00:00:05 <mycroftiv> you dont have to go fast at all, if you go right up to the pole, you can just put your hand on the pole and spin around it in circles til you get dizzy
00:00:17 <pikhq> mycroftiv: The secret of eternal youth is going to a black hole.
00:00:30 <mycroftiv> pikhq: and balancing on the event horizon?
00:00:33 <ehird> pikhq: also secret of becoming thin
00:00:51 <pikhq> ehird: And also the secret of how not to be seen.
00:01:13 <mycroftiv> pikhq: not true - you are *eternally* visible to external observers if you fall into a black hole
00:01:23 <ehird> hmm invisibility, eternal youth and thinness
00:01:26 <AnMaster> <pikhq> ehird: And also the secret of how not to be seen. <-- is it just me or does that remind you of monty python for some reason?
00:01:28 <ehird> black holes could solve all our problems.
00:01:30 <ais523> not really, after a while you'll get so faint that people won't be able to see the resulting photons
00:01:31 <mycroftiv> however its true you get so red-shifted its very hard to actually see you
00:01:39 <pikhq> AnMaster: Not just you.
00:01:53 <ehird> you broke the chain of "name: no"
00:01:54 <AnMaster> I can't think of what the sketch was though
00:01:59 <mycroftiv> ^what ais523 said, he understands what i meant
00:02:01 <ehird> even pikhq continued it though accidentally
00:02:18 <ehird> [00:01] pikhq: mycroftiv: No.
00:02:18 <ehird> [00:01] pikhq: AnMaster: No[…]
00:02:23 <ehird> not starts with no
00:02:35 <AnMaster> ehird, true, but thought you meant the whole message
00:02:38 <pikhq> AnMaster: "How Not to be Seen" is the name of the sketch.
00:02:44 <ehird> then pikhq would have broken it with "No."
00:02:46 <ehird> since that has a dot
00:04:21 <AnMaster> ehird, but didn't notice the . first time
00:05:13 <ehird> mycroftiv: have you seen the absurd LoseThos?
00:05:19 <ehird> http://www.losethos.com/
00:05:27 <mycroftiv> ehird: of course, i actually tried to download and run it awhile ago even
00:05:27 <ehird> he has 680x480 16 colour graphics... yet 12GB of RAM
00:05:32 <ehird> i can't ramble about it
00:05:35 <ehird> i looove rambling about it and him
00:05:44 <ehird> especially his markov chain-esque godspeak
00:05:44 <mycroftiv> ive even seen him in forum debates
00:05:53 <ehird> http://www.loper-os.org/?p=46#comment-819
00:05:55 <ehird> "Your writing sounds like a rant of a person more crazy than I am."
00:06:57 <ehird> http://www.reddit.com/r/programming/comments/97qbw/12_men_went_to_the_moon_using_an_understandable/c0br1sq?context=1
00:07:00 <ehird> It's got God for on-line support. (random words or passages on a plug-in hot key.)
00:07:11 <ehird> he is seriously advocating using his babble program to talk to god for help using the OS
00:07:11 -!- FireFly has quit ("Later").
00:07:13 -!- BeholdMyGlory has quit (Remote closed the connection).
00:07:31 <mycroftiv> im crazy enough to appreciate the genius of that idea, but not crazy enough to think it would work
00:08:01 <mycroftiv> however the idea that sampling the 'random numbers supplied by the universe' at a given point is a way of finding hidden mechanism is something that 'most people' seem to actually believe
00:08:21 <ehird> what's that meant to mean
00:08:43 <mycroftiv> ehird: thats what astrology, tarot, i ching, etc, all have as their idea - you get some random numbers, interpret them by rule, the universe hides meaning in them
00:09:05 <ehird> mycroftiv: i'm fairly sure adherents don't consider them random per se, even with hidden meaning
00:09:08 <mycroftiv> i dont believe in it myself, but if you press people on it, they think there is 'some truth' to those things, that they 'can work sometimes, for some people'
00:09:11 <ehird> that is, the meaning isn't hidden in them, they ARE the mening
00:09:24 <ehird> also, it's hopeless talking to such people
00:09:35 <mycroftiv> such people fill our world, we dont have much choice but to talk to them
00:09:42 <ehird> their reality is always full of hopeless contradiction and rampant subjectivism of absolutely everything; solipsist-level
00:10:01 <ehird> mycroftiv: but that's what the internet is for! :P
00:10:25 <AnMaster> what a shock everyone would get if it turned out to be true! (not that it is likely to ever turn out that way)
00:10:27 <mycroftiv> well, subjectivist philosophy is hopefully a bit more respectable than solipsism...but its true that the average person certainly retreats to patently solipsistic style reasoning if you try to engage them on the topic
00:10:59 <mycroftiv> a statement like "everyone has their own truth and words dont really mean anything" is where trying to establish a consistent framework of definitions for people's own statements usually gets to, and quickly
00:11:08 <ehird> nobody would get a shock
00:11:27 <mycroftiv> just asking people to provide their own freely chosen definitions for the terms they use in their statements will make them very angry if you press the point, generally
00:11:27 <ehird> i don't think you understand what solipsism means.
00:11:37 <ehird> AnMaster: if solipsism is correct, you don't exist.
00:11:40 <ehird> you are not conscious.
00:11:40 <AnMaster> ehird, I think you are referring to something else that I did
00:11:43 <ehird> my brain made you up.
00:11:48 <AnMaster> ehird, I meant i ching and such
00:11:50 <ehird> in fact, everyone but me
00:11:59 <ehird> AnMaster: your referent was very vague and delayed
00:12:04 <pikhq> My brain made me up.
00:12:14 <pikhq> I am a figment of my own imagination.
00:12:38 <ehird> i'm a figment of my toes
00:13:05 <pikhq> Your toes are a figment of your mom.
00:14:16 <mycroftiv> shhhh, you guys are talking to loud - you guys are gonna wake the colorless green ideas, they are busy sleeping furiously
00:15:04 <ehird> mycroftiv: what vm host do you use for plan 9 btw?
00:15:11 <ehird> it's always been slow when I've tried it
00:15:22 <mycroftiv> both qemu and vmware server free beer
00:15:34 <ehird> qemu was the one that was dog slow
00:15:42 <ehird> also graphics drivers
00:15:43 <mycroftiv> i bet you were using qemu to do the graphics though
00:15:46 <ehird> yeah, a vm with driver problems
00:15:49 <ehird> mycroftiv: erm yes
00:16:07 <mycroftiv> i never have qemu do the graphics, i run my VMs as headless CPU servers and always drawterm in or just import from them
00:16:16 <ehird> that doesn't give me rio.
00:16:36 <ehird> but my middle mouse button doesn't work!
00:16:48 <mycroftiv> did you not set the mousetype correctly?
00:16:48 <ehird> i mean in general :P
00:17:15 <mycroftiv> my middle mouse button does scrolling in plan9, as well as the 'button 2 click' functionality that is the original mouse vocabulary
00:17:24 <ehird> i don't think you understand mycroftiv
00:17:27 <ehird> the middle mouse button on my mouse
00:17:49 * ehird notes: to talk to mycroftiv, repeat line a lot
00:18:27 <mycroftiv> um, well, usually when someone says that 'X doesnt work' in a computer context, its at the software level - because most hardware that is broken gets unplugged and different component substituted
00:18:42 <ehird> all my other usb mice are unusable though
00:18:49 <ehird> either they don't have a scrollwheel or they're not usb mice.
00:18:58 <ehird> apart from my mighty mouse
00:19:10 <ehird> but in its touch-sensitive glory, the right click has become temperamental
00:19:13 <mycroftiv> if someone is installing ubuntu and says "my monitor isnt working" im gonna assume they need help with an X server problem, not a coupon to wal-mart
00:19:16 <ehird> and after enough time, just stops working
00:19:27 <ehird> also, this mouse feels nice.
00:20:43 <mycroftiv> that reminds me, indirectly, i think im gonna find (since someone probably already did it) or write the patch to make the horrible scrollbars in plan9 windows behave conventionally
00:20:58 <ehird> i've never seen a scrollbar in plan 9
00:21:17 <ehird> i never use scrollbars though
00:21:42 <ehird> hmm rio's interface could be very interesting with one of those fancy multitouch trackpads on the macbook pros
00:21:53 <ehird> you could basically eliminate the window menu
00:21:55 <mycroftiv> yeah actually rio could very easily be updated to a multitouch moel and be quite nice
00:22:22 <ehird> problem being the drivers of course
00:22:44 <mycroftiv> hmm that reminds me, when we are talking about fundamental issues...
00:23:03 <ehird> "how do you do drivers?"
00:23:10 <mycroftiv> the fact that despite them being open-source, drivers are hard to reuse without extensive modification, is so frustrating
00:23:36 <ehird> damn i thought i was gonna be able to rant about my os
00:23:42 <ehird> but linux drivers suck.
00:24:45 <mycroftiv> the fact that the linux kernel is so competent now at handling hardware but those free software drivers havent resulted in *every* free os being equally capable is just so damn sad, but its a symptom of the deeply ad-hoc nature of how everything is engineered
00:24:55 <ehird> to hell with kernsl
00:26:04 <ehird> (that was indeed an awkward attempt to segue into talking about my OS!)
00:27:21 <ehird> (instead it killed the chat)
00:27:29 <mycroftiv> uh, were waiting for you to start talking
00:27:38 <mycroftiv> with your brilliantly prepared segue having established the context
00:27:44 <ehird> awwwwwwwkwaaaaaaaar
00:30:11 <ehird> What state are we in
00:30:12 -!- Judofyr has quit (Remote closed the connection).
00:30:27 -!- Judofyr has joined.
00:30:33 <mycroftiv> im in state 'move right on tape until encountering a / '
00:30:37 <ehird> are we still waiting for me
00:31:51 <mycroftiv> well this is a multitasking preemptive IRC so we arent strictly 'waiting', if some other event occurs we can context switch to a new process
00:35:05 * mycroftiv slaps ehird with a large trout until he says whatever the hell he was gonna say
00:35:12 <ehird> but i don't know what context we're in!
00:35:20 <mycroftiv> much *more* awkward to get trout-slapped than just say something, isnt it?
00:36:59 <ehird> we are not amused.
00:39:21 * mycroftiv removes the / from the tape and begins to travel left until encountering a blank space
00:40:12 <ehird> i was going to talk about my os wasn't i
00:40:18 <ehird> hands up if you want to know how hardware/drivers work in my os.
00:41:03 <mycroftiv> if we raise our hands, how can we type on our keyboards to let you know our hands are raised?
00:42:43 <ais523> you can tell because people aren't typing
00:43:49 <ehird> mycroftiv: ok you don't have to raise your hand
00:43:51 <ehird> you can just type it
00:43:57 -!- Azstal has joined.
00:44:25 <mycroftiv> i hope the user interface for getting information out of your os is easier than the user interface you present in IRC for us getting the information from you
00:45:33 <Azstal> you could raise your keyboard with your hand
00:45:42 <ehird> mycroftiv: should i just start blabbing
00:45:50 <mycroftiv> 19 minutes and counting currently on latency between ehird signalling the channel he had information to communicate and the delivery of said information
00:46:02 <mycroftiv> SO TELL US ABOUT THE DRIVER MODEL ALREADY
00:46:19 <ehird> wow this is going to be so anticlimatic
00:46:22 <ehird> should i even bother. i wonder.
00:46:45 <ehird> (mycroftiv tears out my soul.)
00:46:46 <mycroftiv> up to you, do you enjoy the troutslapping?
00:46:47 -!- GregorR-L has joined.
00:46:58 <ehird> GregorR-L enjoys troutslapping.
00:47:15 -!- ais523 has quit (Remote closed the connection).
00:47:16 <mycroftiv> i doubt he would tell me about how the drivers work in your os though.
00:47:29 <mycroftiv> so i dont see how it would maximize my utility
00:48:08 <ehird> drivers are voluntary
00:48:20 <ehird> any object that can get another object with which to communicate directly to the hardware
00:48:26 <mycroftiv> if my hard drive driver doesnt volunteer, i cant make it work?
00:48:28 <ehird> can offer its services to any other object allowed to access it
00:48:34 <ehird> mycroftiv: shush you
00:48:42 <ehird> with these decentralised driver objects
00:48:50 <ehird> we can define eg interfaces
00:48:54 <ehird> and supply relevant driver objects
00:49:04 <ehird> and using some hardware is just getting the relevant object
00:49:16 <ehird> a driver is just something that translates sugary messages into communication with the low-level object
00:49:23 <mycroftiv> sounds like the correct model for your os, indeed
00:49:25 <ehird> nothing special, nothing in the "kernel" (there is no kernel)
00:51:18 <ehird> it pleases me when stuff just fits in my model without any new "kernel" code or whatever
00:51:22 <ehird> reaffirms that it's a good model
00:53:06 <ehird> mycroftiv: you know, if there wasn't the issue of Other People's Things being unreliable, and slowness of the internet, I'd probably encrypt every object and distribute the storage across every other machine
00:53:17 <ehird> alas those are bohh false
00:53:26 <ehird> every other machine running that OS, that is
00:53:47 <ehird> but drives disappear, network nodes disappear, and the internet is slow.
00:53:52 <mycroftiv> there are a couple projects in existence that do that - i forget what the main one is called
00:54:07 <mycroftiv> but i dont think making that your *default* is at all sane - as you correctly have stated
00:54:23 <ehird> sometimes I swear that people have never heard of http://en.wikipedia.org/wiki/Fallacies_of_Distributed_Computing
00:55:08 <mycroftiv> and wow is it maddening to design around all those issues...
00:55:43 <ehird> sometimes i think the notion of a computer is flawed.
00:55:45 <mycroftiv> my first big plan9 project was kind of a distributed DNS-for-9p that ties into the inferno registry, and 99.9% of my development time on it was working out how to handle every possible failure mode
00:56:11 * ehird tries to find a piece of vm software well-suited to playing with plan9
00:56:20 <ehird> vmware fusion/parallels aren't suitable, they're way too windows-centric...
00:56:39 <mycroftiv> im not sure what the best solution is for that
00:56:47 <ehird> well qemu and stuff work but
00:56:52 <ehird> mycroftiv: yeah, it's not reaaaaaaal plan 9 :P
00:57:01 <mycroftiv> i believe there is a pretty decent os x version of 9vx - and yes it is, especially if you use a full tree
00:57:15 -!- Asztal has quit (Read error: 110 (Connection timed out)).
00:57:18 <mycroftiv> you can get the regular plan9 .iso and copy it over to your filesystem and run 9vx from it
00:57:27 <mycroftiv> the 9vx binary distribution is sadly crap, i agree
00:57:33 <ehird> but but but but but but but but
00:57:37 <ehird> it isn't exactly the same!
00:57:44 <ehird> it uses the host fs doesn't i
00:57:52 <mycroftiv> right, although thats kind of an advantage really
00:57:59 <mycroftiv> gives you natural integration between the systems
00:58:04 <ehird> it is my escape from resource forks :-P
00:58:28 <mycroftiv> wait, resource forks still exist in the mac os? are you joking? or am i confused?
00:58:52 <mycroftiv> wow, i thought os x got rid of those...
00:59:08 <pikhq> No, it's somewhat NeXTish as well.
00:59:11 <mycroftiv> how can os x be a UNIX then? sorry, i should know this, i dont keep up on os x like i should
00:59:30 <ehird> here, we use = and + to mean "bastard of x, y and z"
00:59:44 <mycroftiv> ironically, i have a white plastic imac but i run gnu/linux on it, havent booted it to os x in forever and i never really learned what was up with os x before turning it into primarily leenooks box
01:00:33 <ehird> Mac OS X = Darwin (derivative of XNU) + NEXTSTEP (XNU (BSD + Mach) + own stuff) + FreeBSD + Mac OS + own stuff
01:00:47 -!- Sgeo has joined.
01:00:52 <mycroftiv> sure i know that, conceptually - i know the hsitory, carbon, cocoa, blah blah, mach etc
01:00:55 <ehird> well the Mac OS part is basically the fs
01:01:00 <ehird> it uses the mac os fs with changes to make it unix-compatible
01:01:03 <ehird> thus resource forks are retained
01:01:06 <Sgeo> Nightmare website: http://farmingdale.edu/lieoc
01:01:14 <pikhq> ehird: Throw in some GNU for good measure.
01:02:10 <pikhq> ... Fine, mostly just the GNU C compiler.
01:02:15 <pikhq> Which is almost everywhere.
01:02:19 -!- oklopol has quit (Read error: 110 (Connection timed out)).
01:02:19 <ehird> mycroftiv: by default, yes
01:02:24 <pikhq> mycroftiv: Which is almost everywhere.
01:02:30 <ehird> pikhq: they're working on replacing gcc with llvm/clang
01:02:39 <ehird> i know this because i have used os x 10.2
01:02:48 <pikhq> That shit was C shell.
01:02:50 <ehird> it used to be tcsh
01:03:14 -!- Judofyr has quit (Remote closed the connection).
01:05:32 <ehird> mycroftiv: is there a way to get 9vx to use a big file as a block device?
01:05:35 <ehird> if so, then I'll consider using it
01:06:19 <mycroftiv> you can either just run fossil/venti (if thats the idea) from flat files, or you could use devfs to make virtual devices from them
01:06:27 <ehird> yeah, I wanted to use fossil
01:06:40 <ehird> basically I just want a plan 9 system whose kernel doesn't get emulated.userspace
01:06:56 <mycroftiv> you can make a fossil, problem is that without patching that fossil wont be your *boot* fossil which i gather is also what you want?
01:07:42 <mycroftiv> right, thats possible, but it starts to get into the realm of patching-and-hacking
01:07:52 <ehird> i'll just use a fast vm + drawterm i guess
01:08:22 -!- Pthing has quit (Remote closed the connection).
01:08:35 <mycroftiv> since you are using os x, i unfortunately cant really push my own toolchain on you, since i distribute it in linux-centric form
01:09:15 <mycroftiv> otherwise id be saying you should download my gridtools stuff since its all set up as a true modular multinode plan9 system already
01:11:01 <mycroftiv> the thing about plan9 thats frustrating is that the base bell labs .iso distribution is so far removed from a fully configured setup that most people never get to the point of experiencing the whole purpose of the os, since there is a ton of not-very-graceful admin stuff required to make the 'magic' happen
01:13:25 <mycroftiv> ehird: thats why your ideas i think are fundamentally sound - to express things crudely, you want to 'cut out the middleman' of all the annoying implementation/administrative details of running a computer, by making the upper and lower layers more conceptually unified.
01:13:53 <mycroftiv> of course, as a basic concept that isnt new and in some ways that is what everyone (and every failed and awkward system) was trying to achieve :D
01:14:48 <mycroftiv> thats not criticism, just an acknowledgment of the difficulties of the task and a recognition that smart people have been trying to make sane and sensible systems for a long time
01:17:20 <AnMaster> <Sgeo> Nightmare website: http://farmingdale.edu/lieoc <-- doesn't load Sgeo
01:20:47 <ehird> mycroftiv: i do want to cut out the middleman, I think yours is incidental though
01:20:51 <ehird> my main driver is:
01:21:06 <ehird> one, we have so much useless work
01:21:11 <ehird> we separate tasks too much
01:21:16 <ehird> we separate application from application
01:21:20 <ehird> we separate objects from disk
01:21:22 <Sgeo> AnMaster, the page apparently is a 302 status code, which goes to an HTML page that uses a meta tag to go to the proper place
01:21:27 <Sgeo> AnMaster, check the source?
01:21:29 <ehird> mycroftiv: and the second,
01:21:37 <ehird> we have too much duplication
01:21:41 <ehird> so many garbage collectors
01:21:44 <AnMaster> Sgeo, can't. get a "connection reset during loading" message
01:21:44 <ehird> so many address books
01:21:49 <ehird> so many everythings
01:21:56 <AnMaster> Sgeo, so don't even get past HTTP headers I guess
01:22:01 <Sgeo> Does http://www.farmingdale.edu/campuspages/campusaffiliates/lieoc/index.html work?
01:22:01 <mycroftiv> well, all of those things you are saying sound to me like many of them fit conceptually into the broad idea of 'cut out the middleman'
01:22:03 <ehird> as a result of fixing those two everything fits together
01:22:18 <ehird> also, link to your toolchain thing anyway?
01:22:44 <AnMaster> Sgeo, I get headers but no document
01:23:07 <AnMaster> Sgeo, note: cookies and javascript are *off*
01:23:09 <ehird> mycroftiv: even if i can't use it i'm interested
01:23:15 <Sgeo> AnMaster, that may well be it
01:23:18 <mycroftiv> ehird: 9gridchan.org is my website with tons of plan9 related stuff, a full explanation and context is maybe offtopic for this channel but feel free to ask whatever
01:23:26 <AnMaster> Sgeo, not going to turn them on
01:23:36 <ehird> 9gridchan, is that like a 4chan derivative on a grid?
01:23:47 <mycroftiv> ehird: uh, not really, but sort of, yeah
01:24:12 <ehird> how on earth is it sort of :P
01:24:27 <Sgeo> The css for this page: http://pastie.org/576211
01:24:31 <mycroftiv> well, its a grid that anyone can connect to so that is chan-like
01:24:49 <ehird> i was referring more to things like 4chan → 7chan and the like
01:24:52 <mycroftiv> i think the chan imageboards are some of the better stuff on the net conceptually, free speech, simple interface, no barrier to entry, etc
01:24:54 <ehird> distributed pointlessness!
01:25:03 <ehird> mycroftiv: yeah i know all the arguments in favour of them
01:25:41 <AnMaster> mycroftiv, you think /b/ is "[one] of the better stuff on the net conceptually"?
01:25:49 <ehird> AnMaster: ad hominem, strawman
01:25:59 <ehird> (more strawman than ad hominem)
01:26:07 <mycroftiv> AnMaster: well, if you read my actual text above, that isnt what i said, so i think the statement i made was pretty clear
01:26:13 <ehird> the only flaw with your jokes AnMaster
01:26:19 <ehird> is that none of them are ever funny or carry any hint of being jokes.
01:27:09 <AnMaster> <Sgeo> The css for this page: http://pastie.org/576211 <-- and?
01:27:39 <Sgeo> Did you look at it? .style7 ?
01:28:19 * Sgeo was facepalming at the bad names
01:28:21 <AnMaster> mycroftiv, on your website... what is bind -a '#¤' /dev
01:28:27 <AnMaster> that looks like encoding error to me
01:29:07 <mycroftiv> yup, that does indeed look like an error, lets check the original ns to see what it should be
01:29:19 <AnMaster> mycroftiv, and what does the #b and such mean?
01:29:37 <mycroftiv> those are how plan9 talks about binding device drivers into the namespace, pretty much
01:29:57 <ehird> mycroftiv: isn't /srv mostly used?
01:29:59 <ehird> at least that's what i saw
01:30:01 <mycroftiv> ah ok, yup, thats a weird unicode symbol that clearly has a different visual appearance
01:30:22 <AnMaster> mycroftiv, make web server send correct encoding :)
01:30:26 <mycroftiv> ehird: well /srv is where the user level file servers usually post themselves for stuff to mount, but the kernel level drivers work a bit differently i would say
01:30:47 <ehird> mycroftiv: so this drawterm stuff, what do i have to do to get it working?
01:31:06 <mycroftiv> ehird: from what base are you starting? default bell labs .iso installed into a vm, or ?
01:31:25 <ehird> well, I have the iso and nothing else
01:31:37 <ehird> the less friction this takes to get working the better
01:31:51 <AnMaster> ehird, this is plan9... what do you think
01:32:00 <ehird> AnMaster: okay, fuck off, seriously
01:32:10 <ehird> for the last hours all you've done is dis plan 9 and other crap
01:32:11 <AnMaster> as in: it isn't smoothly polished ubuntu.
01:32:23 <mycroftiv> hm, how much plan9 related spam tech/support do you want in this channel?
01:32:29 <ehird> you like it in the most vaguest, tenuous sense possible
01:32:33 <ehird> just enough so you can hate on it all the time
01:32:39 <ehird> and i never said i wanted ubuntu; strawman fallacy
01:32:45 <AnMaster> I just find it amusing that you think it will work out of box more or less
01:32:48 <ehird> i said i wanted the least frictionful way possible
01:32:55 <ehird> AnMaster: i've used plan 9. it worked out of the box
01:32:58 <ehird> and did i ever say that?
01:33:04 <ehird> i never said i wanted it working out of the box
01:33:14 <ehird> you are truly excelling putting words in my mouth
01:33:18 <ehird> new heights of strawman
01:33:53 <mycroftiv> AnMaster: ironically, i actually provide preconfigured preinstalled plan9 systems that instantiate a full 4 functional node grid on your desktop out of the box, but i only provide that for linux ;)
01:34:05 <ehird> mycroftiv: impossible, that doesn't fit his biases
01:34:20 <AnMaster> mycroftiv, awesome and now I really need to sleep *turns off monitor*
01:34:58 <ehird> inferno has always confused me
01:35:02 <ehird> why do you want an OS that can only run virtualised?
01:35:20 <mycroftiv> actually inferno can also run natively, but that isnt very common except on small devices
01:35:53 <mycroftiv> i dunno, you dont like sqweak right? well its not too different from that - or from java in a different way, i mean java is real world popular.
01:36:10 <ehird> java just has a vm underneath
01:36:14 <ehird> that's basically an implementation detail
01:36:16 <mycroftiv> but besides, whats an os, whats software? i think its a false dichotomy
01:36:26 <mycroftiv> well, think of inferno as the vm implementation for the limbo language
01:36:35 <ehird> i'm not making an os, i'm making a system
01:36:45 <ehird> it boots up and comes with a set of base objects "absolutely free"
01:37:00 <ehird> mycroftiv: i've seen limbo.
01:37:02 <mycroftiv> thats good, im tired of paying $5 per boot
01:37:06 <ehird> looks like a pretty boring lang tbh
01:37:43 <mycroftiv> i havent learned it, but a lot of people like it for its concurrency/message passing features i guess - and by 'a lot of people' i mean 'a handful of people with plan9/inferno interests'
01:38:47 <mycroftiv> limbo is actually maybe the most 'on topic' thing from the plan9 universe for this channel - well, that and alef of course
01:39:07 <ehird> mycroftiv: how smoothly does plan9 run with qemu/drawterm?
01:39:16 <mycroftiv> very nicely in my experience, even on moderate hardware
01:39:39 <GregorR-L> I'm looking for more cellular automaton to make into ties. There are only two good ECAs, but I can extend that to size-5 neighborhoods and then the range is almost endless.
01:39:39 <ehird> qemu is basically the slowest emulator apart from bochs
01:39:39 <mycroftiv> much better than with qemu providing the graphics
01:39:39 <ehird> you aren't using the kqemu thing are you?
01:39:57 <ehird> GregorR-L: make a cellular automata whose atoms are cellular automata
01:40:01 <mycroftiv> i use kqemu on some boxes but not all, and even without it, i get performance that i find acceptable
01:40:05 <ehird> and use a CA to generate CAs
01:40:15 <mycroftiv> but for all i know qemu on os x is even slower than slow, i dunno
01:40:15 <ehird> mycroftiv: ok. so how easy is drawterm to set up?
01:40:34 <mycroftiv> ehird: well, let me give you the canonical 'how to' link for how to do it from the default install from the .iso, one second
01:41:04 <mycroftiv> ehird: http://www.plan9.bell-labs.com/wiki/plan9/Configuring_a_standalone_CPU_server/index.html
01:41:23 <mycroftiv> that is Ye Olde Testamente of how to take your default install and make it a standalone cpu auth server you can drawterm into
01:41:23 <ehird> i love how they call it a wiki, it's been read only for like 8945349534 years
01:41:36 <ehird> anyway looks irritatingly complicated.
01:41:37 <mycroftiv> its not read only as a matter of fact
01:41:46 <mycroftiv> and yes, that is exactly why i made the tools i did, because its ridiculous
01:41:47 <ehird> mycroftiv: i mounted it and couldn't save.
01:42:02 <ehird> this was a few months ago
01:42:42 -!- Zonbi has joined.
01:42:55 -!- Zonbi has left (?).
01:44:27 <ehird> mycroftiv: so, any particular recommendation or should i just set up qemu
01:45:05 <mycroftiv> well, im hesitant to make any os x recommendations, honestly - i get the sense from reading 9fans that os x is a bit different in what the optimal strategies are
01:45:45 <mycroftiv> ive found that qemu seems to be most generally reliable plan9 virtualization platform for the oses ive worked with it in, which is various gnu/linux distros, windows xp and vista, and freebsd
01:46:19 <ehird> how big do you recommend i make the disk?
01:46:24 <mycroftiv> if you are going to use qemu, you could try the standalone version of the preinstalled image i distribute btw
01:46:33 <ehird> what does that get me over the stock?
01:46:43 <mycroftiv> drawterm out of the box, additional configuration work done
01:46:54 <ehird> anything to emulate a middle mouse button? :P
01:47:14 <mycroftiv> um, doesnt plan9 have something to do that anyway? some key combo or something?i should know this
01:47:23 <mycroftiv> http://sphericalharmony.com/plan9/ventigridserver.qcow2.img.tgz
01:47:29 <ehird> isn't qcow compressed?
01:47:42 <mycroftiv> that is a .tgz of a single file which is a qcow2 preinstalled hdd image
01:47:56 <ehird> a tar of a single file?
01:47:57 <mycroftiv> is the meaning of that question "why bother to .tgz it?"
01:48:09 <mycroftiv> ehird: no, im not nuts at all! :) want the method for the madness?
01:48:13 <ehird> also, no, it's just "isn't qcow = compressed = slow?"
01:48:32 <mycroftiv> no, the way qemu handles qcow2 is pretty efficient
01:48:34 <pikhq> Qcow is copy on write, not compressed.
01:48:51 <pikhq> And a few other silly things to maximise sparseness.
01:48:52 <ehird> Q.app tells me it's compressed
01:48:52 <mycroftiv> ok, its very simply this - most people want/need to preserve the archival initial copy of the VM
01:49:13 <mycroftiv> if you provide a .tgz, the standard command line tar xzf foo.tgz leaves the original .tgz behind unchanged
01:49:30 <ehird> mycroftiv: don't plan 9 guys oppose hacks
01:49:48 <mycroftiv> this is very convenient and useful, in comparison to a .gz where the ungzip will annihilate the original, and then youll fuck up your vm since you dont understand plan9 and have to redownload...
01:49:51 <ehird> mycroftiv: btw does this image use venti
01:49:58 <mycroftiv> ehird: no, not in that standalone version
01:51:20 <mycroftiv> the file is badly named, its called 'ventigridserver' because its actually the version of the image that can be used to *host* a venti for other nodes to use
01:51:31 <ehird> should it be .qcow2.img or just .qcow2
01:51:32 <mycroftiv> but it doesnt do that by default, or use venti itself as its backing store
01:52:31 <ehird> how much ram should i allocate? I have 2.5gb
01:52:42 <mycroftiv> for drawterm you also need to deal with the port redirections needed to access the vm, im not sure of the details of that in os x - 256mb for the vm is plenty of ram, plenty
01:52:53 <ehird> also, do I really need to redirect ports for localhost?
01:53:06 <ehird> finally, will anything break if I boot this with graphics?
01:53:18 <mycroftiv> the vm has to be able to listen on those ports even for localhost - and no, you can boot the image fine as standard graphical vm
01:53:36 <mycroftiv> it includes the standard initial terminal/glenda setup from bell labs as an option
01:53:58 <ehird> rio started automatically for me when using the iso
01:54:07 <ehird> mycroftiv: acme won't run under the console
01:54:28 <ehird> do i want glenda or gridna?
01:54:48 <ehird> eenie meenie miney mo
01:55:04 <ehird> mycroftiv: btw, what's the done plan9 thing for users?
01:55:10 <ehird> whee, alggy indowing system
01:55:32 <ehird> when using plan 9 it was a pain to get a user running with the ability to administrate
01:55:56 <mycroftiv> bootes is set up on that image though as the admin/root
01:56:05 <mycroftiv> you can add your own user to admin stuff easy though
01:56:29 <ehird> shift+right = middle
01:56:59 <ehird> ok, if I can get a user set up here and then drawterm that'd be ideal
01:57:04 <mycroftiv> btw, are we dragging #esoteric too far offtopic? if so i have a channel for my various plan9 projects on here called #plan9chan
01:57:31 <ehird> dude, we haven't been on topic for years
01:57:45 <ehird> admittedly this diversion is rather *extended*, but
01:57:47 <pikhq> Dude, we're more often off topic than on.
01:57:54 <ehird> Duuuuuuuuuuuuuuuuuuuuuuude
01:58:00 <pikhq> ehird: Actually, we've been on topic on occasion.
01:58:09 <pikhq> But. That's a rare occasion.
01:58:13 <olegfink> speaking of being off topic, does anyone here do K?
01:58:21 <ehird> an idler just spoke
01:58:24 <ehird> what do we do whatdowedo
01:58:30 <ehird> olegfink: oklopol does J
01:58:34 <olegfink> well, that should have happened someday
01:58:47 <pikhq> Okay. Not a demigod, then.
01:58:50 <mycroftiv> have to spend a few minutes on phone, brb
01:58:54 <ehird> Fun fact: more than one person is named Oleg
01:58:59 <olegfink> yes, I'm in almost no way mr. Kiselev
01:59:00 <ehird> olegfink: your new name is "inferior oleg"
01:59:43 <olegfink> again, if I've correctly identified the oleg you're talking about, I doubt he's likely to be found on irc
01:59:57 <ehird> but yeah he doesn't irc.
02:00:38 <olegfink> so yeah! that's what makes me the /other/ one, right?
02:00:46 <ehird> no, the inferior one
02:00:59 <ehird> it's okay inferior oleg
02:01:02 <ehird> you're just inferior.
02:01:40 <ehird> arguing breeds rebellion.
02:02:42 <olegfink> after all, I only do saner stuff with ocaml... actually running ocamlrun on bare hardware was about as far as I got... *bursts in tears*
02:02:52 <ehird> wow you did that? awesome.
02:03:46 <olegfink> anyway, back to k... there are many j programmers, but I just need some place to ask stupid k questions.
02:04:30 <ehird> i don't really get why someone would use k over j
02:05:13 <olegfink> ehird, well, it used(s) das u-boot for the dirty init work, and worked more or less when run in the orienting the board in the right direction.
02:05:28 <ehird> orienting the ... board?
02:06:02 <ehird> i repeat my question
02:06:37 <olegfink> yeah, it was an arm9 thing with fpga, but I never got to actually use it, though the whole project was about playing with metaocaml
02:07:51 <olegfink> re k over j, why someone would use airbuses over sea liners?
02:08:29 <pikhq> When you should be using a interstellar spaceship instead.
02:08:35 <ehird> olegfink: begging the question
02:09:09 <olegfink> j is an executable mathematical notation that is particularly efficient in number theory applications, k is just a cool general-purpose functional language
02:09:24 <ehird> the languages are hugely similar
02:09:28 <ehird> from what i've seen
02:11:05 <olegfink> i have very limited knowledge of j, but from what I know its set of primitives is much larger, the syntax is both more powerful and more complicated
02:12:43 <ehird> j has more primitives?!
02:12:46 <olegfink> I don't know how many people use j as a general-purpose language, as they use the usual ocaml, haskell or c
02:12:47 <ehird> its vocab fits on one page of 3 cols
02:13:03 <pikhq> Ah, C++: [](){}();
02:13:35 <ehird> olegfink: all of it
02:13:46 <pikhq> ehird: That's a noöp.
02:13:52 <ehird> pikhq: oh C++ not C
02:14:51 <olegfink> well, K doesn't have them, it has about 50 primitives
02:15:10 <olegfink> http://jsoftware.com/help/dictionary/vocabul.htm seems to list slightly more.
02:15:55 <ehird> i'm sure k has more actual functions than 50
02:15:58 <ehird> just maybe in libraries
02:16:07 <ehird> the typical k program isn't just going to be compositions of 50 funcs
02:16:30 <olegfink> aye, I'm not counting things like trigonometric functions (yes, K has no o.!)
02:17:33 <ehird> k seems very much more secret and corporate
02:18:05 <olegfink> not counting the fact that it is no longer marketed
02:18:17 <mycroftiv> uggggg 20 minutes of trying to explain the basics of networking to a teenage kid trying to freeload off his neighbor's wireless with a 10 year old laptop
02:18:40 <ehird> mycroftiv: why bother?
02:18:55 <mycroftiv> ehird: i was trying to avoid it, its my gf's kid
02:19:27 <mycroftiv> the fact that i dont even use windows enough to know how to access the networking control panel doesnt help, either
02:19:27 <ehird> olegfink: shit sux
02:19:27 -!- bsmntbombdood has quit (Read error: 104 (Connection reset by peer)).
02:19:37 <olegfink> it's just C with lisp semantics and apl syntax...
02:19:56 -!- bsmntbombdood has joined.
02:20:48 <olegfink> but I can't give an objective comparison as I've never written anything big enough in J
02:21:15 <ehird> j doesn't do big programs
02:21:41 -!- Azstal has quit (Read error: 110 (Connection timed out)).
02:22:30 <ehird> http://freetexthost.com/rvsaogqp2y ← woah man.
02:22:35 <olegfink> I'm taking that into account when using the term 'big'. I think my largest ever J program was 3 lines.
02:24:19 <olegfink> it was that big because I never quite got how to do simple file i/o in J the right way.
02:27:30 <ehird> olegfink: 'j for c programmers' might help and is included in the docs
02:27:37 <ehird> it has file io in the first example
02:30:23 <olegfink> iirc it didn't help me, but maybe I was just too sleepy when I was reading jfc
02:30:44 <olegfink> what I needed was basically read some numbers from a text file, output some numbers
02:30:57 <ehird> the first example has that
02:33:05 <olegfink> well, honestly, I don't remeber what my problem was, maybe it was about parsing the string for integers, but I remeber seeing that example (is http://jsoftware.com/help/jforc/continuing_to_write_in_j.htm#_Toc191734364 what you're talking about?).
02:34:06 <ehird> http://jsoftware.com/help/jforc/a_first_look_at_j_programs.htm
02:36:12 <olegfink> seems the magic I haven't had mastered at that time was a proper ReadFile ;-)
02:38:47 <olegfink> I recall some problems with CRLF confusing J or something... meh, I can't find my code.
02:39:06 <ehird> why are you using crlf :P
02:39:56 <ehird> why is it on highlight?
02:40:07 <GregorR-L> So that I can harass people who use CRLF
02:40:25 <GregorR-L> We must rid ourselves of this vile plague.
02:40:29 <olegfink> iirc CRLF in J is a list of CR and LF
02:40:36 <GregorR-L> I was kidding, I don't actually have CRLF on highlight :P
02:40:46 <ehird> what about classic mac os and its CR
02:41:41 <ehird> tbh in 1983 it wouldn't have been as crazy
02:43:00 <ehird> mycroftiv: i had something queued up to say but i forgot.
02:43:40 <ehird> mycroftiv: incidentally does plan 9 have any text antialiasing mechanisms?
02:43:51 <mycroftiv> ehird: you can download some different subpixel hinted fonts if you want
02:44:00 <ehird> it's... encoded in the font?
02:44:52 <mycroftiv> i honestly dont know the details, but i know if you want smoothed fonts, theres some stuff you can download - im oblivious to fonts having grown up on 40 column all caps displays and thinking that all modern displays post 2000 look fine
02:45:26 * ehird whacks mycroftiv with the typography/display nerd bat
02:45:31 <ehird> It's a bat that you whack people with.
02:46:23 <mycroftiv> i do know that the plan9 fonts that many disdain are in fact also NONFREE for some bizarre reason and that one traditional annoyance of various things is that the fonts arent freely redistributable separate from plan 9
02:47:25 <pikhq> mycroftiv: Learn typography!
02:47:36 <ehird> pikhq: Shut up, X11 user.
02:47:41 <pikhq> Also. What the crap? Bitmap fonts?
02:47:48 <pikhq> ehird: I at least freely admit it's shitty.
02:47:48 <ehird> If you really loved typography, you'd sacrifice every other value for it.
02:47:55 <ehird> Well, freetype, not X11.
02:48:28 <pikhq> ehird: Idea: Display TeX.
02:48:37 <olegfink> ehird: bitmap fonts don't need any antialiasing.. just get a higher resolution display. :-)
02:48:54 <ehird> pikhq: TeX's actual font rendering isn't that good
02:49:01 <ehird> olegfink: (a) that makes them incy wincy
02:49:11 <ehird> (b) i'd spend $$$ for a 600dpi display, hells yeah
02:49:22 <pikhq> ehird: Yeah, but that's the only thing it doesn't do all that well.
02:49:42 <ehird> also, implementing smart quotes by making `` and '' freakin' ligatures
02:49:44 <pikhq> Oh, right. Straight TeX doesn't do that.
02:49:52 <olegfink> I heard TeX doesn't really make coffee as well.
02:51:05 <olegfink> anyway, leaving for the weekend. thanks for the time.
03:00:32 -!- ehird has quit.
03:14:16 -!- amca has joined.
03:42:04 -!- GreaseMonkey has joined.
03:48:28 -!- alexlyoko94 has joined.
03:48:42 -!- alexlyoko94 has left (?).
04:13:00 -!- xim_ has joined.
04:13:38 -!- xim_ has left (?).
04:14:59 -!- Sgeo[Circe] has joined.
04:16:52 -!- Sgeo has changed nick to Sgeo_.
04:16:57 -!- Sgeo[Circe] has changed nick to Sgeo.
04:17:13 -!- Sgeo has changed nick to Sgeo[Circe].
04:17:17 -!- Sgeo_ has changed nick to Sgeo.
04:51:07 -!- oerjan has joined.
04:56:59 -!- bsmntbombdood has left (?).
04:56:59 -!- bsmntbombdood has joined.
04:56:59 <Sgeo[Circe]> Hello bsmntbombdood, and welcome to #esoteric
05:03:06 <Sgeo[Circe]> It's a client with a very easy scripting API
05:03:22 <pikhq> And for that YOU DIE.
05:03:28 <Sgeo[Circe]> Probably the least powerful one ever conceived, but still
05:03:59 <Sgeo> The code for autogreet: http://pastie.org/576319
05:09:06 <Sgeo> ..no comments?
05:09:28 <pikhq> Is it Turing-complete?
05:09:47 <Sgeo> How does that apply to an API?
05:09:58 <pikhq> I think it a valid question at all times.
05:11:09 <Sgeo> Good night all
05:11:33 <oerjan> good turing-complete night
05:15:29 <pikhq> More off-topicness: I have been reading the comic "Sandman".
05:15:30 -!- Sgeo[Circe] has quit ("Circe: http://circe.nick125.com/").
05:19:40 <amca> pikhq: Im curious: are you turing complete?
05:20:18 <oerjan> that is not a valid question
05:20:55 <oerjan> the "Is it" is essential. "are you" does not work.
05:22:07 <pikhq> Thank you for that, oerjan.
05:24:37 <amca> Sorry, you right.
05:24:51 <amca> pikhq: Is it turing complete?
05:25:02 <amca> Where "it" is pikhq
05:25:54 <oerjan> only in the same way as 1 is even, where 1 = 2
05:26:23 <pikhq> amca: For certain values of Turing.
05:30:45 -!- oerjan has quit ("Later, in a non-computable way").
05:34:48 -!- nescience has quit (wolfe.freenode.net irc.freenode.net).
05:34:48 -!- Warrigal has quit (wolfe.freenode.net irc.freenode.net).
05:34:48 -!- Robdgreat has quit (wolfe.freenode.net irc.freenode.net).
05:34:48 -!- Dewio has quit (wolfe.freenode.net irc.freenode.net).
05:34:49 -!- Warrigal_ has joined.
05:34:50 -!- Robdgreat has joined.
05:34:52 -!- Dewi has joined.
05:34:59 -!- nescience has joined.
05:37:21 -!- Sgeo has quit (Read error: 110 (Connection timed out)).
06:50:09 -!- puzzlet_ has quit (Remote closed the connection).
06:50:13 -!- puzzlet has joined.
07:00:18 -!- GreaseMonkey has quit ("HydraIRC -> http://www.hydrairc.org <- Nobody cares enough to cybersquat it").
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:33:04 -!- coppro has quit ("The only thing I know is that I know nothing").
08:39:50 -!- amca has quit ("Farewell").
09:01:06 -!- coppro has joined.
09:10:25 -!- M0ny has joined.
09:13:21 -!- M0ny has quit (Client Quit).
09:13:35 -!- M0ny has joined.
10:11:48 -!- BeholdMyGlory has joined.
10:26:10 -!- KingOfKarlsruhe has joined.
10:32:30 <AnMaster> <GregorR-L> I'm looking for more cellular automaton to make into ties. There are only two good ECAs, but I can extend that to size-5 neighborhoods and then the range is almost endless. <-- ECAs? size-5 of what?
10:53:54 -!- coppro has quit (Remote closed the connection).
10:55:29 -!- coppro has joined.
10:57:46 <fizzie> ECA ("elementary cellular automaton") is at least what Wolfram calls the "rule N" things. Which have a size-3 neighborhood, but you could easily imagine adding the next two neighbors too.
11:28:29 <AnMaster> is the number of possible rules limited btw?
11:28:44 <AnMaster> given the same size of neighborhood I mean
11:40:38 <fizzie> Uh, of course. There are only 256 of the elementary ones, for one thing.
11:44:58 -!- M0ny has quit (Read error: 104 (Connection reset by peer)).
11:48:36 -!- FireFly has joined.
12:19:27 -!- Pthing has joined.
13:04:01 <GregorR-L> fizzie: If you disregard trivial permutations of other rules, there are only 64 ECAs.
13:18:20 -!- MigoMipo has joined.
13:18:48 -!- morenel has joined.
13:18:53 <morenel> whats that language rhat uses indentation for program flow?
13:18:58 <morenel> frightens the shit out of me
13:34:03 <Pthing> python: widely considered the most terrifying language
13:35:26 <Pthing> only if you don't stop to think
13:35:31 <Pthing> about the terrible secret of whitespace
13:35:35 <Pthing> and what is hiding in it
13:36:29 <Slereah_> I am here to protect you from the terrible secret of whitespace
13:36:37 <Slereah_> Whitespace has a terrible power
13:37:03 -!- SimonRC has quit (Read error: 113 (No route to host)).
13:45:36 -!- GregorR-L has quit (Read error: 110 (Connection timed out)).
13:56:08 -!- lament has joined.
14:00:41 -!- ehird has joined.
14:01:23 <ehird> 21:03:59 <Sgeo> The code for autogreet: http://pastie.org/576319
14:01:23 <ehird> 21:09:06 <Sgeo> ..no comments?
14:01:28 <ehird> haven't seen you here before etc
14:01:45 <ehird> my logreading has not yet seen you
14:03:19 <ehird> python frightens you? well me too but for less trivial reasons
14:03:24 -!- jix has joined.
14:07:52 -!- MigoMipo has quit (Nick collision from services.).
14:07:53 -!- MigoMipo_ has joined.
14:08:55 -!- MigoMipo has joined.
14:22:19 <ehird> lament: a U man with wavy arms and a penis?
14:23:28 <ehird> looks like it zoomed in
14:26:53 -!- MigoMipo_ has quit (Read error: 110 (Connection timed out)).
14:28:57 -!- SimonRC has joined.
15:05:05 <ehird> oh is the letter called wat or something
15:06:04 <lament> you live in the past, man
15:24:05 <ehird> <reddit> Best source code ever... [ASCIIPIC?] http://www0.us.ioccc.org/2001/williams.c
15:24:08 <ehird> WOW YOU CAN ADD WHITESPACE
15:25:53 <Deewiant> Slow-ass US mirrors... http://www.de.ioccc.org/2001/williams.c
15:27:36 <ehird> who cares, it's a boring entry
15:28:01 <Deewiant> It's cute but simple ASCII art
15:29:05 <ehird> yes, but shit like that can be automated for chrissake
15:29:57 <ehird> i'm not dissing it
15:30:02 <ehird> but it's at the top of proggit
15:32:26 <Deewiant> There's been worse stuff at the top of proggit
15:56:34 -!- Asztal has joined.
16:10:39 <pikhq> ehird: There's much better IOCCC entries.
16:10:51 <ehird> btw let's start using there're
16:10:52 <pikhq> Why not the bit of code that calculates Pi based on its own area?
16:10:54 <ehird> instead of there's → there are
16:11:29 <pikhq> Or the bootstrapping subset-of-C compiler?
16:25:03 <ehird> no fabrice bellard allowed
16:25:06 <ehird> 's just too clever
16:25:28 <lament> there're a problem with that idea
16:26:22 -!- Asztal has quit (Connection timed out).
16:26:37 <ehird> lament: umm care to clarify are you just on crack
16:27:47 -!- Asztal has joined.
16:29:07 -!- upyr[emacs] has quit (Remote closed the connection).
16:36:15 -!- oerjan has joined.
16:38:36 <oerjan> <morenel> whats that language rhat uses indentation for program flow?
16:38:53 <oerjan> if by "flow" you mean "structure"
16:39:27 <lament> every good language there is!
16:39:37 <pikhq> lament: Not quite.
16:39:43 <pikhq> ... There exists F#.
16:40:11 * oerjan now accuses pikhq of being John D. Harrop
16:40:56 <oerjan> not very probable, i know, he wouldn't be able to mention F# that rarely in this channel
16:41:22 <pikhq> And he wouldn't be derogatory concerning the language.
16:41:32 <oerjan> wait, that was derogatory?
16:41:34 <Asztal> does Whitespace count?
16:41:59 <pikhq> oerjan: "Every good language there is!" "Not quite. There exists F#."
16:42:13 <pikhq> oerjan: I'm at least under the impression that F# has indentation-significant syntax.
16:42:16 <oerjan> pikhq: ok so F# _is_ indentation sensitive, but not good? i don't know it, so i interpreted it the other possible way
16:42:54 <pikhq> oerjan: It's a .Net "functional" programming language.
16:43:03 <ehird> Uhh, pikhq. pikhq.
16:43:10 <ehird> F# = OCaml for .NET.
16:43:18 <ehird> It isn't indentation-significant.
16:43:25 <oerjan> oh i didn't mean "don't know" in the never heard about it sense. i just haven't seen the syntax.
16:43:27 <pikhq> ehird: Okay, then.
16:43:27 <ehird> Harrop is a retard, but stop the FUD :P
16:43:42 <pikhq> F# is still an awful language, though.
16:43:59 <ehird> It's just OCaml with access to .NET stuff.
16:44:09 <ehird> Nothing spectacular, but I find it hard to hate.
16:44:13 <pikhq> "You got .Net in my functional language!" "You got your functional language in my .Net!" "YAY!"
16:44:24 <pikhq> ehird: I find the concept distasteful. ;)
16:44:24 <oerjan> it's a chimera, i guess. since ocaml already sort of is.
16:44:33 <ehird> pikhq: OCaml is object-oriented already.
16:44:33 <oerjan> and it adds .NET to that
16:44:39 <ehird> Binding .NET does not muddy the language at all.
16:45:00 <pikhq> ehird: My complaints are two-fold: OCaml, and .NET.
16:45:06 <oerjan> ehird: i doubt the type systems are compatible.
16:45:11 <ehird> You just hate every language that isn't Haskell.
16:45:21 <pikhq> Not *every* language!
16:45:31 <pikhq> ... There's something to be said for the untyped lambda calculus.
16:45:34 <ehird> have you ever actually used OCaml
16:45:36 <lament> i hate every language that is or isn't Haskell
16:45:45 <pikhq> No, I'm just being silly on IRC.
16:46:02 <oerjan> i recall reading that F# does type inference somewhat unintuitively for classes
16:46:43 <oerjan> (since hindley-milner does not go well with subtyping)
16:47:03 <ehird> Subtyping is nice.
16:47:13 <ehird> I want subtyping a lot when Haskelling.
16:52:37 <oerjan> AnMaster: lambert is being very sensible today
17:14:15 -!- pikhq has quit (Read error: 104 (Connection reset by peer)).
17:27:37 -!- pikhq has joined.
17:35:45 -!- pikhq has quit (Read error: 104 (Connection reset by peer)).
17:39:41 <oerjan> ^ul (S)::^(~:(o)~^~:^):^
17:39:41 <fungot> Sooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo ...too much output!
17:39:44 <ehird> i think i dreamed something
17:39:48 <ehird> then forgot it was a dream later
17:42:39 <ehird> i think i do that a lot
17:44:50 <oerjan> yes, you are already forgetting that this is a dream
17:59:42 -!- pikhq has joined.
18:08:19 -!- pikhq_ has joined.
18:09:20 -!- pikhq has quit (Nick collision from services.).
18:09:25 -!- pikhq_ has changed nick to pikhq.
18:24:47 <ehird> "I am an atheist who wants to become religious again. Ask me anything." // grr @ people who don't make a distinction between dogmatic atheism and rationalism
18:24:59 <ehird> you're running ext3 except with extents!
18:25:11 <ehird> you're running ext2 except with journaling and extents
18:25:59 <pikhq> ehird: And btrees.
18:45:24 -!- ehird has set topic: This haughty infidel says a cross revealed "O, never you must stray to a roaring tessellation saying 'What is this... holy ass! THISACRONYMSTARTSWITHAT!'" | http://tunes.org/~nef/logs/esoteric/?C=M;O=D.
19:00:02 -!- JoelyWoely has joined.
19:00:37 -!- CESSMASTER has quit (Nick collision from services.).
19:00:44 -!- JoelyWoely has changed nick to CESSMASTER.
19:06:29 <augur> ehird: i love you.
19:06:43 <ehird> why thank you augur
19:07:24 <augur> thae topic is brilliant
19:08:37 <ehird> thanks, i wrote that in february i think :P
19:09:17 <oerjan> despite being nearly unable to type due to the biting cold
19:09:58 <augur> oejan: you stole that from me :|
19:10:15 <augur> saying angkor in response to wat
19:10:32 <ehird> angkor in response to wat
19:10:40 <augur> well thats what i thought
19:10:43 <augur> but noone fucking gets it
19:10:47 <ehird> i actually guessed oerjan would do it this time :)
19:10:51 <ehird> augur: i had to google it first time.
19:10:58 <augur> sometimes i say thom
19:11:13 <augur> because angkor thom is both more obscure and more interestng than angkor wat
19:11:30 <augur> and that just confuses people. :D
19:11:48 <augur> SOMETIMES i say "soup" and people are like "huh?"
19:11:53 <augur> so i clarify with "wat soup"
19:11:57 <augur> and they're still confused
19:12:05 <ehird> Fun fact: the OS X system sound "Sosumi" (so sue me) was a reference to Apple Corps v. Apple Computer; it's a xylophone, so technically that counts as distributing music (tenuously)
19:12:15 <ehird> (Proof from Infallopedia: http://en.wikipedia.org/wiki/Sosumi)
19:12:19 <augur> so i explain that its a kind of soup form around angkor wat in cambodia, used to greet people
19:13:35 <oerjan> augur: no such thing, says google
19:13:44 <augur> no such thing as what
19:14:18 <ehird> <augur> Phallopedia
19:14:21 <ehird> let's make Phalluspedia
19:14:35 <ehird> an encyclopedia about cocks
19:14:41 * oerjan swats ehird for mangling greek morphology
19:14:55 * oerjan forgot the swatter -----###
19:14:56 <ehird> * augur's phallus swats oerjan for etc
19:15:47 <ehird> <augur> .... I'm so edgy because I use the wrong number of dots
19:17:42 <ehird> The problem with Twitter, I think/is that though anecdotal evidence/is common as a sink/there is a lack of prudence/and I consider it rude s
19:17:46 <ehird> ense/to cut off a poem on the brink.
19:18:23 <ehird> Also weird rhyming scheme; ABAsortofBBA
19:18:52 <oerjan> !haskell length "The problem with Twitter, I think/is that though anecdotal evidence/is common as a sink/there is a lack of prudence/and I consider it rude sense/to cut off a poem on the brink."
19:19:10 <ehird> !haskell length "The problem with Twitter, I think/is that though anecdotal evidence/is common as a sink/there is a lack of prudence/and I consider it rude s"
19:19:25 <ehird> Bitch. That was the joke.
19:19:25 <oerjan> i was gonna check that
19:19:33 <augur> http://news.yahoo.com/nphotos/Michael-Jackson-Staples-Center/photo//090807/482/014ee302e9be4f0a89a4eeca91a2e595//s:/ap/20090808/ap_en_ot/us_michael_jackson_insurance;_ylt=AgTPYYGRfmU9P0ms7L75FsZY24cA;_ylu=X3oDMTE5dmZwOWFyBHBvcwMxBHNlYwN5bl9yX3RvcF9waG90bwRzbGsDaW50aGlzaGFuZG91 << look at me i'm michael jackson and i rehearse in CG environments that are photoshopped poorly!
19:21:50 <oerjan> some day photoshopping will get so good that no one manages to detect it, at least officially. unofficially mental hospitals start filling up with CG experts who claim they can detect photoshopping artifacts in reality.
19:23:13 <augur> oerjan have i told you my reasoning for why the world might be a simulation? x3
19:25:08 <augur> long story short, 3, maybe 4 of the features of the universe's fundamental design are the sort of thing you'd add to a large scale universe simulation to make it easier to computer
19:26:21 <ehird> the simulation argument is wrong.
19:26:33 <augur> it probably is, but even so its fun
19:27:40 <Pthing> what are these three maybe four features
19:28:02 <ehird> augur being able to spout bullshit
19:28:06 <ehird> and people taking the simulation argument seriously
19:28:12 <Pthing> that is a really good feature
19:28:12 <oerjan> support the factory union
19:28:12 <ehird> thus distracting them from actually thinking and looking
19:28:21 <ehird> which would cause them to find all the flaws in the approximations
19:28:51 <augur> pthing: non-simultaneity, a maximum speed on motion, and probabilistic particle location
19:29:22 <ehird> augur that adds up to five
19:29:28 <ehird> you just had to add one or two more.
19:29:41 <Pthing> actually i think he's saying I'm the fourth fundamental feature of the universe
19:29:47 <Pthing> and he used a colon instead of a comma
19:30:08 <ehird> another argument against the simulation argument: why the heck would you want to simulate this universe?
19:30:17 <Pthing> also "simulating what?"
19:30:19 <oerjan> stealth terrorists from uranus
19:30:23 <Pthing> the answer appears to be
19:30:31 <ehird> they haven't intervened, they've just run one presumably not entirely unlike their own (they couldn't imagine simulating something too different from their universe; almost certainly)
19:30:35 <ehird> what the fuck is the point
19:30:36 <augur> pthing: what do you mean?
19:30:41 <Pthing> <ehird> those are two.
19:30:41 <Pthing> <augur> pthing: non-simultaneity, a maximum speed on motion, and probabilistic particle location
19:30:50 <Pthing> relativity demolished the first two, quantum the second
19:31:00 <ehird> BUT THAT LEAVES THE THIRD
19:31:01 <Pthing> if you remove these "easier to compute" features
19:31:05 <augur> what do you mean "demolished"?
19:31:16 <Pthing> you are left with a 19th century cosmos
19:31:30 <augur> i dont follow what you're saying, pthing
19:31:34 <ehird> this universe would be way way way easier to simulate if we did one thing
19:31:35 <Pthing> which is the more-real-than-real universe that the real universe is just a computational simulation of
19:31:37 <ehird> ELIMINATE QUANTUM MECAHNICS
19:31:53 <Pthing> ehird, no apparently it is a simplification?
19:31:58 <augur> ehird, i dont think it would
19:32:01 <ehird> classical mechanics gets like 99.999999999% of shit right, and fuck the rest
19:32:12 <augur> because then you'd have to actually calculate all particle motion and so forth
19:32:13 <ehird> we would just try experiments and conclude yep that classical shit is the bomb
19:32:31 <Pthing> you point to those as features that are put in to make the universe easier to compute
19:32:32 <augur> its far easier to just leave particle positions as probabilistic things that dont have to be calculated all the time
19:32:41 <augur> pthing: yes, i do.
19:32:45 <ehird> do you know anything about QM
19:32:48 <Pthing> that means in the hyper-real universe
19:32:49 <ehird> no way that shit simplifies vs classical
19:32:52 <Pthing> they're not there, right?
19:33:06 <augur> ehird: i think it does!
19:33:16 <ehird> augur: because you're not a physicist, you're a linguist on crack
19:33:17 <augur> pthing: who knows. the hyper-real might not even be like our universe at all!
19:33:28 <augur> ehird: actually i was a physics major before i was a linguistics major
19:33:36 <ehird> i can see why you switched
19:33:39 <Pthing> augur, well whatever, *a* hyper-real universe that our universe is simulating, right?
19:33:51 <ehird> Pthing: whoa you mixed that up there
19:33:55 <ehird> that sentence is freaky shit.
19:34:06 <Pthing> dont worry little one these ideas are VERY HARD
19:34:14 <augur> if the simulation hypothesis is correct, sure!
19:34:18 <Pthing> real scienticians are afraid of them
19:34:27 <ehird> Pthing: dude dude you said that our universe is simulating our simulator
19:34:28 <Pthing> the hyper-real universe is, in fact, Newtonian
19:34:29 <ehird> that must have been a typo :P
19:34:46 <augur> he's just a Scheme fan
19:34:58 <augur> METACIRCULAR UNIVERSE
19:35:05 <Pthing> is just the same thing as people who thought quantum mechanics and relativity were funny tricks
19:35:12 <Pthing> and Newton got it all right via unaided reason
19:35:37 <augur> pthing, iiii dont think so. you're assuming that the "hyperreal universe" is newtonian, when it could very well be completely different and non-newtonian
19:35:48 <ehird> stop right there for one second augur
19:35:53 <Pthing> The hyper-real universe is like our universe
19:35:54 <augur> i think its a mistake to assume that the only two options are newtonian and relativoquantumic
19:35:55 <ehird> i want you to do some thinking for me
19:35:57 <Pthing> minus these computational aids
19:36:06 <augur> but why would you assume that it's like our universe at all? :o
19:36:13 <augur> ehird: im listening, talk. jesus.
19:36:16 <Pthing> well because that is by definition
19:36:26 <Pthing> I'm not saying our universe is simulating the hyperpeople's universe
19:36:28 <ehird> augur: give me the mechanics of a universe that are consistent, result in complex patterns, and are relatively stable
19:36:30 <Pthing> just *a* hyperuniverse
19:36:36 <ehird> and make them totally unlike anything we've ever dreamt of
19:36:57 <Pthing> and *the particular* hyperuniverse our computer is simulating appears to be our universe - computational flaws
19:37:00 <ehird> congratulations, YOU LOSE
19:37:17 <augur> ehird: congratulations, i dont
19:37:26 <augur> pthing: i dont get what you're saying
19:37:34 <ehird> you lost in fact as soon as you said "the simulation argument"
19:37:43 <augur> what universe you're talking about, for one
19:37:49 <augur> ehird: you said it first :D
19:38:18 <Pthing> i am talking about the universe that the hyperpeople want to learn things about to which our universe is a computational simulation
19:38:29 <ehird> augur stop embarrassing yourself
19:39:03 <oerjan> hyperpurple people eaters
19:39:08 <augur> pthing, no no when i said its a simulation i just mean simulated, not a simulation of /something/
19:39:13 <ehird> nonsense, hyperpeople just make huge simulation projects of wildly different universes that still requires oodles of computation FOR FUN!
19:39:19 <Pthing> oh, but that's implicit!
19:39:25 <Pthing> You don't just have A Simulation
19:39:26 <ehird> we're three year old bobby's first universe
19:39:27 <augur> in the same way that like, The Sims is a simulation, but its not really a simulation of our universe
19:39:33 <augur> i eman, i guess in some sense it is
19:39:44 <Pthing> it's a simulation of a particularly human-sized suburban american part of it
19:39:52 <augur> not all simulations are simulations of preexisting things!
19:40:17 <augur> which was a particularly fun game, btw.
19:40:17 <Pthing> um, that's a puzzle game
19:40:24 <augur> its a simulated environment!
19:40:29 <ehird> wow you lose augur
19:40:32 <ehird> you massively lose.
19:41:12 <augur> watch me not shut up
19:41:22 <augur> you know why i wont shut up? because your opinion is irrelevant
19:41:22 <ehird> actually he can make you shut up.
19:41:25 <ehird> cool feature of irc clients.
19:41:28 <augur> how does that make you feel?
19:41:33 <augur> knowing that you're irrelevant?
19:41:37 <Pthing> "I think the universe is a simulation by some hyperpeople of a newtonian universe"
19:41:54 <Pthing> "slobber slobber sir isaac newton slobber slobber"
19:42:11 <augur> pthing you're so silly
19:42:21 <ehird> i believe you now augur
19:42:24 <ehird> i will have it etched in stone
19:42:30 <ehird> "opposers you're so silly"
19:42:39 <ehird> — augur, founder of the hyperpeoplenewtonianquantumosimulatron theory
19:42:49 <augur> pthing isnt even opposing anything
19:43:00 <oerjan> this opossum is ur-elephant
19:43:22 <augur> ehird: well, seeing as how im not saying the universe is a simulation, no
19:43:33 <ehird> then what the fuck ARE you saying
19:43:42 <augur> maybe you should read, ehird
19:43:53 <ehird> Pthing: isn't that what he said?
19:43:57 <ehird> because i'm fairly sure
19:43:59 <ehird> that's what he fucking said.
19:44:01 <augur> i said like half an hour ago that the universe isnt a simulation.
19:44:08 <Pthing> I do not know what you *are* saying
19:44:21 <Pthing> apart from your points aren't very good
19:44:28 <ehird> augur: if you can't communicate with someone of any intelligence, that's your problem
19:44:31 <ehird> not the listener's
19:44:39 <augur> <ehird> the simulation argument is wrong.
19:44:39 <augur> <augur> it probably is, but even so its fun
19:44:39 <ehird> come back when you can state in a sentence WHAT YOU ARE SAYING
19:44:46 <augur> WHATS THAT? ME AGREEING WITH YOU EHIRD?
19:44:49 <ehird> generally for the purpose of argument,
19:44:53 <ehird> we assume you believe what you're arguing for
19:45:11 <ehird> for instance, "Devil's advocate: Nazis rock" "No they don't, Nazis don't rock, you're stupid" "I NEVER SAID NAZIS ROCK"
19:45:24 <ehird> "...that's what you're arguing for."
19:45:32 <ehird> "No it's not, I said half an hour ago that nazis don't rock"
19:45:47 <ehird> i'm proud to have never made a good analogy in my life.
19:46:32 <augur> pthing, stop telling people to shut up.
19:47:03 <augur> hey i just realized, both pthing and ehird are in england
19:47:09 <augur> and theyre both homosexuals!
19:47:21 <augur> well, ok, ehird is 13, not a homosexual
19:47:23 <ehird> yeah I'm actually Pthing.
19:47:35 <ehird> Pthing is gay because, pee thing, you see
19:47:37 <ehird> it refers to a penis.
19:47:56 <augur> pthing, its "om". your mantra sucks.
19:48:05 <augur> get with the program, gosh
19:48:06 <ehird> it's the skeptic's mantra
19:48:25 <ehird> HOLY FUCK 360 DEGREE HOLOGRAPHIC DISPLAY
19:48:25 <ehird> http://www.youtube.com/watch?v=c2VusJwGTQQ
19:49:25 <augur> its still got a spinning component tho :(
19:49:55 <augur> its got a spinning mirror
19:50:07 <ehird> it must spin fast enough to work
19:50:13 <ehird> if you get your head chopped off, your own fucking fault
19:50:18 <augur> we've had spinning-surface "holograms" for ages
19:50:22 <augur> like 20 years or something
19:50:48 <ehird> who fucking cares, it looks awesome
19:50:59 <augur> its been looking awesome for 20 years or something!
19:51:13 <ehird> add some tactile gloves to that and you've got the most awesome "futuristic" computer ever
19:51:30 <augur> sure, one that you cant actually touch because your hand will get fucked up
19:51:46 <ehird> the mirrors aren't actually in the hologram are they
19:51:57 <augur> the hologram reflects off the mirror
19:52:05 <ehird> i thought it was like
19:52:08 <ehird> you need a circular chamber
19:52:11 <ehird> and the mirror spins around the edges
19:52:15 <ehird> projecting it into the middle
19:52:24 <augur> but that is an interesting idea
19:52:37 <ehird> have to spin really fast though
19:52:46 <ehird> and that'd be uber noisy
19:52:56 <ehird> but, it should work
19:53:03 <augur> itd be difficult to achieve, i think
19:53:09 <augur> and itd be pretty big
19:53:28 <ehird> wouldn't a mirror spinning really fast be equivalent to uh
19:53:32 <augur> and youd have to look down at it
19:53:47 <ehird> augur: not if you put them in the (transparnet) ceiling and floor
19:53:51 <augur> the mirror spins around its center, so that its pointing at different places throughout its spin
19:53:54 <ehird> as well as the walls
19:54:02 <augur> ah, so a big room then
19:54:07 <augur> that would be veeerry tricky
19:54:26 <ehird> augur: just move the mirrors with magic
19:54:29 <ehird> magic can do it fast enough!
19:54:41 <augur> but can it hold the mirror glass together, thats the question!
19:55:14 <augur> theres a planetarium-like thing at some university thats intended for something like that
19:55:17 <augur> there was a TED talk about it
19:55:26 <augur> it doesnt use holography
19:55:35 <augur> but at the scale and distance used, it doesnt matter much
19:55:37 <ehird> but anyway, that + big circular room + antigravity + tactile glove thingies + huge fucking earmuffs to block out all the awful noise from the spinning and levitation = BEST INTERFACE EVER
19:56:11 <augur> sounds like the gunnery interface in that one Babylon 5 movie
19:57:04 <ehird> alas it is not easy to maneuver in the kind of antigravity that doesn't involve a very fast plane
19:57:17 <ehird> or indeed that kind too; if you could just sort of speed up movements it'd be fine
19:57:19 <augur> http://en.wikipedia.org/wiki/LEP#Results
19:58:06 <ehird> just do it in space
19:58:07 <augur> RUININ MAH SPERMENTS
19:58:14 <ehird> no interference at all!
19:58:27 <augur> yeah but then the alien space ships will fuck it up!
19:58:39 <augur> slereah_, you work in the cern area right?
19:58:46 <Slereah_> My quantum physics teacher told us
19:58:48 <augur> or do you just contract out to nantes?
19:59:05 <Slereah_> Although he just told us about the train, though
19:59:44 <ehird> it'll probably be simpler to make good VR than to do the awesome interface :P
20:00:02 <augur> have you seen Lawnmower Man?
20:00:35 <Slereah_> Is it about the retard who gains superpowers?
20:00:36 <ehird> i think i saw it when i was a kiddie
20:00:43 <augur> slereah: not quite
20:00:45 <ehird> Slereah_: if i'm thinking about the same movie, yes
20:00:50 <ehird> well not superpowers but
20:00:53 <augur> retarded kid used in experiments with VR to boost intelligence
20:01:06 <ehird> no no no don't call them experiments with vr
20:01:07 <Slereah_> Come on, he makes a lawnmower in some dude's mind!
20:01:08 <augur> in the sequel he ends up creating a massive virtual world where people can upload their minds
20:01:10 <ehird> what it was is flashing images
20:01:20 <ehird> and that made him clever
20:01:26 <augur> by playing a video game
20:01:42 <augur> the game they played was a lot like the pre-release version of Wipeout, oddly
20:01:50 <augur> which looks pretty interesting
20:01:53 <ehird> wipeout is a good game
20:02:06 <augur> ehird, if youve never seen the pre-release version, check out the Hackers movie
20:02:07 <ehird> erm i dunno if i played wipeout
20:02:14 <ehird> augur: no i'm never watching that movie
20:02:24 <augur> the game they play on the big tv is the prerelased wipeout
20:02:26 <augur> aww ehird why not?
20:03:39 -!- jix has quit ("Lost terminal").
20:03:46 <augur> ehirds worried that hes going to start thinking that hacking involves lots of pretty pictures on old mac laptops
20:04:19 <augur> when everyone knows that it involves lots of pretty pictures on new ubuntu+compiz fusion laptops!
20:04:50 <ehird> they don't call them laptops any more augur
20:04:59 <ehird> alas the lap is now obsolete as NOTEBOOKS are
20:05:14 <augur> i hate my sperm anyway.
20:05:36 <ehird> you can tell they actually like to be on your lap because they get all hot
20:05:36 <pikhq> REPUBLICAN RETARDS THINK THAT OBAMA IS PROPOSING MAKING SOYLENT GREEN REALITY. D':
20:05:48 <ehird> pikhq: Sounds like a modest proposal to me.
20:06:31 <ehird> http://www.youtube.com/watch?v=MR2Jh2i7JgY Wow.
20:08:20 <augur> americans are stupid
20:08:24 <augur> this is a general fact
20:08:36 <ehird> You're such a bitch, bitch.
20:08:39 <augur> republicans as much as democrats
20:08:49 <augur> ehird: but i'm your bitch li
20:09:13 <augur> li is one character left of ;o
20:09:32 <pikhq> augur: I'm neither. The Democrats are too right for my tastes.
20:09:58 <augur> but only because i favor the destruction of the whole political system as it currently exists. :D
20:10:04 <AnMaster> mycroftiv, there? I have a plan9 installer question: How do I get it to use Swedish keyboard layout. I forgot since last time I did a plan9 in qemu
20:10:50 <ehird> Not in the installer you don't, afaik.
20:11:04 <augur> guys: $100/£50 for the first person who can figure out what this translaiton SHOULD have looked like:
20:11:05 <augur> آڈولف میركلے KFC died of her father when took his family business توجدید on the lines while it too much development soon .
20:11:11 <AnMaster> ehird, also where is / on US layout? And +
20:11:19 <ehird> augur: I like to fuck goats
20:11:26 <ehird> AnMaster: / is after .
20:11:35 <ehird> qwertyuiop[] (either enter or \)
20:11:37 <augur> it is infact supposed to have been "When Adolf Merckle took charge of his family's business after the death of his father, he greatly improved the business very quickly through a process of modernization."
20:13:26 <ehird> Also, don't do venti.
20:13:44 <ehird> Unless you have a RAID with a buncha disks.
20:14:06 <augur> i'll raid your disks alright
20:14:20 <AnMaster> I want to try venti that is why I'm doing this :P
20:14:29 <ehird> AnMaster: Well, that's stupid.
20:14:52 <ehird> Because you don't have a lot of disk, and venti/fossil isn't exactly the most compact storage mechanism.
20:15:11 <AnMaster> 15 GB disk image. True not that much
20:15:13 <oerjan> fossils are too thinly spread out
20:15:55 <AnMaster> fuck the screen dimming thingy in ubuntu btw. it doesn't work well
20:16:07 <ehird> Hmm, I don't actually have a spare x86/BIOS box anywhere
20:17:24 <AnMaster> btw. it plan9 cd doesn't even boot under virtualbox
20:18:15 <AnMaster> or rather, with the ICH6 IDE controller (not the default one) it does boot (very slowly) but then freezes when it tries to start the GUI
20:18:34 <ehird> AnMaster: also use his 9grid image
20:18:39 <AnMaster> ehird, tried sata too. didn't work well
20:18:39 <ehird> instead of the install iso
20:18:44 <ehird> that's what he told me to do at least
20:18:49 <AnMaster> ehird, what was the url now again
20:19:10 <ehird> http://sphericalharmony.com/plan9/gridtoolsplus.tgz
20:19:16 <ehird> http://sphericalharmony.com/plan9/ventigridserver.qcow2.img.tgz is just the image
20:19:39 <AnMaster> and qemu can't make use of the cool VT-x thingy
20:20:02 <pikhq> AnMaster: Sure it can.
20:20:20 <pikhq> Just need to have kvm in your kernel.
20:20:22 <AnMaster> ehird, yes but that doesn't use hardware virtualization iirc? Only does some stuff in kernel to speed it up. Since kqemu works on my old sempron too
20:20:37 <ehird> your sempron has hardware virtualisation
20:20:46 <pikhq> ehird: That's not what kqemu does.
20:20:50 <ehird> plan 9 is very frugal with resources
20:20:59 <ehird> so vt-x is totally unneeded
20:21:06 <pikhq> And no, Semprons don't have it.
20:21:10 <AnMaster> flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt lm 3dnowext 3dnow rep_good pni lahf_lm
20:21:16 <coppro> kvm only works if you have a processor that supports it :(
20:21:36 * AnMaster looks around in qemu-launcher for a KVM option
20:21:40 <ehird> AnMaster: Fabrice Bellard also wrote a Linux kernel module (with preliminary ports to FreeBSD and MS Windows) named KQEMU or QEMU Accelerator, which notably speeds up x86 emulation on x86 platforms. This is accomplished by running user mode code directly on the host computer's CPU, and using processor and peripheral emulation only for kernel mode and real mode code
20:22:08 <AnMaster> ehird, yes, but that isn't VT-x :P which was my point
20:22:41 -!- oerjan has quit ("Good night").
20:22:42 <coppro> kqemu can accelerate kernel mode code, too
20:22:53 <pikhq> coppro: It's a hack.
20:22:53 <AnMaster> I wonder if having virtualbox and kvm running at the same time will cause issues
20:23:28 <ehird> I'm typing vertically
20:23:34 <ehird> It is uncomfortable
20:23:41 <ehird> AnMaster: takes like 30m.
20:23:44 <AnMaster> maybe no DMA? though I said yes when it asked if I wanted to use DMA
20:24:13 <AnMaster> ehird, about 15 minutes passed of the "installing file system" phase. And it is at 15%
20:24:22 <AnMaster> I don't think that will add up to 30 minutes
20:24:40 <ehird> have you ever used windows' file copy dialog AnMaster
20:25:36 <AnMaster> ehird, yes. right. but a) this shows ETA in percent, not minutes b) the percent seems to be going at a steady rate, and has all the time
20:27:02 <ehird> i'm trying to imagine a desk too small to fit a notebook.
20:27:10 <AnMaster> I can't fit the throttle/joystick on there when the laptop is also on it
20:27:18 <AnMaster> that is, throttle/joystick connected to desktop
20:27:28 <AnMaster> and keyboard/mouse for desktop there too
20:27:29 <ehird> you'll have to ask augur about joysticks
20:27:33 <ehird> can't advise there
20:28:07 <AnMaster> as in: full size keyboard, mouse, large joystick, large throttle, laptop
20:28:24 <ehird> i already told you that i can't advise about joysticks.
20:29:28 <AnMaster> ehird, well. I don't need advice. there is just no way to fit it :P
20:29:46 <ehird> i'm sure augur is an expert on extra-large joysticks
20:31:37 <AnMaster> ah maybe with laptop near the back of the desktop, since it is just installing atm, don't need to type anything
20:31:51 <ehird> right! it's always important that your joystick fits snugly.
20:32:11 <AnMaster> I want to play vegastrike. that is why
20:32:28 <ehird> sure, sure, ...play is an important part of human existence
20:32:34 <ehird> just don't break anything
20:33:16 <AnMaster> anyway, I don't know about "extra large". the base "box" of the throttle is 17x17 cm. about the same for the joystick
20:34:13 <AnMaster> throttle? depending on what position it is in.. between 14 and 20 cm it seems
20:34:49 <ehird> that's more than average
20:35:12 <AnMaster> joystick: close to 30 cm when in center position. Both these measures include the height of the base "box" too
20:35:23 <AnMaster> which is about 5 cm for either
20:35:42 <AnMaster> ehird, is it? Never owned any other joystick
20:35:51 <ehird> ...i don't think many people do...
20:36:04 <augur> my joysticks are indeed extra large
20:36:24 <ehird> [20:35] AnMaster: joystick: close to 30 cm when in center position. Both these measures include the height of the base "box" too
20:36:29 <augur> i have an 17" and an 18" joystick. ;o
20:36:40 <ehird> i have a 1000" joystick, true story.
20:36:57 <ehird> can get hard to carry around, like to lan parties
20:37:02 <ehird> and people like to play with it a lot there
20:37:05 <ehird> kinda attracts a crowd
20:37:07 <AnMaster> it uses the hall effect for sensing the position of the joystick. Rather than messy potentiometers
20:37:10 <augur> oh sorry i thought we were talking about dildos
20:37:26 <ehird> AND THUS THE EXTENDED INNUENDO ENDS
20:37:31 <AnMaster> augur, no. Joysticks. As in real joysticks
20:37:37 <augur> anmaster has no sense of humor.
20:37:38 <ehird> real joysticks, AnMaster
20:37:38 <AnMaster> and yes I know what ehird was trying to do :P
20:37:40 <ehird> that's what we're talking about,.
20:37:49 * augur plays with ehirds joystick
20:37:50 <ehird> AnMaster: retcon retcon retcon
20:37:56 <ehird> (erection erection erection?)
20:38:02 <ehird> why you gotta do that augur
20:38:05 <AnMaster> ehird, I actually realised it at "<ehird> that's HUGE!"
20:38:15 <AnMaster> because I though: "no I'm pretty sure it isn't"
20:38:15 <ehird> AnMaster: your naivety knows no bounds
20:38:41 <AnMaster> ehird, yes I'm probably naive when it comes to innuendo
20:39:09 <ehird> the joystick is where? in your end? oh.
20:39:43 <AnMaster> ehird, do you like blinkenlights?
20:39:53 <ehird> no finkerpoken or mittengrabbem
20:40:04 <augur> another translation thingy: "In his words : ' is the operation was now are to Provide INFORMATION ABOUT IT ."
20:40:18 <augur> apparently machine translations are Start To YELL AT YOU!!!
20:40:40 <ehird> AnMaster: uhh, howso
20:41:17 <AnMaster> ehird, aren't you supposed to like shiny blinking lights and such?
20:41:27 <ehird> okay, let me get this straight
20:41:29 <ehird> you made a reference
20:41:37 <ehird> and then you accused me of like
20:41:39 <ehird> breaking the reference
20:41:43 <ehird> youuuuuu're stuuuuupid
20:42:30 <AnMaster> ehird, um. I didn't intend it as a reference like that. I intended it as a normal word for the mentioned concept (that of blinking lights)
20:42:38 <ehird> http://en.wikipedia.org/wiki/Blinkenlights
20:42:49 <AnMaster> <ehird> no finkerpoken or mittengrabbem <-- read "no" as "nor"
20:42:55 <AnMaster> as in, you didn't like the reference
20:43:28 <ehird> guh electron microscopes use electrons
20:43:33 <ehird> i thought they looked at electrons
20:44:19 <ehird> it isn't _that_ far-fetched
20:44:33 <AnMaster> ehird, they use electrons due to the shorter wavelength. In order to be able to "see" smaller details
20:44:55 <AnMaster> at least that is what I learnt in the physics course in school
20:45:43 <ehird> listening is never mandatory :P
20:46:23 <AnMaster> qemu emulates a Pentium2? *HUH*
21:20:18 -!- comex has quit (Read error: 113 (No route to host)).
21:21:26 <augur> ive started packing my books
21:21:53 <augur> ive got almost three boxes full already
21:22:08 <augur> and im only done with the shelf along the top part of my wall
21:25:26 <Pthing> electron microscopes do look at electrons, though :|
21:25:39 <AnMaster> ehird, question: as you know my laptop is high dpi. How do I make it possible to actually read what it says in plan9 on it :D
21:25:51 <AnMaster> as in, where do I set DPI/font size stuff
21:27:04 <AnMaster> ok. good luck with finding large enough boxes
21:27:39 <augur> oh i have, dont worry
21:31:42 <AnMaster> I get this across my plan9 screen atm: "err 2; arena arenas00 creation time after last write time"
21:49:54 <ehird> AnMaster: plan9 fonts are bitmaps iirc
21:50:01 <ehird> so tough shit, scale up the window
21:50:11 <ehird> or use a bigger font
22:08:54 * pikhq has a low opinion of bitmap fonts
22:12:22 <ehird> It doesn't exactly matter unless you have good typefaces.
22:12:39 <ehird> The Plan 9 fonts are non-free though, which is dumb.
22:13:10 <ehird> (Of course, just because they're non-free doesn't mean they could just bundle Helvetica or something; they presumably aren't paying royalties on the current fonts.)
22:15:45 <mycroftiv> AnMaster: oh, that issue came up on 9fans recently - let me see if I can find the thread - the basic cause I believe is using venti and not allowing venti data dumps to complete, along with clock sync stuff
22:16:12 <ehird> ha, venti breaks everything!
22:17:22 <mycroftiv> well, heres the thing - plan9 was designed to *not* run everything on the same box. its possible to do so, but stuff like the massive dumps of venti prints to your console because of frequent reboots tend to crop up
22:18:21 <mycroftiv> if this is a qemu vm, when i use venti and plan9 in qemu, i have to set up the venti server as a userspace process on a different qemu VM - trying to stuff a full venti/fossil/auth/cpu/terminal into a single qemu VM has never worked well for me
22:19:03 <mycroftiv> mostly because the qemu virtualization layer cant seem to handle it well - an all-in-one setup seems to work OK on native hardware or in vmware, but qemu just folds under the pressure
22:21:01 <mycroftiv> AnMaster: if you are interested in grappling with the 'arenas00 creation time' prints, look in http://9fans.net/archive/2009/07 and search for 'arenas' and you will find a long thread about exactly your issue
22:21:08 <mycroftiv> the cause, and how to eliminate it
22:21:31 <ehird> <AnMaster> Meh, that's work. *GIVES UP*
22:21:39 <ehird> <AnMaster> Further proof that Plan 9 sucks.
22:22:18 <mycroftiv> in terms of having the system be usable by people without extensive study and experience - plan9 is absolutely awful, that is for sure
22:22:40 <mycroftiv> i would say that learning how to use plan9 in any 'real' way took me almost 9 months of 10 hours a day study
22:23:17 <ehird> mycroftiv: uhh really?
22:23:21 <mycroftiv> i think plan9 is so amazing i devote large amounts of my time trying to figure out how to make the learning/setup curve less insane
22:23:23 <ehird> i did shit with plan 9 in hours
22:23:34 <ehird> got it set up, looked at the
22:23:34 <mycroftiv> ehird: i doubt you did the things that make plan9 'important' in just a few hours
22:23:43 <ehird> used acme to hack out some code
22:23:45 <mycroftiv> if you havent had a cpu/auth server running before, imo youve never 'really' used plan 9 at all
22:23:47 <ehird> tried to use /dev/tcp and failed
22:23:54 <ehird> mycroftiv: meh, that involves having multiple boxes
22:24:17 <mycroftiv> the 'real' thing plan9 is supposed to do is distribute its functions over a minimum of 4 boxes, really
22:24:30 <ehird> multiple boxes = noise + space + money + unportability
22:24:50 <mycroftiv> well, thats why VMs are so important - i have 5 VMs running on an inexpensive desktop box currently
22:25:07 <mycroftiv> so moore's law and virtualization solved those issues
22:25:14 <ehird> distributing across the same hardware is pointless!
22:25:18 <ehird> it's semantically null! it does nothing!
22:25:29 <ehird> it's just extra overhead
22:25:42 <ehird> any OS that requires such a thing just has a bad process/communication model
22:28:29 <mycroftiv> ehird: any electronic circuit that requires more than a few transistors is clearly overdesigned and inefficient.
22:28:35 <mycroftiv> computers were a mistake to begin with!
22:28:51 <ehird> What is it with this channel and strawmen lately?
22:30:02 <augur> deep like the mariana's trench
22:30:33 <mycroftiv> i get confused between the marianas trench and the mindano trench
22:31:06 <augur> one is named after mariana
22:31:12 <augur> the other is named after mindano.
22:31:15 <ehird> is named after mindano
22:31:29 <mycroftiv> that clears things up, i hadnt noticed the difference in names before
22:32:02 <augur> i figured it was probably mindanao, but you never know. :D
22:32:47 <ehird> i ought to write a trivial bit of unexecutable os code so i can decide it's tedious and give up.
22:33:33 <mycroftiv> ehird: let me try to put it historically - the concept of plan9 was that bell labs would own and run all the hardware (cpus/disk servers/etc) and you would just have a cheap terminal that you dialed up the system on - and in your office, same thing, central machine room, cheap user terminals
22:33:50 <mycroftiv> so that was the usage scenario and model the OS was designed for, circa 1990
22:34:00 <ehird> thin clients failed.
22:34:05 <ehird> and that's a good thing
22:34:24 <ehird> thin clients are fundamentally bad, tho
22:34:59 <mycroftiv> well, i dont see how you can say that - i mean, all of us now use our computers often as 'thin clients' given the modern web and 'all i ever use is the browser' style mainstream computer use
22:35:12 <ehird> don't get me started
22:35:49 <mycroftiv> so as a practical matter, the 'thin client' mode of use is now incredibly popular - and i know lots of people who do all their work on remote xen hosted stuff delivered via citrix, also
22:36:01 <ehird> plenty of things are incredibly popular
22:36:16 <ehird> thin clients are so popular right now becaus
22:36:19 <ehird> they're CONVENIENT
22:36:27 <ehird> due to the bloated heap of modern computing
22:36:32 <ehird> but that's not required
22:36:42 <ehird> sure, thin clients are good in a lot of cases
22:36:46 <ehird> not so much thin client
22:36:52 <ehird> as a thick client that gets external resources
22:37:00 <ehird> a thin client has the display dictated to, just does IO
22:37:06 <ehird> a thick client with external objects is much more powerful
22:37:20 <ehird> and, even then, plenty of things could work better locally if only it was *more convenient* to do so
22:37:34 <mycroftiv> anyway, all this for me is subpoint to what im trying to say, which is that recreating an analog of the original plan9 architecture with multiple VMs on a single machine isnt pointless at all, any more than running more than one *application* at a time is pointlesss, because thats all you are doing!
22:38:04 <ehird> it just shows to me that if plan 9 needs an ip address and a protocol to do more than one computation at once effectively, then its model is outdated
22:38:21 <mycroftiv> if i run a venti server VM, and a fossil server VM, and a cpu server VM, and then a terminal Vm (to give the extreme) - im just running 4 different software applications on one computer, its just a 'heavy' way to do so
22:38:28 <ehird> bloody double spaces
22:38:44 <ehird> mycroftiv: a lot of redundancy there too, i see your point,
22:38:55 <ehird> and i get that with plan 9 it's "needed" to emulate the distribution this way,
22:39:02 <ehird> I just don't think it's the good way, as opposed to a hack
22:39:05 <mycroftiv> the technical problem is actually the qemu virtualization layer being inadequate as a direct practical matter
22:39:24 <mycroftiv> so the software that is at fault is qemu, not plan9, in the original source incident i believe
22:39:45 <ehird> ehhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh
22:40:00 <mycroftiv> (though plan9 people would say that you are supposed to 'know' that after you first install plan9, you should probably *not* reboot until your initial archival dump from fossil to venti completes)
22:42:01 <mycroftiv> ehird: but, if you want to find evidence to support your assertion about "if plan 9 needs an ip address and a protocol..." is a comment in the source code near an early boot bring up of IP stack saying "this is such a crock"
22:42:35 <mycroftiv> the fact that the process of trying to accomadate the average modern home user by figuring out how to stuff all the necessary components of a plan9 system into a single box that can be booted all at once - not so perfect, even by their standards
22:42:50 -!- GregorR-L has joined.
22:43:21 <mycroftiv> in the original system, the disk storage servers didnt even run the same kernel as the cpus and terminals
22:44:14 <mycroftiv> it was originally, code-wise, a 'different os' for each of the different functional components - over time that evolved and changed and some of the componenets were swapped out, like venti replacing the original archival data server
22:44:33 <ehird> have i mentioned recently that kernels suck?
22:44:38 <AnMaster> mycroftiv, thanks for that link
22:45:31 <mycroftiv> AnMaster: welcome back, good luck with your setup - as i mentioned venti+fossil boot system inside qemu has always been problematic for me
22:45:44 <AnMaster> mycroftiv, that is what I used yes
22:45:47 <ehird> AnMaster: the explanation involves big words
22:45:59 <ehird> brief summary http://tunes.org/wiki/no-kernel.html
22:46:09 -!- MigoMipo has quit (Nick collision from services.).
22:46:10 -!- MigoMipo_ has joined.
22:46:59 <AnMaster> ehird, how do you do scheduling of CPU time between objects?
22:47:14 * ehird hands AnMaster a "Didn't Read The Page At All" badge
22:47:24 <AnMaster> ehird, I read the first few lines
22:47:29 -!- comex has joined.
22:47:42 <ehird> Well fuck you, I'm not your personal reading assistant.
22:47:51 <ehird> You can't read seven short paragraph
22:48:11 <mycroftiv> AnMaster: long story short - the answer is nontechnical but processes dont schedule themselves, they attach to another 'meta-object' that controls task switching
22:48:28 <ehird> sentence was over 10 words.
22:48:37 <AnMaster> mycroftiv, right... And that bit runs in ring0?
22:48:43 -!- MigoMipo_ has left (?).
22:48:51 <mycroftiv> AnMaster: as i said, the wiki page was nontechnical
22:48:58 <ehird> <AnMaster> Hur hur the hardware is what makes all design decisions hur hur
22:49:05 <ehird> <AnMaster> Abstraction? UNPOSSIBLE!!!
22:49:11 <AnMaster> mycroftiv, well. I'm wondering how the hell this could be implemented on x86 at least....
22:49:34 <AnMaster> due to it being the most common architecture
22:49:48 <AnMaster> (in this case x86 includes 64-bit variants)
22:50:07 <mycroftiv> AnMaster: my opinion as a practical matter is that your OS abstraction layer is above what you might call 'shim code' and that (just imo at least) in some ways the differences are semantic/conceptual more than anything else
22:50:27 <ehird> if you really can't imagine how to not have a kernel and just use attached procedures for task switching...
22:50:34 <ehird> then you're either hugely massively ignorant about how OSes work
22:50:36 <ehird> or just can't hack.
22:50:46 <AnMaster> mycroftiv, sure this looks all nice and fluffy, but whatever code runs in ring0 is "kernel" to me...
22:51:02 <ehird> AnMaster: That's funny, then; you'll class my whole system as kernel.
22:51:04 <AnMaster> and kernel includes stuff like loadable drivers to me.
22:51:12 <AnMaster> ehird, ah, it is like inferno or such?
22:51:21 <ehird> <AnMaster> If I define kernel to mean something, it means that and not what it actually means.
22:51:21 <AnMaster> and yes, that is an unusual case.
22:51:24 <mycroftiv> thats why im saying its mostly semantics and conceptualizing - and some real differences in what functional tasks are located in what coe blocks - but still
22:52:28 <pikhq> My only comment on a no-kernel design is that it is probably very freaking hard to get working initially.
22:52:37 <ehird> ummmmmmmmmmmmmmmmmmmmm
22:52:46 <ehird> we must be reading different articles
22:52:52 <ehird> because it's exactly as easy as a kernel system
22:53:00 <AnMaster> well, maybe a bit more accurate would be: whatever handles cpu scheduling, interrupt management, core memory management, doing the actual ring-0 code to do stuff like DMA and so on is the kernel
22:53:15 <ehird> <AnMaster> If I define kernel to mean something, it means that and not what it actually means.
22:53:38 <AnMaster> ehird, then provide *YOUR* definition of kernel
22:53:55 <pikhq> ehird: But you have to, like, do more design than "Alright, let's toy with the standard UNIX kernel design a bit".
22:54:04 <ehird> Do you want me to ship a copy of Wikipedia to your house?
22:54:06 <ehird> Would you like that?
22:54:15 <ehird> Or is it too fucking long didn't goddamn read.
22:54:38 <AnMaster> ehird,I find my definition quite fitting with the one on wikipedia
22:54:47 <pikhq> Basically, the thing making the design hard (IMO) is that you'd kinda be paving the way.
22:54:58 <AnMaster> ehird, or do your system *trust all "traditionally userspace" code*
22:55:00 <pikhq> Not much of a criticsm, just a comment.
22:55:18 <pikhq> AnMaster: It's like a microkernel only more so.
22:55:41 <ehird> It is nothing like a microkernel at all.
22:55:46 <ehird> In fact, microkernels are possibly the antithesis of it.
22:56:07 <AnMaster> pikhq, he indicated everything ran in ring 0 above: "<ehird> AnMaster: That's funny, then; you'll class my whole system as kernel.". Seems insecure if you are, say, browsing the web
22:56:28 <AnMaster> there will *always* be bugs in anything as complex as a web browser
22:56:29 <pikhq> "Make it so that as little is done in the kernel as possible" is a lot like "Make it so that the 'kernel' is a misnomer."
22:56:37 <ehird> AnMaster: You're truly an idiot.
22:56:58 <mycroftiv> AnMaster: thats right, the web is TEEEMING with exploits targeting a nonexistent operating system ;)
22:57:01 <pikhq> AnMaster: And the OS would be written in Smalltalk.
22:57:08 <ehird> AnMaster: No, because you never understand any of my justifications or rebut them with strawmen and faux-jokes because you're an idiot.
22:57:12 <ehird> So I'll settle for a personal insult.
22:57:15 <AnMaster> mycroftiv, well, assuming it would become popular. Which I also assume is a goal of every OS
22:57:35 <pikhq> The only exploits akin to, say, the crazy shit done on Windows pre-NT would be from bugs of the Smalltalk implementation.
22:57:46 <mycroftiv> nobody with any sense thinks any os is going to be popular unless it starts with a w, ends with s, and has indow in the middle
22:58:29 <AnMaster> mycroftiv, yet linux is not that uncommon nowdays. Popular might even be a good word for it. Though far from as popular as windows indeed.
22:58:31 * ehird remembers that his client has an ignore feature, unlike his previous one
23:00:06 <AnMaster> mycroftiv, anyway as far as I can see, trusting all code that runs seems stupid. Surely running code in a ring != 0 has proven a good (which isn't same as "perfect") solution by now :P
23:01:10 <mycroftiv> AnMaster: i have no opinion, and none of the stuff from tunes or that ehird has said actually seems to make any kind of statement one way or another in that regard - i think ehird was speaking more generally about the 'whole os being kernel' than what ring stuff was assigned to
23:01:11 <AnMaster> iirc also at least on x86 and x86_64 running everything in ring 0 would be suboptimal. As in: stuff like task switching and page tables are designed for a kernel/userspace split.
23:01:29 <AnMaster> at least that is what reading the architectural documentation for AMD64 seemed to indicate
23:02:12 <mycroftiv> AnMaster: from everything ive gathered from ehird, most of the things he talks about are always taking place at a layer of abstraction built up from the native hardware behavior
23:02:50 <AnMaster> mycroftiv, also, memory protection is good for other reasons too. Remember "app crashing crashed whole OS"? Wasn't that long ago that was the norm outside *nix
23:03:10 <ehird> [22:50] AnMaster: mycroftiv, sure this looks all nice and fluffy, but whatever code runs in ring0 is "kernel" to me...
23:03:10 <ehird> [22:51] ehird: AnMaster: That's funny, then; you'll class my whole system as kernel.
23:03:14 <ehird> but that was just trolling him.
23:03:44 <AnMaster> mycroftiv, and I'm mostly interested in the actual implementation details. The low level stuff.
23:03:52 <AnMaster> which ehird seems to hate thinking about
23:04:27 <mycroftiv> AnMaster: yes, its obvious that you and ehird direct your thinking at different topics, and i myself agree that 90% (99%?) of the actual work tends to be struggling with that stuff to get the abstractions you want created and working right
23:04:42 <ehird> actually, my OS maps fairly directly to the hardware.
23:04:48 <ehird> moreso, dare I say, than Linux
23:04:52 <ehird> as it's just simpler.
23:04:55 <AnMaster> mycroftiv, was that agreeing with me or ehird?
23:05:07 <mycroftiv> AnMaster: its saying that you are talking past each other
23:05:25 * ehird attempts to think of a context in which mycroftiv's last line makes sense
23:05:27 <mycroftiv> ehird: do you know about um, whats it called, battlecruiser 3000 or something, and that guy's legendary flame wars on the net?
23:05:27 <pikhq> ehird: Certainly easier to add features to.
23:05:42 <pikhq> Hooray, modularity.
23:05:43 <mycroftiv> ehird: it makes sense when its obvious that you guys disagree on definitions
23:05:46 <ehird> Battlecruiser 3000AD (also known as [BC3K] in Usenet) is one of the longest-developed games in computer game history.
23:05:59 <AnMaster> mycroftiv, I'm a fan of OSes like http://www.coyotos.org/ Imagine how much I will dislike a OS *not* caring about security
23:06:17 <mycroftiv> well, its interesting - and its a cautionary tale - it actually turned out to be a pretty good game, and the guy has a continuing series of later games that seem to have a fan community
23:06:24 <pikhq> AnMaster: ... But he's not not caring about security.
23:06:45 <ehird> is AnMaster accusing my OS of being insecure?
23:06:46 <AnMaster> pikhq, and gets angry when I want to know about how he implemented it
23:07:01 <ehird> does he know I'm doing a full fucking capability based system?
23:07:03 <ehird> with no exceptions?
23:07:12 <mycroftiv> however, the original game took forever to make and was notorious for the guy making it trolling the hell out of usenet by talking about how great his game was, before hed actually written much of it
23:07:15 <pikhq> You'll note that the OS is meant to be programmed in Smalltalk. Who needs hardware memory protection when the language itself guarantees memory safety?
23:07:20 <AnMaster> ehird, why not unignore me if you are going to ask what I just said...
23:07:31 <ehird> HURR BUT WHAT IF YOU HAVE A BUG IN THE INTERPRETER
23:07:41 <AnMaster> pikhq, hah. Well I didn't know that until a few minutes ago when ehird said it
23:07:43 <ehird> Oh, wait, Linux has the exact same problem. So do all kernel systems.
23:07:47 <ehird> Can I say "system calls"?
23:08:05 <pikhq> He's said it in the past, at least.
23:08:17 <AnMaster> pikhq, I have been away for most of the last few days
23:08:21 <mycroftiv> ehird: you should stop talking about your os as if its something that exists and is running now and has definite proven characteristics.
23:08:24 <pikhq> It's pretty damned clear he intends to do it all in Smalltalk, and have the Smalltalk implementation written in Forth.
23:08:33 <AnMaster> ehird, your attempts at trolling fails when you are guessing at what I'm saying
23:08:37 <ehird> mycroftiv: i'm not
23:08:42 <ehird> i'm talking about the fundamental design
23:08:52 <ehird> i.e., if these constraints aren't met, it isn't the OS
23:09:21 <mycroftiv> well, thats why i point to the battlecruiser 3000 example, because its so similar, in terms of the failures of communication between designer and people he was talking to
23:09:43 <mycroftiv> AnMaster: i dunno, i always just google it when it crosses my mind to check in on it
23:10:09 <mycroftiv> it is now, yeah - its day of legendary usenet flame wars was a decade ago or something
23:10:50 <AnMaster> mycroftiv, anyway I'm genuinely interested in how ehird is going to solve the actual implementation on x86/x86_64 while ensuring good performance. And good security.
23:11:43 <mycroftiv> AnMaster: just as i told ehird he shouldnt talk about his os as if it exists yet, i think its also silly for you to expect implementation details given that you know there isnt any code running yet
23:11:51 <AnMaster> mycroftiv, I wonder what will become of DNF then...
23:12:41 <AnMaster> mycroftiv, well surely he must have considered "will this design actually work on hardware that is likely to be available"
23:12:57 <AnMaster> "or do I need to develop my own <ehird OS name>-machine?"
23:13:02 <pikhq> AnMaster: See Smalltalk for more details.
23:14:15 <mycroftiv> AnMaster: Well, you can get whatever chips you want to implement whatever formal behavior you want, thats pretty well established - and I agree with where Fred Brooks ended up about software development:
23:14:43 <mycroftiv> iterate, improve, iterate, improve - always from a working minimal testbed - and do your design 'ahead' of yourself, but constantly with redesign based on the testing and use
23:14:50 <AnMaster> pikhq, I know it is closed world. But you have to consider "will it run well on common architectures" when designing anything. To take an _extreme_ example: Searching memory by using CAM is more efficient than most other search algorithms. Yet it isn't commonly available on most hardware.
23:15:12 <AnMaster> Found in special equipment like network switches and routers mostly
23:15:26 <pikhq> AnMaster: ... But Smalltalk works, and it works sufficiently well for most purposes.
23:15:37 <AnMaster> an OS that depended on it for being efficient would be rather useless for most people
23:16:06 <pikhq> This is like asking if C could be used for an OS...
23:16:06 <AnMaster> pikhq, yes. But does he intend to run it under a host OS hosting the smalltalk implementation forever?
23:16:08 <mycroftiv> AnMaster: anyone doing any work on creating a 'new os' should obviously know as a practical matter most people will never hear of it or care. and thats ok.
23:16:15 <AnMaster> or will it be the "native OS you boot into" at some point
23:16:28 <AnMaster> like the one you select in grub
23:16:38 <AnMaster> (or whatever bootloader you use)
23:17:35 <mycroftiv> AnMaster: ehird has stated clearly that he intends to control the hardware natively and not run simply as hosted 'environment'
23:18:05 <AnMaster> mycroftiv, right. Maybe he did before. But as I mentioned, I have been pretty busy during the days recently
23:18:26 <AnMaster> and I hardly have time to read several hours of fast paced conversation in logs
23:18:53 <AnMaster> so I might have missed details when I was marked away :P
23:19:09 <mycroftiv> i didnt mean for my statement 'ehird has stated clearly...' to carry the implication that you were at fault for not knowing that
23:19:37 <AnMaster> and he intends to code it in small talk?
23:19:46 <AnMaster> won't the small talk runtime system be the "kernel" then
23:20:05 <mycroftiv> not really, he wants to use forth for the lower level components of the system
23:20:13 <AnMaster> since iirc smalltalk is pretty hosted. far from the "portable asm" that C is
23:20:31 <mycroftiv> create a smalltalk implementation hosted within his forth environment
23:21:01 <AnMaster> won't the forth part act as the kernel of the smalltalk part then
23:21:03 <mycroftiv> thats the basic concept he has outlined, which is definitely challenging, but i dont think there is anything at all 'wrong' with it conceptually
23:21:31 <mycroftiv> i cant speak for him, the way the two of you define 'kernel' hasnt been straightened out
23:21:50 <mycroftiv> so i think when you say 'kernel' its fkdjkjaasd to him and when he says 'kernel' its kaoiuweruieu to you
23:22:32 <mycroftiv> you seem to be using an 'operational' definition, in the sense of 'the kernel is whatever code that does x, y, z, so if x, y, z are done, whatever does them, we call a kernel'
23:22:47 <AnMaster> mycroftiv, to me kernel is the lowest level parts of an OS. Mostly those that *must* be run in ring 0 to work. Which on x86 means stuff like task switching and memory management, setting up DMA transfers, managing and handling interrupts
23:23:16 <mycroftiv> AnMaster: yes, and that definition seems to make ehird angry, he seems to regard is as unconventional/incorrect
23:23:28 <AnMaster> mycroftiv, and he referred me to wikipedia
23:23:36 <AnMaster> and that wikipedia page seems to agree with me
23:23:51 <mycroftiv> i have no role as referee as to definitional correctness, i wash my hands
23:24:00 <mycroftiv> i agree with humpty dumpty personally
23:24:04 <AnMaster> "[The kernel's] responsibilities include managing the system's resources (the communication between hardware and software components)"
23:24:10 <mycroftiv> "when i use a word, it means exactly what i want it to mean, neither more nor less"
23:24:35 <mycroftiv> ehird: i was trying to answer some questions, you should verify my statements because i cant speak for you
23:24:56 <mycroftiv> but i believe stuff like 'forth lower layer, smalltalk upper layer' is 'established fact' about your os design now
23:25:47 * ehird unignores AnMaster and reads logs to make sense
23:25:49 <ehird> 15:21:01 <AnMaster> won't the forth part act as the kernel of the smalltalk part then
23:25:56 <ehird> Smalltalk, for instance, handles talking to the keyboard
23:26:06 <ehird> Forth is just for writing the Smalltalk and what's needed for that.
23:26:24 <ehird> 15:22:47 <AnMaster> mycroftiv, to me kernel is the lowest level parts of an OS. Mostly those that *must* be run in ring 0 to work. Which on x86 means stuff like task switching and memory management, setting up DMA transfers, managing and handling interrupts
23:26:36 <ehird> by that definition, both all of my OS and none of my OS and everything in between is the kernel
23:26:47 <ehird> sure you could isolate some bits and say maybe-this-is-the-kernel
23:26:52 <ehird> but it'd be hopeless
23:27:07 <ehird> 15:24:04 <AnMaster> "[The kernel's] responsibilities include managing the system's resources (the communication between hardware and software components)"
23:27:08 <AnMaster> mycroftiv, you can't do certain stuff in userspace on most modern CPU arches. This includes x86, PPC and some more. Userspace being defined as "whatever is the equivalent of ring 3 on the CPU arch". You will get general protection fault for loading the task register while in ring 3 for example on x86.
23:27:10 <ehird> there is no manager in my OS
23:27:15 <ehird> consider the kernel as the government
23:27:20 <ehird> and my OS as a peaceful anarchist commune
23:27:56 <ehird> (depending on your politics, you may have a mental blockage that anything could be peaceful without the implicit global threat of force)
23:28:12 <AnMaster> ehird, are all applications managed, as in running under a VM, like inferno or such
23:28:22 <AnMaster> where VM here would be the smalltalk code I assume
23:28:49 <ehird> if an object has a reference to another object that talks directly to the hardware
23:28:51 <ehird> then it can talk to the hardware
23:28:52 <AnMaster> ehird, what *enforces* the capabilities
23:29:00 <ehird> AnMaster: capabilities don't require "enforcement"
23:30:01 <AnMaster> <ehird> (depending on your politics, you may have a mental blockage that anything could be peaceful without the implicit global threat of force) <-- I don't live in US...
23:30:08 <ehird> AnMaster: loooooool
23:30:16 <ehird> you know how a government derives its authority right?
23:30:31 -!- KingOfKarlsruhe has left (?).
23:30:31 <ehird> there is no such thing as a de jure government, only a de facto one enforced by the threat of violence
23:30:33 <ehird> anyway, technically, an assorted bunch of objects around the system will have methods that run in ring-0, as an implementation detail
23:30:40 <ehird> this does not include drivers
23:30:45 <mycroftiv> government derives its authority from a mandate from the masses, not from some farcical aquatic ceremony
23:30:45 <ehird> drivers just use these objects to talk to the hardware
23:30:58 <ehird> and so you'd have drivers running in userspace according to you
23:31:23 <AnMaster> ehird, ok. I can already see the performance problems ahead on x86. I suspect I know more about this sort of thing on x86 than you do.
23:31:35 <ehird> think what you want
23:31:51 <ehird> you realise that smalltalk as an os is nothing new right
23:31:55 <ehird> that's what the original smalltalk was
23:32:02 <AnMaster> ehird, drivers in user space is certainly possible. But for some stuff it just won't cut it
23:32:18 <ehird> i'm uninterested in your argument by assertion
23:32:39 <pikhq> ehird: Shame that modern CPUs are effectively C machines, isn't it? Makes it a bit hard to imagine anything else.
23:32:50 <AnMaster> overhead of switching between userspace/kernel when handling interrupts from a 10 gbps ethernet card?
23:32:58 <ehird> AnMaster: userspace/kernel?
23:33:23 <AnMaster> and don't even suggest ring 1 or 2 on x86. They are *even* slower to switch to/from
23:33:46 <ehird> taking everything you said as axiomatic, because I can't be arsed to argue: then we run everything under ring 0
23:33:54 <AnMaster> since ring 0<->3 has a "fast path" named SYSCALL/SYSRET and/or SYSENTER/SYSEXIT
23:34:00 <AnMaster> depending on if you ask AMD or Intel
23:34:23 <ehird> i like how you seem to think that an operation on a cpu will somehow not achieve 10gbps?
23:34:31 <pikhq> AnMaster: Actually, there's more than those for ring 0<->3 jumps...
23:34:59 <AnMaster> pikhq, yes there is more to do. I was simplifying for the sake of the discussion
23:35:03 <pikhq> Linux has a function at a fixed memory location that does whatever's fastest, because there's so many ways to do it...
23:35:20 <AnMaster> pikhq, but going to/from rings 1 and 2 is *even* slower
23:35:40 <AnMaster> pikhq, and I know about the vdso :)
23:35:48 <ehird> AnMaster: btw, I'm among good company with my won't-cut-it performance-destroying managed model
23:35:51 <pikhq> AnMaster: Tell that to Xen? :P
23:35:53 <ehird> like microsoft, with their Singularity project
23:36:09 <ehird> it's comforting to know both me and microsoft know less about x86 than AnMaster
23:36:13 <AnMaster> ehird, running everything in ring 0 could work
23:36:23 <pikhq> ehird: Microsoft Research.
23:36:34 <ehird> pikhq: aka the good part of Microsoft
23:36:35 <pikhq> 're the same guys that think Haskell is a decent language. ;)
23:36:41 <pikhq> ehird: Quite right.
23:36:45 <AnMaster> pikhq, Well. That is relative "interpreting/dynamic recompiling in userspace"
23:37:01 <AnMaster> pikhq, then yes using the extra rings is better
23:37:59 <AnMaster> pikhq, iirc virtualbox uses ring 1 too for this. Not sure if that applies when using VT-x/AMD-V though
23:39:37 <pikhq> VT-x/AMD-V is most akin to VM-86. That is, the code runs in ring 0, except with unsafe calls automagically getting shipped on through to the hypervisor.
23:39:43 <mycroftiv> AnMaster: since you mention virtual machines, doesnt that point in the direction of 'not worrying' because I can run a qemu vm with no kqemu on an almost 10 year old freebsd box and put plan9 inside and still have a usable environment?
23:39:43 <ehird> anyway AnMaster, my system absolutely has no *centralised* kernel
23:39:48 <ehird> and that's what a kernel is!
23:39:54 <ehird> sure it has privileged code
23:39:58 <ehird> but a kernel is centralised, that's the definition
23:40:07 <AnMaster> mycroftiv, plan9 doesn't use a lot of resources
23:40:36 <AnMaster> I was thinking more about "server under high load, maybe DDoS" kind of scenario.
23:40:51 <ehird> mycroftiv: not worrying about what?
23:40:53 <ehird> i missed the context
23:41:00 <mycroftiv> im just saying, so far as I can tell, even the 'worst case scenario' of having the slowest possible way of using weak hardware, you can still get something usable out of it
23:41:38 <mycroftiv> ehird: not worrying too much about optimal hardware performance
23:42:02 <AnMaster> ehird, http://en.wikipedia.org/wiki/Capability-based_addressing looks nice, but that requires hardware support, which isn't present on x86 at least.
23:42:04 <ehird> oh you mean the ring0 vs ring3 task switching crap?
23:42:11 <ehird> i ignored that because it's obviously fast enough
23:42:16 <mycroftiv> AnMaster: see, heres my perspective - i started using a computer in 1980, and subjectively, that was still the FASTEST computer ive ever used in terms of time-to-task in many ways
23:42:36 <ehird> you basically never access memory directly anyway
23:42:43 <ehird> so why would I use that?
23:43:00 <ehird> correction: Not in any sane system.
23:43:10 <AnMaster> ehird, do you want to be able to run ported programs written in other languages?
23:43:37 <ehird> No. My OS doesn't have programs, it has objects.
23:43:57 <ehird> Dual-booting or using a VM is infinitely preferable to breaking the model like that, because you'll just end up with something that acts like a VM anyway.
23:44:00 <mycroftiv> definitions, definitions, definitions
23:44:03 <AnMaster> ehird, so you will need to develop your own web browser object replacement instead of creating an object that wraps webkit?
23:44:05 <ehird> If you really need it, POSIX/ELF emulation layer, fully sandboxed.
23:44:24 <ehird> AnMaster: Maybe if I was trying to freakin' make something practical I'd be cloning Linux, have you considered that?
23:44:28 <ehird> mycroftiv: no, fundamental design
23:44:35 <mycroftiv> im gonna make an os where i dont have programs, i dont have objects, i have DELICIOUS CHEESECAKE and ICE CREAM and it will be great
23:44:37 <AnMaster> ehird, I thought something usable was a goal
23:44:52 <AnMaster> as in "this is a OS I will use as my main one in 10 years time"
23:44:53 <ehird> Plan 9 doesn't integrate POSIX apps either.
23:45:00 <mycroftiv> ehird: i dont believe smalltalk and object oriented is actually 'fundamnetal' in the sense you do - i think its a perfetly valid and useful and beautiful set of abstractions
23:45:09 <ehird> mycroftiv: fact is
23:45:13 <mycroftiv> but i dont think it means that a program aint a program any more, its just an objecct
23:45:14 <ehird> a posix program does NOT fit in with my model
23:45:58 <AnMaster> ehird, will you have some sort of window manager for the OS?
23:46:16 <AnMaster> some "this isn't horrible X but it handles the GUI" sort of thing
23:46:35 <ehird> <Leenus Torvalts> I'm working on a hobby OS called Linux
23:46:35 <ehird> <Anne Muster> Does it have a web browser yet?
23:46:49 <ehird> but don't say GUI, there is no command line.
23:46:57 <ehird> well, the bootup forth console could count.
23:48:32 <AnMaster> ehird, I was not saying "has it yet", I was saying "any plans for this"
23:48:37 <AnMaster> please read what I actually said
23:48:46 <ehird> the G is redundant
23:48:49 <ehird> there's just the UI
23:48:55 <AnMaster> I meant about the "Leenus Torvalts" bit :P
23:48:57 <ehird> it happens to require a graphics processor and a colour display
23:49:14 <ehird> AnMaster: window manager is an implementation detail, though
23:49:17 <ehird> who says i'll even use windows?
23:49:30 <AnMaster> ehird, isn't my old vector display enough? WILL I NEED A FRAMEBUFFER?
23:49:38 <ehird> quite so good cheap
23:49:52 <AnMaster> ehird, what would you use instead of windows
23:49:58 <AnMaster> as in, what sort of abstraction
23:50:27 <AnMaster> the abstraction of "windows" while far from perfect seems to be one that "kind of works better than everything else thought of so far"
23:51:00 <ehird> You know that the typical user's workflow consists of a tab bar at the bottom, with applications, and in one of those windows a tab bar at the top, being their pages, right?
23:51:11 <mycroftiv> this has nothing to do with ehird's os (since he doesnt like them) but i believe that we should replace the desktop metaphor with a namespace tree/network map
23:51:13 <ehird> Floating windows aren't used by the average computer user
23:51:24 <ehird> mycroftiv: wait when did i say i didn't like those
23:51:38 <mycroftiv> ehird: you said you didnt like hierarchical file systems
23:51:42 <ehird> well yeah, I don't
23:51:45 <ehird> I don't like trees
23:51:55 <ehird> I like floating things that you can search
23:52:27 <mycroftiv> right, and my visualization is a big 'tree' that shows your file system and you can zoom in and out on the content hanging on the branches, and it also can be zoomed out to show network map of the lan, etc - show the interconnections on the workspace
23:52:27 <AnMaster> <mycroftiv> ehird: you said you didnt like hierarchical file systems <-- that is easily solved. Don't use MacOS
23:52:38 <mycroftiv> not just a blank space full of disconnected icons
23:52:40 <AnMaster> (HFS is short for "hierarchical file system" ;P)
23:53:10 <ehird> Similarly, Mac OS is clearly an operating system for raincoats.
23:53:13 <AnMaster> ehird, you don't like trees? You will hate implementing memory management on x86 then :D
23:53:47 <ehird> I'll just molest x86 until it does what I want.
23:53:55 <ehird> If I have to rewrite all the microcode, so be it.
23:54:10 <AnMaster> ehird, think it is hard wired into the silicon.
23:54:30 <ehird> i'll just find some instructions that heat up the processor so much that it melts
23:54:36 <ehird> and then cool down at just the right time
23:54:38 <ehird> voila! new microcode
23:54:46 <mycroftiv> ok, i meant to get some food 3 hours ago, time to avoid dying of starvation due to os discussions
23:55:05 <ehird> mycroftiv: I WAS THINKING ABOUT MEMORY MANAGEMENT
23:55:12 <ehird> also your inability to eat.
23:57:04 <AnMaster> ehird, memory management is fun. Getting it right is *hard*
23:57:20 <ehird> Doesn't sound fun to me
23:58:22 <AnMaster> ehird, go read some reference manuals for system programming. AMD ones are generally easier to find your way around in than intel ones. Use both for the best result
23:58:42 <ehird> I've been able to avoid any Intel/AMD manuals so far in my life
23:58:51 <ehird> admittedly that makes coding asm nontrivial
23:58:55 <AnMaster> ehird, you won't if coding a full blown OS
23:59:24 <AnMaster> ehird, handing SMP is even funnier
23:59:48 <ehird> please tell me the CPU doesn't have a predefined notion of processes/threads that it uses to do that?
23:59:52 <ehird> please tell me it's lower-level ;_;