dietlibc + SNOW (as opposed to ELF). but didn't have time.
18:23:50 <OrngeTide> SNOW is cool because you can do "dynamic" linking that is as cheap as static linking.
18:23:59 --- quit: Sonarman (Read error: 110 (Connection timed out))
18:24:18 <OrngeTide> you can only do it for the system libraries though. but isn't that what you are linking to 99.9% of the time
18:24:23 * crc recalls some discussion of SNOW from the early days of the VR3 pda
18:24:29 <OrngeTide> yea. I have a VR3 :)
18:24:36 <crc> I want one!
18:24:51 * crc found the company that now produces them... it's just a matter of time
18:25:05 <OrngeTide> sucky battery life. but it has a nice form factor. almost the entire thing is LCD. with just a few tiny buttons on the bottom and edges.
18:25:24 <OrngeTide> runs Xfree86 and is actually pretty zippy.
18:25:38 <OrngeTide> LCD is almost as easy to read as a PalmIII's. (grayscale)
18:25:59 <OrngeTide> it's better for doing in direct sunlight than those color pdas that everyone has now
18:26:14 <crc> what cpu does it use?
18:26:17 * crc forgets....
18:26:19 <OrngeTide> an NEC mips
18:26:24 <crc> ahh
18:26:35 <crc> Another CPU to learn :)
18:26:43 <OrngeTide> it doesn't have a slot for memory and can only talk rs232. so that's kinda too bad. has a little bit of built-in flash. 
18:26:50 <OrngeTide> mips is a nice simple cpu. 
18:26:58 * OrngeTide works at cisco. he deals with MIPS a bit:)
18:27:09 * crc has a PocketPC (ipaq h2210) that's tolerable in sunlight
18:27:12 <Frek> crc: the only CPU you need to learn is java :)
18:27:30 <Frek> or wait 
18:27:37 <sproingie> what on earth is SNOW?  pointers?
18:27:39 <Frek> may I mean. or whatever
18:27:41 <crc> Frek: no!!!
18:27:42 <sproingie> seems like an ungooglable word
18:27:45 <OrngeTide> my Treo 600 is utterly useless in sunlight
18:27:49 * crc stays away from Java
18:27:53 * sproingie figures if he googles for ELF and SNOW, he'll get santa's workshop
18:27:59 <Frek> crc: java is teh ultimate language/vm!
18:28:22 <crc> nah
18:28:38 <OrngeTide> sproingie, you setup well defined offsets for all your libraries. it works identical to static linking, except you have a front-end to only load what you need at predefined positions. 
18:28:58 <OrngeTide> basically it is static but it allows you to share pages between apps using the same libraries.
18:29:11 <sproingie> OrngeTide: sounds kind of like PE
18:29:21 <OrngeTide> also on mips it kicks butt because SNOW means you don't have to do position-independent code. which hurts on a mips r4k (or older)
18:29:22 <sproingie> with PE, there's preferred offsets, and if they're available, they're used
18:29:40 <OrngeTide> sproingie, SNOW has offsets, and they are mandatory.
18:29:45 <sproingie> ow
18:29:54 <sproingie> why would i want to go back to a.out's headaches
18:30:14 <OrngeTide> sproingie, speed and size.
18:30:26 <sproingie> i suppose with a platform like a palm
18:30:35 <OrngeTide> when VR3 switched to snow they actually had a measurable boost in performance
18:30:42 <sproingie> doesn't palmos use a master db of magic offsets?
18:31:12 <sproingie> don't think i'd want to inflict it on a highly dynamic platform
18:31:15 <OrngeTide> sproingie, it's good for tightly integrated systems. since it's not a big deal to rebuild all your apps when your libc goes over the size you allocated.
18:31:49 <OrngeTide> sproingie, yea. palm is the same way. except it does some tricks since you can also link in 3rd party libraries
18:32:13 <OrngeTide> snow is pretty dynamic as long as you don't need two libraries that have overlapping offsets:)
18:33:36 <alexander_> I'd like to see a protected memory forth kernel, where processes load blocks into the memories
18:33:51 <alexander_> like dynamic library loading
18:33:58 <OrngeTide> that'd fit in really well with AT91
18:34:15 <alexander_> yah
18:34:27 <OrngeTide> would be good for a multithreaded forth. since it would narrow down which thread is having issues.
18:34:33 <alexander_> yah
18:34:39 <alexander_> you could even preemptively multitask
18:34:44 <alexander_> if you wanted such a thing
18:34:55 <OrngeTide> you can preemptively multitask on a Z80...
18:35:01 <alexander_> yah
18:35:03 <OrngeTide> don't need memory protection to do that.
18:35:08 <alexander_> but I mean, the combination is interesting
18:35:12 <OrngeTide> oh sure.
18:35:32 <OrngeTide> basically the memory protection is good for debugging and good for running untrusted code. (although I doubt you'd be doing much of the later)
18:35:37 <alexander_> yah
18:35:48 <sproingie> i'm still waiting for one of the free forths to have cooperative multitasking ala PAUSE
18:35:52 <OrngeTide> i like virtual memory though. i'm a virtual memory addict.
18:36:04 <OrngeTide> sproingie, add it to IsForth
18:36:12 <alexander_> I think doing a single-level store with a Forth would be great
18:36:15 <sproingie> beyond my abilities
18:36:21 <OrngeTide> alexander_, single level store?
18:36:53 <OrngeTide> i wrote a quick and dirty user-space cooprative threading library in C in about a weekend. the concepts aren't hard. it's just a pain to debug if you make a mistake
18:36:54 <alexander_> OrngeTide: the memory is all virtual, and there's no sense of a disk.. you just write to RAM, and as things get paged out/synchronized, it gets stored.
18:37:18 <OrngeTide> and frankly testing and diagnosing properly written forth code is way easier than properly written C
18:37:36 <alexander_> if you lose power, you just turn it back on and it comes up in its synchronized state (of course with all external state needing to be recreated)
18:37:51 <OrngeTide> alexander_, oh. that's cool. there is something like that based around lisp. (i think called vapour)
18:37:55 <alexander_> ah
18:38:01 <sproingie> like smalltalk
18:38:13 <alexander_> it's been done in mini and mainframes, and keykos and eros both do it
18:38:13 <OrngeTide> GC'ing stuff that's on disk must be fun..
18:38:15 <sproingie> well with smalltalk you usually save the image
18:38:18 <alexander_> yah
18:38:43 <alexander_> OrngeTide: yah, you need some way to pay attention to what's needed, and what isn't.. usually through an allocator of some such
18:38:57 <sproingie> i always wonder what the use of a tiny forth system would be on a modern PC
18:39:06 <OrngeTide> with a forth (or lisp) kernel designed to do it, you could save quickly since it would only have to save what was still in ram. and it would leave things that were swapped to disk alone.
18:39:10 <sproingie> then i think, well it's so small, you could store an undo buffer for EVERYTHING
18:39:17 <sproingie> you could rewind time quite a ways back
18:39:19 <OrngeTide> linux has sort of an experimental suspend to disk trick that works that way.
18:39:51 <sproingie> modern PC could handle a "temporal buffer" no sweat
18:39:54 <OrngeTide> it's hard to write a forth for a superscalar architecture that takes full advantage of it. :(
18:40:03 <alexander_> sproingie: yah,.. exactly
18:40:15 <alexander_> you have so much disk, that you just keep writing
18:40:16 <OrngeTide> most mini forths for x86 actually are pretty weak.
18:40:25 <alexander_> if you need more space, just drop old crap off of the disk
18:40:27 <OrngeTide> anyways. i have to go or I'll be late for dinner.
18:40:30 <alexander_> late
18:40:56 <sproingie> could turn a knob and tell it what granularity you want the buffer at
18:41:01 --- join: AlexF (~Alex@196-220.suscom-maine.net) joined #forth
18:41:02 --- quit: KB1FYR (Read error: 104 (Connection reset by peer))
18:41:14 <OrngeTide> alexander_, the problem with that model is 32-bit memory space. i guess you could slice the rest of your disk into some block i/o.
18:41:37 <sproingie> snapshot the whole system every  5 minutes just for autosave safety?  or go into debug mode and journal every last instruction?
18:41:40 <OrngeTide> the lisp i saw didn't have any block i/o or file system. it was ALL virtual memory. ideal on a 64-bit architecture. kind of limited on a 32-bit
18:41:45 <sproingie> debugger that could work backward
18:41:51 <alexander_> OrngeTide: yah,. you'd need to play some tricks
18:41:56 <alexander_> it wouldn't come for free
18:42:02 <OrngeTide> sproingie, multiple snapshot forks would be better. then you could go back to a specific revision. or like you said. just have multiple levels of undo
18:42:34 <sproingie> OrngeTide: right, different approaches depending on the task.  a tiny system could actually make use of all that space
18:42:39 <OrngeTide> alexander_, well if you had different process spaces you could hack the page tables to include a little extra info to have a wider address space.
18:42:57 <OrngeTide> i'm late!
18:43:03 --- join: Sonarman (matt@adsl-64-169-95-61.dsl.snfc21.pacbell.net) joined #forth
18:43:15 <sproingie> otherwise, it's really a waste.  i mean hell, i'm chatting about this on an irc client written in *javascript* fer godsakes
18:43:17 <sproingie> and it responds instantly
18:43:38 <alexander_> sproingie: haha
18:43:40 <sproingie> 100 layers of abstraction and i don't notice.  so bare metal doesn't impress me, but the ability to control time itself on my PC would
18:44:12 <sproingie> ever play Prince of Persia: Sands of Time?
18:44:16 <sproingie> imagine typing rm -rf /
18:44:23 <sproingie> and going oops and hitting the rewind button
18:44:27 <alexander_> yah
18:44:45 <sproingie> fuck all of you
18:44:45 <alexander_> I'd mainly want it for development
18:44:48 <sproingie> *oops rewind*
18:44:51 <alexander_> haha
18:44:52 <sproingie> oops, committed
18:45:14 <alexander_> I'd like to be able to have an editor, that shows the state of forth words inside a running forth process
18:45:21 <alexander_> and then if you fuck up, you can rewind on the timeline
18:45:28 <alexander_> like an automatic CVS 
18:45:35 <sproingie> obviously apps that commit to external state like my tourettes gaffe above would have to make note of that
18:45:56 <sproingie> so in some sense, time is relative from app to app
18:46:29 <alexander_> yah
18:46:32 <alexander_> external state is a problem
18:46:53 <sproingie> well-known problem in functional programming circles.  referential transparency
18:46:55 <alexander_> so apps would need to know that they were thawed, and be able to reconfigure themselves (like reconnect to the network)
18:47:02 <sproingie> it gets trickier when the apps intersect
18:47:06 <alexander_> yah
18:47:19 <sproingie> i could rewind my terminal, but not to the point before it installed chatzilla, which i'm now using
18:47:55 <alexander_> yah
18:47:57 <sproingie> i suppose it could just tell me which parts are undoable, though it might create weird assumptions in a replay
18:48:12 <sproingie> good old entropy
18:48:52 <alexander_> okay
18:48:54 <alexander_> time to get dinner
18:48:54 <alexander_> bbl
19:09:38 --- join: solar_angel (~jenni@CPE0000e812679b-CM000a7362da55.cpe.net.cable.rogers.com) joined #forth
19:16:49 --- quit: AlexF ()
19:20:18 --- join: KB1FYR (~Alex@196-220.suscom-maine.net) joined #forth
19:31:01 --- join: AlexF (~Alex@196-220.suscom-maine.net) joined #forth
19:31:02 --- quit: KB1FYR (Read error: 104 (Connection reset by peer))
19:52:47 --- join: asymptote (~dmesg@220.muma.nsvl.chcgil24.dsl.att.net) joined #forth
19:56:46 --- quit: asymptote ("Free the mallocs!")
20:00:20 --- join: Sonarman_ (matt@adsl-67-113-234-126.dsl.snfc21.pacbell.net) joined #forth
20:06:53 --- quit: solar_angel ("The setting sun.")
20:08:24 --- join: LOOP-HOG (~chatzilla@sub22-119.member.dsl-only.net) joined #forth
20:08:29 <LOOP-HOG> hi
20:12:58 --- quit: Sonarman (Read error: 110 (Connection timed out))
20:22:28 <arke> hi
20:26:46 <LOOP-HOG> my brother and myself are going to try some 2 player gba games soon
20:26:57 <LOOP-HOG> i'm waiting for him to get the datalink cable from the store
20:27:13 <LOOP-HOG> also I am updating my links site, putting in descriptions for the links
20:33:31 <alexander_> you doing it in forth?
20:35:13 <LOOP-HOG> my site, maybe :^)
20:40:58 --- part: Frek left #forth
20:49:18 <alexander_> I meant your games
20:51:03 --- join: Frek (~anvil@h208n2fls31o815.telia.com) joined #forth
20:51:31 --- join: snoopy_1711 (snoopy_161@dsl-084-058-128-045.arcor-ip.net) joined #forth
20:59:34 --- quit: Snoopy42 (Read error: 145 (Connection timed out))
20:59:56 --- nick: snoopy_1711 -> Snoopy42
21:00:25 <LOOP-HOG> my game Skampy was written in Forth and Assembly
21:00:52 <LOOP-HOG> If it was on the GBA, it would probably be 100% Forth, as the GBA has actual graphics support hardware, something the Traimels didn't think was important
21:00:57 <LOOP-HOG> bless their kidney stones
21:01:30 <LOOP-HOG> No the GBA games we are going to try are commercial games that we bought
21:05:24 <alexander_> ah
21:20:39 <LOOP-HOG> well, see you later
21:20:40 --- quit: LOOP-HOG ("ChatZilla 0.9.61 [Mozilla rv:1.7.1/20040707]")
21:59:24 --- part: AlexF left #forth
22:13:07 --- quit: Herkamire ("off to bed")
23:25:03 --- quit: sproingie (Remote closed the connection)
23:59:59 --- log: ended forth/05.06.08