00:07:37 -!- cheater- has joined.
00:10:40 -!- cheater00 has quit (Ping timeout: 255 seconds).
00:55:29 -!- pikhq has joined.
00:55:38 -!- pikhq_ has quit (Ping timeout: 250 seconds).
00:57:50 -!- ais523 has quit (Remote host closed the connection).
01:07:00 -!- clog has quit (Remote host closed the connection).
01:17:38 <Gregor> Welp, the libm.so auction has started.
01:23:03 -!- esowiki has joined.
01:23:17 -!- glogbot has joined.
01:26:17 -!- Sgeo has quit (Ping timeout: 255 seconds).
01:30:04 -!- Mathnerd314 has quit (Ping timeout: 264 seconds).
01:53:55 -!- lament has quit (Quit: Lost terminal).
02:08:45 <Gregor> Given everybody else's ping timeouts, maybe my bots aren't actually doing anything wrong :P
02:22:36 -!- zzo38 has joined.
02:24:27 <zzo38> Still, I prefer SIRCL rather than the raw format that glogbot uses, but at least glogbot is not lying about the raw logs in the way that clog is doing.
02:25:13 <Sgeo_> How does clog lie?
02:26:49 <Gregor> zzo38: Do you have a description or example of SIRCL's format?
02:26:53 <zzo38> It says 'These logs are purposely "raw" and are intended to be parsed/reformated/wrapped before viewing.' but it isn't raw.
02:28:09 <Gregor> glogbot's raw format is not really meant to be readable, just bake-able. Its baked format could probably use some TLC.
02:28:32 <zzo38> Gregor: Very simple. UNIX timestamp, TAB, and then the message, terminated by CRLF (always CRLF, LF only is not allowed). Metadata commands have no TAB and have an asterisk as the first character of the line (before the timestamp, if any). Here is an actual log created in that format: http://zzo38computer.cjb.net/irc_log/ADMIN/1291325292
02:29:47 <Gregor> The fact that that lines up the messages does make it more readable I'll admit.
02:31:02 <Gregor> I chose my format to 1) distinguish outgoing messages since it does in fact send messages out on requests, 2) have the totally-unnecessary added precision of gettimeofday and 3) use space delimitation for the simple reason that IRC is already space-delimited (modulo the :-to-EOL rule)
02:33:08 <zzo38> SIRCL format is not intended to be more or less readable than any others, it is just a way that makes sense to me. It uses TAB only after the timestamp; IRC commands are delimited with spaces and colons as normal. The TAB separates the timestamp from the message. It requires CRLF the same as the IRC prtoocol specification says it is. You are not required to use this format; but I do have it.
02:35:33 <Gregor> My log reformatter is extremely-lazily-written, and tokenizes my entire raw-format line as if it was an IRC message, so having spaces saves me one utterly-minor step X-P
02:35:37 <zzo38> (Also note these logs are created directly by the server, so in this case there is no need to distinguish send/receive, it simply logs everything that any client on that channel would receive. One way to use it in client logging if wanted, is to omit the sender part for sent messages. This format also supports metadata, although any program that creates or parses this format is not required to write or use it.)
02:35:51 <zzo38> Gregor: O, that makes sense, if that is how you wrote it.
02:44:10 <zzo38> I would, however, have the filenames a bit different: The raws have no extension (also no "-raw"), while the formatted ones formatted as HTML with the ".html" extension. So, for example, you would have "2011-03-20" and "2011-03-20.html" files. At least that is my opinion; you do not have to agree. But at least I like that you actually have raw logs, unlike clog.
02:45:19 <Gregor> I have considered making HTML ones ... the fact that the raw ones have a "-raw.txt" extension is because in my delusional imagination I postulate that the primary audience cares more about the baked logs, so marking the raw logs as "Here there be dragons" doesn't hurt *shrugs*
02:49:07 <Gregor> ANNOUNCEMENT: I am a superstitious ninny who unplugs my laptop from mains when there's a thunderstorm. Boo me.
02:50:12 -!- pikhq_ has joined.
02:50:25 -!- pikhq has quit (Ping timeout: 252 seconds).
02:51:07 <Gregor> I should add a PM command to glogbot that makes it so it notices you the last few lines when you join a channel. Opt-in of course.
02:52:44 <zzo38> I have invented SIRCL as a possible standard format for IRC logs.
02:53:13 <zzo38> Gregor: Yes, perhaps PRIVMSG glogbot :TAIL 12 for 12 lines, and put 0 to disable it.
02:53:38 * Gregor strongly considers this notion ...
02:54:29 <zzo38> There really also ought to be STATUS and HELP commands, even if you have no TAIL command like this.
02:55:11 <Gregor> I'm not sure what status would do, and there is !glogbot_help, although I could just as well make it respond to more in PM *shrugs*
02:56:40 <Sgeo_> I'm "helping" someone with chemistry by turning chemistry problems into.. stuff about marbles
03:00:52 <zzo38> In PM perhaps it would respond even if "!glogbot_" is not prefixed? And STATUS would display the URL as well as number of channels, possibly data rate and so on? Perhaps some option that the inviter can select whether or not commands sent to the channel are recognized? Some might want the log but not want it to recognize commands sent to the channel, which is why the channel operator should be able to turn it off somehow?
03:02:48 <zzo38> Perhaps if the channel operator sets channel mode +q glogbot!*@* then it will ignore commands sent to the channel completely... it could be implemented by having glogbot attempt to send a NOTICE with the log URL to the channel immediately after being invited....?
03:04:33 <zzo38> You receive a 404 if you send when you have +q that way it can check.
03:06:18 <Gregor> zzo38: What purpose is there in ignoring messages sent to the channel? It always responds with a personal NOTICE now, so it is not the one disturbing the peace.
03:07:02 <zzo38> Gregor: In case you want a channel without bot commands, or if there is confliction. So, two reasons.
03:09:38 <zzo38> At least I think that making it send the NOTICE with the URL and checking for 404, solves two problems at once. But that is just my opinion.
03:11:56 -!- esowiki has joined.
03:12:07 -!- glogbot has joined.
03:12:12 <Gregor> It just doesn't check for errors.
03:12:46 <Gregor> There, now it accepts truncated commands via PM :)
03:13:02 <Gregor> ... and hopefully not directly, as I shall test thusly:
03:13:14 <Gregor> s/directly/in a channel/
03:16:56 <Gregor> And now it has a passive-aggressive !glogbot_status command :)
03:17:41 <zzo38> OK, I can see that works.
03:34:22 <zzo38> Now, I wonder, should they implement a NS SET MODE command?
03:35:34 <zzo38> I can ask them maybe if they know
04:14:30 -!- wareya_ has joined.
04:15:27 <pikhq_> Aaaand the US cellphone market gets smaller still.
04:16:16 -!- wareya has quit (Ping timeout: 250 seconds).
04:16:18 <pikhq_> This will leave us with precisely *one* GSM carrier in the US, as well.
04:16:36 <zzo38> I do not think one is enough?
04:16:56 <pikhq_> And three majors carriers overall.
04:17:18 <pikhq_> Ma Bell is coming back.
04:17:51 <zzo38> Well, it doesn't affect me, I do not use cellular telephones anyways.
04:18:20 <pikhq_> AT&T, Verizon, and Sprint.
04:19:30 <pikhq_> Sorry, I should specify; they're buying T-Mobile US; no overseas operations.
04:21:17 <zzo38> I live in Canada, but regardless of where I live I still do not use cellular phone.
04:28:23 <zzo38> Maybe later I will write more TeX programs, such as macros to make calendar, mailing labels, barcodes, and more. All designed to be used with Plain TeX, DVI output, and no DVI specials.
04:35:00 <zzo38> Many other packages are LaTeX only, PDF only, Type 1 fonts only, e-TeX only, or require external "makeindex" program, DVI specials, SVG, PostScript, or something else. So now I make the one that does not require anything special.
05:10:58 <calamari> oh crap :( guess that means my cell phone bill is going up lol
05:22:55 -!- pikhq has joined.
05:25:40 -!- pikhq_ has quit (Ping timeout: 276 seconds).
05:35:09 -!- asiekierka has joined.
06:10:03 -!- asiekierka has quit (Read error: Operation timed out).
06:19:42 -!- sftp has quit (Remote host closed the connection).
06:34:05 -!- pikhq_ has joined.
06:35:04 -!- pikhq has quit (Ping timeout: 240 seconds).
07:21:01 -!- zzo38 has quit (Remote host closed the connection).
07:32:38 -!- calamari has quit (Quit: Leaving).
08:08:11 -!- augur has quit (Remote host closed the connection).
08:17:29 -!- cheater99 has joined.
08:18:21 -!- cheater- has quit (Ping timeout: 248 seconds).
08:21:21 -!- Wamanuz3 has joined.
08:24:24 -!- Wamanuz2 has quit (Ping timeout: 240 seconds).
08:32:09 -!- augur has joined.
08:35:28 -!- azaq23 has joined.
08:40:26 -!- ais523 has joined.
09:08:34 -!- jcp has quit (Ping timeout: 246 seconds).
09:13:27 -!- jcp has joined.
09:19:45 -!- pikhq_ has quit (Read error: Operation timed out).
09:21:08 -!- pikhq has joined.
10:39:59 -!- oerjan has joined.
10:41:36 * oerjan laughs at today's xkcd hovertext
10:44:02 <oerjan> SORRY I CANNOT HEAR YOU
11:14:01 -!- cheater00 has joined.
11:15:20 -!- cheater99 has quit (Ping timeout: 255 seconds).
11:27:26 -!- iconmaster_ has joined.
11:30:25 -!- iconmaster_ has quit (Client Quit).
11:44:08 -!- Slereah has quit (Ping timeout: 255 seconds).
11:50:03 -!- Slereah has joined.
11:56:59 -!- ais523 has quit (Read error: Operation timed out).
11:59:12 -!- ais523 has joined.
12:04:13 <ais523> oerjan: today's xkcd is quite good generally
12:05:32 -!- variable has quit (Read error: Operation timed out).
12:06:07 -!- sftp has joined.
12:06:44 -!- variable has joined.
12:14:17 -!- ais523 has quit (Read error: Connection reset by peer).
12:18:20 -!- ais523 has joined.
12:23:05 -!- lifthrasiir has quit (Ping timeout: 240 seconds).
12:24:47 -!- lifthrasiir has joined.
12:25:47 <Vorpal> ais523, hm talking of today's xkcd. Is there any special significance to the year 2017? As opposed to any other ones in the near future
12:26:11 <ais523> but 6 years from now is about the right timespan for the joke to work
12:26:17 <ais523> maybe 5 or 7 would have worked too
12:26:43 <Vorpal> just thought it might be some movie reference or such
12:27:22 -!- Slereah has quit.
12:28:18 <fizzie> There's a solar eclipse, that's all I can think of. The year-to-reference would have probably been 2012, but that's both overdone and a bit too soon.
12:28:47 <Vorpal> fizzie, where is that eclipse?
12:29:10 <fizzie> Them Americans have it, I think.
12:30:01 <fizzie> http://en.wikipedia.org/wiki/Solar_eclipse_of_August_21,_2017
12:30:07 <fizzie> Goes right through North America there.
12:31:00 <Vorpal> shouldn't that page have one of those "this article is about an upcoming event" kind of boilerplate
12:31:10 <Vorpal> I seem to remember wikipedia has that variant as well
12:32:47 <ais523> Vorpal: that's basically put there in order to prevent idiots writing things into the article that makes no sense
12:34:04 <Vorpal> ais523, oh? Such as wrong tense?
12:34:42 <fizzie> http://en.wikipedia.org/wiki/Template:Future seems to have been deleted.
12:34:48 <ais523> no, it's mostly to stop people writing vandalism about future events as if they've already happened/already known, then other people believing it as it's Wikipedia
12:34:51 <ais523> I'm not sure how often that happens
12:35:47 <ais523> deleted three times, but first time seems to be via WP:CENT: http://en.wikipedia.org/wiki/Wikipedia:Centralized_discussion/Deprecating_%22Future%22_templates
12:36:11 <ais523> (not that WP:CENT is actually a deletion process, but I don't think anyone cares; it's heavier-weight than most deletion processes, so probably falls under IAR)
12:37:31 <Vorpal> "then other people believing it as it's Wikipedia" <-- it is strange that happens really
12:37:49 <Vorpal> ais523, oha and what is IAR?
12:38:02 <ais523> there are huge debates on how to actually interpret it
12:38:08 -!- Sgeo_ has quit (Ping timeout: 255 seconds).
12:38:29 <Vorpal> ais523, presumably you could just ignore IAR and be done with it
12:39:08 <ais523> although my interpretation is more or less that a) if nobody disagrees with an action, you can do it regardless of what policy says (WP:BOLD's along those lines, too); b) the spirit of the rules are more important than the letter; c) documented rules lag behind accepted practice (i.e. it becomes accepted /then/ it's documented), so it's OK to break an old rule if Wikipedia has moved on since
12:39:10 <Vorpal> but more seriously, isn't it to avoid getting stuck on rules when they are clearly suggesting a bad/stupid way forward?
12:40:01 <ais523> that's a special case, but an important one
12:40:28 <ais523> in the case here, "you can delete a page via WP:CENT even though it isn't a deletion process" seems a perfectly valid use of IAR
12:40:49 <ais523> given that it sort-of trumps other processes, in that it would be a reasonable way to change deletion process
12:41:53 <ais523> anyway, tl;dr of the CENT discussion about {{future}}: it was meant to be a warning that the page was being heavily edited as details came out, but wasn't actually being used that way
12:42:45 <fizzie> ais523: There was a {{future}} in the article "29th century".
12:42:50 <ais523> and the way it was being used was entirely useless
12:42:56 <ais523> fizzie: yes, that's a good example
12:43:11 <ais523> unlikely to have more details coming out rapidly, causing a flurry of edits, on that example!
12:45:30 <Vorpal> the mere existence of that article is pretty ridiculous
12:46:01 <ais523> well, isn't the 29th century more notable than (insert favourite example of niche Wikipedia article here)?
12:51:08 <fizzie> "The CPR [Canadian Pacific Railway] obtained a 999-year lease on the O&Q [Ontario and Quebeck Railway] on January 4, 1884. -- The CPR also leased the New Brunswick Railway in 1890 for 990 years, --" Why exactly a bit less than thousand years (as opposed to some other ludicrous number), I wonder.
12:51:38 <ais523> because 1000 would have looked too large, it's the same principle as charging £9.99 in shops
12:51:48 <fizzie> "A 999-year lease is, under historic common law, essentially a nominal lease of property for life. The lease locations are mainly in Britain, her former colonies and Commonwealth. The longest possible term of a lease of real property is legally a 99-year lease."
12:51:59 <fizzie> Figures that there is an article for "999-year lease" specifically.
12:53:31 <fizzie> (Not a *good* article, though.)
13:08:27 -!- azaq23 has quit (Quit: Leaving.).
13:18:13 -!- oerjan has quit (Quit: leaving).
13:23:01 -!- esowiki has joined.
13:23:14 -!- glogbot has joined.
13:27:38 <Gregor> Why can't these guys stay pinged in.
13:33:24 -!- hagb4rd has joined.
13:34:40 <Vorpal> Gregor, what is glogbot?
13:39:09 <Gregor> In no way has it "replaced" optbot, since it doesn't serve the same function >_>
13:42:13 <Vorpal> Gregor, there is a typo in the topic. Might be intentional
13:42:50 <Vorpal> oh wait, PH set the topic
13:42:52 <Gregor> For some reason PH decided that Roman...ism and socialism are the same thing.
13:43:08 <Vorpal> "GLORIOVS" yeah, how strange
13:46:04 <Gregor> Anyway, glogbot is exactly what it sounds like.
13:46:20 <Vorpal> Gregor, a gnu or gnome logging bot :P
13:46:24 <ais523> that's not a typo in ancient Latin, except it probably isn't a real word in ancient Latin
13:47:34 <Vorpal> Gregor, Gregor Compiler Collection
13:47:48 <Vorpal> out of your various languages
13:48:30 <ais523> it'd be egologbot if it were Gregor's
13:48:44 <Vorpal> oh true, so it must be gnu then
13:48:48 <ais523> (note: assertion may contradict reality)
13:49:41 <Gregor> "ego" is for esorelated things.
13:50:05 <ais523> and a #esoteric logbot isn't esorelated?
13:50:43 <Gregor> It's a general-purpose logbot that just so happens to be being used on #esoteric .
13:52:06 <ais523> meanwhile: http://blog.cleveland.com/metro/2011/03/identifying_photocopy_machine.html
13:52:10 <ais523> non-eso, but pretty funny
13:52:15 <ais523> especially if you like silly legal situations
13:52:40 <ais523> it's basically a case where someone tried to answer a question about if there was a photocopier in an office, by repeatedly asking the questioning lawyer to define a photocopier
13:57:05 <Vorpal> <ais523> meanwhile: http://blog.cleveland.com/metro/2011/03/identifying_photocopy_machine.html <-- wtf
13:57:08 <Vorpal> ais523, is that a joke?
13:57:55 <ais523> it's not even particularly silly by the standards of typical court cases
13:58:02 <ais523> but it makes a change from SCO
13:59:07 <Vorpal> ais523, SCO is still going at it?
13:59:15 -!- pikhq has quit (Read error: Operation timed out).
13:59:21 <ais523> you'd expect them to have stopped?
13:59:40 <ais523> they recently managed to get a bankruptcy judge to approve them selling all their assets but only a very small number of liabities
13:59:47 <ais523> also, they're bankrupt on paper, again
14:00:12 <ais523> in terms of technically owning a negative amount of cash, not that that seems impossible in bankruptcy court as you can delay payments
14:00:24 <ais523> (they're legally bankrupt already, in that they declared bankruptcy years ago now)
14:00:36 <Gregor> ais523: lololol @ "Xerox"
14:00:37 <Vorpal> ais523, aren't there time limits on that sort of stuff
14:00:50 <ais523> Vorpal: yes, and they've been extended/broken repeatedly
14:01:01 -!- pikhq has joined.
14:01:09 <Vorpal> ais523, aren't there limits on how many times an extension can be granted?
14:01:22 <ais523> Gregor: I can't figure out if the person in question was a) an idiot, b) trolling, c) serious
14:01:34 <ais523> although b) would be epic
14:01:49 <ais523> Vorpal: the CEO of the company was removed and replaced with a bankrupcy-court-appointed trustee
14:01:59 <ais523> who kept on doing exactly the same thing the removed CEO had been doing
14:02:06 <Vorpal> ais523, that sounds very strange
14:02:22 <ais523> Vorpal: this is /SCO/ we're talking about
14:02:38 <Gregor> So, who would find a glogbot "tail" mode useful? (That is, when you join it sends you in a NOTICE the last few messages on the channel)
14:02:55 <Gregor> I'm not sure whether it's worth implementing, since it has some risk of flooding itself into oblivion :P
14:03:00 <Vorpal> ais523, don't they realize it must be futile to continue?
14:03:03 <ais523> I'd find that vaguely useful, although not sufficiently useful to insist someone else should implement it
14:03:06 <ais523> it should be opt-in, at least
14:03:09 <ais523> that'll help to reduce flooding
14:03:14 <Gregor> ais523: Of course opt-in
14:03:37 <ais523> it seems futile from the point of view of saving SCO, but that doesn't seem to be the actual objective
14:03:44 <ais523> and it may be accomplishing whatever the actual objective is quite well
14:03:47 <Vorpal> ais523, what would the objective be then
14:04:10 <ais523> although there's enough money moving around, that someone probably ends up benefiting as a result
14:04:23 <ais523> I'm not entirely sure who, as I lost track
14:04:33 -!- FireFly has joined.
14:04:39 <fizzie> At least the bankruptcy lawyers seem to be being paid.
14:04:45 -!- cheater- has joined.
14:05:00 <Vorpal> didn't they have a negative amount of money?
14:05:19 <Vorpal> so where would they get money from, surely lending it would be hard for them...
14:05:43 <ais523> well, there have been people lending them money, perhaps surprisingly
14:06:18 <fizzie> They did also have some positive money at the beginning of the banruptcy process. At least if you don't count the amount they owe to Novell.
14:06:48 -!- cheater00 has quit (Ping timeout: 255 seconds).
14:06:53 <ais523> it's not quite "owe", it was found that the money was "converted" from Novell, so it technically owns to Novell, just it's currently in SCO's ownership
14:07:04 <ais523> except it isn't, because they already spent it
14:07:25 <Vorpal> what does that mean in this context
14:07:38 <ais523> Vorpal: it's a legal term, I'm not quite sure what it means exactly, but it seems to have a similar meaning to "stolen", or perhaps "defrauded"
14:07:51 <fizzie> I think Ocean Park Advisors have been billing them for about $30000/month for 16 months now.
14:08:33 <fizzie> Mostly for planning and implementing the mythical "restructuring plan".
14:09:46 <Gregor> `addquote <Vorpal> ais523, "converted"? what does that mean in this context <ais523> Vorpal: it's a legal term, I'm not quite sure what it means exactly, but it seems to have a similar meaning to "stolen", or perhaps "defrauded"
14:09:48 <HackEgo> 337) <Vorpal> ais523, "converted"? what does that mean in this context <ais523> Vorpal: it's a legal term, I'm not quite sure what it means exactly, but it seems to have a similar meaning to "stolen", or perhaps "defrauded"
14:10:00 <Vorpal> Gregor, why was that quote-worthy?
14:11:10 <Gregor> Because lawl with emphasis on "law"
14:11:24 <fizzie> (The $30k figure was based on a quick look at the recent bills; seems that the early ones were rather larger, $196,002.50 for the first six weeks and so.)
14:11:48 <Vorpal> fizzie, the chances of those ever being repaid seems to be about zero?
14:11:55 -!- variable has quit (Read error: Connection reset by peer).
14:12:21 <fizzie> No, I think those have been paid, more or less; which is probably one of the reasons why SCO no longer has any of the money it was loaned.
14:12:35 <Gregor> Clearly SCO has done quite well for themselves.
14:13:17 <ais523> also, the loan was secured against everything the company owned, more or less
14:13:31 <ais523> so if it defaults, I'm not entirely sure what will happen, but it could be quite interesting
14:13:45 -!- variable has joined.
14:13:48 <fizzie> Actually isn't today Novell's last day to appeal for the UnXis sale?
14:13:55 <ais523> (bonus points if SCO manages to get multiple loans secured against the same assets)
14:14:05 <ais523> do you think they'll appeal?
14:15:22 <fizzie> I would have guessed "yes" if it was just Novell, but I don't know how much the Novell/Attachmate thing changes the matter.
14:18:34 <fizzie> Of course I haven't been following the whole thing very closely, just read the groklaw posts; those make the sale sound pretty dubious indeed.
14:20:10 -!- ais523 has quit (Read error: Connection reset by peer).
14:21:31 -!- ais523 has joined.
14:21:51 <ais523> hmm, I'm dealing with a system that uses 0 for true, non-0 for false
14:22:12 <ais523> what value should I use, concretely, for false? atm I'm doing -1 as it's easy to write as "not true" in two's complement
14:22:22 <ais523> (also, what value should I use for FILE_NOT_FOUND?)
14:28:52 <Gregor> You have a language/environment where 0 is true, and also you have a bitwise not operator?
14:30:26 <ais523> no, it's what happens when mathematicians try to project their idealized views on programming languages into the real world
14:30:37 <ais523> besides, this is #esoteric, is having 0 for true and bitwise-not /that/ bad?
14:31:58 <oklopol> 0 is used as true in math?
14:32:16 <ais523> and the language just chose the "wrong" mapping
14:32:19 <oklopol> well 1 is often used as true, and 0 as false
14:32:47 <oklopol> actually i used 1 as true, 0 as false around 15 minutes ago
14:36:29 <oklopol> but then again i suppose i could've just not thought of the 1 as true
14:37:07 <cheater-> rofl, as if web 2.0 wasn't bad enough.. now i have found a company which says they're an expert in Enterprise 2.0.
14:37:40 <cheater-> any sort of job you don't want to do :D
14:38:48 -!- copumpkin has quit (Quit: Computer has gone to sleep.).
14:39:52 <oklopol> it's not very hard to find a job i don't want to do
14:40:39 <ais523> how many of those jobs could you successfully apply for, though/
14:41:19 <oklopol> i'm sure i could get any job i wanted to by just going to the interview and telling them i'm the awesome
14:52:37 <cheater-> aww, i just called someone and they told me off for calling them in the office
14:55:22 -!- FireFly has quit (Quit: swatted to death).
14:56:10 -!- FireFly has joined.
14:58:54 -!- olsner has quit (Ping timeout: 248 seconds).
15:05:58 -!- copumpkin has joined.
15:08:57 -!- BeholdMyGlory has joined.
15:11:56 -!- olsner has joined.
15:17:18 <ais523> gah, the instructions for these pills don't have a "what to do if you need eight tries to swallow them whole and manage to spit half a bottle of water onto the floor in the process"
15:17:49 <ais523> I often have trouble with the mechanical aspects of getting a pill down my throat, but that was just ridiculous
15:17:52 <ais523> also, my trousers are wet now
15:18:20 -!- cheater- has quit (Ping timeout: 255 seconds).
15:18:27 -!- cheater00 has joined.
15:19:47 <ais523> well, I can only really practice once every eight hours
15:19:59 <oklopol> you can practise swallowing other things tho
15:20:05 <ais523> and I fear dire things might happen to me if I screw up even worse than that, e.g. biting it in half by mistake
15:20:12 <ais523> and there are not many things around designed to be swallowed whole
15:20:49 <ais523> people swallow those whole?
15:20:54 <ais523> I fear I'd choke if I tried, they seem rather large
15:21:21 <oklopol> but i doubt it's very hard
15:23:07 <Gregor> http://google.com/search?q=learn+to+deep+throat THERE YA GO
15:23:21 <oklopol> deep throating isn't very hard either
15:23:25 <ais523> Gregor: I doubt that link's going to turn up a lot of relevant stuff, even with safesearch on
15:23:32 <Gregor> elliott: THIS IS WHY PEOPLE THINK HE'S GAY
15:24:05 <oklopol> you can control most reflexes with a bit of practise, but i can't seem to get autoblinking off no matter what i do
15:24:20 <oklopol> i spent hours and hours on this last summer
15:24:28 <ais523> this is more an antireflex
15:24:41 <ais523> in the end, I managed it this time by reaching all the way back into my mouth and physically inserting the pill there
15:24:56 <ais523> both it and my mouth were so wet after the previous failures that it actually fell right down
15:25:41 -!- pikhq_ has joined.
15:25:47 -!- pikhq has quit (Ping timeout: 250 seconds).
15:26:12 <oklopol> any of you tried drinking without swallowing? that would be a neat thing to learn, but since you can die of it afaik, i haven't had the guts to go trial and error
15:26:16 <Gregor> <ais523> in the end, I managed it this time by reaching all the way back into my mouth and physically inserting the [censored] there <ais523> both it and my mouth were so wet after the previous failures that it actually fell right down <oklopol> :P
15:26:42 <ais523> oklopol: a sort of controlled drowning?
15:27:21 <oklopol> the liquid goes in your stomach
15:27:50 <ais523> I know, I just wanted to say "controlled drowning"
15:28:02 <Gregor> You should waterboard yourself.
15:28:05 <oklopol> how that's guaranteed, i don't know, i imagine if you just go for it, you prolly either succeed, or nothing goes down.
15:28:08 <Gregor> Seems like the ideal solution.
15:28:21 <oklopol> i never quite got how waterboarding could be dangerous
15:28:36 <ais523> I don't think it's physically dangerous, more it's the psychological issues
15:28:42 <oklopol> would be nice to try that some time, but the serious brain injury part puts me off a bit
15:29:23 <ais523> although I imagine waterboaring /yourself/ would cause different issues than someone else doing it to you
15:29:25 <oklopol> but drowning sounds like an interesting experience
15:29:52 <Gregor> What was it they did on tosh.0, orange-juice-boarding or something
15:30:00 -!- Slereah has joined.
15:30:03 <ais523> presumably triggering the body's drown reflex could be dangerous, too
15:30:31 <Gregor> Does the body actually have a "drowning" reflex per se?
15:30:31 <ais523> note that there's also a case where someone consented to be waterboarded, and afterwards decided it was life-runing torture despite being consented to and sued
15:30:51 <oklopol> it's a horrible thought that i have this whole set of drowning related feelings built in, and i never get to experience them
15:30:53 <ais523> freedivers learn to trigger it deliberately
15:31:05 <ais523> and use it to hold their breath for several minutes at a time
15:31:47 <ais523> perhaps oklopol could learn to do that too
15:31:47 <ais523> `addquote <oklopol> it's a horrible thought that i have this whole set of drowning related feelings built in, and i never get to experience them
15:31:48 <HackEgo> 338) <oklopol> it's a horrible thought that i have this whole set of drowning related feelings built in, and i never get to experience them
15:32:31 <ais523> well, humans have an ability that can only be triggered unconciously to use all their muscle fibers at once (typically they only use about 1/6 or so, because using more does permanent damage)
15:32:56 <ais523> and people who are panicking sufficiently sometimes do it and do really implausible things as a result, like lifting trees
15:33:11 <ais523> (presumably, which aren't attached to the ground at the time, and are quite large, or it wouldn't be implausible)
15:33:50 <oklopol> yeah because lifting trees that are attached to the ground isn't hard at all
15:34:00 <ais523> well, it would be impressive
15:34:06 <ais523> but more implausible than I'm easily willing to consider
15:34:12 <oklopol> the roots go "hey cool i get to see the sun" and help you lift it
15:34:44 <ais523> but that's not what roots are /for/
15:34:46 <Gregor> It's still alive at that point, so it's a willing participant.
15:34:50 <Gregor> It just climbs out of the ground.
15:35:07 <ais523> it's like your heart deciding it wants to do some thinking for a bit, and swapping places with your brain
15:35:17 <ais523> which is a nice metaphor, but seems unlikely if taken literally
15:35:40 <Gregor> Hormones can cause behavior, and are carried in the blood :P
15:36:10 <oklopol> yeah, suddenly you start beating your head against the wall every second or so. and spout fluids everywhere.
15:37:01 <oklopol> so how about someone who can actually lift an incredible amount of shit
15:37:15 <oklopol> if their dog is sick and they get all hulk
15:37:30 <oklopol> or do they already use more than the healthy amount of their muscles
15:38:08 <Gregor> So, to be perfectly clear ...
15:38:18 <Gregor> Their dog usually produces like 75lb turds
15:38:26 <Gregor> Which they carry to the bin with no effort at all
15:38:44 <oklopol> because it's an interesting concept otherwise: you train doing X all your life, and you actually had the theoretical possibility of doing X even better, before you even started learning it
15:38:45 <Gregor> Their dog is now sick and has filled their entire house with dark-matter shit.
15:39:26 <oklopol> yeah that's roughly what i meant
15:39:37 <ais523> I'm not sure how it works with athletes
15:40:27 <ais523> I think possibly the training strengthens their muscles so they can safely use a larger proportion
15:41:51 <oklopol> also often they sniff that what's that stuff before lifting stuff
15:42:42 -!- asiekierka has joined.
15:44:14 <ais523> hmm, I wonder how easy it would be to statistically predict who said a given sentence based on things like word choice and order?
15:44:27 <ais523> it's a game we used to play with optbot sometimes last time it was round here
15:44:33 <ais523> and seems relatively automatable
15:45:22 <oklopol> i can usually spot my own rather easily, i'm usually the ones that make me laugh
15:45:58 <fungot> Gregor: a linear algebra course and fnord by reading euclid. but now it is
15:46:27 <Gregor> fungot: ENTERTAIN US MORE!
15:46:28 <fungot> Gregor: lack of sensible fnord. then you can choose from major and minor alternatively is kinda cool right mabye you should put in a good language for teaching purposes, it seems
15:46:51 <ais523> "The copied elements that contain instructions, such as BREAD and CPIO, might perhaps be trade secrets, but Defendants' experts have argued persuasively that these instructions are either in the public domain or otherwise exempt."
15:48:05 <ais523> Gregor: it's a quote from the BSD court case (where it was determined that BSD UNIX wasn't a trade secret)
15:48:21 <ais523> but I thought cpio was a tar competitor, and who calls an instruction BREAD?
15:48:30 <Gregor> Yeah, what's BREAD X-D
15:50:14 <fizzie> ais523: I did do some authorship attribution experiments on #esoteric logs (using code we did for guessing book authors on a "statistical NLP" course); given a large enough sample it did reasonable, despite being bog-stupid, but single comments not so well.
15:50:37 <fizzie> (Unsurprising since the features were word-length statistics and such.)
15:51:45 <Ilari> APNIC down 0.13: 16k to Taiwan, 1k to Indonesia, 640k(512k+128k)+1280k(1M+256k)+512k+128k to China, 4k+2k+256+/48 to India, 4k to Vietnam, 4k to Australia, 256 to Philippines.
15:51:46 <fizzie> Some author-discriminating features (attribution/punctuation styles and such) would probably help.
15:51:52 <Gregor> fizzie: Why are you making fun of bogs?
15:52:16 <fizzie> Gregor: 'Cause they're not so smart.
15:52:37 <Gregor> fizzie: I've got a bog that could beat you in Chinese Checkers seven times out of twelve!
15:53:13 <fizzie> That's not hard: even I can beat myself that often.
15:53:27 <Gregor> Yeah, I'll bet you "beat" yourself that often.
15:54:28 <Ilari> Pretty quiet day there. Only 7% of their pool gone in single day.
15:54:41 -!- ais523 has quit (Ping timeout: 252 seconds).
15:54:47 -!- ais523 has joined.
15:57:21 -!- Wamanuz3 has quit (Ping timeout: 250 seconds).
15:57:36 <oklopol> every time anyone says anything
15:57:41 <oklopol> it tells who it thinks said it
15:58:04 <Gregor> I shall implement it, and call it AnnoyingSpammerBot
15:58:15 <oklopol> it could do it in pm if you've put that feature on
15:58:31 <oklopol> and sometimes on the channel if it's REALLY sure
15:58:31 <Gregor> It could do it in CTCP CHAT :P
15:58:47 <oklopol> it sends you a text message
15:58:55 <Gregor> <oklopol> hew how about <ReallyAnnoyingBot> oklopol said that!
15:59:44 <oklopol> hey seriously, this has to happen
15:59:51 -!- Phantom_Hoover has joined.
16:05:10 -!- Wamanuz3 has joined.
16:05:20 <Vorpal> oklopol, Gregor: should it just look at the nick or should it instead do some heuristic on the text? Matching writing style I mean
16:05:31 <oklopol> or Gregor, if he has some algo for it easily handable
16:05:36 -!- MigoMipo has joined.
16:05:41 -!- quintopia has quit (Ping timeout: 252 seconds).
16:05:46 <oklopol> Vorpal: what do you think lol
16:05:50 -!- quintopia has joined.
16:06:15 <fizzie> oklopol: Gregor is more of a guy who does; I'm both busy and in a bus, and the other assorted excuses.
16:06:25 -!- ais523 has quit (Read error: Connection reset by peer).
16:06:52 <oklopol> fizzie: those "excuses" are just an excuse
16:07:22 <fizzie> Oh no, I've become invisible?
16:08:12 -!- ais523 has joined.
16:24:14 -!- ais523_ has joined.
16:25:59 -!- ais523 has quit (Ping timeout: 252 seconds).
16:26:31 -!- ais523_ has changed nick to ais523.
16:27:19 -!- cal153 has quit (Ping timeout: 250 seconds).
16:34:22 -!- Wamanuz4 has joined.
16:36:24 -!- Wamanuz3 has quit (Read error: Operation timed out).
16:43:55 -!- boily has joined.
16:47:14 -!- BeholdMyGlory has quit (Remote host closed the connection).
16:48:31 -!- BeholdMyGlory has joined.
16:49:37 -!- boily has quit (Quit: leaving).
16:52:12 -!- ais523 has quit (Read error: Connection reset by peer).
16:52:36 -!- ais523 has joined.
16:57:57 -!- copumpkin has quit (Ping timeout: 276 seconds).
16:59:38 -!- copumpkin has joined.
17:01:13 -!- cal153 has joined.
17:03:35 -!- ais523 has quit (Remote host closed the connection).
17:19:47 <Phantom_Hoover> Silly syntax based on something I read on Reddit: \lambda x.expr = \frac{expr}{x}
17:20:26 <Gregor> What sort of bastard half child of Lambda Calculus is this?
17:21:17 <Phantom_Hoover> Well, it's based on the classic exam fail of thinking cos x/x = cos.
17:22:46 -!- azaq23 has joined.
17:24:49 -!- cheater99 has joined.
17:26:09 -!- cheater00 has quit (Ping timeout: 255 seconds).
17:36:06 -!- pikhq has joined.
17:36:15 -!- pikhq_ has quit (Ping timeout: 260 seconds).
18:00:25 -!- pikhq has quit (Ping timeout: 240 seconds).
18:00:40 -!- pikhq has joined.
18:06:01 <Ilari> As scale of things going on with APNIC: 14 more days like today will deplete APNIC. And what happened today wasn't so extraordinary: 3 of 5 business days last week saw much greater activity.
18:13:51 -!- ais523 has joined.
18:21:21 -!- pumpkin has joined.
18:24:39 -!- sftp has quit (Remote host closed the connection).
18:24:49 -!- copumpkin has quit (Ping timeout: 252 seconds).
18:37:10 -!- BOI-ish has joined.
18:39:17 -!- BOI-ish has left ("ciao.").
18:43:25 -!- sftp has joined.
18:43:47 <Ilari> Last week: 0.76 blocks gone. Oh, and there are only 1.71 left.
18:44:52 <Ilari> Also, starting to fragment: The largest free block is 1M (which there is only 1 of).
18:46:08 <Ilari> 28 748 288 addresses in 2 287 blocks.
18:47:39 <Ilari> (Not counting 103.0/8)
18:49:17 <Ilari> Logaritmic size: /7.223
18:50:17 <Ilari> Up from /7.120 before today. So +0.103.
18:50:22 -!- cheater99 has quit (Ping timeout: 248 seconds).
18:57:06 -!- cpressey has joined.
18:58:12 <cpressey> Deewiant: you know what would be a really nice addition to the Mycology readme? A copy of the complete expected output for a compliant interpreter.
18:58:21 <ais523> isn't it really easy to allocate blocks in a way that avoids fragmentation, as long as they aren't expanded later?
18:58:38 <ais523> cpressey: there's a bunch of UNDEFs, so you couldn't just use diff
18:59:49 <Ilari> Huh. Linux sure nowadays has lots of kernel threads: 98 for me.
19:00:14 <cpressey> ais523: I realize that. Though now that you mention it, I'm not sure why undefined cases are even being tested.
19:00:39 <ais523> it's to inform the user of what the case does
19:00:56 -!- pikhq_ has joined.
19:01:23 <cpressey> That's a nice thought, but it's not a test.
19:01:25 -!- pikhq has quit (Ping timeout: 260 seconds).
19:03:04 <ais523> well, it /is/ a test, isn't it?
19:05:20 <cpressey> Not if a test either passes or fails.
19:20:15 <cpressey> Just saying, there's an argument for putting undefined-behavior-probing in other source file(s), and not mixing them into an otherwise unambiguous test suite.
19:21:17 <cpressey> And for my purposes, I can just download a copy of cfunge and see its output, since it doesn't fail out anywhere according to the results page. Hopefully easier than installing D and cmake and such :)
19:25:52 <cpressey> Sure, except for the "checking it out using yet another different source-control tool" part. I am not at all sure bazaar > D
19:28:04 <Gregor> They're sort of both hell.
19:46:44 -!- augur has quit (Remote host closed the connection).
19:49:35 -!- cheater- has joined.
20:02:52 <cpressey> There is something deeply wack with FBBI's stack management routines. This is not news. What is news is, I think I see what it is.
20:03:38 <cpressey> Also, why does this coffee taste like whole wheat? That's very disturbing.
20:05:11 <pikhq_> Perhaps it's actually mugicha (麦茶)?
20:06:28 -!- oerjan has joined.
20:06:47 <oerjan> 17:22:24: * Phantom_Hoover wonders what algebraic structures are preserved.
20:06:47 <oerjan> 17:22:46: * azaq23 has joined #esoteric.
20:07:04 <oerjan> Gregor: i think it would be a good idea not to use * for non-emotes, it's confusing
20:07:33 * pikhq_ thinks Gregor should use an irssi-esque output format
20:07:51 <pikhq_> Which would entail only using * for emotes.
20:08:07 <pikhq_> -!- for non-PRIVMSG stuff.
20:08:58 <pikhq_> Phantom_Hoover: He also has logs that are essentially a filtered dump of raw IRC traffic, with times prepended.
20:14:23 -!- pikhq has joined.
20:16:26 -!- pikhq_ has quit (Ping timeout: 240 seconds).
20:22:13 -!- rapido has joined.
20:25:59 * cheater- sets mode #esoteric +b oklopol
20:26:31 -!- asiekierka has quit (Ping timeout: 246 seconds).
20:27:38 <rapido> are there an interesting 'collection oriented' language that is not apl/j/k?
20:27:51 -!- augur has joined.
20:29:19 -!- poiuy_qwert has joined.
20:29:48 <rapido> is there something like 'map theory'? I know there is something like 'array theory'
20:32:37 <rapido> array theory: http://www.nial.com/ArrayTheory.html
20:33:16 -!- pikhq_ has joined.
20:33:20 <rapido> ah found something: http://www.mangust.dk/skalberg/papers/gkli-slides1.pdf
20:33:30 <rapido> map theory: http://www.mangust.dk/skalberg/papers/gkli-slides1.pdf
20:35:14 <rapido> wouldn't it be nice to have a map oriented language?
20:35:32 <rapido> everything is a map - data and code
20:36:14 -!- pikhq has quit (Ping timeout: 252 seconds).
20:37:18 <rapido> concrete map: [0=0;1=1;2=4;3=9]
20:37:20 <Gregor> If somebody wants to make an HTMLifier, I will not argue with them :P
20:38:03 <rapido> Phantom_Hoover: yes, that's one way of phrasing it
20:38:26 <Gregor> I thought you meant like a map /operation/, not a map /datatype/
20:38:28 <rapido> total functions would be nice
20:38:42 <Gregor> It's turtles (maps) all the way down.
20:39:27 -!- FireFly has quit (Quit: swatted to death).
20:39:46 <rapido> this would be a lazy map: [x<-[0..10000000];x*x]
20:40:41 <rapido> still finite because the domain is finite
20:41:17 <Gregor> That looks more like a lazy list to me.
20:42:00 <rapido> Gregor: ok, i haven't really settled for a notation
20:42:30 <Gregor> If it's just a notation issue, then I don't understand what that means X-P
20:43:39 <rapido> domain: 0..10000000 : range: x*x
20:44:09 <rapido> Phantom_Hoover: yes - thanks
20:45:07 -!- augur has quit (Remote host closed the connection).
20:45:34 -!- augur has joined.
20:47:09 <rapido> the domain (keys) and range (values) can be maps too.
20:47:48 <rapido> In fact, literals are maps in disguise
20:48:02 <rapido> there should be only maps!
20:51:23 <Gregor> Phantom_Hoover: Oh, OK, I wasn't getting the syntax.
20:51:23 <rapido> I've done something similar with enchilada- but i like to be more restrictive than enchilada (i.e. finitie maps only)
20:51:49 <Phantom_Hoover> rapido, so basically everything is a function from a finite sense?
20:54:29 -!- calamari has joined.
20:54:42 <Phantom_Hoover> http://www.uclassify.com/browse/uClassify/GenderAnalyzer_v5
20:55:46 <oerjan> Phantom_Hoover: i don't see why that makes it non-TC.
20:56:39 <rapido> Phantom_Hoover: say that you have an recursive function that doesn't terminate
20:57:38 <rapido> now let's imagine an interpreter that takes this same recursive function, together with a user-defined 'number of interpreter steps'
20:57:43 <pikhq_> Phantom_Hoover: Seems that all of Hemingway's writing would be classed as female, then.
20:57:46 <oerjan> you could still do cons lists.
20:57:51 <pikhq_> And Hemingway is 200% male.
20:58:09 <rapido> when the interpreter reaches the 'number of interpreter steps' it terminates
20:58:29 <oerjan> [0=1; 1=[0=2; 1=[0=3; 1=[0=4; 1=[0=5; 1=[0=6; 1=
20:59:30 <rapido> oerjan: consing can be done - nice observation
21:00:20 <cpressey> "That this Funge has 80 dimensions"
21:03:30 <cpressey> I'm just impressed that I got it to get that far. I had to rewrite a good chunk of the stack routines. They were confused about when they were supposed to allocate new memory for a stack header versus when that memory was already allocated for them.
21:05:11 <cpressey> Phantom_Hoover: I'm not. FBBI thinks it interprets an 80-dimensional funge. When you ask it. With 'y'.
21:06:03 <rapido> question: how would you give a unique name to a arbitrary block of bytes without hashing (=possible collisions) and without using a central service (thing p2p)
21:07:06 <rapido> oh - the same block of bytes should map always return the same name
21:08:33 <cpressey> rapido: I don't think it's possible.
21:09:11 <cpressey> Well, a perfect hash function would do it. But that's hashing.
21:09:41 <cpressey> And if the blocks are truly arbitrary, the only perfect hash function is identity :)
21:10:27 <ais523> cpressey: incorrect, rot13 is also a perfect hash function
21:10:28 <rapido> cpressey: ok, what about a central service which just increases a counter for each new block that has been issued?
21:10:35 <ais523> as are many compression algorithms
21:10:48 <pikhq_> Any function with an inverse, really.
21:11:38 <rapido> what if we scale the central naming service to log(n) naming services - with n being the number of blocks issued?
21:12:23 <rapido> dns scales pretty good
21:12:50 -!- wareya_ has changed nick to woroyo.
21:15:23 <cpressey> ais523: fine -- the only pefect hash functions are permutations
21:15:37 <ais523> gzip is a permutation?
21:16:08 <cheater-> LOL JAVA: http://www.nupxl.com/wp-content/uploads/2010/10/Fail-and-Funny-Names-batman-bin-suparman.jpg
21:17:11 <pikhq_> cpressey: There's a large class of bijective functions.
21:17:28 <pikhq_> Some of them aren't, strictly speaking, permutations.
21:17:46 <cpressey> Yes, but I'm not allowed to speak the truth.
21:18:12 <pikhq_> Which clearly means that you are permitted to speak the truth, but opt not to.
21:18:16 <rapido> cpressey: i want to achieve (function) memoization - not only within one instance of running program - but globally
21:18:44 <pikhq_> rapido: Universal memoization is not as good an idea as you may think.
21:18:57 -!- MigoMipo has quit (Read error: Connection reset by peer).
21:19:48 <Phantom_Hoover> Hmm, shouldn't compressing a compressed file increase its size?
21:20:03 <rapido> pikhq_: it doesn't need to be persistent always - just the most used functions (structures)
21:20:05 <pikhq_> Phantom_Hoover: Not *necessarily*.
21:20:29 <pikhq_> rapido: Automatic memoization is a *hard* problem.
21:20:38 -!- impomatic has joined.
21:20:40 <oklopol> pikhq_: isn't a bijection called a permutation if the set is infinite?
21:20:49 <variable> Phantom_Hoover: compressing a compressed file may not always be possible
21:20:54 <rapido> pikhq_: 'memoization is a *hard* problem' - i like that!
21:21:06 <variable> one still need to provide a minimum amount of information
21:21:12 <Phantom_Hoover> variable, it arguably *shouldn't* be possible given a perfect compression algorithm.
21:21:24 <pikhq_> Phantom_Hoover: There's no such thing.
21:21:34 <pikhq_> rapido: At least as hard as parallel computing.
21:21:47 <rapido> pikhq_: i'm the author of enchilada - i have done some 'experiments' on the subject.
21:21:50 <variable> Phantom_Hoover: Information theory discuses the absolute minimums and maximums as relate to compression
21:21:59 <pikhq_> Phantom_Hoover: A "perfect" compression algorithm is a bijective mapping from a larger set to a smaller set.
21:22:04 <pikhq_> Phantom_Hoover: Which is of course impossible.
21:22:07 <cheater-> oklopol: a bijection is called a permutation if the set is FINITE
21:22:24 <Phantom_Hoover> pikhq_, hence I'm defining "perfect" far more fuzzily.
21:22:25 <cheater-> oklopol: you usually don't talk about permutations of infinite sets
21:22:34 <oklopol> "<rapido> are there an interesting 'collection oriented' language that is not apl/j/k?" <<< toi
21:22:45 * variable is writing a BF compiler using llvm ... should be fun how to use llvm :-)
21:22:52 <pikhq_> Phantom_Hoover: And I define addition to mean fish blue purple flimble boob.
21:22:53 <rapido> i want to get rid of enchilada's cryptographic hashes - but still scale in a distributed setup
21:23:34 <oerjan> oklopol: that's just your toi project
21:23:59 <oklopol> "<variable> oklopol: ?" <<< sometimes one needs to o
21:25:16 <rapido> oklopol: is there a interesting 'collection oriented' language that is also esoteric ;)
21:25:45 <oklopol> "<cheater-> oklopol: a bijection is called a permutation if the set is FINITE" <<< i was not aware of this
21:26:06 <oklopol> at least in group theory, this convention is not always followed
21:26:26 <oklopol> people speak of permutation representations even for infinite gruplers
21:26:26 <cheater-> every bijection of a finite set is always a permutation.
21:26:38 <cheater-> oh ok yeah, but that's a different sense
21:26:54 <cheater-> then you're talking about permutations of structured sets.. groups etc
21:27:20 <oklopol> permutation and bijection are synonyms, but i wouldn't be surprised if people only use it for say up to countable infinity
21:27:51 <oklopol> "<cheater-> then you're talking about permutations of structured sets.. groups etc" <<< no, sets
21:28:06 <cheater-> it wouldn't feel natural to me to call hilbert hotel a permutation for example
21:28:09 <oklopol> permutation representation = representing the group as a set of permutations
21:28:17 <cpressey> a permutation of a set is not itself a set
21:28:26 <cheater-> btw, permutation is only when it's a bijection onto itself
21:28:29 <oklopol> do you mean the thing where everything is shifted
21:28:36 <oklopol> that's an injection but not a bijection
21:28:42 <oklopol> so of course it's not a permutation
21:29:26 <oklopol> i thought the hilbert hotel thing was that you have a sequence indexed by N and you shift it to obtain empty rooms
21:30:09 <oklopol> "<cheater-> btw, permutation is only when it's a bijection onto itself" <<< oh, well this is certainly true
21:30:19 <cheater-> oklopol: imagine Z^2 where (0, c=const) is injected into (c, 0) and (c, v=var) is hilberted vertically
21:30:38 <cheater-> and then the empty space is collapsed
21:31:00 <oklopol> erm, you're describing a random bijection?
21:31:15 <cheater-> i described a bijection which uses the hilbert hotel principle
21:31:30 <cheater-> to show you it's not always an injection
21:32:34 <oerjan> ...a bijection is always injective.
21:32:56 <oklopol> but okay, you don't consider self-bijections of even countable sets worthy of the name permutation, that's all you wanted to say i suppose
21:33:42 <oklopol> oerjan: he meant using hilbert hotel principle doesn't always give you something that is an injection without being surjective
21:33:49 <rapido> don't surjectively inject your hilbert hotel principle into the discussion - please!
21:34:25 <oklopol> oerjan: what if some REALLY complicated and SURPRISING bijections aren't injective?
21:35:18 <pikhq_> oklopol: A bijection is an injection and a surjection, by definition.
21:35:22 <oklopol> and don't just say "they're injective by definition", things don't become true if you define them to be true
21:35:33 <oklopol> yeah pikhq_ takes the easy way out
21:35:54 <pikhq_> Things do become true if we define them to be true.
21:36:08 <pikhq_> At least in the context where it's defined to be true.
21:36:36 <oklopol> oerjan: in the theory of CA, we say G is time symmetric if there is an involution I such that G^(-1) = IGI, is this a common concept elsewhere?
21:37:15 <oklopol> time-symmetry includes lots of things that consist of particles, for instance, it's an interesting concept
21:37:32 <oklopol> but i don't recall seeing anything exactly like it anywhere
21:37:40 <oklopol> then again what the fuck do i know, that's why i'm asking
21:38:09 <oklopol> pikhq_: yeah right, next u gonna say the reals are uncountable lol
21:38:19 <oklopol> oerjan: sorry I is an involution if I^2 = id
21:38:28 <oklopol> where maybe i should've named things differently
21:38:35 <pikhq_> oklopol: No, I'm going to say the naturals are countable.
21:38:36 <cpressey> if you can apply it to a CA, I would be surprised if you couldn't apply it to any semantics, with enough effort
21:38:37 <oerjan> i know but why not just G^(-1) = IGI^-1 ?
21:38:49 <oklopol> oerjan: well sure, that works too
21:39:02 <oerjan> um are they equivalent
21:39:28 <oklopol> what are equivalent exactly?
21:39:38 <oklopol> i read "um they are equivalent"
21:39:47 <oklopol> no i don't see why that definition would be equivalent
21:40:03 <oerjan> well why then did you say "that works too"
21:40:07 <oklopol> but yeah the idea is you conjugate with I to get G's complement
21:40:16 <oklopol> oerjan: sorry, i meant like "well sure, you can generalize it"
21:40:36 <cpressey> oerjan's looks like a conjugate. oklopol's doesn't.
21:40:43 <cpressey> but i don't know from conjugates
21:40:43 <oerjan> i recall using the word anti-conjugate or something like that
21:41:29 <pikhq_> http://www.space.com/7044-moon-apollo-astronauts-customs.html
21:42:01 <oerjan> it popped up in one of the basic theorems we looked at, an orbit equivalence with continuous cocycles is always either a conjugation or an anti-conjugation
21:42:05 <oklopol> oerjan: you don't recall anything with involutions tho?
21:42:07 <rapido> are there any CA formalism that takes previous (N not just the current) world states as input?
21:42:43 <fungot> Phantom_Hoover: i've only sometimes talked to him and started pulling his boots and kecks off.") well, i'll call is something different, though.
21:42:58 <oklopol> rapido: no, but those are essentially the same thing
21:43:30 <cheater-> btw, i noticed lately that knuth defined some numbers of negative base
21:43:36 -!- MUILTFN_ has joined.
21:43:40 <cheater-> he also proved some fun properties about them
21:44:05 <oklopol> surely someone has defined such a construction (running n generations simultaneously), for instance the block representation, that is, remembering neighbors in cells, and higher block representations, compressing blocks of n into single symbols, are used a lot in symbolic dynamics
21:44:12 <cpressey> finally, something with which I can describe my bank balance
21:44:39 <cpressey> that was in response to cheater-'s comment, but now I want it to have applied to what oklopol just said
21:44:40 <rapido> oklopol: could such formalism be more powerful - not in a TC sense - but in a 'programming' sense - whatever that means
21:44:53 <variable> is there any variant of brainfruck that lets the user make syscalls?
21:45:02 <oklopol> but i've read more about symbolic dynamics than CA, and in symbolic dynamics, you always just run your CA for exactly one step :)
21:45:12 <oerjan> rapido: mcell has some "ca families" that use memory
21:45:15 <cpressey> variable: I don't know, but, yes
21:45:35 <rapido> oerjan: thanks for the pointer
21:45:37 <oerjan> but in theory that just gives a more compact representation
21:46:18 <oklopol> rapido: well i haven't seen them used, at least
21:46:51 <oklopol> but actually, remembering your neighborhood is essentially just what you described, for the shift CA
21:46:57 <cheater-> http://en.wikipedia.org/wiki/Negative_base < haha, a practical use of intercal
21:47:11 <oklopol> so you just generalized one of the key concepts of everything to do with sequences!
21:47:13 <Zwaarddijk> the russians built a computer once that was in balanced ternary
21:47:41 <variable> Zwaarddijk: yes - it was expensive & unstable
21:47:45 <cheater-> wasn't that in the 80s or something
21:47:53 <Zwaarddijk> possibly earlier, but yeah, expensie & unstable.
21:48:04 <variable> cheater-: I already have a practical use of INTERCAL
21:48:21 <variable> whenever someone asks me what language to learn first I have a good answer
21:48:34 <rapido> i like surreal numbers
21:48:51 <rapido> surreal number subsume all numbers
21:49:12 <oerjan> ...well not _all_ numbers
21:49:15 <cheater-> i liked the definition of hyperreal numbers
21:49:25 <oklopol> i don't know their definition
21:49:30 <Gregor> Index finger -> cat nose is like the human<->cat equivalent of a fist-bump.
21:49:50 <cheater-> oklopol: a formal definition of things like dx, dy, 0+, 0-, and so on
21:49:51 <oerjan> the basic definition doesn't include complex numbers, although you can probably complete it
21:50:08 <oerjan> and i doubt it includes p-adics
21:50:08 <oklopol> i also know astrophysics is about space and stuff
21:50:26 <oklopol> and that sex is about doing things under the blanket
21:50:44 <oerjan> although they do indeed include real numbers and transfinite ordinals
21:51:12 <oklopol> well aren't surreals meant to make the real line as long as humanly possible
21:51:37 <oklopol> cheater-: yeah i know the usual definition is based on ultrafilters
21:51:45 <oerjan> oklopol: i'm not sure that was the original motivation...
21:52:05 <oerjan> but including ordinals does pretty much ensure that
21:52:18 <cheater-> http://chc60.fgcu.edu/images/articles/Marczynski.pdf < lol cool!!
21:52:36 <cheater-> RAM made out of acoustically induced mercury. wtf?
21:54:00 <oklopol> "<cheater-> oklopol: that's all u need" <<< you know how fucking annoying it is when people learn about hyperreals and suddenly they can do their calculus without thinking altogether because "hey, this can be made formal, stick that epsilon up your ass!"
21:54:17 -!- rapido has quit (Remote host closed the connection).
21:54:51 <cheater-> oklopol: is it because you wish you had learnt about it earlier?
21:55:00 <oklopol> because that's like saying "hey, there is a god, therefore i don't have to think about the details"
21:55:08 <oklopol> all they have learned is the term hyperreal
21:55:08 <Phantom_Hoover> oklopol, what's the problem with using hyperreals over limits?
21:55:21 <oklopol> Phantom_Hoover: nothing, but no one ever uses them
21:55:37 <cheater-> oklopol: so what you mean is that people learn about hyperreals and suddenly they THINK they can do their calculus without thinking altogether
21:56:55 <oklopol> we bought a whiteboard today
21:56:57 <oerjan> he's finnish. he's sorry by definition.
21:57:52 <oklopol> apparently hyperreals make calculus rather nice and intuitive
21:58:07 -!- rapido has joined.
21:58:29 <cpressey> do they make gabriel's horn make any more sense?
21:58:56 <cpressey> sorry, it's just that you said "nice and intuitive" and "calculus" in the same sentence, and that's the first thing I thought of
21:59:23 <oklopol> oh it's that retarded thing
22:00:39 <Phantom_Hoover> It's the integral from 1 to infinity of 1/x, isn't it?
22:01:12 <cpressey> It's an infinite object with finite volume but infinite surface area, or vice versa, I forget which.
22:01:22 <oklopol> and it's really unintuitive that the infinite line is infinitely long, but that part under it is finite
22:01:59 <oklopol> of course you have to rotate it 2pi to make the "paradox" slightly less ridiculous
22:04:27 <pikhq_> Anyways. Calculus gets some pretty surprising *results*, though many of the concepts actually do make some intuitive sense.
22:04:53 <Phantom_Hoover> Although you can create similarly weird things even easier with fractals.
22:05:53 <rapido> is there a fractal based esoteric language?
22:06:09 <oklopol> (gabriel's horn is not weird)
22:06:10 <cheater-> oklopol: how is it unintuitive?
22:06:16 <rapido> 'living on the edge' which is infinite
22:06:20 <oklopol> cheater-: it's not, i was being sarcastic
22:06:22 <cheater-> oklopol: what about the idea of filling a square with a curve?
22:06:34 <oklopol> cheater-: now that's slightly surprising
22:06:43 <oklopol> at least was for me back in the day
22:07:00 <pikhq_> cheater-: Recent proposal of an alternate circle constant, which is a bit more elegant.
22:07:05 <oklopol> EVERYONE: I'M FINE WITH YOU BEING AMAZED AT HILBERT'S SPACE-FILLING CURVE
22:07:12 <cheater-> well my intuition is that if i take a DIN A4 white page of paper and try to make it completely black with just a ball pen, it will take so much time i will FAIL
22:08:21 <cheater-> http://en.wikipedia.org/wiki/New_Math
22:08:23 -!- pikhq has joined.
22:08:30 <cpressey> oklopol: for your benefit: OH WOW
22:08:31 <cheater-> In the Algebra preface of his book "Precalculus Mathematics in a Nutshell," Professor George F. Simmons wrote that the New Math produced students who had "heard of the commutative law, but did not know the multiplication table."
22:08:37 <Zwaarddijk> Phantom_Hoover: wikipedia agrees with me!
22:08:43 <pikhq> It's just that the ratio of a circle's circumference to its radius makes more sense when we don't care about the diameter.
22:10:14 <rapido> Phantom_Hoover: aaah, a new number system to learn....... how many are there?
22:10:33 <impomatic> http://twitpic.com/3nt8eh/full Hilbert Curve bomber in Core War :-)
22:10:35 <cpressey> < rapido> is there a fractal based esoteric language? <-- I know there were a few that got to the "planning" stage, but I don't know of any complete ones
22:10:43 <oklopol> what the fuck at the "Examples"
22:11:06 -!- pikhq_ has quit (Ping timeout: 240 seconds).
22:11:15 <rapido> i don't like infinite/uncountable stuff - but hey - i'' make an exception
22:11:16 <oerjan> <rapido> is there a fractal based esoteric language? <-- i'm pretty sure there was one but i don't remember the name
22:12:18 <cpressey> I know calamari wanted to make one. And I have a half-idea for one. And you could make an argument that anything based on a tree has self-similarity and whatever
22:12:36 <oklopol> the only known proofs of the undecidability of the reversibility of CA in dimensions greater than 1 are based on the hilbert curve
22:13:53 <cpressey> meaning, if, given an arbitrary CA, if you could decide whether it was reversible or not, you could solve the HP?
22:14:15 <cpressey> oklopol: is that what you mean by "the undecidability of the reversibility of CA"?
22:14:28 <oklopol> the reduction is from the tiling problem
22:14:35 <oklopol> that is, given a set of wang tiles, is there a tiling of the plane
22:14:57 <oklopol> cpressey: that's the problem of deciding, given a CA, whether it is invertible
22:15:12 <oklopol> yes, it means exactly that
22:15:23 <oklopol> i thought HP was hilbert curve :D
22:15:39 <cpressey> NEXT TIME I WILL TYPE THE P MORE CLEARLY
22:15:57 <rapido> i like reversible languages: <shameless plug> enchilada is reversible (modulo hash collisions)
22:17:02 <rapido> wasn't the Halting Problem - essentially Hilbert's Problem?
22:17:06 * oklopol is sad that no one asked for details :'(
22:17:30 <oklopol> but at least one of his great problems turned out to be undecidable
22:18:15 <oklopol> the one about diophantine equations
22:18:39 <oklopol> what the exact problem was, i do not remember, however
22:18:46 <oklopol> presumably whether they have solutions
22:21:23 <rapido> what is this with matrix multiplication?- that the most optimum algorithm for general kxk * kxk multiplication is cannot be algorithmically be constructed? (Ω(k^2logk) and this has to do with group theory?
22:23:22 -!- ais523 has quit (Read error: Connection reset by peer).
22:24:09 <rapido> you cannot find an optimal (generic) matrix multiplication algorithm for an arbitrary size of matrices
22:24:36 <cpressey> that somehow seems unsurprising
22:24:50 <oklopol> oerjan: for time-reversibility you get a fun characterization: the time-reversible CA are exactly the ones that are the composition of two involutions
22:24:59 <oklopol> which is of course rather clear from the definition
22:25:00 <rapido> cpressey: i find it rather surprising
22:26:38 <oklopol> i can't really comment on rapido's statement without knowing more details, with my definition of the concepts, that is trivially true: there is always a faster algorithm
22:27:10 <calamari> cpressey: yeah.. I never go a good idea on how to make a fractal language.. I wanted it all to fit together perfectly, rther than feeling tacked on
22:27:21 <cpressey> oklopol: are you referring to the speedup theorem? yeah, but i never understood that one, either
22:28:18 <cpressey> yeah, it's not a simple thing to design well... i certainly haven't gotten very far with my idea, and it's not even a very ambitious interpretation
22:28:33 <rapido> calamari: what should be the ingredients of a good fractal language?
22:28:34 <oklopol> cpressey: the speedup theorem is a rather trivial consequence of the fact we can add arbitrarily many states to our turing machines, and that we can have arbitrarily big alphabets
22:28:52 <calamari> I also felt it needed to provide a bit of data creation.. what I mean is like mandelbrot where it's self-similar but not exactly
22:29:05 <oklopol> just compress the given word to say 1/1000000 of its original size, and do computation 1000000 times faster
22:29:15 <oklopol> because you can fit 1000000 symbols in one cell
22:29:20 <oklopol> by having a bigger alphabet
22:29:36 <rapido> would we recognize pi if it were in base 31?
22:29:43 <cpressey> oh, is that all. i got the impression it was much more profound than that.
22:30:07 <oklopol> i imagine your cs lecturer has succesfully hidden the fact the speedup theorem is fucking obvious by adding tons of details before mentioning the general idea.
22:30:37 <rapido> calamari: because 31 is insignificant :)
22:31:05 <oklopol> cpressey: there are very profound things that have to do with speeding things up, but this is certainly not one of them
22:31:12 <calamari> did they ever figure out how to convert that arbitrary digit algorithm to base 10?
22:31:27 <calamari> or did they decide it wasn't possible?
22:32:30 <oklopol> there's also a speed-up theorem that says something like, if you have a computable function f, then there exists a recursive language L such that for any algorithm you make for it, there exists another algorithm that is f(n) times faster on an input of size n
22:32:45 <oklopol> or something crazy like that
22:33:14 <rapido> hey! - what about memoization of numbers and functions (that produce numbers), irrespective of base?
22:33:28 <oklopol> "<rapido> would we recognize pi if it were in base 31?" <<< no
22:33:56 <oklopol> the bases up to 100424 have been tried. but maybe 100425 does it, you could be a pioneer.
22:34:35 <rapido> what about a complex base?
22:34:53 <oklopol> we actually just had a lecture series about the expansions of constants
22:35:17 <oklopol> i don't think we know anything about pi
22:35:26 <oerjan> calamari: http://en.wikipedia.org/wiki/Bailey%E2%80%93Borwein%E2%80%93Plouffe_formula#The_search_for_new_equalities is a little confusing
22:35:29 <oklopol> but e has a rather simple continued fraction
22:36:21 <oklopol> sorry i left that kinda open
22:36:39 <oklopol> well whatever, i don't recall the details
22:37:01 <oerjan> ...but pi doesn't have a rather simple simple continued fraction
22:37:17 <rapido> the best compression of pi is to recognize it in disguise.
22:37:17 <calamari> it converges horribly.. but it's simple :)
22:37:48 <oerjan> calamari: note, two "simple" there
22:38:00 <oklopol> it's simple to understand, but it's not simple to work with, at least it doesn't help for the that particular lecturer was using the digits
22:38:05 <oklopol> but let me look up the sequence
22:38:11 <oklopol> maybe i can point out the complication
22:38:19 <calamari> apparently simple is a specific math term
22:38:21 <rapido> i believe you can calculate the n'th digit of pi without calculating it's previous digits?
22:38:32 <calamari> generalized continued fraction
22:38:55 <oerjan> rapido: in binary, what we're talking about is whether you can do it in base 10 or so
22:39:17 <oerjan> that section i linked may or may not imply that base 3 is known...
22:40:13 <cpressey> Deewiant: You'll be happy* to know that FBBI makes it to the end of Mycology, now, with "only" 9 BADs. (*Feel free to substitute the emotion of your choice here.)
22:40:45 <oklopol> i'm finding a very weird continued fraction for pi
22:40:49 <oklopol> i thought it was rather simple
22:41:00 <oklopol> "<oerjan> ...but pi doesn't have a rather simple simple continued fraction"
22:41:48 <rapido> i have an idea for a fractal language
22:42:02 <rapido> when you zoom in to the border
22:42:37 <rapido> the 'result' of your specified program will be more and more accurate
22:43:02 <rapido> but you never reach the 'ultimate' solution of your algorithm
22:43:07 <calamari> rapido: just be sure that it is capable of writing a fractal encryption code unbreakable by the borg.. otherwise the future of humanity is put at risk
22:43:12 -!- Phantom_Hoover has quit (Quit: Leaving).
22:43:13 <cpressey> a rather simple simple simple continued fractal
22:43:33 -!- Sgeo has joined.
22:43:34 <oklopol> the littlewood conjecture is true for all pairs (e, x), since the continued fraction representation of e contains arbitrarily large numbers
22:43:54 <rapido> btw: the algorithm specification is also nearing the optimal algorithm - but not quite - fractal wise
22:44:53 <rapido> hmmm - let me iterate on that
22:45:08 <cpressey> f(s,n)=s/n+f(-s,n+2) is not a continued fraction, but it is rather simple
22:45:51 <rapido> i know continued fractions (vaguely from uni)
22:46:07 <cpressey> i don't like them. they wet their nests
22:46:08 <oklopol> i know very little about them
22:46:14 <Sgeo> Why hasn't elliott been here in a while?
22:46:26 <oerjan> cpressey: um do you have your precedence right there
22:46:31 <oerjan> Sgeo: he left in a huff
22:47:03 <oerjan> f(s,n)=(s/n)+f(-s,n+2) ?
22:47:21 <cpressey> um... does / not come before +?
22:47:40 <oerjan> not in most PLs i am familiar with
22:47:44 <rapido> are we being concatenative?
22:48:13 <oerjan> it's usually the same precedence as *
22:48:14 <lambdabot> help <command>. Ask for help for <command>. Try 'list' for all commands
22:48:18 <lambdabot> http://code.haskell.org/lambdabot/COMMANDS
22:48:49 <oerjan> most definitely not in haskell
22:50:00 <oerjan> cpressey: um i thought you were disagreeing with my adding parentheses above?
22:50:20 <oerjan> "comes before" is not a term i can understand clearly
22:50:41 <oerjan> cpressey: because you were talking about continued fractions and they use a/(b+c)
22:50:54 <cpressey> i also said it was not a continued fraction
22:51:07 <rapido> a precedence and parentheses - you don't have that 'problem' with 'concatenate' languages (or apl/k/j)
22:51:27 <cpressey> CONFUSION REIGNS. ALL HAIL THE OMNIGOAT
22:52:22 <cpressey> of course this means we need an esolang where the precedence varies in a difficult-to-predict way
22:52:27 <cpressey> possibly depending on the values involved
22:52:33 <oerjan> cpressey: also in haskell "is evaluated before" is rather distinct from precedence >:)
22:53:25 <Gregor> Uhh, evaluation order is distinct from precedence in many, perhaps most languages.
22:53:27 <rapido> also in haskell 'is evaluated whenever' is rather not related to precedence ;)
22:53:53 <cpressey> it's kind of hard to evaluate what x+3 is without first evaluating x, though
22:53:56 <oerjan> <cpressey> of course this means we need an esolang where the precedence varies in a difficult-to-predict way <-- i think oklopol had something like that
22:53:56 <oklopol> for a while now i've been trying to make a language where every little change in the program changes EVERYTHING, yet somehow you can program in the language
22:54:08 <oklopol> no locality. what i tried to do with toi, but failed
22:54:22 <Gregor> cpressey: But it's easy to evaluate x+y*z in the order x->y->z instead of y->z->x
22:54:24 <rapido> cpressey: i like x to be free - let us be free
22:54:39 <oklopol> in cise, there's no precedence, the first correct parse is used
22:54:56 <rapido> oklopol: you have - reference - to ..... toi ?
22:55:17 <oklopol> rapido: it should be on the wiki
22:55:24 <oerjan> no, oklopol is fundamentally opposed to revealing his languages in anything other than irc babbling
22:55:30 <oklopol> an oklopol quality article, i'm sure
22:55:41 <oklopol> erm, i'm pretty sure toi has an article?
22:56:03 <oerjan> oklopol: DON'T LET YOUR PUNY FACTS GET IN THE WAY OF MY STEREOTYPING
22:56:05 <oklopol> i'm totally a wiki-using grown up nowadays
22:56:18 <oklopol> sorry didn't realize you were a total noob lol :PPPPp
22:56:27 -!- pikhq has quit (Ping timeout: 255 seconds).
22:56:33 -!- pikhq_ has joined.
22:56:56 <rapido> Toi (the name means nothing
22:57:42 <cpressey> Don't believe it. It's short for T(ower of Han)oi
22:57:48 <rapido> Brackets have to match, otherwise everything's a legal program ?????
22:57:50 <cpressey> At least, that's what I always think of, when I read it
22:58:26 <oklopol> i'm not sure that's entirely true .DS
22:59:11 -!- sebbu2 has changed nick to sebbu.
23:00:18 <rapido> '(A{B}' is a type of for-each. For each s in S the Toi program A is run with s as the context.
23:00:42 <rapido> the context concept is a bit underdeveloped (documentation wise)
23:01:06 <oklopol> each run of a toi program has a context, that just means you recursively call the interpreter with another context
23:01:25 <oklopol> i thought that was clear but yeah i guess it's a weird way to explain things
23:01:50 <oklopol> maybe "the toi program A is run with s as the initial context"
23:02:11 <cpressey> what, all that drama over a bot?
23:03:55 <cpressey> and who is Herobrine? wait, I forgot -- I don't really care
23:04:14 -!- cpressey has quit (Quit: leaving).
23:05:13 <rapido> cpressey is in CET timezone? : +00 hours here in the Netherlands - morning folks - later
23:05:40 -!- rapido has quit (Quit: rapido).
23:10:23 <cheater-> oklolol: what do you find ultrafilters are most useful for?
23:11:34 <oklopol> cheater-: what's an ultrafilter?
23:12:11 <cheater-> <oklopol> cheater-: yeah i know the usual definition is based on ultrafilters
23:12:59 <oklopol> erm, just before that you said they are based on sequences
23:13:15 <oklopol> granted, i'm sure you know what sequences are
23:13:20 <oklopol> but i don't think that's relevant
23:13:36 <oklopol> that was the point i was making
23:13:39 <cheater-> you think sequences are irrelevant to hyperreals?
23:13:55 <oklopol> no, ultrafilters are some kind of sequences, i assumed you meant them but didn't know the big word
23:14:13 <oerjan> ultrafilters are not sequences
23:14:20 <oklopol> whereas i talked to this prof once who does nonstandard analysis and i managed to copypaste a pretty cool word.
23:14:37 <oklopol> oerjan: oh? what were they then?
23:14:45 <oerjan> not even nets, although they are related a bit
23:14:45 <oklopol> i have this really vague recall
23:15:02 <oklopol> actually it's so vague i can't really put it into words
23:15:10 <oklopol> something about the natural numbers and somthing
23:15:11 <cheater-> We are going to construct a hyperreal field via sequences of reals. In fact we can add and multiply sequences componentwise; for example:
23:15:25 <oklopol> ultrapower construction :D
23:15:36 <cheater-> http://en.wikipedia.org/wiki/Hyperreal_number < read here
23:15:56 <cheater-> i think landen or zeno linked me to an old article about that somewhere on the net
23:15:59 <oerjan> an ultrafilter in X is a family of subsets of X such that (1) for every subset of X, either it or its complement is a member (2) a superset of a member is a member
23:16:10 <cheater-> oklopol: are you like being aggressive or something
23:16:25 <oerjan> equivalently, it's a maximal filter
23:16:26 <oklopol> cheater-: no i don't know what i'm doing
23:16:36 <oklopol> really you should not listen to anything i say unless you're in a similar state of mind
23:17:11 <oerjan> hm (3) the intersection of two members is a member. i think.
23:17:24 <cheater-> i was rather thinking of you getting some weed to help with the jitteriness
23:17:39 <cheater-> but i guess you can get the medical kind too
23:17:41 <oerjan> or maybe that's redundant
23:17:59 <oklopol> oerjan: okay how do you use them? also that's a pretty cool set
23:18:10 <oklopol> one damn cool set to say the least
23:18:38 <oerjan> oklopol: a non-principal ultrafilter is one that isn't just the supersets of a fixed point.
23:19:31 <oerjan> requires the axiom of choice to find
23:19:57 <oklopol> principal = cyclic = generated by one thing, usually, here it means we just look at one element to know whether the set is in the ultrafilter
23:20:08 <oklopol> principal ultrafilter is that
23:20:25 <oklopol> then it is obvious that the S or S^c is in the ultrafilter
23:21:01 <cheater-> aren't the consequent iterations of the cantor set an ultrafilter
23:21:47 <oklopol> i doubt it, even though i don't understand it
23:22:07 <oklopol> iterations implies countable, and cantor set implies uncountable space
23:22:09 <oerjan> ...what's a consequent iteration
23:22:17 <cheater-> oerjan: the first iteration, second iteration, etc
23:22:37 <cheater-> those put together as a family of sets would make an ultrafilter, no?
23:23:02 <oerjan> you mean when you take away the middle, then the middle of each interval of that, etc?
23:23:05 <oklopol> and ultrafilters are rather fucking big in uncountable spaces so that obviously cannot happen
23:23:39 <oklopol> i hope that's not what he meant
23:23:41 <oerjan> in which case, no, that's not an ultrafilter, not even a filter although you can make one by adding supersets
23:24:07 <oklopol> filter = closed under supersets and closed under finite meets or what?
23:24:17 <oklopol> kinda mixing up terminology there
23:24:32 <cheater-> The notion of I -ultrafilter was introduced in Baumgartner [1995]: Let I be a family of
23:24:32 <cheater-> subsets of a set X such that I contains all singletons and is closed under subsets. Given a free
23:24:32 <cheater-> ultrafilter U on ω, we say that U is an I -ultrafilter if for any F : ω → X there is A ∈ U such
23:24:32 <cheater-> Baumgartner defined in his article discrete ultrafilters, scattered ultrafilters, measure zero
23:24:35 <cheater-> ultrafilters and nowhere dense ultrafilters which he obtained by taking X = 2ω , the Cantor set,
23:24:36 <cheater-> and I the collection of discrete sets, scattered sets, sets with closure of measure zero, nowhere
23:25:06 <cheater-> seems he's constructing ultrafilters from the cantor set somehow
23:26:00 <oerjan> um the cantor set is the whole space, i take
23:26:02 <cheater-> but it's too late for me to understand the details :D
23:27:12 <cheater-> http://docs.google.com/viewer?a=v&q=cache:K61LyUFCLG0J:home.zcu.cz/~flaskova/research/MArevised.pdf+cantor+set+ultrafilter&hl=en&pid=bl&srcid=ADGEESjixG3WoyeYkUp79ujAGyHYhuM0avEL0uWdPbKyIkndXxySqzxUjkOB1S7WvUBXj-LDrrgwZD8PsqdycbHBuBbqP6YjjvWYvnYwAZQLD5DWfm7Xle-Ha_OcV-qTFCVP4DHI80SN&sig=AHIEtbQZCLkVGCtAe7rB005CWzF7DPbnlQ
23:27:43 <cheater-> tbh i would love to know too oklopol
23:28:01 <cheater-> i fear this secret might end up never being revealed
23:28:21 <oklopol> unless oerjan bothers to tell us lowly noobs what it means
23:28:42 <oerjan> in any case i think non-principal ultrafilters definitely require the axiom of choice
23:29:34 <oerjan> (i think "every filter has a larger ultrafilter" is probably equivalent to AoC)
23:30:18 <oklopol> ultrafiltersssssssssssssssssssssssssssss
23:38:03 <oklopol> btw is it a xor, like exactly one of S and S^c is in the ultrafilter?
23:38:24 <oklopol> because i totally just realized that's kinda important
23:41:58 <oerjan> yes. otherwise the empty set would be a member and that's disallowed for filters
23:46:27 <oklopol> but umm isn't an ultrafilter just a filter you can't add any elements to
23:46:50 <oklopol> so it's enough to show an increasing union of filters is a filter
23:46:57 <oklopol> i suppose that's obvious, but lemme check
23:47:02 -!- poiuy_qwert has quit (Quit: This computer has gone to sleep).
23:47:09 <oerjan> yeah it's easy with zorn's lemma
23:47:36 <oklopol> it's obvious that union of filters is a filter
23:47:50 <oklopol> and yeah zorn's lemma is what i need the increasing unions for
23:48:22 <oklopol> well i said so a few lines above and i was too busy to write that complicated word a second time
23:48:51 <oklopol> yeah i'm not sure it's at all obvious that the union of filters in general is a filter
23:49:03 <oklopol> partly because that's not true
23:49:34 <oklopol> "<oerjan> equivalently, it's a maximal filter" <<< whoops
23:49:55 <oklopol> it's not true, i'm pretty sure
23:50:08 <oerjan> ...just take two different principal ones
23:50:21 <oklopol> yeah, that's why i'm pretty sure
23:50:42 <oklopol> because i had just proven it by taking two different principal filters
23:51:29 <oklopol> i'm not nearly as stupid as i look
23:53:39 <oklopol> i wish i had one of those practise tennis balls
23:53:59 <oklopol> a ball that you can throw against the wall without a sound being made our of this event
23:55:25 <oklopol> then i could throw it against the wall from morning till dawn