00:00:10 * augur gives ehird a lollipop.
00:01:09 <augur> its one of those cream lollipops that are like .. oranges and cream/peaches and cream/strawberries and cream/etc.
00:01:17 <augur> its also in the shape of a penis.
00:01:24 <augur> put it in your mouth, ehird.
00:01:30 <augur> put the creamy penis lollipop in your mouth.
00:01:36 <augur> dont you want it in your mouth?
00:01:56 <ehird> You know, I don't think penises are anything like lollipops aside from being vaguely stick-shaped.
00:02:14 <augur> except when the lollipop is shaped to look like a penis, ehir.
00:02:17 <pikhq> ehird: This one is emphatically a phallic lollipop.
00:02:47 <ehird> You know, I don't think even phallus-shaped lollipops are much like penises.
00:02:59 <augur> not MUCH but enough!
00:03:18 <pikhq> Except when 'lollipop' is being used as a euphemism for phallus.
00:03:24 <ehird> From this we can conclude that the most important aspect of a penis is its shape. If it looks vaguely phallus shaped, then it's a good penis.
00:03:26 <ehird> Otherwise, it's not.
00:04:33 <ehird> Hey, blame augur, not me.
00:05:01 <pikhq> From this we can conclude that Freud was right, and everything can be interpreted as either a phallus or a receptacle thereof.
00:05:28 <oerjan> but what about breasts?
00:05:32 <ehird> Bricks are phalluses.
00:05:38 <ehird> Or are they phallus receptacles?
00:05:40 <Warrigal> Is the number 0 a phallus or a receptacle?
00:05:47 <pikhq> ehird: Breasts are worked in there somehow.
00:05:55 <ehird> I didn't ask about breasts, oerjan did.
00:05:56 <pikhq> Warrigal: What does it look like?
00:06:00 <ehird> Breasts are phallus receptacles if you turn them inside out.
00:06:05 <ehird> Make them hollow first.
00:06:20 <ehird> pikhq: It's both vaguely cigar-shaped and a hole.
00:06:31 <Warrigal> It looks like the point right in the middle of the number line, when centered the obvious way.
00:07:18 <Warrigal> I'm not sure what color it is. White or gray, I think.
00:07:43 <pikhq> ehird: Clearly, it is in a superposition of imagery states.
00:07:59 <pikhq> And can be said to represent unclear genders.
00:12:08 <Warrigal> It's like there are little two little tendrils that move toward each other, but they shy away at the very last moment, leaving the tiniest possible gap in between them.
00:12:36 <Warrigal> So, uh, maybe zero represents conversion to heterosexuality.
00:13:05 <pikhq> Warrigal: Conversion from one to the other, more like?
00:13:36 <Warrigal> One is the same, except that one of the partners is precariously perched so that its head is below its feet. It's still staring the other in the face.
00:14:01 <Warrigal> And the other is lying down on a sort of ramp that slopes upward a bit.
00:15:05 <Warrigal> And e, there's a simple description of e.
00:15:13 <Warrigal> If 1 were a penis, e would be the same penis, erect.
00:17:16 <Warrigal> By that logic, I guess 0 is the lack of any external reproductive organs at all.
00:17:31 <ehird> How is e an erect 1?
00:17:35 <ehird> Wouldn't it be the other way around?
00:21:25 <Warrigal> pi represents masturbation, but in a really silly way.
00:21:59 <Warrigal> Or maybe it represents the female sex, or the act of sex itself.
00:28:31 <EgoBot> EgoBot is a bot for running programs in esoteric programming languages. If you'd like to add support for your language to EgoBot, check out the source via mercurial at https://codu.org/projects/egobot/hg/ . Cheers and patches (preferably hg bundles) can be sent to Richards@codu.org , PayPal donations can be sent to AKAQuinn@hotmail.com , complaints can be sent to /dev/null
00:29:21 <ehird> ALL THESE NICKNAMES
00:29:23 <ehird> THEY MAKE NO SENSE
00:29:42 <AnMaster> ehird, I just was surprised he didn't use his usual
00:29:53 <ehird> A 3 second google suggests that it's an old nickname for someone.
00:30:01 <Gracenotes> IN THE GRIM FUTURE OF HELLO KITTY THERE IS ONLY WAR
00:30:03 <ehird> I postulate that GregorR is too lazy to replace his paypal account.
00:30:10 <ehird> Also, I postulate that he doesn't give a fuck.
00:30:18 <ehird> I postulate a lot of things.
00:31:25 <oerjan> Gracenotes: sickeningly cute war, but nevertheless?
00:31:59 <ehird> http://onastick.net/sitz/images/HK40K.jpg
00:32:26 <oerjan> I postulate that AKAQuinn is actually a terrorist organization which GregorR supports
00:32:35 <ehird> I postulate anuses for a living.
00:33:36 <Gracenotes> so I'm thinking I should watch the Hellsing OVAs
00:36:09 <Gracenotes> and also this ratatat song is reminding of Tubular Bells
00:36:18 <ehird> tubular bells kinda sucks
00:36:43 <Gracenotes> actually most ratatat songs reminds me of Tubular Bells
00:36:55 <Gracenotes> remind. and how can you say that about tubular bells! .;[
00:37:06 <ehird> cuz mike oldfield sucks.
00:37:39 <ehird> erm mike oldfield made tubular bells
00:38:00 <AnMaster> GregorR, there? A hg bundle for you!
00:39:07 <Gracenotes> your attack on the composer cannot discredit tubular bells
00:39:20 <ehird> it's a shit album Gracenotes
00:40:10 <Gracenotes> seriously, if you had to commit suicide, wouldn't you want to be listening to tubular bells? That's how good of an album it is
00:42:52 <AnMaster> GregorR, I assume you can handle signing by PGP/MIME. If not please tell me so I can resend the mail unsigned :P
00:43:28 <ehird> why on earth does a commit need to be signed
00:43:35 <ehird> it's not like he'd blindly trust all code coming from you
00:44:14 <ehird> ...do you seriously have to ask that?
00:44:24 <AnMaster> ehird, anyway I have it set as default.
00:44:51 <ehird> the time it took to ask >= the time it would take to disable
00:45:03 <AnMaster> ehird, nah, I already sent it by then
00:45:49 <AnMaster> ehird, plus, every commit of cfunge is signed. I have bzr set up that way.
01:00:27 -!- iano has joined.
01:06:38 -!- BeholdMyGlory has quit (Remote closed the connection).
01:27:16 -!- FireFly has quit ("Later").
01:27:32 <Gracenotes> okay ehird, tubular bells is only okay if feel in the mood for it
02:44:04 -!- upyr[emacs] has quit (Read error: 110 (Connection timed out)).
02:44:33 -!- iano has quit.
03:49:45 <Warrigal> How does the initial "h" of the Spanish noun "hermandad", "brotherhood", come from the initial "g" of the Latin "germanitas"?
03:52:35 <oerjan> also, i recall "hijo" comes from "filius"
03:53:03 <Warrigal> And "halcón" from "falco" (genitive "falconis").
03:53:22 <Warrigal> Does Latin have any doubled vowels? If not, I will freely write that "falcoonis".
03:54:12 <oerjan> latin has long and short vowels, not reflected in spelling. it also has doubled ones, like in "filii" (genitive of "filius")
03:55:54 * Warrigal looks up the Spanish word "afiliado", to make sure it exists.
03:56:10 <Warrigal> Not having that word would simply be inexcusable.
03:58:44 <oerjan> also "vacuum", i cannot recall any with "oo"
03:59:19 <Warrigal> u and i are the vowels you would expect to be doubled.
04:00:57 <Warrigal> My mom says, "I wonder if I've got a plenum."
04:01:04 <oerjan> "creem", subjunctive present of "creo", create
04:01:20 * Warrigal ponders Spanish doubled vowels.
04:01:56 * Warrigal notes that words such as "cocinándoos" don't exist, according to his fourth-year Spanish teacher.
04:02:09 <Warrigal> (Which is a good thing, as that word is a pretty evil word.)
04:07:27 -!- Gracenotes has quit (Read error: 60 (Operation timed out)).
04:28:42 -!- Gracenotes has joined.
04:46:32 -!- oerjan has quit ("Good night").
05:20:43 -!- Pthing has quit (Remote closed the connection).
05:20:57 -!- GreaseMonkey has joined.
05:40:32 <lament> spanish double vowels are weird, it's not supposed to have any
05:41:30 <lament> Warrigal: coordinación
05:44:51 <Warrigal> Why isn't that prefix con- instead of co-, anyway?
05:45:51 -!- dbc has quit (Client Quit).
05:46:15 -!- dbc has joined.
05:53:12 -!- Gracenotes has quit (Read error: 104 (Connection reset by peer)).
05:54:54 -!- Gracenotes has joined.
06:12:19 <Warrigal> You know, only a linguist would come up with a writing system involving such things as m̥̄ and h̥₁.
06:12:30 <Warrigal> My font doesn't even have ₁ in it.
06:17:18 <pikhq> Warrigal: Funny, it seems to me to be close to what I'd come up with.
06:17:51 <pikhq> (well, I'd use different graphemes, no doubt, but... It's the most elegant solution, IPA...)
06:18:39 <Warrigal> Would you use the letter h with three different subscript numbers rather than just coming up with three different letters for that sound?
06:19:01 <pikhq> ... You're not discussing IPA.
06:19:42 <pikhq> IPA is what linguists came up with. And it has different characters for different phonemes.
06:19:54 <pikhq> (it also, last I checked, could represent all phonemes)
07:49:36 -!- Sgeo has quit (Read error: 110 (Connection timed out)).
07:52:26 <lament> what's the IPA character for queefs?
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:18:57 <augur> lament: a small m-width character spelling "ehird" in tall thin letters.
10:27:32 -!- FireFly has joined.
10:46:02 -!- Judofyr has joined.
10:54:05 -!- BeholdMyGlory has joined.
11:14:58 -!- GreaseMonkey has quit (Read error: 110 (Connection timed out)).
11:16:59 <AnMaster> GregorR, I can't enter initial space?
11:33:32 <fizzie> I would guess some part of it trims the space between !command and argument.
11:34:14 <AnMaster> !befunge98 aaaaaaaaa********.a,@
11:34:43 <AnMaster> (yes it can be sorter, just checking I got the number right)
11:36:10 <AnMaster> !befunge98 aa*:*aa*:*a**** .a,@
11:36:19 <AnMaster> !befunge98 aa*:*aa*:*a*** .a,@
11:36:47 <fizzie> !befunge98 a::*:**.a,@
11:37:26 <fizzie> !befunge98 a::*:*:**.a,@
11:37:35 <fizzie> I always forget to print.
11:38:29 <fizzie> Yours has an extra * at the end.
11:38:36 <fizzie> !befunge98 aa*:*:*a* .a,@
11:39:16 <fizzie> !befunge98 'd:*:*a* .a,@
11:40:12 <fizzie> Don't think it will go much shorter than that, since it's just "(100^2)^2*10" which is very simple already.
11:42:50 <fizzie> !haskell 5*(100^2)^2*10-2^32
11:43:21 <AnMaster> fizzie, hm a question, to reverse a text for use in befunge rev(1) is useful. Like producing something like ".elbahcaernu eb dluohs sihT :DAB"
11:43:28 <AnMaster> but what about producing a vertical one?
11:43:34 <AnMaster> is there any pre-written app for it?
11:46:18 <fizzie> I don't know of one. Something like perl -ne 'print "$_\n" foreach split //;' seems to work, with a ... split //, reverse($_); variant if one wants the string the other way around. Unless you want a multi-line piece of text into a multi-column vertical version, but that doesn't sound so very useful either.
11:46:48 <AnMaster> fizzie, what about non-cardinal deltas?
11:47:13 <fizzie> But feel free to programmize an arbitrary-delta string-outputter.
11:47:14 <AnMaster> anyway, it needs to handle " " -> " ":
11:47:56 <fizzie> A good Befunge IDE should be able to do it.
11:48:00 <AnMaster> fizzie, going to make a cardinal one first that also handles the SGML spaces and newlines (adding ..."a"... where needed)
11:48:25 <ais523> AnMaster: there's an Emacs mode for typing in the orthogonal and diagonal directions
11:48:28 <fizzie> Well, a hypothetical good Befunge IDE. I'm not aware of any.
11:48:37 <fizzie> Yes, I almost wrote "I'm sure Emacs can do it".
11:49:17 <fizzie> I think some of the Befunge systems with editor included also let you set the delta, though possibly only to <>v^ if it was a -93 one, which sounds likely.
11:50:49 <AnMaster> ais523, hm. What mode is that?
11:54:18 <fizzie> Aw, even in picture-mode the mark/point-region highlight is still line-based; I was hoping for an actual highlight rectangle, like you get for ^v in Vim.
11:56:23 <AnMaster> ^v <-- avoiding thread collisions? ;P
11:56:34 <AnMaster> ooh that might be a fun variation of concurrent befunge
11:57:33 <AnMaster> if an IP is hit from the side for example it's course might alter according the *roughly* physical laws.
11:57:54 <AnMaster> head on? Hm I wonder how elastic IPs are. Either they bounce or they come to a dead stop
12:02:31 <fizzie> Ooh. "If you use both transient-mark-mode and picture-mode, you will probably realize how convenient it would be to be able to highlight the region between point and mark as a rectangle. Have you ever wished you could see where exactly those other two corners fell before you operated on a rectangle? If so, then this program is for you."
12:36:56 -!- KingOfKarlsruhe has joined.
13:06:59 -!- MigoMipo has joined.
13:17:49 -!- augur has quit (Read error: 110 (Connection timed out)).
13:18:29 <ehird> 10:43 AnMaster: is there any pre-written app for it?
13:18:30 <ehird> 10:43 AnMaster: if not I guess I'll write one
13:18:31 <ehird> it is after all a heavy duty task...
13:21:29 <ehird> ais523: being a microchip sorta persony thingy, what do you think of http://colorforth.com/vlsi.html? it's chuck moore's chip design to production environment from scratch written in 500 lines of colorforth, and there's actually a mini supercomputer (9 chips with 40 computers, so 360 computers, running at native silicon speed, using 1W total: http://colorforth.com/S40.htm and http://colorforth.com/haypress.htm) and is being used to design the http://colo
13:21:32 <ehird> rforth.com/GA.htm chips (more "massively parallel, native silicon, low power usage" stuff)?
13:21:41 <ehird> i mean, aren't most chip design environments huge? :D
13:22:09 <ais523> they don't need to be huge, but all the major ones are controlled by companies with the Visual Studio mentality
13:22:51 <ehird> ais523: still - from scratch to producing a tiny supercomputer in 500 lines? (remember that colorforth lines are, say, 40 chars are less)
13:23:15 <ais523> you can do a simple microprocessor in about 100 lines of VHDL
13:23:17 <ais523> and VHDL is really verbose
13:23:29 <ehird> ais523: oh, that's not at all what it is
13:23:39 <ais523> ah yes, a design environment
13:23:48 <ehird> it's like implementing vhdl
13:23:54 <ehird> and then giving it a design UI to boot
13:24:55 <ehird> ais523: i just can't imagine it being so small and yet able to do such a thing so easily
13:25:15 <ehird> so much for AnMaster's "yeah but colorforth isn't usable for anything practical" iguess :)
13:25:17 <ais523> people have written C compilers in the IOCCC's byte limit before
13:25:38 <ehird> ais523: i would have thought a complete chip design environment good enough to design a supercomputer would be quite large
13:25:58 <ais523> I wonder how good the resulting supercomputer is?
13:26:00 <ehird> a choice quote from the page, on the layout step: "This requires megabytes of RAM."
13:26:03 <ehird> ais523: i gave you links
13:26:11 <ais523> yep, which to me implies that it's doing layout by bruteforcing
13:26:18 <ais523> admittedly, that's how the commercial systems do it too
13:26:30 <ehird> ais523: the blog shows him getting a lot of results, though
13:26:44 <ehird> ais523: and them being desynchronized in speed because they don't have clocks is cool
13:26:54 <ehird> chuck moore is awesome
13:27:12 <ehird> i wonder how come it only uses 1w? ("It's providing 1.8 V on Supply A with some huge current capability. I've long lost the user's manual. As long as it gives me 500 mA, I'm fine for Haypress Creek. Hey DARPA, here's your 1 W supercomputer!")
13:27:17 <ais523> actually, what shocks me is how limited the commercial systems are for how bloated they are
13:27:31 <ais523> and it's probably relatively slow if it has a power consumption that low
13:27:36 <ais523> just due to the laws of physics
13:27:42 <ehird> ais523: "Total of 360 computers running at 700 Mips or 250 Gips."
13:27:55 <ehird> doesn't sound slow to me
13:28:04 <ais523> that's slow for a supercomputer, isn't it?
13:28:13 <ehird> ais523: it's a mini supercomputer. it's 8x8
13:28:31 <ehird> ais523: it's about as fast as a graphics card, I think; except standalone, and you could daisy chain them
13:28:46 <ehird> get 200 and you have something way more powerful than a graphics card with the same power consumption
13:28:54 <ehird> ais523: the idea is to be an _embedded_ supercomputer
13:29:08 <ais523> and I think he could pull it off
13:29:10 <ehird> same thing that greenarrays: http://colorforth.com/GA.htm are doing
13:29:20 <ehird> (considering that's the continuation after the company f'd him over)
13:29:34 <ehird> "The GA4 uses a few milliwatts while running and a few microwatts asleep."
13:29:40 <ehird> 4 computers, 2mm^2
13:30:13 <ehird> ais523: it seems that the main thing about the lower poer is "Low power results from our computer (one node of an array) being asynchronous (unclocked)"
13:30:38 <ais523> aha, so it can slow down to conserve power when not running at 100% CPU
13:30:55 <ehird> ais523: from what i've read, they're saying it simply doesn't have a clock
13:31:00 <ehird> and runs however fast the electrons want it to
13:31:03 <ais523> and it's very rare for all the nodes of a supercomputer to be at full CPU all the time, as they have to wait for memory
13:31:11 <ehird> ofc i could be vastly misinterpreting
13:31:22 <ais523> asynchronous design has a sort-of clock
13:31:34 <ais523> it's sort of a chain of one signal to the next, rather than a regular clock sent everywhere
13:31:54 <ais523> it's a lot more hectic and incoherent than a synchronous design
13:32:24 <ehird> i wish Moore and Knuth weren't so old (same age, actually)... TAOCP will probably never be finished :p
13:34:29 <ehird> http://secamlocal.ex.ac.uk/people/staff/mrwatkin/zeta/caveney.txt // i'm convinced
13:34:41 <lament> i wish YOURMOM wasn't so old!
13:41:15 <AnMaster> <ehird> so much for AnMaster's "yeah but colorforth isn't usable for anything practical" iguess :) <-- I never claimed that... What I might have said is that it is impractical for general purpose programs running under a normal OS. IIRC.
13:41:37 <ehird> that may have been what you meant; that's not what you said.
13:41:52 <AnMaster> ehird, for an embedded environment it might make sense
13:42:23 <ehird> he designed the chips on a regular computer.
13:42:32 <ehird> you know. a regular system.
13:42:41 <ehird> AnMaster: but feel free to reimplement it in, say, 500 lines of C
13:42:42 <AnMaster> ehird, an embedded system might have a monitor though. Just think of mobile phones,
13:42:44 <ehird> like, just call me back
13:42:53 <ehird> (great, I've finally got rid of AnMaster for good)
13:43:13 <AnMaster> ehird, iirc that colorforth implementation I saw ran on "bare hardware"
13:43:26 <AnMaster> as opposed to running under, say, linux, windows, *BSD or whatever
13:43:32 <ehird> it's an operating system
13:43:52 <AnMaster> ehird, yes. But a rather primitive such currently at least.
13:43:59 <AnMaster> with primitive here I mean hardware support
13:44:10 <ehird> as i said, http://colorforth.com/vlsi.html
13:44:22 <ehird> 500 lines of C where the lines are 40 characters or less
13:44:41 <AnMaster> ehird, but stuff like DMA for disks, good ethernet support for most common on-board ones and PCI ones and so on
13:45:09 <ehird> AnMaster: let me make sure I understand your statement.
13:45:12 <AnMaster> not supporting 3D acceleration is not included in this set of course.
13:45:43 <ehird> <AnMaster> ColorForth is useless for designing microchips because it doesn't have good ethernet support (which I assert without proof or source, presumably finding it obvious whether it's true or not).
13:45:49 <AnMaster> ehird, as a general purpose OS that any random user could use on his/her pc colorforth is rather lacking.
13:46:18 <ehird> so it turns out that talking to a brick wall is more productive than talking to AnMaster.
13:46:22 <AnMaster> ehird, how where they the same thing?
13:47:01 <AnMaster> ehird, you claimed "<ehird> so much for AnMaster's "yeah but colorforth isn't usable for anything practical" iguess :) <AnMaster> <ehird> so much for AnMaster's "yeah but colorforth isn't usable for anything practical" iguess :) <-- I never claimed that... What I might have said is that it is impractical for general purpose programs running under a normal OS. IIRC. <ehird> he designed the chips on a r
13:47:01 <AnMaster> egular computer." <-- and I'm clarifying what I meant exactly there.
13:47:10 <AnMaster> ehird, you are the brick wall here
13:47:18 <AnMaster> who continues talking about this specific case
13:47:30 <AnMaster> even when the discussion shifted.
13:48:23 <AnMaster> ehird, yeah, with your mentality it is more like you having a monologue...
13:49:18 <AnMaster> ehird, anyway, I hope you now understand what part I meant colorforth was lacking in. But if you disagree... why not switch from OS X to colorforth right now? Install it on your mac alongside OS X...
13:49:24 <ehird> i wonder when AnMaster will realise i never actually talk to him with an intent to discuss because he has a piece of wetware that puts itself into an eternal feedback loop after N time
13:49:43 <ehird> ooh, i see AnMaster's figured out how to use the strawman fallacy
13:51:09 <AnMaster> ehird, I'm not saying colorforth is useless for everything. Just not as a general purpose OS. That meaning something like linux, windows, OS X, *BSD and even plan9 (plus a few more)
13:51:25 <AnMaster> and that is what I said all along
13:51:44 <ehird> you're claiming that plan 9 works as a general purpose os where colorforth doesn't?
13:51:49 <ehird> either you understand one or neither.
13:51:57 <ehird> *misunderstand one or both
13:52:10 <AnMaster> ehird, in theory, it could need a bit more drives and such to support modern hardware
13:53:18 <AnMaster> it has quite decent support for a bit older hardware. But when I last looked (about a year ago or so iirc) colorforth wasn't even close
13:56:23 <AnMaster> ehird, porting a posix program to run under plan9 would probably be way easier than doing the same for colorforth for example. And that suddenly means there are lots of apps you could port with a moderate amount of work. Wikipedia even claims there is some "linuxemu" being worked on to run linux binaries under plan9.
14:12:51 <ehird> "Plus, if you generate more energy than you use, the electricity company will pay you!"
14:12:55 <ehird> http://www.power4home.com/index.php?hop=ttinc1
14:13:09 <ehird> I especially like the juxtaposition of the solar panels and a ton of snow.
14:13:15 <ehird> well okay not a ton of snow whatever
14:31:58 -!- upyr[emacs] has joined.
14:45:09 <ehird> Hey, a fysh.org link.
14:45:12 <ehird> Not Zefram, though.
14:53:52 <fungot> Selected style: agora (a large selection of Agora rules, both current and historical)
14:53:55 <ehird> fungot: Punish comex.
14:53:55 <fungot> ehird: this rule. this rule takes precedence over any rule which would cause that entity
14:53:57 <ehird> fungot: Punish comex.
14:53:58 <fungot> ehird: the damages from the
14:53:59 <ehird> fungot: Punish comex.
14:54:00 <fungot> ehird: there is an order. it shall be its judge by announcement. quorum for an appelate judge is recused, and an indication is a contract
14:54:05 <ehird> fungot: Punish comex.
14:54:05 <fungot> ehird: ii) one second before the promotor may make agreements among themselves with the clerk of the organization would require
14:54:08 <ehird> fungot: Punish comex.
14:54:19 <ehird> `echo fungot: butt
14:54:19 <fungot> ais523: ( d) the identity of the group itself, must specify the currency holdings contained in the
14:54:19 <fungot> ehird: as president of the situation at hand with respect to any
14:54:31 <ehird> fungot: Cunish pomex.
14:54:31 <fungot> ehird: the ambassador to conduct auctions for positions in the order of succession. the author), in violation of one or more
14:54:34 <ehird> fungot: Cunish pomex.
14:54:34 <fungot> ehird: as soon as possible after this rule defers to them by the players ( plus 1 for 10 days ( or failure to perform
14:54:36 <ehird> fungot: Cunish pomex.
14:54:37 <fungot> ehird: there shall exist the office is filled ( held) by announcement ( invalid unless the proposal. whenever a player puts self back on active status by writing a message to the
14:54:48 <ais523> are any of those reasonably recognisable as a punishment?
14:54:59 <ehird> the ambassador to conduct auctions for positions in the order of succession. the author), in violation of one or more as soon as possible after this rule defers to them by the players ( plus 1 for 10 days ( or failure to perform
14:55:06 <ehird> you can pastiche and make stuff up :P
15:25:33 <AnMaster> $ ~/bin/btext '>' $'Testing this thing yay spaces newline\n'
15:25:33 <AnMaster> a"enilwen secaps":::" yay"::" gniht siht gnitseT"
15:25:52 <AnMaster> !befunge98 a"enilwen secaps":::" yay"::" gniht siht gnitseT">:#,_@
15:25:53 <EgoBot> Testing this thing yyyay sssspaces newline
15:27:15 -!- ais523 has quit (Remote closed the connection).
15:32:47 -!- ais523 has joined.
15:37:26 <ehird> “We’ve reviewed Instapaper Pro and determined that we cannot post this version of your iPhone application to the App Store at this time because it is not appropriately rated. Our review indicates that the application content is not consistent with the current rating. Instapaper Pro allows unfiltered access to the internet, where content with mature or suggestive themes can be accessed. Applications must be rated accordingly for the highest level of c
15:37:30 <ehird> ontent that the user is able to access.”
15:37:53 <ehird> I wonder when Apple will cause a mature content warning to appear when you start Safari.
15:48:17 <ehird> it's a bookmarking app or something
15:48:26 <ehird> it lets you USE THE INTERNEEEEEEEEEEEEEET
15:48:45 <ehird> <ehird> hmm, looks like north korea have bombed south korea <AnMaster> ehird: just live in Sweden :P
15:49:06 <ehird> ...and other nominees for the "worst analogy of the week" award
15:50:05 <AnMaster> ehird, iphone is like a text book example of vendor lock-in though
15:50:31 <ehird> anyway i bought my iphone before all that
15:50:38 <AnMaster> ehird, is that the name for moding it so you can change carrier?
15:52:51 <ehird> i bought it when you couldn't even put any fuckin' native code on it unless you jailbreaked
15:52:51 <ehird> AnMaster: no, that's unlocking
15:52:51 <ehird> jailbreaking is letting you run unsigned code etc
15:52:52 <ehird> the iphone is an excellent alarm clock and portable web browser/irc client :-P
15:53:07 <ehird> 15:50 AnMaster: ehird, is that the name for moding it so you can change carrier?
15:53:07 <ehird> 15:50 ehird: i bought it when you couldn't even put any fuckin' native code on it unless you jailbreaked
15:53:09 <ehird> 15:50 ehird: AnMaster: no, that's unlocking
15:53:11 <ehird> 15:50 ehird: jailbreaking is letting you run unsigned code etc
15:53:13 <ehird> 15:51 ehird: the iphone is an excellent alarm clock and portable web browser/irc client :-P
15:53:15 <ehird> 15:52 AnMaster: or something else?
15:53:48 <AnMaster> ehird, err... not that order here...
15:53:52 <AnMaster> <ehird> anyway i bought my iphone before all that
15:53:52 <AnMaster> <AnMaster> ehird, is that the name for moding it so you can change carrier?
15:53:52 <AnMaster> <ehird> i bought it when you couldn't even put any fuckin' native code on it unless you jailbreaked
15:53:54 <ehird> Thus "that's some lag".
15:54:04 <AnMaster> ehird, yes. over a minute there for a while
15:54:08 <ehird> 15:50 ehird: AnMaster: no, that's unlocking
15:54:08 <ehird> 15:50 ehird: jailbreaking is letting you run unsigned code etc
15:54:10 <ehird> 15:51 ehird: the iphone is an excellent alarm clock and portable web browser/irc client :-P
15:54:37 <ehird> they reduced the price from like $500 to $300 a few months later
15:54:39 -!- Judofyr has quit (Read error: 110 (Connection timed out)).
15:54:41 <AnMaster> ehird, "<ehird> i bought it when you couldn't even put any fuckin' native code on it unless you jailbreaked" to "<ehird> the iphone is an excellent alarm clock and portable web browser/irc client :-P" all arrived the same second
15:55:00 <AnMaster> ehird, temporary lag spike I guess...
15:55:32 <ehird> http://blog.cocoia.com/2009/pocket-monster-brands/ ← "Here we go: corporate brand design, the Pokémon edition."
15:55:43 <AnMaster> since it didn't happen on other networks I guess it was due to server side issues
15:55:53 -!- oerjan has joined.
15:55:55 <AnMaster> or backbone problems or whatever
15:56:55 <AnMaster> oerjan, seems IWC didn't update today.... During normal time for update browser said connection timed out
15:57:20 <AnMaster> now that is some bad handling scheduled events that happened during downtime...
15:58:28 <oerjan> he's always corrected those manually i think
15:58:49 -!- FireFly has quit ("Later").
16:00:29 <AnMaster> oerjan, on_startup { for each scheduled event during(previous timestamp, now) where job_run == false { run_it_now_dammit() }
16:00:41 <ehird> "Lentic Catachresis" isn't a very senseful title. Misapplication of a word meaning still waters? What?
16:00:50 <ehird> *"of still waters", rather.
16:02:41 -!- Judofyr has joined.
16:02:58 <oerjan> yeah, that sounds like misapplication of those words ;D
16:04:38 <ehird> Wow. So catachresis can be contextually homoapplicative?
16:04:55 <ehird> Autechre are amazing linguists.
16:07:46 -!- Pthing has joined.
16:10:19 -!- KingOfKarlsruhe has left (?).
16:15:05 -!- FireFly has joined.
16:17:39 -!- FireFly has quit (Client Quit).
16:24:06 <oerjan> AnMaster: hm the poll page isn't loading either
16:24:21 <AnMaster> oerjan, oh? send an email or something I usually ignore the poll
16:25:14 <oerjan> i'm sure he'll notice soon enough
16:25:24 <ehird> he couldn't possibly check his own website
16:25:56 <oerjan> i am more thinking about the fuss from people on the forums, actually :D
16:26:19 <oerjan> well, plus everyone _else_ who sends emails
16:27:01 <ehird> http://www.irregularwebcomic.net/ is down anyway.
16:28:53 <oerjan> i guess it's the same error then
16:29:38 <oerjan> maybe he finally got reddited or something
16:30:00 <oerjan> (or again, don't recall)
16:30:05 <ehird> someone should make a compilation album where a bunch of artists cover 4'33"
16:30:19 <ehird> make it 43 minutes and 30 seconds long
16:31:00 <ehird> i'm thinking like a punk band's version being 30 seconds long and then an ambient drone one that's 10 minutes
16:31:51 <AnMaster> <ehird> someone should make a compilation album where a bunch of artists cover 4'33" <ehird> make it 43 minutes and 30 seconds long <-- one of them only half-covered?
16:31:59 <ehird> 16:31 ehird: i'm thinking like a punk band's version being 30 seconds long and then an ambient drone one that's 10 minutes
16:32:09 <ehird> artistic license, dood.
16:33:17 <AnMaster> Am I still connected? No ping reply from oerjan yet...
16:33:17 <AnMaster> Oh there it came. 20 seconds...
16:33:28 <AnMaster> ehird, ah right, missed that line.
16:33:39 <AnMaster> irritating with temporary lag...
16:50:04 -!- FireFly has joined.
17:09:42 -!- FireFly has quit ("Later").
17:13:05 -!- FireFly has joined.
17:43:01 -!- oerjan has quit ("leaving").
18:01:01 <AnMaster> if anyone is interested in that formatter script (not yet for flying deltas, but handles 98-spaces and newlines): http://paste.lisp.org/display/83792
18:01:16 <AnMaster> (I thought I pasted it much earlier, then noticed I didn't)
18:01:35 <ehird> isn't that just like 3 emacs modes
18:01:51 <AnMaster> ehird, um. "but handles 98-spaces and newlines"
18:03:07 -!- Sgeo has joined.
18:03:38 <pikhq> Note: the piece 4'33" does not need to be 4'33". ;)
18:04:10 <Sgeo> It doesn't need to be 4 feet 33 inches
18:04:29 <ehird> the text of 4'33" actually notes the time, iirc
18:04:36 <pikhq> Sgeo: ' and " are also units of time (and sub-degree units).
18:04:44 <ehird> but it's easy to do an interperative cover :P
18:04:57 <pikhq> It's been done before, ehird.
18:05:12 <ehird> i'm just saying that 4'33' does actually require it to be 4:33
18:05:16 <ehird> if you play it strictly
18:07:57 <ehird> The original Woodstock manuscript (August 1952): conventional notation, dedicated to David Tudor. This manuscript is currently lost. Tudor's attempt at re-creating the original score is reproduced in Fetterman 1996, 74.
18:07:58 <ehird> The Kremen manuscript (1953): graphic, space-time notation, dedicated to Irwin Kremen. The movements of the piece are rendered as space between long vertical lines; a tempo indication is provided (60), and at the end of each movement the time is indicated in minutes and seconds.[22] Edition Peters No. 6777a.
18:08:02 <ehird> The so-called First Tacet Edition: a typewritten score, lists the three movements using Roman numbers, with the word "TACET" underneath each. A note by Cage describes the first performance and mentions that "the work may be performed by (any) instrumentalist or combination of instrumentalists and last any length of time." Edition Peters No. 6777 (out of print).
18:08:07 <ehird> The so-called Second Tacet Edition: same as the First, except that it is printed in Cage's calligraphy, and the explanatory note mentions the Kremen manuscript. Edition Peters No. 6777 (i.e. it carries the same catalogue number as the first Tacet Edition)
18:08:32 <Sgeo> Dear Windows: Please stop betraying me
18:08:42 <ehird> In 1962, Cage wrote 0'00", which is also referred to as 4'33" No. 2. The directions originally consisted of one sentence: "In a situation provided with maximum amplification, perform a disciplined action." The first performance had Cage write that sentence.
18:08:43 <ehird> The second performance added four new qualifications to the directions: "the performer should allow any interruptions of the action, the action should fulfill an obligation to others, the same action should not be used in more than one performance, and should not be the performance of a musical composition."[10]
18:08:46 <Sgeo> Dear Firefox, please make this the last time you load slowly
18:08:50 <ehird> Sgeo: the point of windows is to betray you.
18:10:35 -!- Judofyr has quit (Remote closed the connection).
18:11:10 -!- Sgeo has quit (Read error: 104 (Connection reset by peer)).
18:12:37 -!- Sgeo has joined.
18:12:54 <Sgeo> XChat crashed around the same time that Firefox loaded
18:14:03 <ehird> 18:08 ehird: Sgeo: the point of windows is to betray you.
18:14:23 <AnMaster> hm? Xchat crashed? That never happened when I used it.
18:14:31 <Sgeo> Well, Silverex
18:15:07 <AnMaster> Sgeo, that may be different. Haven't used windows for a long time outside a VM. and even inside a wm was quite long ago.,
18:16:03 <Sgeo> Hm, should try AW in Win98 in VirtualBox at some point
18:16:48 <GregorR> Sgeo: Xchat and Firefox aren't friends.
18:17:20 <pikhq> Firefox just isn't very friendly to people.
18:17:57 <Sgeo> Therefore, XChat is a people.
18:18:22 <GregorR> XCHAT IS PEEEEEEEEEOOOOOOOOOOOPLEEEEE
18:19:45 <AnMaster> if you are already on windows I mean
18:19:57 <Sgeo> So I don't have to be on Windows to run it
18:20:11 <AnMaster> Sgeo, what is this "Active Worlds" thingy?
18:20:47 <Sgeo> It's a 3d worlds thingy similar, but different, to SL. It's much older than SL
18:21:08 <pikhq> I remember using it in '97.
18:21:14 <AnMaster> Sgeo, weren't you a SL fan recently?
18:21:24 <AnMaster> what happened to that. haven't heard you mention it for a bit
18:21:25 <Sgeo> I can't be a fan of more than one thing?
18:21:34 <Sgeo> I tend to lose interest in things randomly
18:21:44 <Sgeo> Then regain interest eventually, then lose interest.
18:21:52 <AnMaster> right, I know what you mean about losing interest
18:26:40 * Sgeo is planning on making a video of a game that died in 2005
18:53:12 <AnMaster> didn't someone I wrote very linear befunge code in here once?
18:54:18 <AnMaster> Hm. I think he/she should take a look at the lower parts of http://pastebin.ca/1499346
19:11:14 <ehird> i'm going to make some sort of toy actual-game with js/html5/canvas.
19:11:25 <ehird> maybe a super mario clone. those are easy. :P
19:12:15 <AnMaster> ehird, what about that game with jewels falling down and you have to drag them around to get 3 in a row
19:12:29 <ehird> you mean like bejeweled?
19:12:42 <ehird> good idea actually.
19:13:00 <AnMaster> bejeweled ? aspell thinks it should be bejewelled hm
19:13:24 <AnMaster> http://en.wikipedia.org/wiki/Bejeweled though
19:13:42 <AnMaster> so that actual name of this game. But aspell wouldn't have that
19:14:12 <ehird> popcap games are irritatingly addictive
19:14:23 <ehird> makers of bejeweled, peggle etc
19:15:07 <AnMaster> anyway, I always disliked bejeweled because of the short time limit. You can't spend time planning your next move
19:15:10 <ehird> actually, doing bejeweled would be pointles
19:15:13 <ehird> you don't even need canvas
19:15:17 <Sgeo> I first played a Bejeweled clone on BYOND
19:15:19 <ehird> AnMaster: there are infinite-time ones
19:15:32 <Sgeo> Didn't even know it was a clone grr
19:15:36 <AnMaster> ehird, mhm, will look next time
19:15:46 <AnMaster> ehird, you don't need canvas for it
19:15:50 <AnMaster> you can just move images around
19:15:55 <ehird> AnMaster: not really
19:15:57 <ehird> canvas is smoother
19:16:07 <ehird> also, i'd still use html5
19:16:14 <ehird> Bleep. Bop. Boing.
19:16:16 <AnMaster> ehird, what about doing some 3D demo with canvas?
19:16:34 -!- coppro has joined.
19:16:38 <ehird> 3d in canvas = dog slow. canvas3d = well they've practically built my demo in then haven't they
19:16:57 <AnMaster> ehird, like a software ray tracer at 5 fps rendering two bouncing balls or something
19:17:11 <AnMaster> ehird, meh there is a "canvas3d"? :(
19:17:36 <AnMaster> ehird, well I meant a sofware ray tracer demo sort of thingy :P
19:17:42 <ehird> been done, and way too slow.
19:17:46 <ehird> ray tracing hasn't been done
19:17:49 <ehird> are you fucking nuts?
19:17:57 <ehird> AnMaster: the best hardware ray tracers get 5fps
19:18:00 <AnMaster> ehird, software ray tracing demo. haven't you seen some?
19:18:02 <ehird> and you want me to do it in javascript?
19:18:13 <AnMaster> ehird, I want to see how many fps you mange to get
19:18:21 <AnMaster> use chrome or something to get decent js speed
19:18:24 <pikhq> ehird: Ray tracing has been done.
19:18:42 <ehird> (safari 4 beta was already faster; the final is even moreso)
19:18:46 <pikhq> I believe those guys with the freaking FPGA ray tracer got 30FPS.
19:18:57 <pikhq> Granted, it was only at like 640x480, but.
19:19:03 <Sgeo> Does O3D count as a canvas3d?
19:19:03 <ehird> on complicated models
19:19:25 <Sgeo> Or does canvas3d refer to something else?
19:19:31 <pikhq> ehird, that's the speed of software ray tracing...
19:19:34 <AnMaster> pikhq, link to this being done in javascript? I mean, ray tracing
19:19:43 <pikhq> Oh, in Javascript?
19:20:02 <AnMaster> pikhq, to do it on a canvas in HTML5 with javascript!
19:20:17 <pikhq> Where's Gregor's Javascript ray tracer... :P
19:20:22 <ehird> pikhq: I'm talking high-quality models here.
19:20:30 <ehird> pikhq: As in, professional raytraced image quality.
19:20:33 <ehird> Final product quality.
19:20:33 <pikhq> ehird: Define 'high-quality'.
19:20:46 * Sgeo thinks a canvas3d API should be roughly compatible with O3D apis, such that if O3D becomes a wide-spread standard, it's easy to take old work and just have it use O3D
19:20:48 <ehird> That's Caustic's game.
19:21:11 <ehird> Sgeo: would you shut up? canvas3d already exists
19:21:13 <ehird> it's not a hypothetical
19:21:13 <Sgeo> AnMaster, a plugin by Google that allows Javascript to make 3d stuff
19:21:16 <ehird> https://wiki.mozilla.org/Canvas:3D
19:21:41 <Sgeo> Oh, so they're basically competing...
19:21:51 <Sgeo> Either Mozilla or Google should be slapped. I vote Google.
19:21:55 <ehird> http://www.zazzle.com/three_wolves_howling_at_the_moon_tshirt-235961755604154510
19:22:05 <GregorR> I assume canvas3d is just a 3D extension to Apple's canvas?
19:22:09 <pikhq> ehird: Are you reading reports from the 90s?
19:22:11 <ehird> GregorR: "apple's canvas"
19:22:18 <ehird> GregorR: You mean everyone's by now.
19:22:24 <ehird> pikhq: Dude, this is from THEIR PROMOTIONAL MATERIAL.
19:22:25 <GregorR> Sure, but it was created by Apple.
19:22:34 <AnMaster> http://en.wikipedia.org/wiki/ISO/IEC_8859-1#ISO-8859-1_and_Windows-1252_confusion says "However, the draft HTML 5 specification requires that documents advertised as ISO-8859-1 actually be parsed with the Windows-1252 encoding."
19:22:41 <ehird> pikhq: The point is high-resolution.
19:22:47 <ehird> You can edit ray traced images at their final quality at 5-10fps.
19:22:52 <AnMaster> ehird, how can you love HTML5 with THAT!
19:22:54 <ehird> As in, what you see is what you'll get, exactly.
19:23:09 <ehird> AnMaster: because the point is that you can use HTML5's algorithms and render every existing page on the web as good as current browsers do
19:23:16 <ehird> that's why its parsing is so elaborate, too: it handles -everything-
19:23:32 <ehird> an HTML5 implementation will Just Work with every page in existence
19:23:51 <Sgeo> http://www.tapper-ware.net/canvas3d/ this isn't what's meant by canvas3d, is it?
19:23:53 <AnMaster> ehird, it will fail on a page sending the correct encoding...
19:24:05 <pikhq> It's processor-intensive (4-socket, 4-core system), but Enemy Territory: Quake Wars can run using software ray tracing for rendering.
19:24:10 <ehird> AnMaster: "whereas they are control codes in ISO-8859-1"
19:24:16 <ehird> Nobody puts control codes in a web document.
19:24:22 <AnMaster> ehird, so what encoding do you send in HTML5 when you actually mean "ISO-8859-1" and wants those control codes...
19:24:27 <ehird> pikhq: Uhh, lol, no.
19:24:31 <ehird> pikhq: They did that with a CLUSTER.
19:24:35 <ehird> It was 24 machines or something.
19:24:40 <ehird> AnMaster: you don't.
19:24:47 <ehird> AnMaster: because that's useless, stupid and never happens
19:24:49 <ehird> pikhq: no, they used a cluster
19:24:52 <ehird> i looked into it a while back
19:25:03 <pikhq> STOP ARGUING AGAINST FACTS, BITCH.
19:25:09 <ehird> They used a cluster.
19:25:14 <pikhq> When did you read this?
19:25:32 <ehird> the page describing it.
19:25:40 <pikhq> I said when, not where.
19:25:50 <ehird> maybe a month or two
19:25:56 <pikhq> http://www.tgdaily.com/html_tmp/content-view-37925-113.html
19:26:06 <pikhq> "For the record, the demonstration ran on a 16-core (4 socket, 4 core) Tigerton system running at 2.93 GHz."
19:27:19 <Sgeo> Is canvas3d still being worked on?
19:28:08 <Sgeo> Are the canvas3d and O3D apis compatible in any way?
19:28:11 <pikhq> I seem to recall similar performance being obtained via 2 PS3s, but I don't have a link handy.
19:28:30 <ehird> i can't find the page but i'm fairly sure it was a cluster
19:28:44 <AnMaster> * Sgeo thinks a canvas3d API should be roughly compatible with O3D apis, such that if O3D becomes a wide-spread standard, it's easy to take old work and just have it use O3D <Sgeo> Is canvas3d still being worked on? <Sgeo> Are the canvas3d and O3D apis compatible in any way? <-- you seem to have something against canvas3d?
19:28:48 <pikhq> ehird: I ALREADY CITED MY SOURCE FOR SAYING IT WAS A SINGLE SYSTEM YOU IDIOT.
19:28:58 <pikhq> READ THE MOTHERFUCKING ARTICLE.
19:28:58 <Sgeo> AnMaster, I have something against two competing APIs
19:29:03 <ehird> Hey, now I'm tempted to ignore pikhq. Cool.
19:29:04 <Sgeo> There should be one standard API
19:29:23 <pikhq> ehird: Please, just stop going 'Lalala'. ;)
19:29:39 <AnMaster> ehird, pikhq provided proof. Now if is up to you to prove him wrong.
19:29:49 <pikhq> Also, hardware ray tracing: http://graphics.cs.uni-sb.de/~woop/rpu/rpu.html
19:29:53 <ehird> AnMaster: I'd note that I don't give a flying fuck.
19:30:13 <AnMaster> ehird, you have acted like this way too many times before... So I suggest you just shut the fuck up instead of trying to pretend you weren't wrong when you actually are.
19:30:22 <ehird> 19:29 pikhq: Also, hardware ray tracing: http://graphics.cs.uni-sb.de/~woop/rpu/rpu.html // this doesn't even give fps figures
19:30:26 <AnMaster> or even better: admit you were wrong
19:30:40 <ehird> AnMaster: Waah. Do you want a tissue? You must be remarkably emotional with all this fighting against my tyranny.
19:30:51 <ehird> AnMaster: Poor thing.
19:31:02 <ehird> pikhq: said video wants me to download a video codec.
19:31:08 <AnMaster> ehird, *unrelated* personal attacks is not an acceptable way.
19:31:17 <pikhq> Also, that hardware ray tracing? Running on an FPGA.
19:31:22 <ehird> AnMaster: keep going. One day, you will stop my evil unacceptability.
19:31:24 <pikhq> ehird: ... MPEG-4.
19:31:26 <ehird> We must fight. We must stand.
19:31:36 <ehird> AnMaster: And we will triumph!
19:32:02 <pikhq> ehird: What, and your system doesn't handle every video format known to man?
19:32:36 <AnMaster> pikhq, where is the direct link to that video? Let me try mplayer on it.
19:33:19 <AnMaster> iirc ehird mentioned that he uses mplayer?
19:35:21 <Sgeo> Canvas3d isn't working for me :( :( :(
19:35:30 <ehird> Did you do the thing?
19:35:51 <Sgeo> I installed the extension
19:36:00 <Sgeo> And allowed the page to use Canvas3d
19:36:11 <Sgeo> All I get on c3dl.org 's demo is blueness
19:36:42 <ehird> Like Dylan; amirite GregorR?
19:36:45 <Sgeo> When I go to the demos, I get errors
19:37:26 <ehird> Everything see three dee ell dot org is Dylan to me.
19:38:40 <ehird> If it's not Intel or OCZ Vertex, it sucks hard and will give you worse random performance than a hard drive.
19:39:17 <ehird> pikhq: So, they're out now?
19:39:21 <ehird> I told you they would come out.
19:39:25 <ehird> How much do they cost?
19:39:47 <GregorR> And is it measured in $/gig or gig/$? :P
19:39:58 <ehird> All SSDs are $/gig, don't be silly.
19:40:02 <pikhq> They're coming out in two weeks; no prices announced yet, but the new 80G and 160G drives will be cheaper.
19:40:12 <ehird> That's been "news" for ages.
19:40:17 <ehird> I told you that and you doubted it, iirc.
19:40:24 <ehird> But we found that out like, last month.
19:40:37 <pikhq> ... No, I didn't doubt that.
19:40:43 <pikhq> You're the one who doubts things.
19:40:51 <ehird> No I'm not... you... person.
19:41:14 <pikhq> BTW, a single system with 4 chips is not a cluster. It's just silly. ;)
19:41:38 <ehird> For some definition of silly.
19:41:42 <ehird> Nehalem-EX can do 8 chips.
19:41:46 <ehird> 8 chips with 8 cores each.
19:41:55 <ehird> Add in hyperthreading, and we get 128 threads at once.
19:42:06 <pikhq> And they're not on the market yet.
19:42:13 <ehird> pikhq: That's totally irrelevant though.
19:42:22 <ehird> They'll be on the market soon enough.
19:42:28 <ehird> If you have that much money, buy some patience.
19:42:34 <pikhq> Freow. The 80G and 160G drives will be a full $100 cheaper than previous.
19:42:44 <ehird> pikhq: Anyway, 8xNehalem-EX + 4xASUS Mars 4GB.
19:42:58 <ehird> 16GB of graphics RAM? Yes. 8 GTX 285 GPUs? Yes.
19:43:12 <pikhq> ehird: And 3xIntel Postville. :P
19:43:17 <ehird> 24GB of RAM? sure, why not.
19:43:46 <pikhq> RAID of SSDs? Sure.
19:43:48 <ehird> pikhq: Eh, just buy a RAMdrive for the OS and program.
19:44:04 <ehird> With 24GB of RAM and 16GB of graphics RAM, the disk will be neglegible
19:44:07 <pikhq> ehird: ... In addition to the SSDs.
19:44:13 <ehird> pikhq: I'm talking about practical here.
19:44:40 <ehird> pikhq: Not everyone is an individual
19:45:00 <pikhq> Most companies don't by single systems that powerful.
19:45:24 <ehird> pikhq: If you want to do some serious supercomputing, this could be a good path.
19:45:37 <ehird> Instead of getting 100 commodity machines, get 10 of these; it'll probably be cheaper and faster.
19:46:20 <pikhq> You don't buy normal desktops for most supercomputing, you buy xU units or a few blade units...
19:46:30 <ehird> "Beowulf cluster", sir.
19:46:38 <ehird> The term begs to differ with you.
19:46:47 <pikhq> Beowulf clusters are built from cheap commodity machines.
19:46:56 <ehird> I was offering an alternative to a Beowulf cluster.
19:47:06 <ehird> Let's ignore ASUS Mars, and go for a cost-effective small-unit-count.
19:47:09 <pikhq> If you're actually spending megabucks on something, you get a bunch of blades or something.
19:47:17 <ehird> ASUS Mars only has 1,000 produced anyway, so it's not long-term viable.
19:47:26 <ehird> $400 for a 2GB GTX 285.
19:47:43 <ehird> Four of em = $1,600.
19:47:50 <ehird> Let's say a Nehalem-EX processor costs $2,000.
19:48:28 <ehird> ~$2000 worth of RAM, $600 worth of SSD (for cache)
19:48:34 <ehird> Well, $700 of RAM.
19:48:38 <ehird> (4GB DDR3 is costly)
19:48:48 <ehird> = $18,900 per system.
19:48:51 <ehird> And what a system it would be.
19:48:52 -!- augur has joined.
19:49:32 <ehird> Get 3 of them = $56,700 and I dare you to find a system that fast for that cost. We're assuming a certain type of computation here:
19:49:35 <ehird> lots and lots of data.
19:49:41 <ehird> In a gigantic beowulf cluster, the bottleneck is the connections between nodes.
19:49:50 <ehird> So I'd say this is likely to turn out faster.
19:51:12 <ehird> But no, this isn't usually practical.
19:51:23 <ehird> I'm bored now so I'm going to assemble a cost-effective supercomputer node in my text editor.
19:52:06 <ehird> I should go work for GreenArrays; I seem to have some sort of obsession with supercomputers :P
19:52:16 <ehird> I went WTF when I found out RoadRunner ran Fedora.
19:52:25 <pikhq> For comparison, $18,900 could reasonably get you a filled rack of computers.
19:52:26 <ehird> I just imagined one monitor connected to it, booting up to Gnome.
19:52:40 <ehird> pikhq: True, but when computer-to-computer is your bottleneck...
19:52:43 <ehird> I never said it was practical.
19:52:47 <ehird> I just said it could have a use case.
19:53:18 <pikhq> ehird: When computer-to-computer is your bottleneck, you're spending a bit more and getting rather crazy fiber links going. :P
19:54:26 * ehird starts off by picking the process for these nodes (3.2GHz Nehalem) and graphics cards (2 x 2GB GTX 285)
19:54:56 <ehird> why won't someone give me a budget for making one of these :E
19:55:36 <Sgeo> http://labs.flog.co.nz/raytracer
19:56:03 <ehird> "I suggest using Opera, as it has a great (fast) JavaScript processor."
19:56:31 <ehird> It doesn't even do antialiasing. :P
19:57:00 * Sgeo uses it in Chrome
19:57:41 <Sgeo> Also, wouldn't worker threads make it not make the browser unresponsive
19:57:48 <ehird> you can't do threads in js
19:58:24 <Sgeo> Then what's https://developer.mozilla.org/En/Using_DOM_workers ?
19:58:56 <ehird> "Applies to Firefox 3.5, SeaMonkey 2, and Thunderbird 3 and later "
19:59:17 <ehird> Creating a new worker is simple. All you need to do is call the Worker() constructor, specifying the URI of a script to execute in the worker thread, and, if you wish to be able to receive notifications from the worker, set the worker's onmessage property to an appropriate event handler function.
19:59:21 <ehird> as opposed to, you know
19:59:23 <ehird> passing a fucking function
20:01:54 <Sgeo> Ok, Firefox's address bar doesn't seem to want my text
20:02:06 <Sgeo> Ok, now it's working again
20:08:02 <ehird> pikhq: I have the main components for a massively powerful supercomputer node for $5,689.81 in a text file here.
20:08:26 <ehird> All on the market now, and basically the fastest thing ever.
20:08:45 <ehird> Ten of them would probably make an entry on the top 200 supercomputer list.
20:09:48 <ehird> The amount of cruft (motherboard, chassis etc) you have to buy for each node makes powerful nodes appealing, IMO.
20:10:40 <ehird> (Note: i was being hyperbolic about the supercomputer list. :P)
20:34:12 <AnMaster> pikhq, I suspect that within maybe 5 years, GPUs able to ray trace at a decent FPS for gaming will be available. Possibly less time.
20:34:32 <ehird> AnMaster: you know that most modern fpses are only palyable at ~40fps, right?
20:35:18 <pikhq> ehird: And there's real-time ray tracing available now. Unfortunately, it's on a freaking FPGA, so it's not all that great, quality-wise.
20:35:28 <pikhq> Stick that design on an ASIC and you're golden. ;)
20:36:30 <AnMaster> pikhq, ray tracing can take advantage of parallelism very well... So I think with enough FPGAs their slow speed would not be much of an issue.
20:37:02 <pikhq> AnMaster: They designed it to scale with additional FPGAs, and got nearly linear performance increase.
20:37:02 <AnMaster> a larger issue is memory bandwidth...
20:37:08 <pikhq> (only tested with up to 4 FPGAS)
20:37:19 <pikhq> The FPGAs also had rather low memory bandwidth.
20:37:20 <AnMaster> pikhq, yes, until you hit memory bandwidth issues.
20:37:30 <ehird> http://www.caustic.com/
20:37:33 <AnMaster> pikhq, I read about half of the paper so far
20:37:54 <AnMaster> and yes I see they mention trying to combine memory requests and so on
20:38:07 <AnMaster> which will of course help a lot but... far from always
20:38:58 <pikhq> AnMaster: Sticking it on an ASIC would help a lot.
20:39:06 <pikhq> ehird: Not what I was discussing.
20:39:18 <ehird> caustic's stuff is newer and better.
20:39:33 <AnMaster> pikhq, with a complex enough scene, with enough incoherent rays, not even an ASIC will help. Since then memory bandwidth will be the main issue. More and larger caches yes...
20:39:51 <pikhq> "Someone built a rocket to go to space on a few million!" "Bu, but NASA spends BILLIONS OF DOLLARS on their space program! That's bullshit!"
20:40:07 <pikhq> AnMaster: They had megabytes/sec memory bandwidth.
20:40:07 <AnMaster> pikhq, anyway yes, increasing the speed they operate on will help quite a bit. But do you imagine that they can afford it? For a research project?
20:40:32 <ehird> pikhq: that quote is totally irreleavnt
20:40:50 <ehird> is your logic really "caustic aren't better because they're a company, not a research project"?
20:40:55 <pikhq> What I'm saying is that it is perfectly practical now to make an RPU with good performance, and nobody's doing it.
20:40:56 <AnMaster> pikhq, Even for production, the production will be so limited that AISC might not be worth it...
20:41:19 <pikhq> ehird: ... No, I'm saying that you were being silly when I was discussing Intel's Quake Wars on a single system.
20:41:35 <AnMaster> pikhq, there is one commercial implementation iirc?
20:41:44 <ehird> pikhq - commenting, without context hints, on things from hours ago since 2009
20:41:50 <pikhq> I'm just saying "We have the technology".
20:42:10 <ehird> hmm, the evidence seems to point to AnMaster ignoring me
20:42:16 <ehird> AnMaster: You fuck sheep.
20:42:35 <ehird> (this is basically like jumping out of the window as soon as you determine you're in a lucid dream.)
20:43:20 <ehird> AnMaster: Caustic.
20:45:19 <pikhq> ehird: I'm never forgetting that quote.
20:45:22 <AnMaster> pikhq, what about being able to access ram in more parallel? I'm not sure what I mean, but memory seems like the bottleneck in that implementation you linked.
20:45:35 <ehird> pikhq: Ambiguous referent "that".
20:45:40 <ehird> Please correct & resubmit.
20:45:43 <pikhq> AnMaster: Yes, and they had absurdly slow memory.
20:45:49 <pikhq> ehird: <ehird> You sheepfucker.
20:46:04 <AnMaster> pikhq, please, I have him on ignore, I do not wish you to relay :)
20:46:06 <ehird> Okay, pikhq. (You sheepfucker.)
20:46:32 <ehird> AnMaster: Your stupidity is like a colossal mountain of sheep, all fucking each other, with you as their liver in every single one.
20:46:35 <ehird> Romeo and Juliet are sheep.
20:46:44 <ehird> Your brain twists and turns into a vaguely pea shape.
20:46:50 <AnMaster> pikhq, 512 MB SRAM? That would rock :D But be insane
20:46:54 <pikhq> AnMaster: I'm just reminding him of his stupidity. :)
20:46:55 <ehird> At this point it is at its most potent, and decides to continue the current action: fucking sheep.
20:47:05 <pikhq> AnMaster: And considering an ignore myself.
20:47:21 <AnMaster> pikhq, is he still claiming it was a cluster and not 16 cores?
20:47:40 <pikhq> AnMaster: Actually, he's just being a troll now.
20:47:50 <pikhq> AnMaster: Something about fornication with sheep.
20:47:51 -!- zzo38 has joined.
20:48:00 <ehird> I think I'll start an initiative where everyone ignores everyone else but one person.
20:48:06 <AnMaster> pikhq, ah that phase. Expected behaviour after he lost a discussion.
20:48:10 <ehird> Truly, it would be internet - relay - chat.
20:48:24 -!- zzo38 has left (?).
20:48:25 -!- zzo38 has joined.
20:48:27 <AnMaster> pikhq, soon he will part in what I believe he calls "a huff"
20:49:17 <ehird> "AnMaster psychic services - almost as accurate as Uri Geller with more than 10x the biases!"
20:50:15 -!- zzo38 has left (?).
20:50:17 -!- zzo38 has joined.
20:50:55 <AnMaster> zzo38, hi, what are you doing?
20:51:07 <zzo38> I have the index of free software/open source project I make, but the #freenode channels don't help me.
20:51:41 * GregorR tries to imagine what the #freenode channel could POSSIBLY help you with in this respect ...
20:52:13 <zzo38> Someone on #freenode channel asked me before, but now I answered them and they might be gone possibly?
20:52:35 <ehird> zzo38: so what drugs are you on today?
20:52:36 <AnMaster> I'm still completely lost about what the hell you are talking about.
20:53:00 <ehird> I rate the probability of that statement as extremely low, zzo38. :p
20:53:01 <GregorR> AnMaster clearly has ehird on ignore again :P
20:53:33 <ehird> GregorR: Yes, with the usual ignorer schtick of "I'm ignoring him. Oh, him? I'm ignoring him by the way. What was that? Oh it was to him? I'm ignoring him, you know."
20:53:47 <ehird> "PAY ATTENTION TO ME GUYS I AM _IGNORING_ HIM"
20:53:55 <zzo38> I made a index of the free software projects I maintain (because I was asked by #freenode to do so), and I have IRC channel, but now what am I supposed to do about it? Nothing?
20:54:00 <ehird> Thank you for the demonstration, pikhq.
20:54:20 <AnMaster> GregorR, correct. Once he started arguing that a 16 core system was actually a cluster and refused to accept it even when pikhq provided the relevant link.... And then ehird started trolling. Read log if you for some odd reason want to see more...
20:54:26 <AnMaster> I can't imagine what reason though
20:54:43 <AnMaster> zzo38, who was it on there that asked you?
20:54:52 <AnMaster> zzo38, better check your irc logs if you don't remember
20:54:56 <ehird> I claimed that (a) there was a raytraced game and (b) it used a cluster of Intel CPUs.
20:55:01 <AnMaster> (I assume you have your client set to log)
20:55:03 <ehird> Both were correct; it was just not the example pikhq pointed out.
20:55:04 <zzo38> If I remember I would go to directly private message to them.
20:55:09 <zzo38> I didn't enable the log, sorry.
20:55:31 <zzo38> I don't remember if it was some staffer or not, either.
20:55:33 <AnMaster> zzo38, btw, you are in luck, I *do* log all channels, and I'm in #freenode. If I was there at that point maybe I can help you?
20:57:47 <zzo38> Did you find it yet?
20:58:00 <AnMaster> zzo38, 34 MB logfile. Takes a bit...
20:58:51 <AnMaster> zzo38, everything matching zzo38 with one line of context before/after from the last two days: http://pastebin.ca/1499435
20:59:07 <AnMaster> zzo38, timestamps are in UTC+2
20:59:15 <AnMaster> so adjust for your timezone as needed
20:59:57 <AnMaster> zzo38, doesn't look like a staffer anyway
21:00:00 <zzo38> Now I know who it was.
21:00:26 <AnMaster> zzo38, so probably not an official request of any kind (thus no one would know what on earth you were talking about later on there...)
21:01:22 <AnMaster> zzo38, indeed not a staffer (just checked)
21:02:37 <AnMaster> wow. 999MB in logs before compressing from last month
21:03:01 <zzo38> I will write my own IRC server and move them to that one once I write it
21:03:21 <AnMaster> zzo38, what is wrong with existing ones? Like charybdis and inspircd?
21:03:28 <zzo38> Is PHP good enough for writing daemons on Windows?
21:03:54 <pikhq> PHP is not recommended for writing anything on anything.
21:04:23 <pikhq> And it is especially poorly suited for the writing of daemons.
21:04:59 <zzo38> I wrote a IRC client in PHP. (I don't write most programs in PHP but if this uses internet connection I use PHP because I know how to socket function with it)
21:05:48 <zzo38> I wrote a Gopher server in FreeBASIC, but that's running on inetd and it is simpler than a IRC server. FreeBASIC is not sufficient for writing a IRC server
21:05:50 <AnMaster> zzo38, for an irc server you want to use a single thread and async IO. Use something like epoll or kqueue to be able to wait on lots of sockets in an efficient way
21:06:18 <zzo38> I use a lot of different program languages, including C.
21:06:22 <AnMaster> I believe windows has something similar to epoll (linux) and kqueue (freebsd)... forgot what the windows one was called
21:06:35 <pikhq> For Haskell is greater than it all.
21:06:42 <AnMaster> zzo38, C is probably one of the saner languages for ircds. Erlang will probably also be good
21:07:03 <zzo38> I don't know how to write internet server software in C.
21:07:08 <AnMaster> pikhq, can you do efficient async IO waiting on thousands of sockets at once?
21:07:10 <pikhq> (Erlang is also recommended for anything where asynchronous message passing makes sense)
21:07:32 <pikhq> AnMaster: Yes, though I don't know the library that does that.
21:07:42 <AnMaster> pikhq, erlang will use kqueue, epoll and such automatically, that is why it makes sense for ircds.
21:08:12 <pikhq> Erlang also makes sense for the job, I will freely admit.
21:08:30 <pikhq> I just like the Haskell. ;)
21:08:43 <AnMaster> This is really the most important thing when it comes to good performance in ircds. DON'T USE select() OR poll()! Use whatever the OS provides which is more efficient.
21:09:06 <AnMaster> since select() and poll() requires you to send the complete list of fds to the kernel *every time you wait*
21:09:11 <ehird> threads perform better than async io.
21:09:22 <zzo38> Can it be cross-platform supporting?
21:09:23 <AnMaster> and you wait very often in ircds. and on many fds
21:09:28 <ehird> see http://www.usenix.org/events/hotos03/tech/full_papers/vonbehren/vonbehren_html/index.html
21:09:56 <zzo38> PHIRC uses socket_select
21:09:58 <AnMaster> zzo38, epoll and kqueue? Just provide your own abstraction for these, so you just can do something like -DUSE_KQUEUE at compile time or such. That is what other good ircds do
21:10:24 <AnMaster> zzo38, php is NOT a good idea for ircds. Nor anything else. But especially not stuff like ircds and other daemons.
21:10:29 <zzo38> AnMaster: Thanks for telling me. That seems sensible to me
21:10:56 <zzo38> PHIRC is a IRC client, not server
21:11:00 <AnMaster> zzo38, however, I suggest you try some of the better existing ircds first. Such as charybdis and inspircd.
21:11:21 <AnMaster> and inspircd is very configurable, which should fit you
21:11:24 <zzo38> If source-codes is available I can make modification?
21:11:34 <AnMaster> zzo38, both are open source. So not an issue
21:11:48 <AnMaster> charybdis is in C, inspircd is a in a mostly sane subset of C++
21:12:32 <AnMaster> but I haven't ever needed to modify either except when I was on the QA team of inspircd some years ago
21:13:35 <zzo38> When trying to access the web-site for charybdis ircd I get a protocol mismatch error.
21:13:36 <AnMaster> and with "modify" here, I actually mean "apply patches from devs to test them"
21:14:28 <zzo38> http://www.ircd-charybdis.org/Main_Page
21:14:28 <AnMaster> http://www.stack.nl/~jilles/irc/#charybdis right?
21:15:12 <AnMaster> zzo38, I haven't heard of that one before? Nor is it listed in the topic of #charybdis (not on freenode, another network)
21:15:59 <AnMaster> zzo38, the charybdis devs are good developers, but they don't spend much time on their website.
21:16:49 <AnMaster> Unlike one of the guys on the inspircd team
21:16:50 <zzo38> What was the first IRC server software written, and is it free software/open source?
21:17:05 <ais523> probably "ircd", nothing else would have a name that simple
21:17:28 <AnMaster> Open and free? No idea. I believe the source is lost nowdays for the very first versions
21:17:38 <AnMaster> besides, those used an incompatible irc protocol
21:17:48 <AnMaster> it didn't stabilise until a bit later
21:18:12 <zzo38> It doesn't matter, if it is available at all and FOSS, I would fix it
21:18:30 <ehird> zzo38: it won't even run on modern systems...
21:18:36 <ehird> wasn't it made on a timesharing system?
21:19:23 <AnMaster> zzo38, Remember IRC is very old. I know someone tried some of the older ones and couldn't get them to compile on modern OSes
21:19:27 <zzo38> My channels are currently on IRCNET (us.ircnet.org:7000) but I plan to move them to zzo38computer.cjb.net:194 when the IRC server software is ready and I have written it and set it up
21:19:57 <AnMaster> I think either Brain or w00t made an ircd collection. Maybe they have some of the older ones...
21:20:03 <AnMaster> I shall ask next time I see them
21:20:59 <AnMaster> zzo38, the first ircd was written around 1988-1989 or so...
21:21:15 <AnMaster> zzo38, so yeah, the oldest versions are lost
21:22:06 <zzo38> Because if I can get the oldest and simplest one available (preferably in C) then I can work from that to fix things and add stuff.
21:22:27 <AnMaster> zzo38, then you would just be reinventing the wheel
21:23:55 <zzo38> I can also write it completely by myself, if I can know how to make internet server softwares in C and MinGW.
21:24:16 <AnMaster> MinGW is just GCC ported to Windows isn't it?
21:24:19 <zzo38> But I want it to be cross-platform.
21:24:56 <AnMaster> zzo38, your best bet if you want it to run on windows as well as Linux, freebsd and so on is inspircd. Charybdis won't work on windows
21:25:02 <AnMaster> inspircd does have a port to windows
21:25:18 <AnMaster> zzo38, and I'm unable to find these old ircds versions currently.
21:25:36 <zzo38> I probably ought to write my own so that I can learn more about internet servers softwares in C.
21:29:43 <zzo38> The channel of my projects is +Vonkeror +gophserv +MegaZeux +PHIRC
21:30:42 <zzo38> Not on freenode though.
21:31:09 <AnMaster> zzo38, + is local to server isn't it? Not widely supported due to being pretty useless.
21:31:22 <zzo38> No, & is local to server
21:33:11 <AnMaster> zzo38, ircnet is really one of the worst networks out there for many reasons. 1) no services 2) bad ircd 3) policies and rules varies between servers.
21:33:28 <AnMaster> efnet has the same problem basically, except it does have services, but not usual ones
21:33:33 <ehird> ircnet was the -original- network
21:33:39 <ehird> what you're saying amounts to "IRC sucks"
21:33:51 <AnMaster> (and their service "CHANFIXER" is one of the worst possible ones I can think of)
21:36:59 <zzo38> IRCNET is the closest network I could find to what I needed.
21:37:21 <AnMaster> I wonder what your needs are then...
21:39:02 <zzo38> Mostly support for + type channels
21:39:16 <AnMaster> zzo38, what did + ones do now again?
21:39:27 <zzo38> + type channels are modeless
21:39:34 <AnMaster> zzo38, and why do you want that?
21:39:47 <zzo38> The other features of IRCNET are also the ones I was looking for, too.
21:40:31 <zzo38> I want to use modeless channels and that's why my projects are not hosted on freenode.
21:41:35 <ais523> arguably, services are a hax
21:41:59 <AnMaster> ais523, agreed. But they integrate quite well these days. Work well too
21:42:18 <AnMaster> ais523, the alternative would be the same functionality built into the ircds. And distributed
21:42:26 <zzo38> I don't need services
21:42:27 <AnMaster> but then you would have various problems syncing them and so on
21:42:55 <AnMaster> and what about netsplits? Same nick could be registered on both sides then
21:43:17 <AnMaster> there are ways to work around this and so on, but all uglier hacks than services as they are done now
21:43:54 <AnMaster> (such as forbidding registration through services unless more than 50 % of all servers are connected to the current server atm.
21:44:34 <fizzie> The alternative is not "nick registration built into the server" but "no nick registration".
21:45:15 <AnMaster> ais523, also, I agree most classical services are backs. But atheme, which freenode uses since several months ago (switched during 2008 iirc?) handles nick changes much better.
21:45:22 <AnMaster> "* [ais523] is signed on as account ais523"
21:45:30 <AnMaster> that would follow you around even if you changed nick
21:45:45 <ais523> some ircds have rules against reserving nicks in any way
21:45:49 <ais523> and encourage sniping other people's nicks
21:46:02 <AnMaster> older services needed you to always have a nick registered to your account or you were logged out
21:46:16 <AnMaster> ais523, um... Not any of the big ones at least.
21:46:30 <ais523> same for channels IIRC
21:46:37 <ais523> that's closer to the original spirit of IRC
21:46:47 <ais523> although, original IRC used ident as a means of authentication
21:47:00 <ais523> it was supposed to be able to tell who you were from your IP, or logged-in-ness, or something
21:48:17 <AnMaster> http://irc.netsplit.de/networks/top10.php <-- I know at least 6 of those use services. IRCnet does not. Efnet doesn't use normal services, so didn't count it in those 6 (google chanfixer if you want to know more), webchat and that ustream.tv I have no idea about
21:48:50 <AnMaster> almost all networks use services these days
21:50:30 <AnMaster> ircnet and (possibly, depending on how you define services) efnet are the only big ones I know lack services. Almost all smaller networks have services.
21:50:59 <AnMaster> ais523, services are overwhelmingly more popular than "no services"
21:52:27 <AnMaster> http://irc.netsplit.de/networks/top10.php?year=2009 <-- I wonder why the user graph for most networks look like a sawtooth?
21:52:39 <fizzie> How many different ways are you going to keep saying that same thing?-)
21:52:56 <AnMaster> fizzie, I can possibly add a few more.
21:53:11 <fizzie> It looks like it's something week-periodic there, since it's about four periods per month.
21:53:47 <fizzie> I can't really tell from that graph. Rise would perhaps be more logical.
21:54:12 <AnMaster> fizzie, the server count (third graph) is really strange for dalnet...
21:54:29 -!- _AnywhereIs_ has joined.
21:54:34 <AnMaster> fizzie, wasn't dalnet the first network to introduce services iirc?
21:54:52 <ais523> for all the Rubicon fans over here, I just made a new puzzle: http://kevan.org/rubicon/game.php?level=dinasog
21:55:09 <zzo38> How do I get help for writing the internet server software with C and cross-platform?
21:55:10 <AnMaster> ais523, is that the game based on RUBE?
21:55:24 <fizzie> The server count graph to me looks like that irc.netsplit.de is gathering their dalnet statistics from some leaf server that has a bit flaky connection.
21:55:30 -!- _AnywhereIs_ has left (?).
21:55:30 <AnMaster> zzo38, you look at existing software and learns how it works?
21:55:54 <zzo38> Which software should I look at mostly?
21:55:58 <AnMaster> fizzie, iirc netsplit.de use three different servers.
21:56:04 <pikhq> Write it for UNIX.
21:56:13 <AnMaster> zzo38, for ircds? probably inspircd and/or charybdis
21:56:21 <AnMaster> they are the best ones currently
21:56:23 <pikhq> Test it on multiple Unixes.
21:56:29 <pikhq> Then, port to Windows.
21:56:35 <zzo38> No, just internet server software in general. And it has to work on MinGW and on UNIX.
21:56:37 <AnMaster> performs best, most modern features, least messy code bases.
21:56:39 <pikhq> (I suggest using cygwin or mingw for that)
21:56:55 <AnMaster> zzo38, a httpd would be written very differently from an ircd for example
21:57:08 <pikhq> zzo38: Assume UNIX, then make it work for Windows.
21:57:18 <pikhq> That's the single easiest way to get it working everywhere.
21:57:26 <zzo38> However I cannot test it on UNIX, I have to test it on Windows only.
21:57:34 <AnMaster> you might want threads in a httpd even. You will never want a threaded ircd.
21:57:51 <AnMaster> (possible exception is if it is using erlang threads, which are rather different from normal ones)
21:57:56 <fizzie> If you look at the weekly graph -- for example, QuakeNet at http://irc.netsplit.de/networks/details.php?net=QuakeNet&submenu=weeks -- it actually looks as if there's a bit less than average people on weekends.
21:57:58 <zzo38> I can try to look at codes of charybdis and so on
21:58:21 -!- MigoMipo has quit ("Ich schlafe. Ich bin sehr müde.").
21:58:45 <zzo38> But still, the *client* PHIRC is written in PHP (and I am using it right now)
21:59:04 <pikhq> AnMaster: Hmm. Yeah, you'd basically want a fork() with message passing for an ircd.
21:59:08 <fizzie> It might be that there's some sort of thing that is not IRC that people do. I've heard rumours, though I haven't experienced such a thing myself.
21:59:17 <AnMaster> zzo38, a client makes a handful connections. Maybe 5-10 or so, depends on how many networks you are on
21:59:20 -!- Sgeo has quit (Read error: 110 (Connection timed out)).
21:59:24 <AnMaster> but an ircd have thousands each
21:59:36 <AnMaster> * Current local users: 3987 Max: 4291
21:59:36 <AnMaster> * Current global users: 48593 Max: 57036
21:59:36 <AnMaster> * Highest connection count: 4293 (4291 clients) (340841 since server was (re)started)
21:59:49 <AnMaster> varies between which server you are connected on of course
21:59:59 <zzo38> Each instance of PHIRC connects to only one server at a time (but you can have multiple instances running at once, in separate windows)
22:00:08 <pikhq> Hmm. fork() with message passing...
22:00:27 <pikhq> AnMaster: Evil, to be sure.
22:00:33 <AnMaster> pikhq, I bet you won't be able to handle more than maybe 100-200 clients then. At most.
22:01:04 <zzo38> The only way I understand is the way I wrote PHIRC in.
22:01:26 <pikhq> zzo38: Don't write an IRCd.
22:01:33 <pikhq> Learn how to do network code in C first.
22:01:41 <AnMaster> pikhq, an ircd that can't handle over 10 000 clients in a single channel (all connected to a single ircd) all sending data all the time isn't really worth the it's LOC in gold
22:01:57 <AnMaster> and yes I know inspircd handles that, as I said, I was on the inspircd QA team a few years ago :)
22:02:01 <pikhq> AnMaster: I didn't say it was a good idea.
22:03:14 <AnMaster> pikhq, oh and that build was a -O0 -ggdb3 build. and that was C++. So for inspircd the bandwidth would be the limiting factor in most cases.
22:04:56 <zzo38> PHP is not even using as much memory as Apache and stuff with PHIRC running.
22:06:35 <pikhq> zzo38: You should do that in a different language.
22:06:50 <zzo38> Why is PHP such a bad way to do it, Apache uses ten times as much memory as PHP.
22:07:19 <pikhq> Nice non sequitur.
22:07:51 <pikhq> "Why is a Ford Model T a bad car? The Space Shuttle uses several thousand times as much fuel as it."
22:08:31 <GregorR> More like, "Why is the Ford Model T a bad car? A truck towing a Ford Model T would probably use more gas."
22:13:28 <zzo38> Does PHIRC even missed anything important? I think I have everything necessary now.
22:13:58 <pikhq> You made RawIRC and are asking if it needs anything else.
22:14:29 <zzo38> It isn't the same as RawIRC, it is a bit different. I can't even get RawIRC to compile.
22:14:51 <zzo38> Did you even look at it? Maybe if you look at the codes you might notice something I missed
22:16:39 <zzo38> http://pastebin.ca/1499497
22:19:06 -!- zzo38 has left (?).
22:19:41 <GregorR> "The ability to not inexplicably part channels with no warning"
22:21:20 <AnMaster> GregorR, zzo definitely lacks it yes
22:22:05 <AnMaster> GregorR, he also lacks "The ability to not join the channel and starting talking like you were in the middle of a conversation, thus leaving you very confused"
22:23:57 <GregorR> But can we build that into an IRC client? ;)
22:24:23 <AnMaster> GregorR, I wasn't aware you planned to build it into an irc client
22:25:38 <AnMaster> GregorR, btw, I have the "The ability to not inexplicably part channels with no warning" in my bouncer
22:26:02 <AnMaster> means it will block parts and automatically rejoin me if I make a mistake
22:26:11 <AnMaster> like clicking the wrong button
22:26:22 <AnMaster> I have this channel set to sticky :)
22:26:51 <GregorR> We're all very proud to be so ... sticky?
22:27:04 <GregorR> I've been reading too much ... dinosaur comics?
22:27:31 <AnMaster> GregorR, oh? I missed that reference... Even though I occasionally read it...
22:27:43 <GregorR> T-Rex seems to form a lot of sentences ... like this?
22:28:20 <AnMaster> GregorR, none in the last one it seems *looks a bit back*
22:28:35 <GregorR> Naw, not every comic certainly, but I think it's indicative.
22:31:50 <AnMaster> http://www.qwantz.com/index.php?comic=1503 <-- heh, graphic change!
22:32:32 <AnMaster> anyway found one example between 1503 and 1512
23:10:47 <ehird> 21:57 AnMaster: you might want threads in a httpd even. You will never want a threaded ircd.
23:11:40 <ehird> 22:01 AnMaster: pikhq, an ircd that can't handle over 10 000 clients in a single channel (all connected to a single ircd) all sending data all the time isn't really worth the it's LOC in gold
23:11:45 <ehird> such channels... don't exist.
23:19:50 -!- coppro has quit ("Reconnecting…").
23:20:15 -!- coppro has joined.
23:21:43 <ehird> http://kitenet.net/~joey/code/etckeeper/ ← hells yeah
23:38:00 <ais523> hmm... I recognise that URL, the first bit at least
23:38:06 <ais523> isn't that by Debian's former C-INTERCAL maintainer?
23:40:41 <ehird> ais523: he also made ikiwiki.
23:40:57 -!- MizardX has quit ("reboot").
23:41:06 <ehird> but he might have done that
23:41:48 <ehird> "I've developed an xmonad configuration file for using this window manager with the Palm Pre smart phone. (I described how I run X on the Pre in VNC here.)"
23:41:54 <ehird> He also appears to be batshit insane
23:42:12 <ehird> ais523: he implemented debconf
23:42:35 <ehird> and started debian-installer
23:45:15 -!- MizardX has joined.
23:57:54 -!- ais523 has quit (Remote closed the connection).