18:47:55 -!- esowiki has joined.
18:47:56 -!- glogbot has joined.
18:47:57 -!- HackEgo has joined.
18:47:58 -!- EgoBot has joined.
18:47:59 -!- esowiki has joined.
18:48:00 -!- esowiki has joined.
18:49:01 -!- Gregor has joined.
19:03:02 <zzo38> Different point of reference, different wavefunction, and all that stuff, is part of what my point is when I was saying those things.
19:16:04 -!- sebbu2 has joined.
19:17:32 -!- sebbu3 has joined.
19:18:30 -!- AnotherTest has left.
19:19:43 <Sgeo> I think this may be the dumbest thing I have ever seen http://motherboard.vice.com/blog/introducing-the-nsa-proof-font
19:19:45 -!- sebbu has quit (Ping timeout: 268 seconds).
19:20:03 <Sgeo> They think it's impossible to have OCR software trained to recognize a new font?
19:20:25 -!- sebbu2 has quit (Ping timeout: 246 seconds).
19:20:33 <Sgeo> Also: Their "false" font, which has a large letter and a small letter where the small letter is the real letter, guess what? real a maps to fake z
19:21:50 <Sgeo> Captchas don't use one specific font though. I think if there was a captcha that always turned a into one of 6 shapes, it would be defeated easily
19:21:58 <Sgeo> (There are six 'fonts' in this package)
19:23:00 <Sgeo> If it were software that created a new font for each document based on this concept, that might be useful, but as-is?
19:23:00 <FreeFull> Where would I be able to find a 100x100 nonogram?
19:24:02 <elliott> [[Sang has no illusions that even a clever cryptographic font—which you can use in email messages to shield them from snoops and font-recognition bots—will remain encoded for long. They're not meant to be long-term tools with which to combat the NSA. Rather, he views them as an awareness-raising measure.
19:24:07 <elliott> Read more: http://motherboard.vice.com/blog/introducing-the-nsa-proof-font#ixzz2WyW8HBod
19:24:14 <elliott> stupid copy-paste hijacker
19:25:18 <Sgeo> The video doesn't exactly make it clear
19:25:45 <Bike> encoding as images seems kind of inefficient if you're just using it for email
19:26:11 <Sgeo> I have no idea where I got that link from
19:29:46 <zzo38> Why does Verilog use a list of I/O ports rather than a single bit vector as the I/O port of a module?
19:30:48 <zzo38> I want to make up "HWPL" which does not have these and the other problem of Verilog and other hardware programming languages.
19:43:27 -!- sebbu3 has quit (Read error: Connection reset by peer).
19:43:46 -!- sebbu has joined.
19:47:54 <shachaf> Bike: why are you in that channel help
19:48:41 <Bike> elliott's scheming.
19:48:56 <elliott> Bike has been in #haskell before! I'm innocent!
19:49:06 <elliott> although I am scheming. constantly
19:53:22 <Sgeo> (Is where I got that link from)
19:54:48 <Sgeo> http://www.fark.com/comments/7809821/Introducing-NSA-Proof-Font-a-typeface-that-would-be-unreadable-by-text-scanning-software-whether-used-by-a-government-agency-a-lone-hacker-misdirecting-information-sometimes-not-giving-any-at-all
19:54:55 <Sgeo> I think most Farkers understand this
20:02:29 -!- oerjan has joined.
20:05:51 <oerjan> <-- is it just me or does that subreddit have css that places a red dot in a fixed position on the screen, so it looks like something's wrong with it...
20:06:02 <oerjan> *<Sgeo> http://www.reddit.com/r/AdviceAnimals/comments/1gu52n/back_in_my_day/ <--
20:07:40 <FreeFull> oerjan: I don't see the red dot
20:07:51 -!- Bike has quit (Ping timeout: 268 seconds).
20:09:36 -!- Bike has joined.
20:10:46 <kmc> helloerjan
20:10:59 <kmc> but no he's right there
20:11:21 <kmc> what a terrible person (the jokes is that i also left #haskell)
20:11:31 <kmc> lambdabot: indeed
20:17:13 <lambdabot> America!! I saw it all!! Vomiting! Waving! JERRY FALWELLING into
20:17:13 <lambdabot> your void tube of UHF oblivion!! SAFEWAY of the mind ...
20:17:20 <lambdabot> Succumb to natural tendencies. Be hateful and boring.
20:26:52 -!- ChanServ has set channel mode: +o oerjan.
20:27:08 -!- oerjan has set channel mode: +b *!*@gateway/web/cgi-irc/kiwiirc.com/ip.77.*.
20:27:37 <ion> *!*@* would be better.
20:27:49 -!- oerjan has set channel mode: -o oerjan.
20:28:00 <elliott> you could remove the 77. on the grounds that nobody should use a client that advertises itself as "hand-crafted" :P
20:28:17 <elliott> probably 77.* includes, like, a ton of stuff, but I doubt it'll come up
20:28:25 <oerjan> elliott: it's all mediaways
20:28:51 <elliott> huh. (how do you look that up? I don't really know where to get this information except the ARIN whois stuff)
20:29:10 <oerjan> host 77.0.0.0 on the command line...
20:29:41 <elliott> 255.255.255.77.in-addr.arpa domain name pointer 77-255-255-255.adsl.inetia.pl.
20:29:49 <elliott> you don't really want .0.0.0 there do you?
20:29:53 <elliott> if you want information for the whole of the range
20:30:13 <oerjan> ...i always thought 0 was the wildcard there
20:30:35 <ion> Err, what’s wrong with whois?
20:31:37 <Taneb> I would like a hug
20:31:50 <lambdabot> http://hackage.haskell.org/trac/ghc/newticket?type=bug
20:32:03 <zzo38> Some things I decided can be: Numbers are binary by default (and can have X and Z in them too), unless you put # or $ to indicate decimal or hexadecimal numbers. All user-defined words must have prefix, so it won't conflict built-ins and binary numbers. The operators + - * / = < > are static operators so it is only used at compile-time; you could check equal at runtime by &(.X~^.Y) instead though.
20:33:58 <oerjan> elliott: hm maybe there isn't actually a wildcard system as i've always thought
20:34:11 <elliott> ion: how can you use whois to get information like "$ISP owns this range in 77.*"?
20:34:57 <ion> whois 77.0.0.0, see where the range ends, add one, whois that, rinse, repeat.
20:35:14 <ion> Someone might have made a tool to automate that.
20:35:37 <oerjan> elliott: ok i just tried whois with the exact ip
20:36:37 -!- ChanServ has set channel mode: +o oerjan.
20:36:43 <kmc> zzo38: what are X and Z for
20:36:50 -!- oerjan has set channel mode: -b *!*@gateway/web/cgi-irc/kiwiirc.com/ip.77.*.
20:36:59 -!- oerjan has set channel mode: +b *!*@gateway/web/cgi-irc/kiwiirc.com/ip.77.180.*.
20:36:59 <zzo38> kmc: X is for an unknown or don't care value, and Z is for high impedance.
20:37:06 <kmc> did hagb4rd try to evade a ban or osmething?
20:37:06 -!- oerjan has set channel mode: -o oerjan.
20:37:10 <kmc> zzo38: ah, for a HDL, I see
20:37:13 <Bike> wow i thought GmbH was a company
20:37:13 <kmc> makes sense
20:37:14 <Bike> i am the worst
20:37:17 <oerjan> that's what whois said the range was, so.
20:37:21 <kmc> it's a type of company
20:37:35 <oerjan> the famous Inc company
20:37:36 <kmc> but yeah I thought that too at one point
20:38:08 <kmc> there's a magazine named Inc but the company is Mansueto Ventures
20:38:17 <Bike> that's terrible
20:38:21 <elliott> kmc: utoneq/untoneq was hagb4rd
20:38:25 <oerjan> mind you 77.181 is owned by the same company. oh well.
20:38:39 <zzo38> The things I have written about above aren't sensible unless it is a hardware programming language anyways.
20:39:00 <kmc> I should have guessed
20:39:07 <kmc> from patterns of talking
20:39:21 <elliott> yeah I realised it was a regular but I had to look up the IP to realise it was hagb4rd
20:39:25 <kmc> 05:09 < utoneq> what exactly is a monad.. and where is the difference to a set or a tuple?
20:39:40 <Taneb> who even is hagb4rd
20:39:48 <zzo38> Why else would arithmetic operators only at compile-time and numbers being in binary notation by default?
20:39:49 <ion> kmc: I’m looking forward to the monad tutorial he will inevitably write.
20:39:50 <elliott> oerjan: imo op kmc so I can bother him instead. this is a serious proposal.
20:40:12 <kmc> i love it when people demand an explanation for X by comparison to some Y or Z they erroneously think is related
20:40:31 <elliott> kmc: the difference is that a monad is a triple
20:41:00 <Taneb> I thought a monad had but one element!
20:41:01 <kmc> i am willing to serve as an #esoteric op if my service in such capacity is desired
20:41:11 <elliott> damn that's some campaign speech
20:41:42 <Taneb> zzo38, I wasn't being serious :P
20:41:51 <Taneb> But, perhaps, the identity element
20:42:35 <Taneb> And now I am sleepy
20:42:42 <Taneb> But it is not even 10 o'clock
20:43:06 <Bike> "really, a monad is just a category with seven elements where each morphism has two inverses"
20:44:54 <kmc> each inverser than the other
20:46:01 -!- sprocklem has joined.
20:48:48 <FreeFull> Bike: Is that unique to monads though?
20:48:55 <nooodl> `addquote <kmc> 05:09 < utoneq> what exactly is a monad.. and where is the difference to a set or a tuple? <elliott> kmc: the difference is that a monad is a triple <elliott> one more element
20:49:05 <HackEgo> 1059) <kmc> 05:09 < utoneq> what exactly is a monad.. and where is the difference to a set or a tuple? <elliott> kmc: the difference is that a monad is a triple <elliott> one more element
20:49:28 <FreeFull> Well, a set is nothing like a tuple
20:49:45 <nooodl> they're just in a file right i could sed it
20:49:59 <kmc> a tuple is a set, if that's how you define tuples
20:50:22 <Bike> a set is a tuple if that's how you define sets! check and mate
20:50:23 <FreeFull> (3,3) is a perfectly valid tuple though
20:50:36 <kmc> check before you mate
20:50:41 <nooodl> {3, {3, 3}} is a valid set
20:50:41 <Bike> it's pretty easy to define ordered pairs with sets, dude
20:50:58 <FreeFull> nooodl: Isn't that {3, {3}} though
20:51:02 <oerjan> `run sed -i '1059s/ <[^ ]/ \&/' quotes
20:51:14 <HackEgo> 1059) <kmc> 05:09 < utoneq> what exactly is a monad.. and where is the difference to a set or a tuple? &lliott> kmc: the difference is that a monad is a triple <elliott> one more element
20:51:14 <nooodl> nah you can't define tuples like that, because
20:51:23 <elliott> btw defining types like tuples in terms of sets is terrible for mathematics. :(
20:51:36 <oerjan> `run sed -i '1059s/ <[^ ]/ &/g' quotes
20:51:37 <nooodl> if you define (a, b) = {a, {b}}
20:51:40 <nooodl> ({0}, 0) would be {{0}, {0}}
20:51:45 <Bike> elliott: here, let me define types as products of prime powers
20:51:46 <HackEgo> 1059) <kmc> 05:09 < utoneq> what exactly is a monad.. and where is the difference to a set or a tuple? <elliott> kmc: the difference is that a monad is a triple <elliott> one more element
20:51:56 <FreeFull> A tuple is some elements, where they have an inherent order and can repeat
20:52:22 <Bike> defined in terms of primes.
20:52:57 <Bike> i've rewritten the HoTT source as a diophantine. no need to thank me.
20:53:40 <katla> "no need to thank me" that's for sure
20:53:50 <nooodl> my dream: asking a SO question that becomes a #1 google search result, gaining fifty thousand rep for nothing
20:54:14 <elliott> it's ok, I make a point to thank Bike as little as possible
20:54:48 <elliott> nooodl: https://www.google.co.uk/search?client=safari&rls=en&q=call/cc+implementation&ie=UTF-8&oe=UTF-8&redir_esc=&ei=EQ_GUbmTIZDwhQetsYCoDg bam
20:54:52 <elliott> maybe it's not #1 for everyone
20:55:41 <elliott> my voting rings expands further
20:55:53 -!- nooodl has quit (Quit: Leaving).
20:58:56 <kmc> `run sed -i '1059s/ utoneq/utoneq/' quotes
20:59:01 <kmc> `quote 1059
20:59:03 <HackEgo> 1059) <kmc> 05:09 <utoneq> what exactly is a monad.. and where is the difference to a set or a tuple? <elliott> kmc: the difference is that a monad is a triple <elliott> one more element
20:59:05 <elliott> that's falsifying the quote
20:59:07 <kmc> elliott.....
20:59:12 <elliott> what you should do is fix your client to not include that awful space
20:59:14 <elliott> and then we can repeat the exchange
20:59:21 <kmc> `run sed -i '1059s/utoneq/ utoneq/' quotes
20:59:25 <Bike> "Well, okay, it's enlightening if you already know what it means. " good writing here
20:59:26 <kmc> `quote 1059
20:59:27 <HackEgo> 1059) <kmc> 05:09 < utoneq> what exactly is a monad.. and where is the difference to a set or a tuple? <elliott> kmc: the difference is that a monad is a triple <elliott> one more element
20:59:32 <kmc> Bike: monad tutorial?
21:00:11 <Bike> call/cc tutorial
21:00:29 <elliott> kmc: my rambling SO answer
21:00:55 <ion> bike: http://www.vex.net/~trebla/haskell/cont-monad.xhtml
21:01:34 <ion> elliott: URL?
21:01:48 <Bike> http://stackoverflow.com/questions/9050725/call-cc-implementation
21:02:12 -!- Taneb has quit (Quit: Leaving).
21:04:41 -!- MindlessDrone has quit (Quit: MindlessDrone).
21:10:37 -!- nooodl^ has joined.
21:10:50 -!- ChanServ has set channel mode: +o oerjan.
21:11:02 -!- epicmonkey has quit (Ping timeout: 246 seconds).
21:11:19 -!- oerjan has set channel mode: -b *!*@gateway/web/cgi-irc/kiwiirc.com/ip.77.180.*.
21:11:36 -!- oerjan has set channel mode: +b *!*@gateway/web/cgi-irc/kiwiirc.com/ip.*.
21:11:44 -!- oerjan has set channel mode: -o oerjan.
21:13:21 <nooodl^> what's kiwiirc and why's it banned
21:13:42 <Bike> it's hand-crafted
21:16:00 <fizzie> It's also "trusted" by freenode, whatever that means.
21:22:02 <fizzie> Re the earlier discussion on networks, you can ask RIPE whois queries like "77/8 except not the exact match but all first-level more specific inetnum/route objects" to see what all is in a thing. (Though it's a long list for that particular query.)
21:22:59 <fizzie> (Including, among others, some Norwegians, which would be a tragic loss indeed.)
21:23:29 <oerjan> anyway i just extended it to all ips since i found a completely different range in the logs
21:23:48 <oerjan> and afaiu no one else in the channel has ever used that client.
21:24:38 <oerjan> (approximately, anyway)
21:24:40 <nooodl^> who's ban evading on kiwiirc, spammers?
21:24:42 <fizzie> "Thor-Henrik Kvandahl" (the admin contact for Telenor Norge's 77.16.0.0/14 block) sounds incredibly stereotypically Norwegian.
21:24:57 <fizzie> At least the Thor-Henrik part.
21:26:18 <oerjan> fizzie: the kvandahl is a stereotypically norwegian toponym.
21:26:40 <nooodl^> now there's an objective reason to ban him
21:28:37 <elliott> that's what's great about ban evasion
21:28:45 <elliott> like, if I ban someone in #haskell, I hope they evade the ban
21:28:57 <elliott> because it means I can ban them forever and not have to worry about them appealing
21:29:25 <fizzie> Philosophy question: why are Thunderbird's once-a-day builds codenamed "Daily", while Firefox's are called "Nigthly"?
21:29:46 <myname> http://adit.io/posts/2013-04-17-functors,_applicatives,_and_monads_in_pictures.html this is probably the best explanation i've read so far
21:29:48 <fizzie> It would be even more puzzling if they were called Nigthly.
21:33:48 <nooodl^> I should write a monad tutorial!!
21:34:17 <nooodl^> I know exactly what analogy ill use too!!
21:34:56 <kmc> nooodl^ analogy
21:35:25 <elliott> actually it would work better for comonads
21:35:31 <elliott> duplicate goes from noodle to nooodle to noooodle
21:35:33 <nooodl^> its gonna compare IO to sheet music
21:35:48 <shachaf> myname: That explanation looks like it's full of misleading things. :-(
21:36:07 <myname> shachaf: at what point?
21:36:45 <shachaf> For example all 24 points that it talks about "wrapped values".
21:36:58 <kmc> what's my clever quote about IO String?
21:37:07 <kmc> @quote IO.String
21:37:07 <lambdabot> shachaf says: getLine :: IO String contains a String in the same way that /bin/ls contains a list of files
21:37:13 <kmc> oh that's your clever quote
21:37:20 <kmc> I think I had one too, equally clever though
21:37:23 <lambdabot> kmc says: it is not hard to troll #haskell for real; you just have to get confused and confrontational about how to convert IO String to String
21:37:33 <myname> shachaf: what's wrong with it?
21:37:47 <kmc> an IO String is not a String that's been wrapped or "tainted" somehow
21:38:03 <kmc> it's a recipe for how to produce a String by doing some IO
21:38:16 <kmc> the String doesn't exist yet; you might never execute the recipe, or you might execute it more than once and get different Strings
21:38:24 <kmc> none of which contradicts the idea that the recipe itself is an inert, pure value
21:38:27 <elliott> imo shachaf's quote is cleverer than kmc
21:38:34 <shachaf> @quote monochrom IO.String
21:38:34 <lambdabot> monochrom says: How do I extract the IO out of IO String?
21:38:39 <elliott> not even cleverer than kmc's quote
21:38:51 <kmc> myname: but for example a Maybe String *is* a wrapped String (or Nothing)
21:39:05 <kmc> from which we conclude that this idea of "wrapping" is not fundamental to monads, but a property of the implementation of *some* particular monads
21:39:38 <kmc> there's very little you can say about all monads in general, because it's such a general interface
21:39:50 <kmc> people have trouble "understanding monads" partly because they expect there to be more to it than there is
21:40:08 <kmc> because of all the stupid bullshit hype by detractors and overexcited beginners alike
21:40:16 <katla> is there actually anyone who doens't understand this stuff
21:40:37 <kmc> nah i just go off on autopilot
21:40:42 <Bike> how is a monad like a writing desk
21:40:43 <nooodl^> isn't there like a table of common monad instances and their implementation of return, bind, join
21:40:51 <kmc> but maybe I've at least convinced myname why "wrapping" is not a good analogy
21:41:11 <kmc> a monad is like a butt
21:41:32 <myname> kmc: i agree in the part that "wrapping" as in "putting something around something other" is not a good analogy
21:42:33 <shachaf> So which part is good about that explanation you liked?
21:42:39 <katla> do you know about data IO a = Return a | Bind b (b -> IO a)
21:42:54 -!- nooodl^ has quit (Read error: Connection reset by peer).
21:42:57 <myname> the visualisation of the types is well made imo
21:43:03 <kmc> | PutChar Char (IO ()) | GetChar (Char -> IO ()) | ...
21:43:06 -!- oerjan has quit (Remote host closed the connection).
21:43:15 -!- nooodl^ has joined.
21:43:33 <kmc> (but I think that Bind isn't quite right, also it's existentially quantified or something?)
21:43:49 <katla> you would write it as a GADT in a real .hs
21:44:04 -!- nooodl^ has quit (Read error: Connection reset by peer).
21:44:10 -!- nooodl^ has joined.
21:44:46 <elliott> it would be | forall b. Bind (IO b) (b -> IO a)
21:44:56 <katla> given that IO String is a string wrapped up in the sense that you need to perform an interactive computation to get it out
21:45:56 <shachaf> "wrapped" meaning "monad"??
21:47:05 -!- nooodl^ has quit (Read error: Connection reset by peer).
21:56:45 <kmc> katla: no, there is no String in there
21:56:56 <kmc> it's a description of some IO which you could perform in the future to get a String
21:57:06 <elliott> well, katla has a point; there often is.
21:57:31 <elliott> if there exists a sequence of responses to its requests such that it terminates in a Return then there is a String "inside" the IO String in the same way that Bool -> String is a container of two Strings
21:57:39 <kmc> sometimes, yeah, but that doesn't mean that "wrapping" is a useful way to think about IO String
21:57:48 <elliott> but I think it mostly shows that "inside" is too vague to really be useful...
21:57:51 <elliott> but I hate talking about monad tutorials
21:58:01 <kmc> we can make this analogy work only if we torture the definition of "wrap" to mean something completely alien to its English meaning
21:58:07 <kmc> which is how a lot of monad tutorials work out
21:58:23 <kmc> monads are just containers! as long as you forget everything you know about what the word "container" means
21:58:40 <elliott> @quote kmc just.containers
21:58:40 <lambdabot> No quotes match. Maybe you made a typo?
21:58:42 <lambdabot> kmc says: [After discussing monads, containers, and tortillas] therefore the key difference between a container and a monad is delicious carne asada
21:58:44 <lambdabot> kmc says: monads are like containers, as long as you forget everything you know about the meaning of the word "container" and take it to be a totally abstract word synonymous with "monad"
21:58:47 <shachaf> monads are like slip covers
21:58:51 <elliott> kmc: i accuse you of plagiarising yourself
21:59:20 <kmc> the "analogy" just exists to trick people into learning a new abstract concept without being scared off by omg math words
22:01:36 <kmc> "A monad has two functions, return and (>>=) --" "OMG math!" "Fine, a monad is like a burrito. A burrito has two functions, return and (>>=) ..."
22:02:51 <shachaf> what does kmc say to vegetarian burritos
22:03:17 <kmc> https://bugzilla.wikimedia.org/show_bug.cgi?id=50029
22:03:48 <zzo38> You can use Free (CoYoneda f) instead of putting all of the PutChar GetChar in that IO data, which has some problem anyways it isn't a monad (unless it is a private data type), but Free (a functor) is monad, always.
22:04:19 <shachaf> hey kmc do you like CoYoneda
22:05:06 <FreeFull> I am listening to something arond 357 kHz and I have no idea what it is
22:05:52 <kmc> frames of hitler interleaved with plans to build a wormhole endpoint
22:06:33 <zzo38> Something I thought of is "oracle sequent calculus", add a "oracle operator", for example if it is called # then you add an axiom schema |- #x if and only if |- x is not provable.
22:07:24 <Bike> so the lesson i'm getting here is: why the hell are CS people scared of math
22:07:57 <Bike> "wrapping" as in "putting something around something other" <-- also, lol
22:08:00 <myname> math is great, but i like my math with less than 3 abstraction layers
22:08:19 <FreeFull> Oh, seems to be aeroport related?
22:08:44 <zzo38> myname: Why do you think 3 abstraction layers is too much?
22:08:44 <kmc> FreeFull: what kind of radio do you have
22:08:53 <kmc> FreeFull: I got one of the cheap DVB dongle software-defined radios
22:08:53 <Bike> monads as in haskell seem like a pretty straightforward abstraction
22:09:01 <FreeFull> Monads are like monoids in endofunctors
22:09:08 <kmc> it's pretty neat. doesn't go down to 357 kHz though
22:09:22 <myname> zzo38: not "too much", but way to abstract if not told carefully
22:09:31 <FreeFull> http://websdr.ewi.utwente.nl:8901/
22:11:49 <kmc> oh it samples the /whole/ spectrum?? wow
22:12:36 <kmc> whole shortwave or something
22:12:56 <elliott> I was hoping you meant, like, the /whole/ spectrum.
22:13:14 <FreeFull> It'd be cool to listen to gamma rays
22:13:53 <Bike> wouldn't they be pretty repetitive
22:13:55 <Bike> alt. white noise
22:16:11 <kmc> wonder if any neutrino detectors are online in this manner
22:16:20 <elliott> shachaf is a neutrino detector
22:16:58 <shachaf> elliott has a neutrino allergy
22:17:18 <kmc> that would be bad
22:17:57 <Bike> i thought neutrino detectors didn't detect very much.
22:18:30 <fizzie> Bike: I'm sure you could make some sort of a social media game out of one.
22:18:41 <fizzie> It goes "ping" in your facebooks and whatnot.
22:18:46 <shachaf> elliott: for every neutrino we detect there are trillions which go undetected
22:22:56 <FreeFull> They can only interact through the weak interaction
22:28:05 <shachaf> help there's such a thing as a monadic functor
22:37:42 <kmc> mathematicians are the worst
22:37:56 <elliott> mathematicians come in second
22:39:52 <elliott> katla: hey did you hear about the homotopy type theory book? I don't know if you're interested
22:41:08 -!- FreeFull has quit (Read error: Operation timed out).
22:41:32 -!- sacje has joined.
22:45:19 -!- FreeFull has joined.
23:26:55 <elliott> lots of people are bored in #esoteric today
23:27:02 <kmc> is that so
23:27:20 <kmc> who is the third
23:28:23 <shachaf> so much time, so little to do
23:29:28 <katla> you ned to think of something new
23:30:08 <elliott> shachaf: because life sucks, hope this helps
23:30:31 <shachaf> elliott: instead of being bored you could read some more of that hott book you're such a big fan of
23:31:25 <elliott> you forgot to consider the part where I'm lazy
23:33:24 <shachaf> have you considered not being lazy
23:54:22 -!- Bike has quit (Ping timeout: 276 seconds).
23:56:54 -!- Bike has joined.