←2011-03-16 2011-03-17 2011-03-18→ ↑2011 ↑all
00:01:24 -!- pikhq_ has joined.
00:01:27 -!- pikhq has quit (Ping timeout: 252 seconds).
00:01:35 -!- copumpkin has joined.
00:13:37 -!- elliott has joined.
00:13:44 -!- elliott has quit (Remote host closed the connection).
00:22:24 -!- ais523 has quit (Remote host closed the connection).
00:22:41 -!- BeholdMyGlory has quit (Remote host closed the connection).
00:30:41 -!- pikhq_ has quit (Read error: Operation timed out).
00:31:10 -!- pikhq has joined.
00:38:11 -!- nathy has joined.
00:56:33 -!- pikhq has quit (Ping timeout: 240 seconds).
00:56:33 -!- hagb4rd has quit (Ping timeout: 255 seconds).
00:56:51 -!- pikhq has joined.
01:03:05 -!- cheater- has joined.
01:06:00 -!- cheater00 has quit (Ping timeout: 255 seconds).
01:13:56 -!- cal153 has quit.
01:39:05 -!- nathy has left (?).
01:41:48 <pikhq> Fuck.
01:41:52 <pikhq> http://www.theatlanticwire.com/politics/2011/03/michigan-declares-financial-martial-law/35861/
01:42:25 <pikhq> Michigan is now under fascist rule.
01:44:37 <pikhq> No, literally, fascism.
01:45:35 <Gregor> pikhq didn't notice what with the partial nuclear meltdown ;P
01:47:26 -!- Alex_Megaroide has changed nick to wareya.
01:48:57 -!- augur has quit (Ping timeout: 246 seconds).
01:49:27 -!- cal153 has joined.
01:52:06 -!- GreaseMonkey has joined.
01:52:45 -!- HackEgo has quit (Remote host closed the connection).
01:53:13 -!- EgoBot has quit (Remote host closed the connection).
01:56:40 -!- augur has joined.
02:15:44 -!- wareya has quit (Read error: Connection reset by peer).
02:16:30 -!- wareya has joined.
02:24:14 -!- EgoBot has joined.
02:24:48 <Gregor> !echo hi
02:24:48 <EgoBot> hi
02:24:51 <Gregor> !quote
02:25:45 <Gregor> Hm
02:26:06 <augur> pikhq: no shit sherlock
02:26:11 <Gregor> Oh, wrong bot durp :P
02:26:29 <pikhq> augur: Wut?
02:26:36 <augur> about fascism in michigan
02:27:12 -!- HackEgo has joined.
02:27:40 <Gregor> `quote
02:27:40 <HackEgo> 123) * Warrigal refuses to say goodbye to Quas NaArt, as he is coming closer, not going farther.
02:29:27 <Gregor> Moved my bots off Codu for now.
02:32:25 <Gregor> `echo blehboop
02:32:26 <HackEgo> blehboop
02:44:41 <Gregor> !bf_txtgen Textadupadup
02:44:43 <EgoBot> 114 ++++++++++++[>+++++++>++++++++++>+++++++++>+<<<<-]>.+++++++++++++++++.>.----.<----.+++.>+.>++++.<<---.+++.>.>.>--. [397]
02:44:49 <Gregor> WOW that's fast :P
02:46:12 <Gregor> Timing buffered disk reads: 4 MB in 4.34 seconds = 942.73 kB/sec laaaaaaaaaaawl
02:52:03 <Gregor> Yay, they're moving me to a different server.
03:02:29 <quintopia> wait
03:02:36 <quintopia> prgmr is?
03:02:52 <Gregor> Yes
03:04:01 <quintopia> does your bf_txtgen implementation tune the number of values that it plays with and their initial values until it hits a local minimum?
03:04:46 <Gregor> It's not mine, and it's a genetic algorithm, and that's all I know about it :P
03:05:01 <Gregor> Everybody assumes it's mine just because my bot contains it, even though it also contains dozens of interpreters I didn't write :P
03:05:49 <quintopia> well, i just meant "your" in the sense of "the one you use"...and i assumed you knew something more about it than that because it's the one you use >.>
03:06:02 <Gregor> I know ... how to call it :P
03:06:11 <quintopia> do you know whether it runs a per-string GA, or the whole program was GA-generated?
03:06:35 <Gregor> Per-string GA.
03:06:51 <Gregor> The program is human-written Java.
03:07:19 <quintopia> ah
03:10:35 <quintopia> do you know what the numbers it outputs mean?
03:10:42 <quintopia> i assume the first one is the length
03:13:08 <Gregor> One is length, the other is the generation.
03:13:15 <Gregor> I choose the shortest one within the first 1000 generations or something.
03:19:06 <quintopia> is the "cool factor" here more important than getting the shortest string possible?
03:52:26 <Gregor> Uhh, no?
03:52:43 <Gregor> The shortest string possible isn't very easy to calculate ...
03:52:57 <quintopia> i don't mean actual optimal
03:53:14 <quintopia> i just meant if there were an algorithm that yielded better results than the current one, would you switch to it
03:53:23 <Gregor> Of course.
03:53:35 <Gregor> Although I'd prefer that somebody just give me an hg bundle so I don't have to do any switching myself :P
03:53:39 <quintopia> what has been tried?
03:53:53 <Gregor> Pretty much nothing :P
03:53:55 <Gregor> This has done quite well.
03:54:16 <quintopia> you know i held a coding competition for this problem a few years ago, right?
03:54:41 <quintopia> i wonder if a "try a bunch of tuned heuristics and pick the best one" couldn't outdo the GA
03:56:33 <pikhq> Well, you could make a much better GA. :P
04:08:09 <quintopia> do you know how the GA works pikhq?
04:09:24 <pikhq> Yeah, it's just tweaking a single setting loop and a string of +-<>. to output the string.
04:09:32 <pikhq> Pretty dang naïve.
04:29:27 <quintopia> gregor: at some point in report.c do you store the values of "number of times p beat q" (bfjoust)?
04:29:39 <quintopia> if not, where can i extract them from?
04:30:00 <Gregor> Yes, that's there.
04:30:09 <Gregor> IIRC it's what's in the global var "scores"
04:33:59 <quintopia> aha
04:37:36 <quintopia> so it's a 2D array of signed chars? so scores[p][q] is the value i'm looking for?
04:38:46 <Gregor> I think it's a 2D array in the C sense; double-indirection is for pussies.
04:39:31 <quintopia> scores[p*programCount+q] then
04:39:35 <quintopia> but one problem
04:39:53 <quintopia> or maybe not
04:39:57 * quintopia looks harder
04:41:12 <quintopia> it looks like it only saves the score of p vs. q in there if q>p
04:41:23 <quintopia> am i misreading?
04:45:12 * Gregor needs to go check :P
04:45:57 <Gregor> No, it always stores it there.
04:46:43 <Gregor> Nowait
04:46:51 <Gregor> Oh, sorry, I misread your misreading X-P
04:47:04 <Gregor> Yeah, it only stores it for the lower program.
04:47:10 <Gregor> But the score for the other is just the inverse of that.
04:47:39 <Gregor> The winner() function will (contrary to its name) always give you the proper score.
04:49:30 <Sgeo> http://www.youtube.com/watch?v=QHVrE1NTgxI
04:51:21 <quintopia> oh]
04:51:43 <quintopia> what is the score exactly?
04:51:48 <quintopia> (remind me)
04:52:00 <quintopia> i seem to recall it is pwins-qwins
04:52:24 <quintopia> oh
04:52:25 <quintopia> no
04:52:29 <quintopia> hmm
04:52:32 <Gregor> It's rwins-lwins
04:52:37 <quintopia> yeah
04:52:44 <Gregor> So if left wins it's negative, if right wins it's positive.
04:52:46 <Gregor> It's like strcmp
04:52:46 <quintopia> i'll need to modify winner then
04:52:58 <quintopia> or make a new thing like it
04:53:08 <quintopia> because i need the actual values pwins and qwins
04:53:26 <Gregor> Ohohohoh
04:53:33 <Gregor> Sorry, I didn't know you needed exactly that.
04:53:49 <Gregor> The fact that that's not reported isn't even report.c's fault, the interpreter doesn't report it.
04:54:02 -!- mycroftiv has quit (Ping timeout: 255 seconds).
04:54:05 <quintopia> shit
04:54:17 <quintopia> so i have to hack that too
04:54:28 <quintopia> gearlance.c right?
04:56:32 -!- sftp has quit (Remote host closed the connection).
05:00:23 <quintopia> ah, found the relevant score summarization data
05:10:37 <pikhq> Oh, hey.
05:10:59 <pikhq> Franklin's daylight savings time proposal was almost certainly meant as a joke.
05:11:12 <pikhq> 'Franklin's essay also "suggested" ringing church bells and firing cannons at dawn to ensure that nobody would be excluded from his benevolent gesture.'
05:11:32 <pikhq> So, I don't have to blame *him* for it.
05:11:39 <pikhq> Just everyone who failed at reading comprehension.
05:15:59 <Sgeo> "Your readers, who with me have never seen any sign of sunshine before noon, and seldom regard the astronomical part of the almanac, will be as much astonished as I was, when they hear of his rising so early; and especially when I assure them, that he gives light as soon as he rises. I am convinced of this."
05:16:49 * pikhq still adores these guys: http://pdos.csail.mit.edu/scigen/
05:17:06 <pikhq> I'm especially fond of the randomly-generated *speeches* they gave.
05:28:08 -!- poiuy_qwert has quit (Quit: Leaving).
05:37:00 -!- mycroftiv has joined.
05:54:37 -!- oerjan has joined.
05:57:30 -!- asiekierka has joined.
05:57:31 -!- asiekierka has quit (Read error: Connection reset by peer).
05:57:37 -!- asiekierka has joined.
05:58:05 <oerjan> <quintopia> so easy to get rid of noobs
05:58:14 <oerjan> HEY, STOP THAT.
05:58:20 <quintopia> twasn't me!
05:58:23 * quintopia points at PH
05:58:40 * oerjan swats phantom_hoover in absentia -----###
05:59:21 <oerjan> 16:13:37 --- join: elliott (~elliott@unaffiliated/elliott) joined #esoteric
05:59:21 <oerjan> 16:13:44 --- quit: elliott (Remote host closed the connection)
05:59:23 <oerjan> i see.
05:59:31 <oerjan> :(
06:01:46 <quintopia> do you miss him? :P
06:01:56 <oerjan> yes.
06:02:17 <oerjan> also i have a bad conscience about this.
06:02:18 <quintopia> cry a bit then
06:03:30 <oerjan> not _that_ bad.
06:04:20 <quintopia> hehe
06:06:46 <oerjan> 21:11:32 <pikhq> So, I don't have to blame *him* for it.
06:06:46 <oerjan> 21:11:39 <pikhq> Just everyone who failed at reading comprehension.
06:07:54 <oerjan> just be glad they didn't all misinterpret swift in the same way :D
06:08:37 <pikhq> Hey, I'm an atheist now. Aren't we supposed to eat babies? :P
06:08:51 <oerjan> SO I'VE HEARD
06:09:01 <pikhq> Omnomnom.
06:09:11 <pikhq> Do put the baby on the plate!
06:10:55 <augur> oerjan!
06:11:05 <oerjan> g'day
06:11:18 <augur> figure the puzzle out yet?
06:11:30 <oerjan> i asked you a couple more questions in the logs
06:11:40 <augur> i dont log read. ask them again
06:12:06 <oerjan> well one was if you can always add another 4-cycle
06:12:22 <oerjan> another was if 1 2-cycle + 1 6-cycle is a member
06:12:22 <augur> you mean just take an arbitrary graph and tack on a 4-cycle?
06:12:31 <oerjan> an arbitrary member
06:12:34 <augur> no.
06:12:41 <oerjan> counterexample?
06:12:52 <oerjan> or was that to the other.
06:12:57 <augur> its to both, actually
06:13:45 <augur> well, the adding a 4-cycle thing might actually be true
06:13:47 <augur> lemme think about this
06:14:04 <oerjan> well i more or less assumed you'd have included it. is the new example list also still with all subgraph examples included?
06:14:10 <augur> no, it's not true.
06:14:16 <oerjan> oh.
06:15:10 <augur> yes, all the puzzles are complete. no subgraphs are ever missing the way im doing this
06:15:21 <oerjan> ok.
06:15:38 <oerjan> hm, does that mean you have to check subgraphs to check a graph
06:16:05 <oerjan> ?
06:16:41 <augur> actually, maybe the adding-a-4-cycle is true. its hard to know :p
06:17:09 <oerjan> i didn't see anything contradicting it in the examples, at least
06:20:24 <augur> so any ideas about whats underlying this?
06:20:32 <oerjan> not yet
06:20:55 <augur> theres some stuff that seems plausible to ask that you're not asking
06:21:15 <augur> also did you see the second puzzle showing a natural subset of the first puzzle?
06:21:18 <oerjan> undoubtedly
06:21:35 <oerjan> i'm looking at it but there are no bells ringing yet
06:21:53 <augur> the bells dont have to be clearly related
06:22:44 <oerjan> does the second list include everything from the first that is also in the natural subset?
06:22:54 <augur> what do you mean
06:23:17 <oerjan> you didn't leave out anything from the first puzzle list that should be in the second list?
06:23:22 <augur> no
06:23:57 <augur> im going to try to keep that sort of thing synchronized, so if i reveal more in one puzzle, and another puzzle is a natural subset, i'll reveal more there too
06:24:34 <augur> similarly, if i reveal more about the subset, i'll reveal some more of the superset (in a natural way) that keeps them synched
06:25:39 <oerjan> does the second puzzle also have the disjoint union property?
06:26:07 <augur> really what i could do is just mark the first puzzle's subset members with some sort of adornment
06:26:23 <oerjan> hm
06:26:39 <augur> nah, it wont be as easy to see
06:26:53 <oerjan> if you replace a 6-cycle by a 2-cycle, is that still a member?
06:26:53 <augur> also, no, regarding the second puzzle
06:27:00 <oerjan> (the first puzzle)
06:27:12 <augur> maybe. im not sure.
06:27:40 <augur> you're thinking too much about the details. you need to step back a bit.
06:27:50 <oerjan> is the membership related to whether the graph can be embedded in some larger structure?
06:28:36 <augur> its possible, but i have no idea.
06:29:11 <oerjan> is it related to linguistics?
06:29:14 <augur> no.
06:29:49 <oerjan> related to anything other than math?
06:29:57 <augur> depends on what you eman by math.
06:30:26 <oerjan> something non-abstract?
06:30:35 <augur> yes, i suppose.
06:31:06 <augur> if i may ask a leading question
06:31:09 <oerjan> hm what about flow?
06:31:17 <augur> rephrase
06:31:31 -!- pikhq_ has joined.
06:31:36 -!- pikhq has quit (Ping timeout: 246 seconds).
06:31:41 <oerjan> obviously there are no sinks or sources, but is it related to something flowing between the vertices?
06:31:50 <oerjan> (other than just the edges)
06:32:00 <augur> ehhhh sort of.
06:32:32 <oerjan> ok lead away
06:33:11 <augur> well, let me just state
06:33:23 <oerjan> hm you pointed out initially they were digraphs. is that an intrinsic part of the property or just a corollary?
06:33:50 <oerjan> *of the construction, perhaps
06:33:59 <augur> well, lets say that being digraphs is what you'd expect given how they're generated
06:34:26 <augur> in that they're not generated from some formula about graphs, as such.
06:35:16 <augur> think of what graphs are
06:35:20 <augur> beyond merely graphs
06:35:39 <oerjan> they are diagrams, also relations
06:36:04 <augur> what sorts of diagrams, what sorts of relations
06:36:45 <oerjan> _any_ relation is a graph, really. diagrams with dots and arrows.
06:36:58 <augur> yes graphs are diagrams with dots and arrows
06:37:05 <augur> but you dont just have diagrams floating around in space
06:37:11 <augur> devoid of, shall we say, meaning
06:37:49 <augur> given how little i know about the mathematical properties it should be clear that im not cranking a graph theory machine, here
06:38:12 <oerjan> do the vertices correspond to two sets of natural objects?
06:38:19 <augur> what do you mean
06:39:22 <oerjan> they're digraphs so the vertices can be divided into two sets with all edges between the sets. but is this division based on some natural distinction between vertices in the construction?
06:40:00 <augur> you mean can the vertices be partitioned into two sets which each constitute some natural class?
06:40:05 <oerjan> do the vertices represent people somehow?
06:40:09 <augur> no
06:40:17 <oerjan> and yes
06:41:13 <augur> let me reiterate that this puzzle is unusually appropriate for this channel
06:41:25 <oerjan> ...do they represent esolangs?
06:41:30 <augur> no
06:41:49 <augur> well i suppose that depends on what you mean by esolang!
06:41:51 <oerjan> flow control?
06:42:01 <augur> these days almost anything can be an esolang!
06:42:16 <oerjan> indeed
06:42:17 <augur> flow control, not as such
06:42:36 <oerjan> grammar?
06:43:10 <augur> again i suppose that depends on what you mean by grammar, given the broadness of this concept
06:43:15 <augur> you're in the right area
06:43:26 <augur> now look at the damn graphs
06:43:36 <augur> dont pay attention to the members of the graphs
06:43:39 <augur> just look at what they _are_
06:43:43 <augur> digraphs
06:43:50 <augur> arrows from nodes to nodes
06:45:30 <augur> if there is some non-graph-theoretic rational behind having graphs at all
06:45:34 <augur> what could that rational be
06:45:58 <oerjan> a relation between the vertices.
06:46:04 <augur> no shit sherlock
06:46:17 <augur> but that just is what a graph is
06:46:30 <augur> its not some sort of prior rational for having graphs at all
06:47:05 <oerjan> i'm sorry but the construction workers have started _and_ the neighbors' dog is barking. my brain shall now be unusable for the next 8 hours.
06:49:19 <quintopia> what does <defunct> mean after a process?
06:49:32 <quintopia> aka, what likely happened to the process
06:50:00 <augur> im going to drive home. while im doing so, contemplate what would naturally lend itself to yielding directed graphs as a way of visualization
06:50:03 -!- augur has quit (Remote host closed the connection).
06:50:38 <oerjan> ...augur didn't get that part about unusable brain, i take.
07:03:45 <oerjan> they're supposed to eventually start drilling as well. i don't know what i will do then but you will probably hear about it on the news.
07:10:56 -!- azaq23 has joined.
07:14:29 -!- augur has joined.
07:22:29 * quintopia gifts oerjan some noise-cancelling headphones
07:28:06 <oerjan> did i mention the vibrations in the ground? probably not, as they hadn't started yet.
07:28:19 <augur> oerjan: earth quake?
07:28:29 <oerjan> no...
07:28:39 <augur> loud bassline?
07:28:48 <augur> thumpa thumpa thumpa
07:28:55 <oerjan> chainsaw, possibly.
07:29:03 * augur plays you music
07:35:16 <oerjan> fortunately i shall have some respite from horror as i have a dentist appointment today.
07:35:37 <augur> x.x
07:35:43 <augur> oerjan: so, any ideas?
07:36:00 <oerjan> you think i'm _joking_ about my brain being unusable?
07:36:02 <oerjan> ->
07:36:05 <augur> o ok
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:25:47 -!- Sgeo has quit (Read error: Connection reset by peer).
08:26:16 -!- Sgeo has joined.
08:27:55 -!- azaq23 has quit (Ping timeout: 252 seconds).
08:45:14 -!- azaq23 has joined.
08:55:52 <cheater-> oerjan: the only thing that can help you now is LOUD MUSIC
08:56:38 <oerjan> i guess i'm doomed, then.
08:57:08 * quintopia slaps oerjan around a bit with a DOOMSDAY DEVICE
09:00:47 -!- oerjan has quit (Quit: leaving).
09:29:35 <fizzie> My Emacs just started to bleed. That can't be a good sign.
09:30:11 <fizzie> All other windows are just fine, but any Emacs windows I open get their backgruond replaced by this flickery red mess.
09:30:18 <fizzie> Which doesn't show up in screenshots.
09:39:19 <Vorpal> <quintopia> i see no reason bots shouldn't sabotage each other. <-- ooh Bot Wars. I wonder if we could sell the concept to some TV station!?
09:39:47 <Vorpal> fizzie, what, seriously?
09:40:03 <Vorpal> fizzie, what did you change that might have caused it?
09:40:09 <fizzie> Nothing.
09:40:17 <Vorpal> fizzie, no system upgrades?
09:40:17 <fizzie> I mean, I didn't even start a new Emacs.
09:40:20 <Vorpal> oh
09:40:23 <Vorpal> wtf
09:40:34 <Vorpal> fizzie, does it happen if you start a new instance?
09:40:37 <fizzie> Yes.
09:40:44 <Vorpal> that is even weirder
09:40:47 <fizzie> I did xlock the screen when I went to lunch; when I came back, it was like this.
09:40:58 <Vorpal> fizzie, maybe some emacs virus?
09:41:10 <Vorpal> I mean, if it is only emacs, hardware issues sound unlikely
09:41:19 <fizzie> Since it doesn't show up in a "xwd" dump at all, I'm inclined to believe some sort of closer-to-display-drivers thing.
09:41:29 <fizzie> Don't know what Emacs might be doing differently to trigger it, though.
09:41:32 <Vorpal> hm
09:41:57 <fizzie> Hmm.
09:42:11 <Vorpal> fizzie, you said it flickered to red? Do you mean flicker from normal to red? In that case maybe the dump got it just between the flickers?
09:42:14 <fizzie> If I xwd the Emacs window, then xwud-view the image, the resulting window flickers too.
09:42:34 <Vorpal> wha...
09:42:44 <Vorpal> fizzie, what if you xwud other windows?
09:43:00 <fizzie> Oh, it even flickers if I convert the .xwd to a .png and then open it in eog, but only the pixels of the background color.
09:43:07 <Vorpal> do you have some strange bg colour there normally?
09:43:09 <fizzie> I didn't notice this before because it only flickers in one screen.
09:43:15 <Vorpal> maybe it bugged out on that specific colour
09:43:23 <fizzie> Yes, that seems likely.
09:43:50 <fizzie> That specific color on this specific screen.
09:44:25 <Vorpal> fizzie, what about dimensions of the emacs window?
09:44:33 <fizzie> Doesn't matter.
09:44:44 <fizzie> Gimp's color-picker "Current" block flickers too if I set it to the background color.
09:44:45 <Vorpal> power cycling that monitor might help
09:45:04 <fizzie> Aw, that fixed it. :/
09:45:08 <Vorpal> heh
09:45:16 <Vorpal> fizzie, didn't want it fixed?
09:45:22 <fizzie> Well, it was sort of interesting. :p
09:45:30 <Vorpal> true
09:45:39 <Ilari> What color it was that flickered?
09:46:15 <fizzie> Ilari: #1e1e27.
09:46:21 <Vorpal> Ilari, out of interest, is that Finish word order? ("it was" in a question)
09:48:07 <fizzie> I think both word orders ("-- mikä väri oli se joka --" and "-- mikä väri se oli joka --") are understandable Finnish, though maybe "what color was it that" would be more canonical.
09:48:48 <fizzie> Some other pixels in the Gimp color-browser gradient did flicker a bit too, so it might not have been limited to exactly that one value.
09:50:06 <fizzie> Now that my screen no longer bleeds, I guess I should try to figure out some way around this "probabilities are no longer finite" error message this GMM tool is giving me.
09:57:24 <Vorpal> fizzie, what is GMM?
09:57:42 <Vorpal> fizzie, also that error reminds me of a certain book
09:57:46 <Vorpal> I'm sure you know which one
10:04:07 <fizzie> Gaussian mixture model. You know, https://chart.googleapis.com/chart?cht=tx&chl=p%28\mathbf{x}%29=\sum_{k=1}^K%20\pi_k%20\mathcal{N}%28\mathbf{x};\,\,\mathbf{\mu_k},\mathbf{\Sigma_k}%29 sort of stuff.
10:04:32 <fizzie> (Couldn't resist a chance to use that chart thing.)
10:05:17 <Vorpal> fizzie, since when does google render tex?
10:05:23 <fizzie> Since ages ago.
10:05:26 <Vorpal> heh
10:06:16 <Vorpal> fizzie, also I have no idea what a Gaussian mixture model is. What is the fancy-looking N in that equation?
10:06:18 <fizzie> At least from 2010-11 onwards.
10:06:24 <fizzie> It's the normal distribution.
10:06:28 <Vorpal> aha
10:06:41 <Vorpal> okay it does make a vague sort of sense now
10:06:43 <fizzie> Sort of a messy notation; the pdf of it.
10:07:37 <fizzie> Just a weighted sum of normal distributions. I left out the part that https://chart.googleapis.com/chart?cht=tx&chl=\sum_{k=1}^K\pi_k=1 of course.
10:08:04 <Vorpal> that the sum is 1?
10:08:13 <fizzie> Yes, but I again couldn't help myself.
10:08:24 <Vorpal> XD
10:09:01 -!- GreaseMonkey has quit (Quit: The Other Game).
10:09:23 <fizzie> Still don't know why the toolbox doesn't want to generate a model for this data.
10:09:48 <Vorpal> fizzie, probably because of probabilities are no longer being finite!
10:10:51 <Vorpal> fizzie, now if we could only get infinite improbabilities...
10:12:25 <fizzie> There is one sort-of known issue with EM training a mixture model like this, which is that it can happen that one component of the mixture converges to cover a single data point with 0 variance. I would think the code would take that into account, though.
10:13:06 <Vorpal> fizzie, is it mathlab?
10:13:39 <fizzie> It's a third-party pile-of-code for MATLAB, from http://www2.it.lut.fi/project/gmmbayes/downloads/src/gmmbayestb/
10:13:41 <Ilari> APNIC this month (16 days): 18 458 112 addresses allocated (1.100 blocks). On IPv6 front: 1 114 118 /48s. I think records on IPv4 allocations are going to get slammed again.
10:14:03 <Vorpal> Ilari, I assume the events in Japan had an effect on the allocation rate?
10:14:58 <Ilari> No idea. Japan has highly spiky allocation profile, making changes difficult to detect.
10:15:09 <Vorpal> Ilari, heh, anyone know why that is the case?
10:15:19 <fizzie> They did get that /9 not long ago, shouldn't that last them for at least a few moments?
10:15:48 <Vorpal> fizzie, a /9 is very large for a single country to get
10:16:19 <fizzie> Single company, even, though I guess they're bit of a monopoly there.
10:16:25 <Ilari> It was for NTT Japan.
10:16:30 <Vorpal> NTT?
10:16:38 <Ilari> Or was it NNT?
10:16:41 <Vorpal> NNT?
10:16:54 <fizzie> Nippon Telegraph and Telephone Corporation, apparently.
10:16:56 <Vorpal> ah
10:16:56 <Ilari> NTT Japan apparently.
10:17:00 <Vorpal> fizzie, a single company getting a /9 these days seems absurd!
10:17:13 <Ilari> APNIC has allocated /8 at once twice.
10:17:51 <Vorpal> seen during mkinitcpio: ERROR: Root file system type detection failed.
10:18:01 <Vorpal> I'm not too worried, since I don't use the distro kernel for booting
10:18:14 <fizzie> My coauthor for this stuff is in Japan, only realized yesterday evening I hadn't though about asking her how it's going. Though I guess Nagoya is reasonably far from the disaster area.
10:18:48 <Vorpal> I only use the distro kernel because the nvidia X drivers package depends on the nvidia kernel driver package which depends on the kernel package.
10:19:49 <fizzie> "log-likelihood diff 179769313486231570814527423731704356798070567525844996598917476803157260780028538760589558632766878171540458953514382464234321326889464182768467546703537516986049910576551282076245490090389328944075868508455133942304583236903222948165808559332123348274797826204144723168738177180919299881250404026184124858368 on round 1"
10:19:57 <fizzie> I'm no expert, but that sounds like a rather large number.
10:20:14 <Vorpal> heh
10:20:19 <Vorpal> fizzie, still finite though
10:20:23 <Ilari> For comparsion: January (31 days): 23 735 040 (1.415) and February (28 days): 22 589 440 (1.346).
10:21:12 <fizzie> >1 blocks a month is already quite a burn rate.
10:21:44 <Ilari> Extrapolation to 31 days gives 35 762 592 (2.132).
10:22:12 <Vorpal> Ilari, so when will it run out?
10:22:32 <Vorpal> fizzie, btw what is it that you are doing this GMM stuff over
10:23:17 <Ilari> Current best estimates are in mid-April. Formal models tend to guess late (underestimating the growth of allocation rates).
10:23:54 <fizzie> They're going to go to that "last /8" mode first, though, and isn't that policy going to last for a while?
10:23:57 <Vorpal> Ilari, yeah wasn't it supposed to be in September just a while ago?
10:25:25 <Ilari> It was supposed to be October-November some time ago. Then September. Then May-July. Now it seems even May is too optimistic.
10:26:47 <fizzie> Vorpal: The GMM is for converting some heuristic (read: don't have much of a theory behind them) numbers in one domain to something that would correlate with observation uncertainties in another domain, to summarize it in one sentence.
10:27:55 <Vorpal> fizzie, pretty vague eh
10:28:06 <Ilari> Heck, it wasn't very many months ago when the estimated IANA depletion was in June. Then came the surprise allocations to AFRINIC, ARIN and RIPENCC.
10:28:07 <Vorpal> but I guess you don't want to reveal too much before publishing
10:31:30 <Ilari> Exceeding 2 blocks in calendar month would be pretty wild (2 blocks in 30 day window has IIRC already been exceeded).
10:31:47 <fizzie> Vorpal: This part of it is (up to some degree, anyway) already published stuff, actually, from INTERSPEECH 2010. Unfortunately their proceedings are behind a "ISCA Members Area (membership starts from 15 Euros)" HTTP authentication request. :p
10:32:34 <Vorpal> fizzie, no clue if the proxy at my university allows accessing that one
10:32:39 <Vorpal> *shrug*
10:33:16 <fizzie> Ours doesn't, which is somewhat surprising; it tends to have a good coverage otherwise.
10:33:47 <Ilari> Heck, picking the last available 30-day window: 2.089 blocks.
10:34:11 <fizzie> (Actually the first author seems to have a copy on his page, but I doubt you're *that* interested.)
10:34:14 <Vorpal> fizzie, can't find any matching *ISCA*
10:34:30 <Vorpal> fizzie, indeed I'm not.
10:34:39 <Vorpal> fizzie, what about a layman description of it? ;P
10:37:56 <fizzie> Well, um; we take our source-domain heuristic numbers for training data, concatenate them with the "correct" answers (since it's training data we constructed, we have those too), build a GMM model for the distribution of the concatenated vectors, then for test-data use the model to give MAP predictions for the latter half (target domain values) given the first half (source domain values).
10:38:11 <fizzie> Or that's what I'd do if this silly thing would construct a model for me.
10:38:19 <fizzie> Instead it's just all
10:38:21 <fizzie> log-likelihood diff NaN on round 20
10:38:21 <fizzie> fix cycle 20, fix loops 1 1 1 1 1
10:38:21 <fizzie> ??? Error using ==> gmmb_em at 161
10:38:21 <fizzie> Probabilities are no longer finite.
10:38:23 <Vorpal> hm
10:38:56 <Vorpal> fizzie, what sort of domain is that? Something corresponding to some physical application?
10:39:41 <fizzie> The source domain is a (warped) spectral one again; target is the cepstrum thing our recognizer likes to eat.
10:41:05 <Vorpal> fizzie, oh sound?
10:41:53 <fizzie> Speech stuff; it's what I do, after all.
10:42:07 <Vorpal> fizzie, yeah and irc log statistics :P
10:42:21 <fizzie> That's just a hobby. :p
10:42:24 <Vorpal> true
10:53:38 <Ilari> Also, APNIC has 5-day FIFO policy, so the request for that 4M block that was recently allocated to Japan was made before the 9.0 earthquake and following tsunami.
11:12:18 -!- variable has quit (Ping timeout: 260 seconds).
11:13:16 -!- variable has joined.
11:21:36 -!- asiekierka has quit (Read error: Operation timed out).
11:22:09 -!- asiekierka has joined.
12:47:02 -!- ais523 has joined.
12:59:14 -!- ais523 has quit (Read error: Connection reset by peer).
12:59:37 -!- ais523 has joined.
13:15:41 -!- augur has quit (Remote host closed the connection).
13:17:18 -!- FireFly has joined.
13:30:29 -!- ais523 has quit (Ping timeout: 260 seconds).
13:35:06 -!- HackEgo has quit (Ping timeout: 252 seconds).
13:35:24 -!- EgoBot has quit (Ping timeout: 252 seconds).
13:37:34 -!- sftp has joined.
13:41:06 -!- cheater00 has joined.
13:41:48 -!- ais523_ has joined.
13:43:25 -!- cheater- has quit (Ping timeout: 255 seconds).
13:47:25 -!- asiekierka has quit (Read error: Operation timed out).
13:47:37 -!- asiekierka has joined.
13:52:05 -!- ais523_ has quit (Quit: Page closed).
13:59:05 -!- augur has joined.
14:11:51 -!- jix has quit (Quit: rebooting).
14:17:08 -!- copumpkin has quit (Quit: Computer has gone to sleep.).
14:23:19 -!- Lymia has quit (Quit: ==(>^w^)> ==(> >.<)>).
14:30:57 -!- copumpkin has joined.
14:31:57 -!- jix has joined.
14:32:55 -!- copumpkin has quit (Client Quit).
14:40:05 -!- ais523_ has joined.
14:49:05 -!- Mathnerd314 has quit (Ping timeout: 255 seconds).
14:50:04 -!- Sgeo has quit (Ping timeout: 240 seconds).
15:02:41 -!- copumpkin has joined.
15:05:20 <Gregor> Sent an email to the zlib guys asking if they have any use for libz.so :P
15:06:01 <ais523_> what does libz.so do?
15:06:05 <fizzie> Are you going to make BUCKETS OF MONEY with your domain dealies?
15:06:11 <ais523_> oh, the domain name
15:06:15 <ais523_> who got libc.so in the end?
15:06:26 <Gregor> ais523_: It hasn't ended yet.
15:06:32 <Gregor> It hasn't started yet :P
15:07:28 <Gregor> fizzie: I'm probably going to LOSE buckets of money on libc.so :P
15:07:33 <Gregor> But GAIN buckets of geek cred if I win it.
15:07:45 <Gregor> That's right. BUCKETS OF CRED.
15:08:31 <fizzie> With a name like that, you can probably name the price. I mean, facebook bought fb.com for $8.5 million.
15:09:16 <fizzie> Surely every group developing a libc will be fighting over it.
15:09:57 <Gregor> Right, but if I DO get it, seeing as that it's a closed auction, then that means they WEREN'T fighting over it :P
15:10:05 <Gregor> And besides that, the groups developing libcs have no money.
15:10:17 -!- ais523_ has quit (Quit: Page closed).
15:10:19 <Gregor> Because they're either profitless Unix ventures or F/OSS.
15:11:21 -!- azaq23 has quit (Quit: Leaving.).
15:11:32 <quintopia> so if you get it, whatcha doing with it?
15:12:10 -!- azaq23 has joined.
15:12:14 <quintopia> also, what country has so as their TLD?
15:12:19 <Gregor> Vanity domain names and I'll see if somebody who's already got a decent man pages site wants to collaborate. Somalia.
15:12:28 <Gregor> Erm
15:12:28 <Gregor> *vanity email addresses
15:12:32 -!- azaq23 has quit (Client Quit).
15:12:42 <Gregor> Basically I don't want to think too far ahead because my chances of getting it are low :P
15:19:11 -!- ais523 has joined.
15:23:30 -!- ais523 has quit (Ping timeout: 252 seconds).
15:24:44 -!- pumpkin has joined.
15:25:56 -!- copumpkin has quit (Ping timeout: 260 seconds).
15:28:41 -!- copumpkin has joined.
15:28:59 -!- pumpkin has quit (Ping timeout: 248 seconds).
15:32:21 -!- myndzi has quit (Ping timeout: 260 seconds).
15:34:13 -!- BeholdMyGlory has joined.
15:53:39 -!- yorick has joined.
15:55:37 -!- pumpkin has joined.
15:55:37 -!- pumpkin has quit (Changing host).
15:55:37 -!- pumpkin has joined.
15:56:50 -!- copumpkin has quit (Ping timeout: 260 seconds).
16:14:31 -!- Phantom_Hoover has joined.
16:15:30 -!- MigoMipo has joined.
16:17:46 -!- azaq23 has joined.
16:22:27 -!- asiekierka has quit (Read error: Connection reset by peer).
16:25:30 <Phantom_Hoover> *Still* no elliott?
16:25:57 <oklopol> when elliott decides to do something, he sticks to it
16:26:28 <Phantom_Hoover> He's online, by the looks of things.
16:26:42 <oklopol> he shall return in three years, speaking perfect japanese
16:26:55 <oklopol> well find him
16:27:03 <oklopol> i wanna know where he is
16:27:04 <Phantom_Hoover> That's rather trickier.
16:27:22 <oklopol> yes
16:27:31 <Phantom_Hoover> There's an elliottt in #haskell.
16:27:45 <Phantom_Hoover> But that's someone entirely different.
16:27:51 <oklopol> i don't think that's the same right
16:28:10 <Phantom_Hoover> Hmm.
16:28:25 <Phantom_Hoover> I suppose I could look through all the Tunes logs and see if he's joined any of those channels.
16:28:42 <oklopol> yes, do that!
16:29:03 <Phantom_Hoover> GTG, will continue upon return.
16:30:15 <Ilari> APNIC down 0.21: 1k to Bangladesh, 1k+/32 to Malaysia, 2x512k+2x128k+64k+32k to China, 32k to Pakistan, 2M(!!!) to India, 2k to Japan, /32 to New Zealand.
16:30:47 <Ilari> Largest available block is now 1M (which there are 5 of).
16:31:52 -!- asiekierka has joined.
16:33:07 <Ilari> Total 256k block count is now 77.
16:35:05 <Ilari> That's 1.20 blocks of space.
16:37:40 <Ilari> That count dropped 12 today.
16:39:48 -!- cal153 has quit (Ping timeout: 252 seconds).
16:40:18 <Ilari> 0.024 from small blocks space.
16:51:52 <Gregor> Somebody accused me of wearing a radioactive tie in recognition of Japan's nuclear problems :P
16:52:58 <Ilari> During last 31 days, 125 of 256k blocks have been allocated. Those blocks are expected to run out sooner than APNIC (and then allocations really start to fragment the address space).
17:03:45 -!- oerjan has joined.
17:04:52 <lifthras1ir> Ilari, i've just traced the history of APNIC depletion since January. awesome.
17:05:02 <lifthras1ir> more than /8 in a week? :/
17:10:28 <oerjan> <Gregor> Somebody accused me of wearing a radioactive tie in recognition of Japan's nuclear problems :P
17:10:47 <oerjan> brilliant business idea or brilliant business idea?
17:10:55 <Gregor> X-D
17:14:16 -!- cal153 has joined.
17:14:17 -!- myndzi has joined.
17:14:23 <lifthras1ir> \o/
17:14:24 <myndzi> |
17:14:24 <myndzi> /<
17:14:34 -!- lifthras1ir has changed nick to lifthrasiir.
17:15:18 <lifthrasiir> uhm, it is better changing a spare nickname...
17:15:31 <lifthrasiir> lifthras1ir and lifthrasiir are way hard to distinguish.
17:16:39 -!- myndzi has quit (Client Quit).
17:16:44 -!- myndzi has joined.
17:20:07 <tswett> Whoever pinged me, if anyone: pong.
17:21:08 <oerjan> no no, we were just talking about the new contact sport of tswetting.
17:21:14 <tswett> Oh, okay.
17:21:21 <tswett> I am a fan of that sport.
17:21:28 -!- ais523 has joined.
17:22:57 <Gregor> It is a very "contact" sport.
17:23:00 <Gregor> Where by "contact" ...
17:23:05 <Gregor> I mean "banned in 47 states"
17:23:49 <tswett> Huh. Must involve some breaches of esoteric contract law.
17:26:34 -!- cheater- has joined.
17:28:18 -!- cheater00 has quit (Ping timeout: 255 seconds).
17:31:27 <Gregor> Whenever I use spreadsheets I realize how terrible they are and want to implement my own, then I go "OH GOD NO NOT THAT"
17:37:13 <quintopia> implement concealsheets instead. it's what all corporations /really/ want. you'll make a fortune.
17:44:29 -!- ais523 has quit (Ping timeout: 255 seconds).
17:44:53 -!- ais523 has joined.
17:50:56 <quintopia> ais523: is there a name for the class of languages in which it is possible to specify any bijection between sets (and only such functions)?
17:51:07 <Phantom_Hoover> Just checked today's clog logs; he's not on any of them.
17:51:25 <ais523> quintopia: I'm not sure
17:51:29 <ais523> and I'm probably the wrong person to ask
17:52:22 <oerjan> quintopia: hm i think i came up with exactly that once when i tried to think of a semantic foundation for reversible languages
17:52:57 <oerjan> so that's my answer
17:52:58 <quintopia> oerjan: I'm going to call it "Permutation-complete" if not
17:53:31 <oerjan> well for finite sets, it's the same, for infinite you need to consider computability and stuff
17:53:52 <quintopia> (the oh
17:53:57 <quintopia> i meant to say finite
17:54:01 <quintopia> i was thinking finite
17:54:08 <quintopia> but that disappeared from my message somehow
17:54:09 <oerjan> aha.
17:54:39 <oerjan> if you add phase shifts you essentially have quantum computation
17:55:15 <quintopia> i thought the phase in QC was just to make it a vector space...
17:55:21 <quintopia> i should read that book again
17:55:33 <oerjan> the phase is to allow you to get more than just permutations
17:55:56 <oerjan> well from a qc perspective
17:56:10 <quintopia> i have a copy here of "quantum computation and quantum information" by Nielsen and Chuang
17:56:45 <oerjan> i've never read an actual book about it, but i know the basic hilbert space formalism, and what a qubit is
17:58:05 <augur> oerjan!
17:58:50 <quintopia> "the Hadamard, phase, CNOT, and pi/8 gates form a family of gates from which any unitary operation can be approximated, and thus is a universal set of gates"
17:59:10 <quintopia> what is the phase gate
17:59:20 -!- cheater- has quit (Ping timeout: 255 seconds).
18:01:24 -!- oerjan has quit (Quit: leaving).
18:03:01 <quintopia> oh aha, it's [1,0;0,i]
18:03:41 -!- cheater00 has joined.
18:04:49 <quintopia> so it's like flip across the Re(x)=Im(x) line, i guess
18:19:18 -!- impomatic has joined.
18:23:45 -!- asiekierka has quit (Ping timeout: 260 seconds).
18:33:34 -!- Phantom_Hoover has quit (Quit: Leaving).
18:33:52 -!- Phantom_Hoover has joined.
18:37:01 <cheater00> hello eso
18:37:04 <cheater00> how are you
18:37:35 <Gregor> Why are my bots down >_<
18:38:38 -!- EgoBot has joined.
18:38:41 -!- HackEgo has joined.
18:39:04 <Gregor> `echo I would be better if my bots weren't down :P
18:39:06 <HackEgo> I would be better if my bots weren't down :P
18:40:32 <Gregor> Since both bots use multibot, there is literally no reason for them to be separated. Combining them would be as simple as cp -R. AND YET I DON'T.
18:45:25 <quintopia> i suppose it's because you want to retain control of the non-hackable things
18:45:28 <quintopia> control freak
18:45:38 <Phantom_Hoover> Multibot?
18:45:45 <Gregor> quintopia: In no way would it compromise control of the non-HackBot things.
18:46:11 <Gregor> Phantom_Hoover: multibot is my simple IRC bot that calls out to shell (or any other kind of) commands when presented with queries.
18:46:20 <quintopia> Gregor: oh you would have two separate folders running on the same instance of multibot?
18:46:34 <Phantom_Hoover> So EgoBot uses it too?
18:46:44 <fizzie> Gregor: Would you call the combination HackEgoBot, then?
18:46:49 <Gregor> quintopia: The commands for ! and for ` would simply not refer to the same directory.
18:46:53 <Gregor> Phantom_Hoover: Yes.
18:46:55 <Gregor> fizzie: I spose :P
18:47:29 <quintopia> Gregor: well that's the reason not to combine them then. having two command prefixes for the same bot instance is annoying.
18:47:35 * quintopia eyes lambdabot fumingly
18:48:06 <lifthrasiir> !echo hello
18:48:06 <EgoBot> hello
18:48:22 <Gregor> quintopia: ... that logic is silly, the two prefixes are already in use, they would just be responded to by the same nick :P
18:48:38 <lifthrasiir> !uname
18:48:40 <quintopia> my logic is impeccable
18:48:45 <Gregor> `uname
18:48:46 <HackEgo> Linux
18:48:47 <lifthrasiir> hmm.
18:48:57 <lifthrasiir> !help
18:48:57 <EgoBot> help: General commands: !help, !info, !bf_txtgen. See also !help languages, !help userinterps. You can get help on some commands by typing !help <command>.
18:49:00 <Gregor> lifthrasiir: ! only response to a specific set of commands, ` is just system()
18:49:08 <Gregor> (Well, it's actually even simpler than system(), but anywho)
18:49:24 <lifthrasiir> Gregor, okay, i thought egobot now uses multibot...
18:49:36 <lifthrasiir> (well it doesn't)
18:49:37 <Gregor> lifthrasiir: multibot is not what enables HackEgo to do the things it does.
18:49:43 <lifthrasiir> hmm?
18:50:18 <Gregor> lifthrasiir: multibot is just a framework for connecting IRC commands to shell commands. The fact that HackEgo goes beyond that and runs requested commands in a protected environment is what's implemented within the commands that multibot calls, not multibot itself.
18:50:23 <quintopia> because what you're saying is "this bot will actually be two bots running under the same nick. there is not one set of features, but two distinct sets of features." at which point i would say "then why not combine the features into the same list if its the same bot?"
18:50:45 <fizzie> !sh uname -a
18:50:45 <EgoBot> Linux gdeskgor 2.6.37-0.slh.2-aptosid-amd64 #1 SMP PREEMPT Sun Jan 9 20:40:11 UTC 2011 x86_64 GNU/Linux
18:50:46 <quintopia> (actually, i think moving all egobot's stuff to hackbot would be pretty sane.)
18:50:51 <lifthrasiir> okay, so multibot itself does not contain any sandboxing capability nor executing arbitrary commands without authorization?
18:50:52 <fizzie> "gdeskgor"?
18:51:01 <Gregor> fizzie: That's my desktop.
18:51:11 <fizzie> What an interesting name.
18:51:18 <fizzie> Do you have a "glapgor" too and so on?
18:51:20 <Gregor> lifthrasiir: Right. All multibot can do is e.g. when encountering a TOPIC command from IRC, run multibot_cmds/TOPIC.cmd with the proper arguments.
18:51:23 <fizzie> "gpadgor".
18:51:39 <Gregor> fizzie: glapgor, gwirgor, grandroidgor
18:51:47 <Gregor> fizzie: Had a gwatchgor for a while.
18:52:05 <lifthrasiir> `echo -ne 4; echo -ne 2
18:52:05 <HackEgo> -ne 4; echo -ne 2
18:52:16 <lifthrasiir> no shell expansion?
18:52:25 <Gregor> `run echo -ne 4; echo -ne 2
18:52:25 <HackEgo> 42
18:52:41 <Gregor> It doesn't do expansion without `run simply so that things like `addquote work as expected.
18:52:54 <Gregor> But anyway, yeah, maybe I'll just merge everything EgoBot does into HackEgo's actual filesystem.
18:53:13 <Gregor> Then EgoBot will be utterly redundant.
18:54:57 -!- cheater- has joined.
18:54:59 <quintopia> i support this plan
18:55:33 -!- cheater00 has quit (Ping timeout: 255 seconds).
18:55:39 <Gregor> That wasn't very doable before, but now that (hopefully) my FS access speed will be in the "non-shit" range, it should be doable ...
18:55:52 <quintopia> non-shit is a good place to be
18:56:08 <Gregor> I had 1MB/s read and 100KB/s write 8-D
18:56:27 <Gregor> That's right: My HD speed was slower than my network speed. Pretty awesome, prgmr! Pretty awesome!
18:56:28 -!- bitmsk has quit (Ping timeout: 252 seconds).
18:56:35 <fizzie> Gregor: So what did you end up doing; switched providers or got them to fix stuff?
18:56:53 <Gregor> fizzie: They claim they're going to move me to a different server.
18:57:33 <Gregor> fizzie: Stupid TOSes keep me stuck to very few VPS providers.
18:57:56 <Gregor> But I'm still glad I found http://www.bestdealvps.com/tos , which is the greatest work of comedy writing I've seen in a while.
18:58:32 <Gregor> There was another great TOS that said all you were allowed to do was run a web server and POP/IMAP/SMTP server, which makes me wonder why you would get a VPS at all :P
18:58:36 <fizzie> Yeah, I tried to do a survey-murvey too, and quite many did indeed have that stupid IRC thing. (And a few I couldn't find anything approaching terms of service for, at least without clicking a "send order" button.
19:00:23 <Gregor> User may not: b) Run stand-alone, unattended server-side processes at any point in time on the server. This includes any and all daemons, such as IRCD. // honestly, did no one with any technical knowledge read this at all?
19:00:38 <fizzie> Gregor: Do not that that part is for "Resource Usage for shared hosting accounts".
19:00:42 <lifthrasiir> Gregor, hm, i've just made a mutually recursive quine for egobot and hackego. how about it? :p
19:00:43 <fizzie> s/not/note/
19:01:14 <quintopia> lifthrasiir: do you think it will work?
19:01:24 <Gregor> lifthrasiir: I believe one of them is configured to ignore the other, but I can't remember which >_>
19:01:31 <quintopia> we can find out
19:01:36 <lifthrasiir> !echo `run X='!echo `run X=xxXXxx;X=${X//x""x/"xx"};echo ${X/X""X/${X//"xx"/x""x}}';X=${X//x""x/"'"};echo ${X/X""X/${X//"'"/x""x}}
19:01:36 <EgoBot> `run X='!echo `run X=xxXXxx;X=${X//x""x/"xx"};echo ${X/X""X/${X//"xx"/x""x}}';X=${X//x""x/"'"};echo ${X/X""X/${X//"'"/x""x}}
19:01:37 <HackEgo> !echo `run X='!echo `run X=xxXXxx;X=${X//x""x/"xx"};echo ${X/X""X/${X//"xx"/x""x}}';X=${X//x""x/"'"};echo ${X/X""X/${X//"'"/x""x}}
19:01:43 <Gregor> fizzie: Except that they don't OFFER shared hosting accounts.
19:01:54 <lifthrasiir> so EgoBot ignores HackEgo. EPIC WIN
19:01:57 <fizzie> Gregor: Yes, but it's still only for shared hosting accounts.
19:01:58 <quintopia> yep
19:02:07 <fizzie> fungot ignores both EgoBot and HackEgo; it's antisocial like that.
19:02:08 <fungot> fizzie: you know 1.7 is out. this erc buffer remarkably similar to the last page of clinger's paper shows up for " grond"
19:02:27 <Gregor> fizzie: ... *brain axplote*
19:02:45 <Gregor> All bots should ignore any nick with 'bot' in it, but then HackEgo doesn't have 'bot' in it :P
19:03:26 <fizzie> Neither does fungot.
19:03:27 <fungot> fizzie: i suspect there being some kind of priorities between your rules. ha ha ha ha
19:03:30 <fizzie> ^ignore
19:03:30 <fungot> ^(EgoBot|HackEgo|toBogE|Sparkbot|optbot|lambdabot)!
19:03:40 <fizzie> This "Sparkbot", I have no idea where it came from.
19:03:51 <fizzie> I've tried to grep my #esoteric logs for it, and can't find anything.
19:04:18 <lifthrasiir> does clog respond to any command?
19:04:38 <Gregor> lifthrasiir: No
19:04:38 <ais523> it responds to ctcp version, IIRC
19:04:46 <ais523> [CTCP] Received CTCP-VERSION reply from clog: CLOG v0.
19:05:21 <Gregor> ais523: Pfff
19:05:23 <fizzie> fungot doesn't ignore myndzi, but I guess you can't set up a loop with just \o/-style stuff since fungot won't respond to that.
19:05:23 <fungot> fizzie: ( google it) to the window
19:05:50 <Gregor> ais523: So glad captain pedantry was here to save the day :P
19:06:10 <ais523> well, most bots don't
19:06:13 <lifthrasiir> `run echo -e '\001ACTION jumps over the lazy dog\001'
19:06:13 * HackEgo jumps over the lazy dog
19:06:42 <Gregor> HackEgo quite intentionally has no protection to prevent that.
19:06:45 <ais523> `run echo -e '\001VERSION\001'
19:07:08 <Gregor> And yes, you could use that to do the world's most awkward and inefficient DDoS attack.
19:07:30 <lifthrasiir> Gregor, well, theoretically.
19:07:42 <Gregor> `echo !echo Why won't EgoBot listen to me *sobs*
19:07:42 <HackEgo> !echo Why won't EgoBot listen to me *sobs*
19:07:49 <Gregor> That's the way :P
19:07:53 <lifthrasiir> in reality it's just slightly annoying though
19:08:28 <Gregor> !echo -e '\xFF\xFF\xFF'
19:08:28 <EgoBot> -e '\xFF\xFF\xFF'
19:08:31 <lifthrasiir> `run echo -e 'x\000y'
19:08:32 <HackEgo> x
19:08:33 <Gregor> ... fail :P
19:08:37 <Gregor> `run echo -e '\xFF\xFF\xFF'
19:08:37 <HackEgo>
19:09:04 <lifthrasiir> anything after \0 will get removed, expected.
19:09:08 <ais523> `run echo -e '\002BOLD\002'
19:09:09 <HackEgo> BOLD
19:09:13 <ais523> this channel strips bold
19:09:23 <fizzie> ^echo !echo `echo
19:09:23 <fungot> !echo `echo !echo `echo
19:09:23 <EgoBot> `echo !echo `echo
19:09:24 <HackEgo> !echo `echo
19:09:42 <Gregor> Classy
19:10:21 -!- cheater00 has joined.
19:10:36 <Gregor> Shhhhhhhhhhhhhhh nobody talk, cheater's back
19:13:06 -!- cheater- has quit (Ping timeout: 255 seconds).
19:13:30 <lifthrasiir> okay there is one cheater in and one cheater out, the net cheater is zero
19:19:12 * Gregor golf clap
19:23:54 <Phantom_Hoover> TV Tropes have outsourced their entire Fetish Fuel section.
19:23:57 <Phantom_Hoover> To *Wikia*.
19:25:01 <Gregor> lolwut
19:28:35 <ais523> Phantom_Hoover: they were in severe trouble because Google threatened to - and actually did at one point - withdraw adverts because they thought there was too much adult content
19:28:47 <Phantom_Hoover> Oh, of course, that.
19:28:49 <ais523> so they've been busy trying to make the whole thing SFW in order to keep their advertising revenue
19:29:11 <Phantom_Hoover> Dear god, why can't we have an internet run by adults.
19:30:01 <Zwaarddijk> we should have an internet running on adultery
19:30:04 <Zwaarddijk> that's what we need
19:30:09 -!- impomatic has quit (Ping timeout: 252 seconds).
19:30:59 <Gregor> Also, golden idols.
19:37:36 <Gregor> Why haven't I ported gelfload to DOS ...
19:37:41 <Gregor> Oh yeah, no good mmap implementation.
19:41:46 <Gregor> Hm, do 32-bit DOS programs not use the MMU?
19:42:13 <ais523> they use DOS extenders which hide all the details
19:42:39 <ais523> and typically they set the MMU to a really simple mapping, like mapping all physical memory into virtual memory with the same addresses, or the same addresses backwards
19:42:51 <Gregor> Mmm.
19:43:04 <Gregor> So unless the extender provided an mmap implementation, you're hosed.
19:43:12 <Gregor> I could still load PIE binaries though ...
19:44:48 <olsner> Note that for Digital Bill, "mainstream audience" apparently means "people who are not interested in technology", including for example those who are interested in podcasts about didgeridoos.
19:44:50 <fizzie> DPMI does mmap: http://www.delorie.com/djgpp/doc/dpmi/api/310508.html
19:45:01 <fizzie> More accurately, DPMI 1.0.
19:45:12 <fizzie> And nobody relly does more that 0.9.
19:47:27 <Gregor> fizzie: Uhh, that doc does not look like mmap ...
19:47:31 <Gregor> It looks like device mapping.
19:47:55 <Gregor> Oh, I was thinking anonymous mmaps, not file/device-backed mmaps >_>
19:47:58 <Gregor> Should've mentioned that.
19:49:06 <fizzie> Well, the whole DPMI api is at that site, you can see if it does what you want.
19:50:16 <fizzie> You can mungle with the LDT to some degree.
19:50:18 <Gregor> Allocate Linear Memory Block [1.0] Allocates a block of page-aligned linear address space. The base address of the block may be specified by the client, and pages within the block may be committed or uncommitted.
19:50:43 <Gregor> This is, like, exactly mmap-anon :P
19:51:52 <fizzie> "While Windows 3.0 includes support for DPMI 0.9,[2] version 1.0 was never fully implemented in Microsoft Windows, so many programs and DOS extenders were mostly only written for version 0.9.
19:51:56 <fizzie> The most famous separate DPMI kernel is probably CWSDPMI; however, it only fully supports DPMI 0.9 and no undocumented "DOS API translation"."
19:52:13 <fizzie> So [1.0] support might be spotty.
19:52:14 <Gregor> djgpp has its own though, doesn't it?
19:52:31 <fizzie> CWSDPMI is what it uses, IIRC.
19:52:32 <Gregor> I don't give a splat if I need a "special" extender, I'm talking about loading ELF binaries on DOS here X-P
19:52:43 <Gregor> Bleh
19:52:49 <Gregor> Then why is this documented at delorie.com X-P
19:53:12 <ais523> because it works with multiple DOS extenders
19:53:17 <ais523> it just recommends one
19:53:20 <fizzie> You might get DOS4GW for free nowadays, it's what Watcom C used to use and there's OpenWatcom.
19:53:38 <ais523> e.g. I use it with HDPMI32 as JPC-RR doesn't emulate CWSDPMI correctly
19:53:42 <fizzie> CWSDPMI might do that part of 1.0 for all I know.
19:54:13 <fizzie> "HDPMI (part of HX DOS Extender) provides "DOS API translation" and almost complete DPMI 1.0 implementation."
19:54:19 <fizzie> That sounds viable too.
19:54:44 <Gregor> HX is friggin' crazy though, I could probably run WinELF under it :P
19:54:49 <Gregor> Almost assuredly in fact.
19:55:20 <Gregor> Put differently, "that's a bit heavy"'
19:56:16 <Gregor> I suppose I could use just the DPMI host part of it though.
19:57:31 -!- Sgeo has joined.
20:13:11 <Ilari> lifthrasiir: More than /8 is week was week of 28th February to 4th March. Of course, there has been those /8 allocations (two of them).
20:13:29 <Phantom_Hoover> Incidentally, elliott responded to a query but he showed no signs of getting over himself/
20:14:54 <ais523> I think he started thinking about Feather
20:14:57 <ais523> I hope he recovers soon
20:27:10 -!- asiekierka has joined.
20:27:28 <Phantom_Hoover> Yay Feather.
20:32:00 -!- Sgeo has quit (Ping timeout: 276 seconds).
20:35:51 -!- asiekierka has quit (Ping timeout: 255 seconds).
20:43:59 -!- Sgeo has joined.
20:50:56 <Gregor> I've once again forgotten what Feather is :P
20:52:39 <Gregor> BEARTATO HAS MY TIE!
20:53:40 * Phantom_Hoover wonders how TV Tropes ended up with an administration so heavy-handed.
20:54:02 <Phantom_Hoover> Discussion threads are _routinely_ nuked, and changes are made by fiat so often I've lost count.
20:55:29 <pikhq_> Spring break achieved.
20:55:38 <Gregor> pikhq_: Huzzah?
20:57:10 <ais523> <Gregor> I've once again forgotten what Feather is :P <- be thankful
20:57:53 <quintopia> why is it called feather anyway?
21:00:27 <Gregor> Same reason why Scape🐐 is called Scape🐐.
21:01:52 <pikhq_> Fuck you and your UTF-Goat.
21:03:07 <quintopia> and what reason is that?
21:03:20 <quintopia> also, is there a scp-alike that does file-resume?
21:03:31 <quintopia> or does scp do it?
21:05:07 <Phantom_Hoover> quintopia, ISTR that it was something to do with starting off as a lightweight Smalltalk.
21:05:36 <quintopia> Phantom_Hoover: nvm that. what about a resumable scp?
21:05:48 <Phantom_Hoover> No idea.
21:05:53 -!- copumpkin has joined.
21:08:24 -!- pumpkin has quit (Ping timeout: 240 seconds).
21:08:33 <pikhq_> Aaah, I finally figured out why aptitude ended up freaking out so much.
21:09:01 <pikhq_> OpenOffice to LibreOffice switch.
21:09:06 <pikhq_> dist-upgrade does it.
21:09:16 <quintopia> http://www.webupd8.org/2011/03/gtk-32-will-let-you-run-any-application.html how many people are already planning on seeing how many levels deep they can recurse firefox?
21:15:24 <cheater00> 0just spent 5 minutes searching for a piece of bread that fell off my plate on the floor. it was in my beard.
21:16:01 -!- ais523 has quit (Remote host closed the connection).
21:17:29 -!- Sgeo has quit (Ping timeout: 276 seconds).
21:25:42 -!- Lymia has joined.
21:26:19 -!- cheater- has joined.
21:28:56 -!- cheater00 has quit (Ping timeout: 255 seconds).
21:28:57 -!- cheater99 has joined.
21:31:10 -!- cheater- has quit (Ping timeout: 255 seconds).
21:38:31 * Phantom_Hoover wonders what the upper and lower bounds on votes on a single reddit comment are.
21:51:14 -!- pikhq_ has quit (Ping timeout: 240 seconds).
21:51:16 -!- sebbu has quit (Ping timeout: 276 seconds).
21:51:29 <olsner> http://www.reddit.com/r/AskReddit/comments/9ogpi/what_is_the_most_upvoted_comment_in_the_history/ might have the answer
21:51:30 -!- pikhq has joined.
21:52:05 <Phantom_Hoover> Gregor, http://twitpic.com/4aciog
21:54:11 <Phantom_Hoover> Wow, 4000+
21:57:02 <fizzie> quintopia: You can do rsync-over-ssh to resume a broken scp transfer; I don't think it does natively.
21:57:46 <quintopia> i've never used rsync
21:58:21 <fizzie> "scp blah remote:bleh" + continue with "rsync -P blah remote:bleh".
21:59:07 <Phantom_Hoover> Wait, there's another with 9500.
21:59:56 <olsner> the one that's a reply to one with -7500?
22:00:00 <Phantom_Hoover> Yes.
22:00:04 <fizzie> Should work reasonably well as long as there's just one file, though it might waste some time/bandwidth in checking that the files match on both sides of the fence.
22:00:27 <Phantom_Hoover> 14,000. What.
22:00:30 <fizzie> (Assuming a rsync that defaults to SSH.)
22:00:59 <olsner> Phantom_Hoover: ooh, where?
22:01:21 <Phantom_Hoover> Don't know.
22:01:31 <Phantom_Hoover> The link is to a user page, not the comment.
22:02:05 <olsner> aha, look_of_disapproval? the *sum* is 14000, motivated by every comment being the same comment
22:02:18 <olsner> or was, it's obviously changed since that post was made
22:02:27 <quintopia> fizzie: do i need to make a tunnel for it or will it just know?
22:02:41 <Phantom_Hoover> olsner, hmm, ah.
22:05:05 <fizzie> quintopia: If it's configured to use ssh, it's going to use ssh just fine. Alternatively pass "-e ssh" to it.
22:08:35 <fizzie> Possibly "--append" instead of "-P" would make it just append without checking the contents of what is already there.
22:12:53 <quintopia> neat
22:16:50 -!- sebbu has joined.
22:16:50 -!- sebbu has quit (Changing host).
22:16:50 -!- sebbu has joined.
22:17:19 <Phantom_Hoover> It annoys me to no end that the kind of people who think ESR is onto something have hijacked the glider?
22:17:23 <Phantom_Hoover> s/?/./
22:19:54 <Gregor> Phantom_Hoover: http://sss.cs.purdue.edu/projects/dynjs/javascript_the_evil_parts.png
22:20:08 <Phantom_Hoover> Seen it.
22:21:58 <Gregor> I'm makin' myself some corned beef and cabbage 8-D
22:22:38 -!- sebbu has quit (Ping timeout: 248 seconds).
22:25:03 <Phantom_Hoover> There isn't a decent way of expressing my reaction to that.
22:25:55 <Gregor> JEALOUSY?
22:27:28 <cheater99> someone just came to my door, handed me money, and told me i'm getting more later
22:27:45 <Gregor> Run for your freaking life.
22:28:09 -!- MigoMipo has quit (Read error: Connection reset by peer).
22:28:48 <Phantom_Hoover> Gregor, it's, like, terror mixed with incredulity mixed with curiosity all in a bag.
22:29:05 <Gregor> ... due to ... corned beef and cabbage?
22:29:48 -!- FireFly has quit (Read error: Connection reset by peer).
22:30:51 <cheater99> Gregor: well, the person does have clinically certified psychosis!
22:31:17 -!- FireFly has joined.
22:35:59 -!- sebbu has joined.
22:35:59 -!- sebbu has quit (Changing host).
22:35:59 -!- sebbu has joined.
22:38:14 <Phantom_Hoover> Gregor, yes/
22:38:37 <Phantom_Hoover> It's, like, combining my least favourite thing with something I have had only mildly negative experiences with.
22:39:26 <Gregor> Phantom_Hoover: You are a bad person.
22:40:24 <Phantom_Hoover> CABBAGE IS A BAD PERSON
22:42:07 <pikhq> Phantom_Hoover: What is your least favorite thing?
22:42:14 <Phantom_Hoover> CABBAGE
22:42:23 <fizzie> BABBAGE.
22:42:26 <pikhq> Ah, good, so not corned beef.
22:42:36 <pikhq> Which is, of course, DELICOUSNESS ITSELF
22:42:58 <Phantom_Hoover> fizzie, THE BASTARD
22:43:09 <Phantom_Hoover> SCUMBAG BABBAGE
22:43:13 <Phantom_Hoover> INVENTS COMPUTERS
22:55:56 * Phantom_Hoover → sleep
22:55:58 -!- Phantom_Hoover has quit (Quit: Leaving).
23:01:26 -!- Sgeo has joined.
23:20:30 -!- fungot has quit (Read error: Operation timed out).
23:34:50 -!- FireFly has quit (Quit: swatted to death).
23:36:04 -!- azaq23 has quit (Quit: Leaving.).
23:55:12 -!- pumpkin has joined.
23:55:47 -!- pikhq has quit (Ping timeout: 246 seconds).
23:55:52 -!- pikhq_ has joined.
23:56:31 -!- copumpkin has quit (Ping timeout: 240 seconds).
←2011-03-16 2011-03-17 2011-03-18→ ↑2011 ↑all