00:00:00 --- log: started forth/07.05.07
01:37:09 --- join: grub_booter (n=charlie@d515301E0.access.telenet.be) joined #forth
02:31:37 --- join: ygrek (i=user@gateway/tor/x-0e6b820176b88403) joined #forth
02:49:29 --- nick: wanderer__ -> probonono
03:13:08 --- quit: ecraven ("bbl")
03:38:07 --- join: luptenschteiner (i=fallback@ppp6-168.lns10.adl2.internode.on.net) joined #forth
03:38:40 --- join: wanderer__ (i=fallback@ppp41-226.lns10.adl2.internode.on.net) joined #forth
03:54:27 --- quit: probonono (Read error: 110 (Connection timed out))
03:54:52 --- quit: yumehito (Read error: 145 (Connection timed out))
03:55:35 --- quit: luptenschteiner (Connection timed out)
04:12:25 --- join: AndyArmstrong (n=andy@82.152.157.85) joined #forth
04:12:25 --- quit: AndyArmstrong_ (Read error: 104 (Connection reset by peer))
04:33:10 --- join: gnomon (n=gnomon@CPE0050eb372bdb-CM001692f57b56.cpe.net.cable.rogers.com) joined #forth
04:33:12 --- quit: ygrek (Remote closed the connection)
04:34:32 --- join: ygrek (i=user@gateway/tor/x-0e4744f4c4ae2b13) joined #forth
05:04:27 --- join: ecraven (i=nex@eutyche.swe.uni-linz.ac.at) joined #forth
05:27:08 --- join: yumehito (n=yumehito@b-internet.87.103.254.70.snt.ru) joined #forth
05:39:08 --- quit: yumehito (Read error: 145 (Connection timed out))
06:14:39 --- quit: ecraven ("bbl")
06:25:04 --- join: yumehito (n=yumehito@b-internet.87.103.254.70.snt.ru) joined #forth
06:35:19 --- join: segher_ (n=segher@dslb-084-056-184-074.pools.arcor-ip.net) joined #forth
06:43:59 --- quit: segher__ (Read error: 110 (Connection timed out))
06:50:19 --- join: yumehito_ (n=yumehito@b-internet.87.103.254.70.snt.ru) joined #forth
06:51:31 --- quit: Quartus__ (Read error: 104 (Connection reset by peer))
07:02:33 --- quit: yumehito (Read error: 113 (No route to host))
07:36:02 --- join: luptenschteiner (i=fallback@ppp2-90.lns10.adl2.internode.on.net) joined #forth
07:37:17 --- join: probonono (i=fallback@ppp15-229.lns10.adl2.internode.on.net) joined #forth
07:53:06 --- quit: wanderer__ (Read error: 110 (Connection timed out))
07:56:00 --- quit: luptenschteiner (Read error: 110 (Connection timed out))
09:11:41 <JasonWoof> jackhammers can be annoying when you're trying to sleep
09:15:09 --- join: perlster (n=chatzill@d142-59-63-133.abhsia.telus.net) joined #forth
09:18:26 <perlster> whois perlser
09:36:31 --- join: luptenschteiner (i=fallback@ppp5-58.lns10.adl2.internode.on.net) joined #forth
09:37:37 --- join: wanderer__ (i=fallback@ppp19-104.lns10.adl2.internode.on.net) joined #forth
09:40:58 <JasonWoof> There is no such nick perlser
09:53:29 --- join: ygrek_ (i=user@gateway/tor/x-e51567e83b4ca317) joined #forth
09:53:33 --- quit: ygrek (Remote closed the connection)
09:56:12 --- quit: probonono (Read error: 110 (Connection timed out))
09:56:25 --- quit: luptenschteiner (Read error: 110 (Connection timed out))
09:59:12 --- join: yumehito (n=yumehito@b-internet.87.103.254.70.snt.ru) joined #forth
10:10:22 --- quit: yumehito_ (Read error: 110 (Connection timed out))
10:13:39 --- join: snoopy_1711 (n=snoopy_1@dslb-084-058-119-232.pools.arcor-ip.net) joined #forth
10:22:07 --- quit: Snoopy42 (Read error: 145 (Connection timed out))
10:22:20 --- nick: snoopy_1711 -> Snoopy42
10:43:26 --- join: yumehito_ (n=yumehito@b-internet.87.103.254.70.snt.ru) joined #forth
10:46:22 --- quit: yumehito (Read error: 60 (Operation timed out))
10:53:02 --- join: yumehito (n=yumehito@b-internet.87.103.254.70.snt.ru) joined #forth
10:56:11 --- quit: yumehito_ (Read error: 145 (Connection timed out))
11:37:09 --- quit: wanderer__ (Read error: 110 (Connection timed out))
11:42:34 --- part: perlster left #forth
11:50:34 --- join: kardinal (n=kvirc@e180147167.adsl.alicedsl.de) joined #forth
11:50:43 --- join: jacereda (n=jacereda@220.Red-80-59-211.staticIP.rima-tde.net) joined #forth
11:59:51 --- quit: kardinal (Read error: 104 (Connection reset by peer))
12:07:21 --- join: ccfg_ (n=ccfg@dsl-roigw1-fe8ade00-21.dhcp.inet.fi) joined #forth
12:14:33 --- quit: yumehito (Connection timed out)
12:15:14 --- join: kardinal (n=kvirc@e180147167.adsl.alicedsl.de) joined #forth
12:19:58 --- quit: ccfg (Read error: 110 (Connection timed out))
12:32:03 --- join: |kardina| (n=kvirc@e180169068.adsl.alicedsl.de) joined #forth
12:49:27 --- quit: kardinal (Read error: 110 (Connection timed out))
12:59:18 --- quit: ygrek_ (Remote closed the connection)
13:24:09 --- nick: OrngeTid1 -> OrngeTide
13:30:21 --- nick: |kardina| -> kardinal
13:36:44 --- join: Quartus__ (n=Quartus_@209.167.5.2) joined #forth
13:43:24 --- join: DocPlatypus (i=skquinn@gateway/tor/x-73ccd592afc2408f) joined #forth
13:44:38 --- quit: jacereda (Read error: 110 (Connection timed out))
14:21:25 --- quit: DocPlatypus (Remote closed the connection)
14:23:09 --- join: DocPlatypus (i=skquinn@gateway/tor/x-f7ee8690de6d1f18) joined #forth
14:26:33 --- quit: cmeme ("Client terminated by server")
14:26:36 --- join: cmeme (n=cmeme@boa.b9.com) joined #forth
14:36:32 --- join: perlster (n=chatzill@d142-59-63-133.abhsia.telus.net) joined #forth
14:45:25 --- part: perlster left #forth
15:04:00 <Quartus__> hi
15:04:06 <madwork> lo
15:45:09 <crc> good evening
15:54:54 <DocPlatypus> is there a portable way to get size of the screen? (like GNU Forth's FORM)
16:02:28 <JasonWoof> wow,
16:02:39 <JasonWoof> my laptop went to sleep as it was halting
16:02:46 <JasonWoof> the battery was dying, so I typed "halt"
16:03:00 <JasonWoof> but before it finished shutting down, it fell asleep
16:03:25 <JasonWoof> (it auto-sleeps when the battery is very low)
16:04:50 <crc> DocPlatypus: none that I am aware of
16:05:55 <DocPlatypus> ok guess I may have to live with some GNU Forth-specific code then
16:10:37 --- join: ttuttle (n=tom@pdpc/supporter/student/gentoo.contributor.ttuttle) joined #forth
16:10:46 <ttuttle> Quartus: Hey.
16:11:01 <ttuttle> Quartus: Do you know iRobot, the company that makes the Roomba robot vacuum cleaners?
16:16:47 --- join: skas_wk (n=skas@121.127.198.32) joined #forth
16:34:16 --- quit: Shine (Read error: 113 (No route to host))
16:51:37 --- quit: slava ()
17:26:21 <DocPlatypus> ttuttle: I'll bite, what about iRobot?
17:34:14 <ttuttle> DocPlatypus: I've got a summer job there, as a programming intern ;-)
17:37:25 <DocPlatypus> rockin'
17:38:27 <DocPlatypus> is it just me or would you be annoyed as well if the source code release of a GPL'd Forth engine was labeled as the "x86 Linux" version?
17:40:58 <DocPlatypus> and what would really be nice is an iRobot Create that could be programmed in Forth ... men can dream
17:59:42 --- join: ttuttle_ (n=tom@pdpc/supporter/student/gentoo.contributor.ttuttle) joined #forth
18:00:57 --- quit: ttuttle (Read error: 104 (Connection reset by peer))
18:19:46 <ttuttle_> 323/quit
18:19:47 --- quit: ttuttle_ ("leaving")
18:51:42 <Quartus__> doc, why would that be annoying?
18:53:12 --- quit: madgarden (Read error: 104 (Connection reset by peer))
18:54:03 --- join: madgarden (n=madgarde@bas2-kitchener06-1096652233.dsl.bell.ca) joined #forth
18:54:19 <DocPlatypus> Quartus__: labelling like that would imply it's binaries for GNU/Linux on i386
18:55:29 <Quartus__> I don't follow what you mean.
18:57:18 <DocPlatypus> maybe it's the cynic in me... but I tend to think anything that has an architecture-specific label on it is a binary
18:57:34 <DocPlatypus> and that source code would be labeled as, well, sources
18:57:47 <Quartus__> why?  Source can certainly be platform-specific.
18:58:22 <DocPlatypus> I'm not saying there aren't cases where it can be, but most of the time in the free software world it's really not
18:58:50 <Quartus__> but if it is, it should certainly be so labelled.
18:59:16 --- quit: madgarden ("?OUT OF DATA  ERROR")
19:00:13 <Quartus__> it may have endian requirements, or embedded assembler instructions, or specific register allocations.
19:00:17 --- join: madgarden (n=madgarde@bas2-kitchener06-1096652233.dsl.bell.ca) joined #forth
19:01:31 --- quit: madgarden (Client Quit)
19:01:55 <DocPlatypus> and somehow I'm reminded of something else... is nobody currently making stack machine CPUs? the info I found today while researching it showed UltraTechnology and Patriot Scientific, neither of which are making CPUs (the latter is just licensing patents)
19:02:05 --- join: madgarden (n=madgarde@bas2-kitchener06-1096652233.dsl.bell.ca) joined #forth
19:02:45 <DocPlatypus> I have grand plans that would ideally involve stack machine or "designed for Forth" CPUs
19:03:37 <Quartus__> why is existing commodity hardware unsuitable for your dream?
19:06:26 --- quit: madgarden (Client Quit)
19:06:37 <DocPlatypus> power consumption comes to mind for one
19:07:21 --- join: madgarden (n=madgarde@bas2-kitchener06-1096652233.dsl.bell.ca) joined #forth
19:07:58 <Quartus__> Forth chips are not low-power simply by virtue of being forth chips, any more than x86 stuff is high-power for being x86.  It's independent of the design.
19:08:01 --- join: mrsbrisby (i=nimh@216.169.144.42) joined #forth
19:08:36 <DocPlatypus> right... well maybe that was a poor way of putting it, but power consumption is a huge issue
19:09:14 <Quartus__> So one of your requirements is low power. There's low power commodity hardware.
19:09:33 <DocPlatypus> the main concern I have is Intel and AMD being too close to Microsoft for my purposes in this project
19:10:38 <DocPlatypus> this would be, in essence, an OS and user environment written from the ground up in Forth, using, probably, 1/4 to 1/10 the computing power on a brand new desktop PC, and probably consuming a comparable amount less power
19:10:45 <Quartus__> Political preferences inform your technical design?
19:10:45 <DocPlatypus> it's something a lot like what Chuck Moore would do
19:11:11 <DocPlatypus> except I'm planning on making it close enough to what some people expect on a "standard desktop PC" to be palatable to the masses
19:11:53 <DocPlatypus> (I'm going to wait until that ping comes back)
19:11:56 --- quit: madgarden (Client Quit)
19:12:06 <Quartus__> ping?
19:12:20 --- join: madgarden (n=madgarde@bas2-kitchener06-1096652233.dsl.bell.ca) joined #forth
19:13:33 <DocPlatypus> I'm lagging horribly in case you haven't noticed
19:13:52 <DocPlatypus> ok
19:14:02 <DocPlatypus> I'm guessing you think I'm just full of crazy ideas
19:14:38 <DocPlatypus> and that's not too surprising, I mean, I've seen information about at least two or three other OSes based around a native Forth implementation, so I know it's been done before
19:15:49 <DocPlatypus> and I've been known to say "keep the PC, get rid of Windows" ... not surprisingly I'm starting to second guess myself
19:15:49 <Quartus__> it sounds misguided and naïve, but everyone should have a hobby :)
19:16:51 <DocPlatypus> well think about this for a second
19:16:53 --- part: littledan left #forth
19:17:12 <DocPlatypus> usually, the first reply someone will get when someone says "I hate my PC" is "get a Mac"
19:17:30 <DocPlatypus> not even "do you really hate your PC, or do you hate Windows?"
19:18:36 <DocPlatypus> sidenote -- from a free software activist standpoint, I can't in good faith recommend buying anything from Apple, in the modern era if you want an Apple iMac, you *will* buy a copy of MacOS, no way around it
19:18:45 --- join: edrx (n=Eduardo@201.5.11.23) joined #forth
19:18:59 <DocPlatypus> this would be as bad as Microsoft being the only source of new PCs, not just new Windows licenses, new PCs
19:19:29 <Quartus__> I'm afraid you've lost me in the tall grasses.
19:20:08 <DocPlatypus> okay. then let me ask this. what type of computer and operating system do you use?
19:20:36 <Quartus__> I use several.  How is that relevant?
19:20:41 <madgarden> I just downloaded Minix 3.
19:21:03 <madgarden> Why not make an Open Source Forth shell for some existing *nix?
19:21:40 <DocPlatypus> I'm trying to figure out your perspective
19:22:11 <DocPlatypus> I use mainly GNU variants with some OpenBSD for security sensitive uses
19:22:15 <Quartus__> my perspective?  Do you think my worldview is warped by what computers I own?
19:23:19 <DocPlatypus> Quartus__: it can change the way one looks at certain computing issues, I will admit
19:23:44 <Quartus__> goodness.  More tall grass.
19:23:58 <DocPlatypus> madgarden: that's a possibility
19:24:43 <DocPlatypus> Quartus__: I have to mentally shift gears when someone talks about their computer having problems, going on to mention viruses, spyware, and corrupted registry
19:25:03 <DocPlatypus> (unlike most people I don't call those computer problems, I call them Windows problems)
19:25:09 <Quartus__> are you asking whether I'm new or uninformed in some way?
19:27:37 <DocPlatypus> no, I'm not
19:27:37 <DocPlatypus> I know you've been around a while
19:28:01 <DocPlatypus> but at the same time I've known people who don't consider anything with an i386 processor a "real computer"
19:29:59 <madgarden> Does it matter what people think?
19:30:03 <DocPlatypus> my mom programs big IBM boxen for a living (AS/400, before that System/38, /36, /34, etc.)
19:30:35 <DocPlatypus> her perspective is a bit different as well
19:30:55 <DocPlatypus> unfortunately in order for social goals to be met it does
19:32:09 <DocPlatypus> there was a time when computing communities were built around sharing ideas and software
19:32:09 <madgarden> What social goals are you trying to meet?
19:32:20 <Quartus__> I don't believe all software should be free, so you'll lose me going off down that road.
19:32:27 <DocPlatypus> now, we have a typical thought process of "getting software" -> "buying software" -> "buying a shrinkwrapped box with a disc containing only binaries and a restrictive EULA"
19:32:46 --- quit: edrx (Remote closed the connection)
19:33:04 --- join: edrx (n=Eduardo@201.5.11.23) joined #forth
19:33:06 <DocPlatypus> I'm not saying I agree with that, but that's what the unwashed masses usually think of when they think of getting more software for a computer
19:34:09 <Quartus__> ok, I don't know who you think you're addressing, but I washed today and I'm not all that massive.
19:34:41 <DocPlatypus> Quartus__: I share a lot of viewpoints with RMS, so don't be surprised
19:34:41 <DocPlatypus> sheesh
19:34:42 <DocPlatypus> it's an expression
19:35:19 <DocPlatypus> my point was... they don't even think of downloading from the Internet, much less getting source code and compiling
19:35:43 <Quartus__> my point is, you're not on Dateline NBC trying to sell a point of view to Joe Sixpack, you're here talking with us.
19:37:51 <DocPlatypus> okay... "I don't believe all software should be free" <-- you're referring to free as in freedom I would assume? even I know making all software free as in price is not sustainable
19:38:20 <Quartus__> I have no problem with closed source.
19:38:40 <DocPlatypus> but I also think "buying software" should not have the strong mental connection to making a trip down to CompUSA/Micro Center/Best Buy and/or ordering a shrinkwrapped box from Newegg
19:39:39 <DocPlatypus> sigh... the question wasn't about source code, it's about freedom, they are two different issues
19:40:04 <Quartus> sigh, eh?
19:42:06 <DocPlatypus> anyway... I would like to think nothing bad will happen to the PC hardware platform as we know it today
19:42:50 <DocPlatypus> but at the same time, it's a reality that Microsoft gets in the faces of PC manufacturers and says "you will put this in your PC or Windows will not run properly"
19:42:55 <DocPlatypus> translated, that means "you will put this in your PC"
19:44:29 <DocPlatypus> and they do, and they will
19:45:08 <DocPlatypus> the GNU project is on to something, but when you have Microsoft influencing hardware design, they can easily shut out any attempts to run an unauthorized kernel or OS if they want to badly enough
19:45:21 <Quartus> ok. Sounds like you have a mixture of beliefs and notions and generalizations, many of which I don't buy into.
19:45:37 <Quartus> So as I say, you lost me in the tall grasses.  Which is fine.
19:46:23 <DocPlatypus> I'd like to think Microsoft will wake up and realize they are competing with what is legally a charity (the FSF is 501(c)3, as much a charity as the Red Cross)
19:48:09 <DocPlatypus> I'm going to reconnect... my lagmeter is showing 306.4 seconds
19:48:12 --- quit: DocPlatypus ("Leaving")
19:49:46 --- join: DocPlatypus (i=skquinn@gateway/tor/x-eb00d8efc103f59e) joined #forth
19:51:28 <DocPlatypus> sorry if that was a bit too rant-ish ... I'm not known for my mainstream points of view or for being shy on certain subjects
20:05:43 --- quit: edrx (Remote closed the connection)
20:14:26 --- quit: DocPlatypus ("Leaving")
20:19:22 --- join: DocPlatypus (i=skquinn@gateway/tor/x-1079efe53a24ab3c) joined #forth
20:58:03 --- join: yumehito (n=yumehito@b-internet.87.103.254.70.snt.ru) joined #forth
21:00:41 --- join: probonono (n=probonon@ppp103-111.static.internode.on.net) joined #forth
21:07:13 <Quartus> well, that's ok. I don't agree with all or part of a number of the things you are on about, so you really do lose me after a couple of leaps there.
21:23:48 <JasonWoof> DocPlatypus: I suspect there is comodity hardware powerful enough for your forthy desktop that isn't x86 and hence free of the microsoft tax
21:24:28 <JasonWoof> though I'm not sure exactly how much computing power you want
21:24:43 <Quartus> You forget the all-important "cannot begin work because of an artificial and arbitrary pre-condition that is not met" consideration. :)
21:25:10 <JasonWoof> do you really think you need the equivilent processing power of a 500-1000 MHZ x86 cpu?
21:25:58 <JasonWoof> Quartus: I don't think it's a bad idea to choose your hardware before you start writing an OS
21:26:23 <Quartus> Sure, I agree.  Not starting until a cheap, low-power, stack-based CPU is made available is something else entirely.
21:28:44 <Quartus> However, there's a huge amount of work involved in creating a whole end-user computing environment, a vast amout of which is completely independent of the choice of CPU.
21:29:56 <JasonWoof> agreed on both points
21:30:26 <Quartus> This is all aside from whether I consider it either something that can be done by a single person, or whether it has a remote chance of succeeding at what its goals appear to be.
21:31:07 <JasonWoof> right
21:31:17 <JasonWoof> most likely it'll just be a learning experience
21:31:29 <JasonWoof> but there is always that chance that something useful to others will result
21:31:36 <Quartus> Linux has been quite successful, but I will warrant that it wouldn't have been if it wasn't something you could run on the computer you already had.
21:32:38 <JasonWoof> yeah
21:32:41 <Quartus> JasonWoof, perhaps -- if he ever even starts.  When I see somebody set unrealistic preconditions, I doubt that they'll ever reach takeoff speeds.
21:32:44 <JasonWoof> I started making an OS in ppc asm
21:32:57 <JasonWoof> that's never going to be finished
21:33:11 <JasonWoof> now I'm working inside a virtual machine
21:33:19 <JasonWoof> so it'll work on the computer (and even the OS) you have
21:33:23 <Quartus> So that's the muffled shouting and thumping.
21:33:41 <JasonWoof> maybe some year it'll run without much of a host system, but I can see that that is so remote, and doesn't really matter at this point
21:34:34 <JasonWoof> "OS" probably isn't the right word for it right now
21:34:39 <Quartus> No, it doesn't.  If you had some beautiful dream of writing a Forth-based OS, you'd write it to a VM to at least show proof of concept.
21:34:40 <JasonWoof> it just runs as an application
21:35:39 <JasonWoof> I figure the way to make it into an OS, is to implement the VM on top of L5 or some existing kernel
21:36:02 <Quartus> Sure.
21:36:17 <JasonWoof> or program an FPGA to do it
21:36:52 <Quartus> Setting unrealistic preconditions is an advance form of procrastination.  As I recall virl doesn't write code because he can't find 'the perfect Forth', the definition of 'perfect' shifting to avoid accidentally finding one that would let him go ahead.
21:37:08 <JasonWoof> lol
21:37:24 <JasonWoof> gotta start somewhere
21:37:26 <Quartus> PoppaVic doesn't write code because, as far as I can decipher his blither, he feels that no computer language is a sufficiently accurate abstraction of his mental model.
21:37:36 <JasonWoof> if you want a perfect forth, decide what that is and make it
21:39:01 <Quartus> Sure, but that would be actual work, and the point is to avoid it. :)
21:39:06 <JasonWoof> I understand not liking the available/popular languages and/or hardware
21:39:24 <JasonWoof> I looked around at my options, didn't like them, so I wrote a VM
21:39:36 <JasonWoof> so I can pretend I'm working on the machine I wish I had
21:39:37 <Quartus> Right, you actually wrote code.
21:39:43 <Quartus> That's the differentiator.
21:39:51 <JasonWoof> yeah, I'm a bit more practical than some
21:40:05 <JasonWoof> still less than you though :)
21:40:31 <JasonWoof> eg I'm chatting instead of getting my work done :)
21:40:36 <Quartus> So am I. :)
21:40:50 --- part: skas_wk left #forth
21:40:54 <JasonWoof> heh
21:41:39 <Quartus> It's an advanced form of procrastination to come up with vague, unrealistic, or impossible prerequisites.  The truly advanced form forgoes all apology and explanation.
21:41:52 <JasonWoof> oh, and I was going to release another version of vor with my latest bug fixes
21:42:08 <Quartus> "vor" immediatly makes me think of "Gor", those wretched novels
21:42:55 <JasonWoof> what novels?
21:43:06 <Quartus> John Norman, I believe is the name of the hack in question.
21:43:25 <JasonWoof> last night I wrote an "autopilot" for vor which randomly presses the movement keys for you
21:43:25 <Quartus> http://en.wikipedia.org/wiki/Gor
21:43:41 <JasonWoof> and the "start a new game" key
21:43:53 <JasonWoof> I left it running all night, and no crashes
21:44:09 <Quartus> Like the Palm OS Emulator's "gremlin" testing feature
21:45:34 <Quartus> Runs a random sequence of input events against your code; you set the random number seed, and the count./
21:45:55 <Quartus> That way you can run the same sequence again to help ensure you've eliminated a given bug.
21:47:43 --- quit: DocPlatypus (Read error: 104 (Connection reset by peer))
21:48:42 <JasonWoof> cool
21:48:44 --- join: DocPlatypus (i=skquinn@gateway/tor/x-9f48877c84d39d43) joined #forth
21:50:28 <JasonWoof> I'm suprised how many bugs I've found lately. between that and valgrind
21:50:42 <JasonWoof> we've been running this game for years
21:50:55 <Quartus> Heisenbugs.
21:51:02 <JasonWoof> and it pretty much only crashes right after I make a stupid mistake in the code
21:51:32 <DocPlatypus> you can be sloppy writing code and think you got away with it
21:51:40 <DocPlatypus> then boom, it blows up on you when you least expect it
21:51:46 <JasonWoof> but after we fixed the bugs with floats rounding inconsistently on x86 we started getting crashes even on ppc
21:52:21 <DocPlatypus> ugh, floats
21:52:30 <JasonWoof> yeah
21:52:44 <Quartus> Useful when used carefully
21:52:45 <JasonWoof> at this point it seems like it would have been easier to do it all with fixed point math
21:53:01 <JasonWoof> I've heard people say that before, and thought they were nuts, like CM
21:53:31 <DocPlatypus> the concept of a separate floating point stack in Forth is a bit strange and takes much getting used to
21:53:47 <DocPlatypus> I find floats to be an unfortunately necessary evil
21:53:52 <Quartus> I heard Knuth say once that the original programmer-mathematicians would have said if you can't solve your problem with integers, you don't understand it properly
21:54:00 <JasonWoof> DocPlatypus: yeah, I used floats in gforth a couple months ago
21:54:09 <JasonWoof> DocPlatypus: the seperate stack thing threw me off
21:54:16 <JasonWoof> DocPlatypus: I had quite a few bugs because of that
21:54:53 <JasonWoof> eg:    : stop ( addr -- ) 0.0 f! ;
21:55:00 <DocPlatypus> with the appropriate support, you should be able to use two numbers in a 2variable with */ in many cases
21:55:02 <Quartus> 0.0 is not a float
21:55:03 <JasonWoof> not: : stop ( addr -- ) 0.0 swap f! ;
21:55:10 <DocPlatypus> yeah 0.0e
21:55:14 <JasonWoof> oops
21:55:16 <JasonWoof> sorry
21:55:26 <DocPlatypus> *that* threw *me* off when I first tried it.
21:55:29 <JasonWoof> should be
21:55:34 <Quartus> 0e  is
21:56:35 <JasonWoof> right, my point was that there's no "swap" in there
21:56:40 <Quartus> Right.
21:56:47 <DocPlatypus> in this statistics program I'm writing I use floats for the average and standard deviation only because they'd be a real bitch to do with integer fractions
21:57:32 <DocPlatypus> one can always, say, 100e f* f>d d>s if they want to
21:57:33 <JasonWoof> the most irritating problem we have with floats is something I never thought would happen
21:57:39 <DocPlatypus> and get the result to two places
21:58:07 <JasonWoof> and it doesn't happen at all on PPC
21:58:14 <DocPlatypus> that would be...?
21:58:20 <JasonWoof> we check if a number is less that 600, and it is.
21:58:27 <JasonWoof> but later when we go to use it, it's 600
21:58:41 <JasonWoof> ie the rounding is inconsistant, even with the same value
21:59:03 <DocPlatypus> oh that's nasty
21:59:31 <JasonWoof> yeah, so every 5-20 minutes vor would draw a pixel off the bottom of the screen
21:59:37 <JasonWoof> (well, at our crappy frame rates)
21:59:47 <Quartus> is that something to do with 80-bit internal storage of a value, vs. 64-bit external?
22:00:18 <JasonWoof> happens pretty dern quick for these guys with new computers who get 15X higher frame rate than me
22:00:43 <JasonWoof> Quartus: probably. though I have no idea how such truncation would make a positive number bigger
22:01:01 <DocPlatypus> drawing off the bottom = crash?
22:01:03 <DocPlatypus> or is it just annoying?
22:01:24 <JasonWoof> drawing of the bottom crashes
22:01:30 <JasonWoof> which is a very good thing
22:01:33 <DocPlatypus> oy
22:01:52 <JasonWoof> because then we can find our bugs
22:02:07 <JasonWoof> it's hard to find a bug when you don't know it's there
22:02:07 <DocPlatypus> advice from me: convert to integer before drawing
22:02:13 <JasonWoof> I do
22:02:23 <JasonWoof> the trouble is it doesn't always convert to the same integer
22:02:40 <JasonWoof> I check that it's within the range 0..599
22:02:49 <JasonWoof> and delete/wrap anything that's not
22:02:56 <DocPlatypus> okay do the compare after converting to integer as well, dup it first, and use that to draw
22:03:10 <JasonWoof> then when I go to draw it, I sometimes get one that rounds to 600 (which before rounded to 599)
22:03:23 <JasonWoof> nah
22:03:46 <JasonWoof> easier to constrain the values to 0.0001 to 599.999
22:03:48 <DocPlatypus> or, blow off one row on each border
22:03:58 <DocPlatypus> and only draw 1-799, 1-599
22:04:13 <DocPlatypus> as much as I don't like that one...
22:04:13 <JasonWoof> no need to kill a whole row
22:04:51 <JasonWoof> can leave a margin much smaller than one pixel and not run a muck of x86's stupid rounding crap
22:06:44 <DocPlatypus> I've noticed that the "ms" word has way more variance than I would really want
22:06:52 <DocPlatypus> even at relatively large values
22:06:56 <Quartus> In Gforth?
22:07:00 <DocPlatypus> yes
22:07:09 <Quartus> Do you know why that is?
22:07:26 <JasonWoof> DocPlatypus: how can you tell?
22:07:28 <DocPlatypus> probably because of task switching
22:07:50 <DocPlatypus> at lower values that would make sense
22:08:56 <DocPlatypus> 5 12000 mstest  ok \ runs "5 ms" 12000 times, feeds the results into an array of cells
22:09:15 <DocPlatypus> statavgstdev f. f. 1015.17315419451 8113.74833333333  ok \ stdev, average
22:09:45 <DocPlatypus> median is 7996 usec
22:10:39 <JasonWoof> what's the max?
22:12:15 <Quartus> and what are you using for a clock? 
22:13:15 --- quit: DocPlatypus (Remote closed the connection)
22:13:40 --- join: DocPlatypus (i=skquinn@gateway/tor/x-528183d5774423a8) joined #forth
22:13:41 <JasonWoof> hmmm... speaking in person does have its advantages
22:13:58 <DocPlatypus> grr that stunk
22:14:07 <Quartus> what are you using for a clock?
22:14:08 <DocPlatypus> what part of that actually came through
22:14:19 <Quartus> [01:09] <DocPlatypus> median is 7996 usec
22:14:32 <DocPlatypus> : mstest { q r -- } r 0 do utime q ms utime 2swap d- d>s >stat loop ;
22:14:36 <DocPlatypus> >stat feeds into my statistics tabulator
22:14:54 <Quartus> The last 3 digits of utime are all zeroes in my version.
22:15:17 <DocPlatypus> they aren't in mine
22:16:06 <Quartus> I'd be questioning the accuracy of the time source as well.
22:17:04 <Quartus> Correction: they're not always zero in mine (here under WinXP at the moment), but they're always the same.
22:17:26 <DocPlatypus> see, I'm running a GNU variant
22:17:36 <Quartus> So this utime has effectively a millisecond resolution.  I don't know what yours has.
22:17:39 <DocPlatypus> I wouldn't put it past Microsoft to not give you decent timing
22:17:59 <DocPlatypus> it may or may not be microsecond accurate, but it should be within at least 1 ms of the target
22:18:17 <Quartus> Decent timing is available under Microsoft, so there's no reason to launch into a rant.  Gforth is not a barometer for the quality of the OS.  It uses C library calls for most of what it does.
22:18:21 <DocPlatypus> I'm seeing things like:
22:18:30 <DocPlatypus> stathighest statlowest . . 4115 20557  ok
22:19:41 <DocPlatypus> to be honest... at 100 ms it's a bit more reliable
22:19:46 <Quartus> Again, verify your timesource is giving you accurate info. 
22:20:06 <DocPlatypus> my stopwatch only does 1/100 sec :-)
22:20:41 <DocPlatypus> statavgstdev stathighest statlowest f. f. . . 188.79870126225 99979.52 98358 100263  ok \ for a 100 x 100 ms run
22:20:50 <Quartus> I was thinking more of finding out what call Gforth makes for utime, and seeing what your system docs say about its reliability
22:21:03 <DocPlatypus> now *that* is a bit more reasonable
22:21:49 <DocPlatypus> I first found this when doing '1000 ms' in a loop that made a full screen clock, and wondering why it skipped seconds
22:22:56 <Quartus> What's a 'GNU variant'?  Linux?
22:23:11 <DocPlatypus> in this case, GNU+Linux
22:25:24 <Quartus> Undoubtedly this has some deep meaning to you that 'Linux' fails to convey on its own.
22:26:36 <DocPlatypus> it's the GNU operating system, plus the kernel, Linux. the GNU project was, and is, about freedom, more than just access to the source code for convenience (the message the "open source" campaign got out was "it's good to have source code for convenience" and unfortunately, it worked damned well)
22:27:23 <Quartus> So, completely useless information in this context.  At any rate, ms appears to call select(), which is probably why it's fuzzy.  You might try nanosleep() instead.
22:27:48 <DocPlatypus> ah-ha
22:28:27 <Quartus> Should be able to ffi out to nanosleep if you need accuracy
22:28:29 <DocPlatypus> and the thing that gets me... there's no easy way to get a simple "number of seconds since the epoch" in GNU Forth
22:29:10 <Quartus> I thought utime was offset from some epoch.
22:29:41 <DocPlatypus> it is. except utime is a double
22:29:45 <Quartus> so?
22:30:39 <JasonWoof> woh, you got less than 99ms once?
22:30:48 <JasonWoof> something isn't right there
22:31:05 <DocPlatypus> JasonWoof: yes, you saw correctly
22:31:21 <JasonWoof> that's bad
22:31:33 <JasonWoof> ms is documented in the ansi draft standard as: Wait at least u milliseconds.
22:31:54 <Quartus> Which it may be doing.
22:32:20 <JasonWoof> yep :)
22:32:31 <JasonWoof> can't tell which timer is off
22:32:36 <JasonWoof> but that's a large margin
22:32:52 <Quartus> Plus gforth is relying on select() having specific behaviour, which it may not.  It's not a clock.
22:33:19 <JasonWoof> documentation is boring ;)
22:34:19 <DocPlatypus> okay. utime does a gettimeofday and then a timeval2us
22:35:13 <DocPlatypus> I don't see where select gets involved at all
22:35:19 <Quartus> ms
22:35:29 <DocPlatypus> ah.
22:35:36 <Quartus> (void)select(0,0,0,0,&timeout);
22:35:48 <Quartus> select() docs say it'll terminate early on an umasked signal.  That's not being checked for.
22:36:02 <Quartus> I'm not sure it'd be worth checking for, but it's one variable at least.
22:36:39 <Quartus> The timeout also specifies the *maximum* interval to wait.  System activity can lengthen the interval.
22:37:12 <Quartus> So I'm not at all surprised at the results.  nanosleep would be a better choice under Linux.
22:37:27 <Quartus> select() being more portable.
22:37:47 <DocPlatypus> OpenBSD has nanosleep
22:38:02 <Quartus> By definition, select() can wait either less or more than the specified time.
22:38:15 --- join: scope (n=junk@203-97-123-216.cable.telstraclear.net) joined #forth
22:38:26 <DocPlatypus> actually... nanosleep is part of the POSIX spec
22:41:04 <DocPlatypus> anyway I am off to bed
23:43:16 --- join: ecraven (i=nex@eutyche.swe.uni-linz.ac.at) joined #forth
23:54:21 --- quit: DocPlatypus (Remote closed the connection)
23:57:40 --- join: DocPlatypus (i=skquinn@gateway/tor/x-1cfc8eef2ccf918d) joined #forth
23:59:59 --- log: ended forth/07.05.07