←2008-10-18 2008-10-19 2008-10-20→ ↑2008 ↑all
00:10:03 -!- KingOfKarlsruhe has quit (Remote closed the connection).
00:16:06 -!- Slereah_ has joined.
00:54:33 -!- Slereah has joined.
00:54:38 -!- Slereah_ has quit (Read error: 104 (Connection reset by peer)).
01:29:58 -!- ryanakca has joined.
01:30:25 <ryanakca> I suppose the sysadmins know that the wiki is kaput?
01:47:13 -!- Slereah has quit (Read error: 113 (No route to host)).
03:07:08 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | ^x x.
03:36:07 -!- pikhq_ has changed nick to pikhq.
05:20:25 <psygnisfive> http://www.peta.org/Sea_Kittens/index.asp
05:20:28 <psygnisfive> save the sea kittens
05:20:29 <psygnisfive> !!!
05:28:34 -!- Sgeo has quit (Read error: 110 (Connection timed out)).
05:38:59 <GregorR> Wow, that's retarded.
05:39:07 <GregorR> That's so retarded it makes me want to go eat fish, and I don't even like fish.
05:40:03 <omniscient_idiot> Yeah, fish is nasty
06:32:58 <pikhq> So retarded that it reminds me of people getting turned into monkeeys.
06:34:16 <pikhq> (Cowboy Bebop episode)
06:43:26 -!- hando has joined.
06:43:35 <hando> hello
06:43:43 <hando> what's wrong with esolangs wiki?
07:11:36 <GregorR> Lurve how people only come to this channel if the wiki is borked :P
07:12:01 <GregorR> hando: Well, the problem is that it's borkleborked, and we don't have our deborklarizing ray warmed up yet (it runs on tubes y'know)
07:12:51 <Asztal> real17m42.983s
07:12:52 <Asztal> user17m18.653s
07:12:54 <Asztal> sys0m1.716s
07:13:06 <Asztal> I think I need to tweak this slightly...
07:13:24 <GregorR> Depends on what your'e doing with those 17 minutes.
07:13:28 <Asztal> I think it's O(n^3 * m^3) now
07:13:36 <Asztal> not a bloody lot :(
07:13:46 <GregorR> That's /probably/ a bad running time :P
07:15:32 <Asztal> it takes every distinct triple from a list of ~1000 strings and finds the longest (longest common substring of the triple), which is really... probably not the right way to go about it
07:15:59 <Asztal> so that's 1e9/6 triples :D
07:16:48 <Asztal> but that 17 minutes figure was with only 140 strings (they're IRC messages, so, not long strings) and it hadn't even finished yet
07:24:00 <Asztal> it kind of works, though... it generates " retarded" given the most recent messages :)
07:29:18 <hando> GregorR: long time no see
07:49:26 <hando> So.. graue runs the wiki, but he's not there, so the wiki keeps being broken?
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:48:30 -!- oerjan has joined.
08:49:22 * oerjan wonders why spammers send him email in german
08:49:39 <oerjan> i mean, if it was random - but i cannot recall any french, say
08:56:12 <oerjan> hando: that looks broken indeed. and graue is very rarely here.
08:57:06 <pikhq> I wasn't even aware that there was a 'graue' person...
08:57:14 <pikhq> And I've been here regularly for 2 years.
08:57:26 <oerjan> oh i've seen him
08:58:02 <oerjan> he's pretty rarely on the wiki too, when i think about it. but for minor problems there are several admins
08:58:16 <oerjan> but i guess only graue can handle a real crash
08:59:06 * oerjan doesn't know as he is not a wiki admin
08:59:50 <oerjan> ais523 is one, but he is away, which means not logged in as he and ehird share an irc bouncer
09:00:55 <oerjan> also, sunday morning is about _the_ slowest time on this channel
09:03:41 <oklopol> oooooooooooooooooooooooooooooooooooooooooooooooooooo
09:03:43 <oklopol> HIYA
09:03:59 <oerjan> good moaning
09:04:00 * oklopol is memorizing processor designs
09:05:05 <oerjan> some day in the far future your memory will be _full_, and you will start regretting it. just before you descend into senility.
09:06:54 <oklopol> doesn't seem too far-fetched, i'm already starting to forget english vocabulary at an alarming rate.
09:07:03 <oerjan> eek
09:07:08 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | err.
09:07:21 <oerjan> even optbot is shocked
09:07:22 <optbot> oerjan: I was going to tell him something!
09:07:35 <oerjan> optbot: well speak out
09:07:36 <optbot> oerjan: near*
09:07:54 <oklopol> for instance, took me about a minute to find the word far-fetched, a word i've known since i was -1
09:08:47 <oklopol> but, usually when i take a break from reading, the words come back, and all the new data stays too
09:09:04 <oklopol> okokokokokokokokokokokokokokokokokokokokokoko
09:09:08 <oerjan> well don't mind me, i cannot even think and chew at the same time
09:09:17 <oerjan> as in, i just bit myself in the mouth
09:10:26 <oklopol> also processor designs aren't that hard to memorize, i know the mips instruction set by heart, and i know exactly how it's translated to opcodes, so all i have to do it traverse the set of operations, and check what lines and registers i need
09:10:26 <oerjan> also, okokokokokokokokokokokokokokokokokokokokoko
09:11:11 <oklopol> the hard part is i have a hard time letting myself draw the lines in slightly different shapes than the ones in the book
09:11:19 * oerjan suddenly wonders if there's a wickedpedia
09:11:37 <oklopol> functionally equivalent, but the exam checker will see i don't have a photographic memory
09:11:42 <oklopol> and he will laugh
09:11:50 <oklopol> and probably spit on my paper
09:11:52 <oklopol> in disgust
09:12:02 <oerjan> apparently there is
09:12:08 <oklopol> :P
09:12:56 <oerjan> it's about disney villains?
09:14:01 <oerjan> (although there is also a wickedpedia.no, which is about erotica)
09:15:32 <oerjan> and people apparently use it as a slur against wikipedia itself. oh well.
09:17:01 <oerjan> there probably is a secret wickedpedia used just by exam checkers, you know
09:20:54 <oklopol> are they wicked?
09:27:50 <fizzie> Wicked morning. I mean, good.
09:29:49 <oerjan> oh noes, fizzie is an exam checker!
09:53:41 -!- oerjan has quit ("Sniffle").
09:59:06 <AnMaster> hm
09:59:49 <AnMaster> morning
10:03:31 -!- Mony has joined.
10:04:31 <Mony> plop
10:14:49 <oklopol> o
10:20:59 <fizzie> o
10:24:52 -!- kar8nga has joined.
10:28:08 -!- oerjan has joined.
11:01:05 -!- oklopol has quit (Read error: 104 (Connection reset by peer)).
11:25:21 <AnMaster> fizzie, there still?
11:26:58 <AnMaster> Would saying something like "If a ATHR thread causes an internal crash in the interpreter, it shall if possible not crash other threads." be a good idea?
11:26:59 <fizzie> Partially. Trying to eat some breakfast.
11:27:35 <AnMaster> I can semi-guarantee that in my implementation. Only thing would be if I hit some internal erlang bug or cause a cascading crash.
11:28:15 <AnMaster> but for example if I had a bug in some fingerprint crashing division by zero it would get logged but not affect other threads.
11:28:28 <AnMaster> Or would that be hard to implement in most other languages?
11:28:37 <ais523> hi ehird
11:28:42 <AnMaster> hi ais523
11:28:58 <ais523> hi AnMaster
11:29:01 <ais523> can't be here long
11:29:10 <ais523> this Internet connection is really dodgy, too
11:29:19 <ais523> a while ago it was only giving me one second of connectivity before breaking
11:29:24 <AnMaster> ais523, http://rafb.net/p/AQQhs539.html
11:29:32 <ais523> also, both my main email and backup email are down at the moment
11:29:35 <ais523> which is annoying
11:29:38 <AnMaster> ais523, but I assume you will be back later today?
11:30:05 <ais523> probably not
11:30:07 <ais523> it's Sunday
11:30:13 <ais523> I will probably be back tomorrow though
11:30:13 <AnMaster> ais523, well see that link now then
11:30:22 <ais523> I have skimmed it
11:30:29 <ais523> very quickly
11:30:35 <ais523> what in particular should I be looking at?
11:30:37 <AnMaster> ais523, and see scrollback for the last few lines before you said hi
11:30:49 <AnMaster> ais523, that ATHR implementation is making progress :)
11:31:55 <AnMaster> ais523, poke?
11:31:55 <fizzie> It's not hard to implement if you only say "if possible".
11:32:06 <ais523> fizzie: things can be possible and still difficult
11:32:09 <oerjan> ais523: anything you can do about the wiki or are we just going to have to wait for graue?
11:32:18 <ais523> oerjan: what in particular needs doing?
11:32:18 <fizzie> ais523: Well, it's also a SHOULD, not a MUST.
11:32:22 <ais523> I can do a lot but not everything
11:32:30 <oerjan> it's crashed, basically
11:32:37 <ais523> I can't fix that...
11:32:41 <AnMaster> fizzie, ah. Hm also I need a function to give back thread id, may not be useful in ATHR on it's own, but could potentially in future fingerprints extending ATHR
11:32:56 <ais523> generally speaking Graue fixes it within a day if you email them, though
11:33:35 <ais523> hmm... same error as last time
11:33:35 <AnMaster> "from within function "MediaWikiBagOStuff::_doquery". MySQL returned error "1194: Table 'mw_objectcache' is marked as crashed and should be repaired"
11:33:36 <AnMaster> ouch
11:33:39 <AnMaster> really ouch
11:33:40 <oerjan> i don't have his (their?) email
11:33:52 <ais523> I know the fix, but it requires MySQL root access on the server the wiki's running on
11:33:53 <ais523> so I can't do it
11:34:07 <ais523> oerjan: well, with a name like Catatonic Porpoise, the gender isn't obvious
11:34:12 <AnMaster> ais523, which is why you use postgre instead of mysql
11:34:35 <oerjan> you can say that. i had forgotten that nickname.
11:34:37 <AnMaster> ais523, since postgre seems to be less messy on crash in my experience
11:36:00 <ais523> graue@oceanbase.org, anyway
11:39:59 <ais523> can anyone hear me?
11:40:15 * oerjan is hearing (and mailing)
11:40:24 <AnMaster> hm
11:40:24 -!- oklocod has joined.
11:40:28 <AnMaster> ais523, no I can't
11:41:26 <AnMaster> fizzie, ais523: Should F (flush signal queue) return any messages that were flushed?
11:41:53 <AnMaster> like count at the top of the stack, followed by that number of signals that were in the queue
11:42:00 <AnMaster> or should it just discard them all?
11:45:26 <AnMaster> I think it should return them
11:47:30 <AnMaster> ais523, fizzie: http://rafb.net/p/PB3f2o65.html
12:00:19 -!- Corun_ has joined.
12:01:23 <AnMaster> ais523, fizzie: I think it may be quite possible to do hot code change of parts of efunge without stopping :D
12:01:29 <AnMaster> not that anyone would ever want it
12:01:37 <AnMaster> in fact I did that on the input server just now
12:01:50 <fizzie> I don't have a very specific opinion on F returning the signals; the name "flush" doesn't really sound like it'd return them. But since you can just use "k$" to get rid of them values, it doesn't really hurt.
12:02:16 <AnMaster> indeed, and it can be useful to get them
12:02:24 <ais523> hmm... figuring out how to hot-change parts of a Feather program while it's running is a major problem
12:02:52 <AnMaster> ais523, oh? Well I could potentially hot-change _any_ part of efunge.
12:03:00 <oerjan> AnMaster: well erlang is known for being able to do that...
12:03:11 <AnMaster> actually the ets tables may cause a tiny amount of issues there
12:03:26 <AnMaster> oerjan, true, but you need to be careful with your design for it to work in erlang too
12:03:50 <AnMaster> following the otp design principles
12:04:01 <AnMaster> (sp?)
12:04:15 -!- oklopol has joined.
12:04:50 <oerjan> there is something called Erlang/OTP, so must be right
12:05:54 <AnMaster> oerjan, no I mean "principles"
12:05:59 <AnMaster> if that was the correct spelling
12:06:03 <oerjan> sounds right too
12:06:05 <AnMaster> I know how to spell OTP of course
12:06:46 <oerjan> i'm sure i've seen it spelled wrong (principals)
12:06:53 <AnMaster> haha
12:08:27 -!- Slereah_ has joined.
12:08:42 <AnMaster> oerjan, http://www.erlang.org/doc/design_principles/part_frame.html
12:09:15 <AnMaster> oh and OTP is probably the most misnamed technology ever.
12:09:40 <AnMaster> it stands for Open Telecom Platform, except these days erlang and OTP is used for lots of stuff not related to telecom...
12:10:45 <AnMaster> actually SASL may be worse since it is confusing with another technology...
12:10:53 <AnMaster> in erlang context it means "System Architecture Support Libraries"
12:11:06 <AnMaster> elsewhere it tends to mean "Simple Authentication and Security Layer"
12:11:22 <AnMaster> oh and OTP usually means "one time password" outside erlang context
12:13:00 <oerjan> yeah noticed that
12:13:25 <AnMaster> so erlang had quite bad luck with acronyms
12:14:14 <oerjan> we need to find an acronym for this concept
12:14:25 <AnMaster> oerjan, what about the erlang table viewer tool? tv
12:14:45 <AnMaster> (a debugging tool really)
12:15:12 <oerjan> O_O you'd think they're doing it on purpose :D
12:15:20 <AnMaster> there is also the event tracer et
12:15:22 <AnMaster> :D
12:16:07 <AnMaster> oerjan, a database called mnesia (the story goes some boss said they couldn't call a database "amnesia". "You can't have a database that forgets things!" so they just dropped the a).
12:16:30 <oerjan> well that's clever enough
12:16:36 <AnMaster> yeah
12:16:58 <oerjan> except i'm dubious on there being an a in the first place
12:17:29 <AnMaster> oerjan, well Joe Armstrong (one of the original designers of Erlang) claims that story in one of his books about erlang
12:18:12 * AnMaster got it as an ebook
12:18:40 <AnMaster> oerjan, http://rafb.net/p/u5s4kE75.html (copy and paste from said ebook)
12:19:19 <oerjan> oh well
12:20:36 <AnMaster> anyway sasl and otp are poorly selected acronyms indeed. Especially since they mean something else within computer context. (while tv doesn't).
12:20:56 -!- oklocod has quit (Read error: 110 (Connection timed out)).
12:21:07 <AnMaster> But sasl and opt are both good and will be used in efunge (otp supervisor tree is already used)
12:21:33 <AnMaster> otp*
12:21:37 <AnMaster> (not opt)
12:23:42 <oerjan> quite common misspelling that, right optbot?
12:23:42 <optbot> oerjan: that isn't an earth shattering problem is it? you can still implement continuations easily with coroutines...
12:25:18 <AnMaster> heh
12:45:41 -!- Corun_ has quit ("This computer has gone to sleep").
12:46:23 -!- Corun has joined.
12:59:56 <ehird> hi ais523
13:00:02 <ehird> oh
13:00:03 <ehird> he's gone
13:00:13 <oerjan> neener neener
13:02:09 <oklopol> neeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeer
13:04:13 <oerjan> those finnish long vowels are really scary
13:07:01 <Slereah_> Nner
13:09:51 <oerjan> that's not finnish, i think it may be estonian
13:11:01 <oerjan> oh wait, albanian
13:39:45 <oklopol> albanian has long ë's?
13:41:05 * oerjan checks
13:44:30 <oerjan> ah, only the Gheg dialect has long vowels
13:44:55 <AnMaster> heh?
13:45:33 <ehird> what's heh about that
13:45:43 <oerjan> albanian has two main dialects, Gheg and Tosk. Tosk is the basis of the standard language
13:47:58 <AnMaster> HUH
13:48:09 <AnMaster> this is crazy :D
13:48:26 <AnMaster> the erlang table viewer got an option to make it output error messages in Haiku.
13:51:07 <ehird> Question:
13:51:24 <ehird> How should I differentiate "wrapped text" and "linebreaked text"?
13:51:26 <ehird> This is a poem,
13:51:27 <ehird> indeed it is,
13:51:30 <ehird> and it should use linebreaks.
13:51:31 <ehird> --
13:51:34 <AnMaster> "The selected table is unreadable! Only table information may be viewed!" (normal) vs. "Table protected.\nThe answer that you're seeking\nwill remain unknown."
13:51:35 <AnMaster> hehe
13:51:37 <ehird> But this is just some long, wrapped text, and
13:51:39 -!- Corun has quit ("This computer has gone to sleep").
13:51:43 <ehird> it should not use linebreaks because it is not
13:51:44 <ehird> a poem.
13:52:26 <AnMaster> ehird, well what about mixing CR and LF, like LF for wrapping and CR for line break?
13:52:29 <AnMaster> or something like that
13:52:39 <AnMaster> that is the only pure text way I can think of
13:52:48 <ehird> AnMaster: I mean, without modifying the original text. :-)
13:52:56 <ehird> A heuristic is fine, of course.
13:53:03 <AnMaster> ehird, ah you mean to *detect*?
13:53:08 <ehird> *nod*
13:53:09 <AnMaster> well that is trickier...
13:53:33 <ehird> I'm pondering writing a little program that turns EmailML (as in, the conventions people use when emailing) into HTML.
13:53:39 <ehird> Since it comes naturally to me already. :-P
13:53:41 <AnMaster> and I can't think of any simple way, if it seems the text is always cut off near column 70-80 then maybe?
13:53:49 <ehird> AnMaster: Yeah, probably.
13:53:56 <ehird> I don't think there are many poems with lines this very long,
13:54:02 <ehird> Because they would not be very fun to read, indeed, they would not,
13:54:11 <ehird> Yep, this is certainly not a poem, indeed, yepidoodle, blah blah blah.
13:54:15 <AnMaster> ehird, if you restrict yourself to some specific type of poem and a specific language you could analyse word structure and so on
13:54:27 <AnMaster> should be possible to detect for example Haiku that way
13:54:32 <AnMaster> and rhymes
13:54:36 <ehird> Well, yeah, that's a little overblown though. :-P
13:54:42 <AnMaster> but a generic solution I can't think of
13:54:59 <ehird> I'm basing stuff on http://inamidst.com/topic/avocet
13:55:02 <oerjan> shakespearian meters have pretty long lines don't they?
13:55:05 <ehird> because you do have to cave in somehow -
13:55:14 <ehird> For example, @links that span many words (http://google.com/)
13:55:17 <AnMaster> oerjan, was just about to mention that and other forms
13:55:23 <ehird> I think i've chosen @ for the marker of the start of those.
13:55:26 <AnMaster> for example Greek mythology
13:55:27 <ehird> It seems pretty natural.
13:55:29 <AnMaster> what's the english name
13:55:35 <AnMaster> hexameter in Swedish
13:56:24 <oerjan> i think it's the same, i seem to recall iambic pentameter for shakespeare
13:59:40 <Deewiant> aye, and occasionally trochaic tetrametre
13:59:59 <Deewiant> with the faerie in Midsummer Night's Dream, IIRC
14:03:41 * ehird concludes that lines under 60 chars will be linebreaked
14:06:06 <AnMaster> ehird, which could be due to it being a header, or because there is a new section
14:06:12 <AnMaster> say a new paragraph
14:06:44 <AnMaster> so last line of a paragraph could just be a few words or whatever
14:06:57 <ehird> well, yes.
14:06:59 <ehird> but
14:07:00 <ehird> I mean
14:07:08 <ehird> the last line won't be breaked
14:07:09 <AnMaster> ehird, also a line could be shorter to avoid breaking a long url on the next line into several parts
14:07:10 <AnMaster> or such
14:07:11 <ehird> because there is no following line anyway
14:07:27 <AnMaster> or a long inline math expression or whatever
14:07:28 <ehird> AnMaster: well, you can't do this:
14:07:31 <ehird> This is an awesome poem yeah yeah:
14:07:32 <ehird> foo bar
14:07:33 <ehird> quux
14:07:34 <ehird> dead
14:07:38 <ehird> Yeah I love it man it is amazing I love it yeah man
14:07:40 <ehird> it is great man yeah.
14:07:48 <ehird> "is-linebreaking" is a property of -paragraphs-
14:07:55 <ehird> so only the first line has to be over 60 chars
14:07:55 <AnMaster> ehird, hm
14:08:23 <AnMaster> what if it is shorter due to a long url didn't fit onto the same line?
14:08:41 <ehird> in the case that you need to break before 60 chars to stop a URL being huge, well, just let it go over 80 columns 'cause when would that ever happen
14:08:44 <ehird> AnMaster: as i said - how many paragraphs have a URL before the first 60 chars?
14:08:46 <ehird> on the first line
14:08:48 <ehird> not just any URL
14:08:50 <ehird> a _long_ URL
14:08:53 <ehird> i'd say... not many.
14:09:04 <ehird> in those rare cases, just let the line go over 80 chars
14:09:07 <AnMaster> Please see the documentation at
14:09:07 <AnMaster> http://www.erlang.org/doc/design_principles/spec_proc.html#6.2 next time blah
14:09:07 <AnMaster> blah blah
14:09:09 <AnMaster> for example
14:09:13 <AnMaster> that is clearly wrapped
14:09:22 <ehird> AnMaster: tough shit
14:09:24 <AnMaster> because it would have been wider than 80 chars
14:09:24 <ehird> put the URL on the first line
14:09:31 <AnMaster> ehird, doesn't fit into 80 chars
14:09:35 <ehird> and?
14:09:55 <ehird> yes, i'm aware you don't use any programs made after the 70s, but everyone else does
14:10:13 <oerjan> to wrap, or not to wrap, that is the question
14:10:19 <AnMaster> ehird, the majority of email clients (including Thunderbird and most other GUI ones I seen) defaults to breaking at 80
14:10:36 <ehird> why are you writing a document in a mail clien
14:10:36 <ehird> t
14:10:46 <AnMaster> ehird, "<ehird> I'm pondering writing a little program that turns EmailML (as in, the conventions people use when emailing) into HTML."
14:10:54 <AnMaster> I thought that was about email?
14:11:10 <ehird> No, it just so happens that the pseudo-plaintext used by humans most commonly appears in email.
14:11:16 <AnMaster> ah
14:11:32 <AnMaster> ehird, also the email specs states some max line length, can't remember what exactly atm
14:11:43 <ehird> That's not the user's problem.
14:12:18 <AnMaster> ehird, hm? It means the email client has to wrap before that
14:12:34 <ehird> why are you writing a document in a mail client
14:13:32 <AnMaster> ehird, well pseudo-plaintext is common there
14:13:39 <ehird> So?
14:13:44 <AnMaster> ehird, and even in text documents I wrap at 80
14:13:48 <AnMaster> and so does ais523 I bet
14:14:30 <ehird> Yes, so do I, for readability (well, actually at around 70). My point stands: The amount of texts in which a long URI is before first 60 characters of the first line of a paragraph is minimal.
14:14:39 <ehird> And in those texts, I'm sure you can handle going over 80 chrs for just one line.
14:15:02 <AnMaster> ehird, right you use bibtex for them instead ;P
14:15:54 <AnMaster> some other detection you may want to add:
14:16:01 <ehird> Hm.
14:16:02 <ehird> Actually:
14:16:03 <AnMaster> leading space and *
14:16:05 <ehird> A line (B) after a line (A) is part of a wrapped paragraph if
14:16:05 <ehird> len(A+' '+B) > 60.
14:16:07 <AnMaster> to detect bullet lists
14:16:13 <AnMaster> same for numbers
14:16:14 <ehird> AnMaster: Yes, yes, I've got that all down.
14:16:25 <ehird> But I think [[A line (B) after a line (A) is part of a wrapped paragraph if
14:16:25 <ehird> len(A+' '+B) > 60.]] is reasonable
14:16:27 <AnMaster> since sometimes they do end near the 80th column
14:16:50 <AnMaster> ehird, hm...
14:17:13 <AnMaster> well this will be tricky any way you do it, but should be fairly interesting.
14:17:22 <AnMaster> You may have to try for different values certainly.
14:17:25 <ehird> Also, I'm not purely plaintext anyway, for linking.
14:17:31 <ehird> The syntax is as such:
14:17:43 <ehird> Foo bar baz (uri) => Foo bar <a>baz</a>
14:17:50 <ehird> Foo @bar baz (uri) => Foo <a>bar baz</a>
14:19:22 <AnMaster> ehird, if you plan to parse email you may want to auto detect gpg headers/footers and strip them or such
14:19:30 <ehird> I do not plan to parse email.
14:19:31 <ehird> I do not plan to parse email.
14:19:31 <ehird> I do not plan to parse email.
14:19:31 <ehird> I do not plan to parse email.
14:19:33 <ehird> :p
14:19:34 <AnMaster> ok
14:19:47 <AnMaster> ehird, you should not have mentioned email in the first place then ;P
14:20:00 <ehird> Well, that's the place where the pseduo-markup is used most.)
14:20:22 <AnMaster> ehird, this thing would be very hard to accurately parse I suspect: http://rafb.net/p/vMDvmM72.html
14:20:40 <AnMaster> especially the lists in the two last sections
14:20:51 <ehird> Apart from the alignment shit, I should be able to handle that fine.
14:21:08 <ehird> ==/-- of same length = header, (*|-|whatever)-space = list
14:21:13 <ehird> and that's about it
14:21:17 <AnMaster> ehird, you plan to convert it to html you said? <pre>message here</pre> ;P
14:21:24 <ehird> Heh.
14:21:32 <AnMaster> of course that is cheating
14:21:44 <ehird> I must say, if I manage to get this working, putting it on an email archive _would_ be nice
14:21:44 <Deewiant> no it's not, it's a perfectly valid conversion to HTML :-P
14:21:50 <ehird> but I imagine:
14:21:56 <ehird> Foo bar baz (uri) => Foo bar <a>baz</a>
14:21:57 <AnMaster> ehird, what about this type of list: http://rafb.net/p/9LtLDv21.html
14:21:59 <ehird> would trip most things up
14:22:11 <ehird> AnMaster: That will not be handled./
14:22:18 <AnMaster> oh ok
14:22:20 <ehird> also:
14:22:28 <AnMaster> Parameters (top of stack first):
14:22:29 <AnMaster> Vpos A vector describing what cell to operate on in Funge-Space.
14:22:29 <AnMaster> Old The value to compare the existing value in said cell to.
14:22:29 <AnMaster> New The new value to write in said cell if Old compares equal to the
14:22:29 <AnMaster> current value.
14:22:29 <AnMaster> that?
14:22:56 <ehird> "This section is a list of known issues (and solutions) with other fingerprints when using ATHR."
14:23:07 <ehird> AnMaster: I imagine I could parse that as a definition list.
14:23:30 <AnMaster> ehird, ah instead of "ATHR have far-ranging effects..."?
14:23:47 <ehird> It should be "has" there anyway, but yes, mine is easier to read. :-P
14:23:57 <AnMaster> since after the general considerations it goes on to list specific fingerprints and so on yeah
14:24:05 <oerjan> AnMaster: Do not use if pregnant
14:24:21 <AnMaster> oerjan, ... heh
14:26:02 <AnMaster> * A C99 compiler, or one that supports a large subset of C99, like GCC.
14:26:02 <AnMaster> + GCC 3.4.6, 4.1.2, 4.2.1 and 4.3.2 are known to work, other versions may or
14:26:02 <AnMaster> may not work.
14:26:02 <AnMaster> + ICC 10.1 is known to work too.
14:26:02 <AnMaster> - TCC 0.9.24 is known to fail at certain C99 constructs used in cfunge.
14:26:05 <AnMaster> what about that one ehird :D
14:26:11 <ehird> AnMaster: patches welcome.
14:26:21 <AnMaster> haha
14:27:09 <AnMaster> * Imported two new libraries into the code:
14:27:09 <AnMaster> genx - An XML output library (used by TURT).
14:27:09 <AnMaster> stringbuffer - Some utilities to build strings in an easy way, code was taken
14:27:09 <AnMaster> from crossfire.
14:27:11 <AnMaster> that?
14:27:28 <ehird> Go to hell. :-P
14:27:36 <AnMaster> well even one that could handle 75% of the cases or so would be very useful
14:27:49 <AnMaster> so if you get this working okish I hope you release the source
14:27:53 <ehird> I will. :p
14:28:05 <AnMaster> python?
14:28:11 <ehird> Yeah, but I haven't actually written it yet.
14:28:32 <AnMaster> right
14:29:49 <oerjan> snake gas
14:31:38 <ehird> Tada.
14:31:40 <ehird> A line (B) after a line (A) is part of a wrapped paragraph if len(A+' '+B) > 60.<br>
14:31:47 <ehird> Is the output for the current docs. :P
14:32:29 * AnMaster sneaks a space and a / into that <br> of ehird
14:32:47 <AnMaster> afk making food
14:33:06 <ehird> AnMaster: Hi. Would you like the 20 pages worth of text on why XHTML is broken and obsolete from the start?
14:33:26 <ehird> Or would you like to continue acting as if you're on the cutting edge by using broken technology.
14:33:42 -!- Slereah_ has quit (Read error: 113 (No route to host)).
14:36:24 -!- Slereah_ has joined.
14:43:16 -!- deveah has joined.
14:43:28 <deveah> when are you guys going to fix the wiki?
14:43:49 <deveah> hi btw
14:44:32 <oerjan> when graue reads his email, probably
14:46:18 <oerjan> aha! "his" is correct
14:46:22 <ehird> uh
14:46:24 <ehird> i could have told you that
14:46:38 <oerjan> well ais523 apparently couldn't :D
14:46:56 <ehird> Hi, I'm Scott Feeney of Fairfax, Virginia and you can contact me by emailing graue@oceanbase.org. If I know you, or should, other means of contact are available, but write there first.
14:47:02 <ehird> -- http://oceanbase.org/graue/
14:47:21 <oerjan> that's what i just discovered
14:48:23 <oerjan> "Ben the Benly Benis"?
14:48:39 <ehird> it seems to be a Pokey ripoff
14:55:34 <AnMaster> ehird, about xhtml... I guess I forgot ~...
14:55:42 <AnMaster> but I thought that was pretty clear
14:55:51 <AnMaster> and I think both xhtml and html are pretty broken
14:55:56 <ehird> general rule of sarcasm - it's meant to be funny
14:56:00 <AnMaster> we should use S-html
14:56:05 <ehird> also, no
14:56:05 <AnMaster> (html
14:56:07 <ehird> sexps are good for data
14:56:08 <AnMaster> (head
14:56:13 <ehird> sgml/xml are good for markup
14:56:21 <ehird> if you can't see why, well, stop talking about markup languages immediately :|
14:56:38 <AnMaster> ehird, I do see why yes, but that doesn't prevent me from having an esoteric viewpoint
14:57:44 <deveah> what are you talking about?
14:57:50 <ehird> deveah: Boring stuff.
14:57:58 <deveah> like what?
14:58:06 <ehird> markup languages.
14:58:08 <ehird> not esolang-related.
14:58:10 <deveah> oh
14:58:22 <deveah> I'm developing a textmode browser right now
14:58:34 <deveah> web browser
14:58:37 <oklopol> why is xml better for markup than sexps?
14:58:55 <oerjan> oklopol: sexps are not family friendly, duh
14:59:14 <ehird> oklopol: it's a bit complicated
14:59:25 <AnMaster> oerjan, hehe
14:59:38 <ehird> but basically the redundancy of sgml/xml, its support for attributes and its leniency in handling trivial plain bulk text
14:59:42 <ehird> is what makes it superior to sexps for markup
15:00:23 -!- oklopol has left (?).
15:00:29 -!- oklopol has joined.
15:00:30 <ehird> bye oklopol.
15:00:32 <ehird> hi oklopol.
15:00:41 <oklopol> hi ehird.
15:01:08 <oklopol> i wonder what the shortcut for parting is... can't be too complicated, since i keep on doing it
15:02:09 <AnMaster> INSERT ELEMENTS head, body INTO ELEMENT html; INSERT ELEMENT title INTO ELEMENT head OF ELEMENT html; INSERT TEXT "this is a horrible idea for markup" INTO ELEMENT title OF ELEMENT head OF ELEMENT html;
15:02:13 <AnMaster> heh
15:02:28 <oklopol> that's quite pretty
15:02:37 <oerjan> EXERT BRAIN
15:02:47 <oklopol> BRAAAAAINS
15:02:52 <AnMaster> it isn't really sql even, just SQL look-alike
15:03:13 <oklopol> ehird: i agree as much as i'm capable of.
15:03:14 <ehird> AnMaster: die :-P
15:03:17 <AnMaster> (with some ideas from apple script, mainly the "of")
15:03:20 <AnMaster> :P
15:03:32 <AnMaster> ehird, hah
15:03:43 <AnMaster> well if anyone did that seriously I would agree with you ehird
15:04:02 -!- deveah has left (?).
15:04:03 <ehird> it's impossible to be serious about that
15:04:03 <ehird> i hope
15:04:09 <AnMaster> so do I
15:04:33 <AnMaster> and I'm sure it is possible to make a even worse syntax
15:04:36 <AnMaster> oh yes
15:05:21 <AnMaster> WITH ATTRIBUTE href HAVING TEXT VALUE "http://example.com"
15:05:41 <AnMaster> however on bright side: You could easily do complex SELECT
15:07:08 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | does your assembly language look like a 'normal' computer? I mean registers, labels you can use as storage and such?.
15:07:54 <AnMaster> SELECT TEXT OF ELEMENT p WHEN ATTRIBUTE id OF ELEMENT p IS EQUAL TO TEXT VALUE "top";
15:07:54 <oerjan> optbot: let's hope not
15:07:55 <optbot> oerjan: Yay!
15:07:55 <AnMaster> XD
15:08:20 <oerjan> COBOSQL
15:08:41 <AnMaster> oerjan, yes and for html documents...
15:08:48 <AnMaster> + some ideas from AppleScript
15:09:26 <oerjan> COBOSML then
15:09:27 <AnMaster> So maybe COBOSQL Script?
15:09:35 <AnMaster> oerjan, hn
15:09:40 <AnMaster> hm*
15:10:49 <AnMaster> SELECT TEXT OF ELEMENT p WHEN ATTRIBUTE id OF ELEMENT p IS EQUAL TO TEXT VALUE "top" AND ALSO TEXT OF ELEMENT p STARTS WITH TEXT VALUE "ehird";
15:10:57 * ehird kills AnMaster
15:11:28 -!- LinuS has joined.
15:11:32 <oerjan> it's for the good of humanity!
15:11:36 <oerjan> the killing, i mean
15:11:38 <AnMaster> ehird, I should actually spec this language
15:11:42 <AnMaster> oerjan, I agree
15:12:16 <AnMaster> Structured Text Markup Definition and Query Language
15:12:19 <AnMaster> STMDQL?
15:13:02 <ehird> Find an acronym that ends up as RAPEFIRE, because that is approximately the feeling using it will give.
15:13:24 <AnMaster> ehird, I think either COBOSQL or COBOSML are good names for it though
15:13:33 <ehird> But misleading.
15:13:38 <AnMaster> also does sql actually need upper case?
15:13:38 <ehird> COBOSQL sounds pretty easy. :-P
15:13:41 <ehird> and no
15:14:05 <AnMaster> ehird, and no I probably won't be able to spec it
15:14:15 <AnMaster> but AND turns into AND ALSO, and OR into OR ELSE
15:15:05 <oerjan> commands should however end with ", OR ELSE!"
15:15:31 <AnMaster> oerjan, somewhat like the opposite of INTERCAL's "PLEASE"?
15:15:36 <oerjan> yay
15:16:45 <AnMaster> oerjan, hm?
15:17:27 <AnMaster> you also want: BUT NOT and MAYBE operators
15:17:28 <oerjan> means: i didn't think of that, but yeah
15:18:40 <AnMaster> oh top element must always be referred to with THE
15:18:53 <AnMaster> like ELEMENT body OF THE ELEMENT html
15:19:42 <AnMaster> oerjan, well I suggest you write the spec, you seem to have some good ideas :)
15:19:50 <AnMaster> I wouldn't be able to to.
15:19:54 <AnMaster> it is too horrible
15:20:04 <oerjan> i'm sorry no can do
15:20:19 <AnMaster> oerjan, due to preserving sanity?
15:20:25 <oerjan> also, wrists
15:20:29 <AnMaster> ouch
15:20:53 <AnMaster> oerjan, hope your wrists get better (not related to this language)
15:20:54 <oerjan> not to mention i have a cold with fever
15:21:23 <AnMaster> oh and all that messy java script changing document object model crap you know?
15:21:27 <AnMaster> we can get rid of it
15:21:29 <AnMaster> just:
15:22:14 <AnMaster> UPDATE TEXT OF THE FIRST ELEMENT p OF ELEMENT body OF THE ELEMENT html SETTING NEW VALUE TO TEXT "Blergh...";
15:22:17 <AnMaster> or something like that
15:22:45 <oerjan> clearly this is not a language for bad wrists
15:22:58 <AnMaster> oerjan, you mean because it is so verbose?
15:23:02 <oerjan> yeah
15:23:04 <AnMaster> indeed
15:23:33 <AnMaster> oerjan, also it is considered bad style to use caps lock, you should hold down shift instead all the time ;P
15:23:41 <AnMaster> recommended coding style
15:23:45 <oerjan> i was thinking of that
15:24:05 <AnMaster> oerjan, since it is case sensitive
15:24:31 <oerjan> "This language is insensitive to everything, except case"
15:24:39 <AnMaster> :D
15:27:23 <AnMaster> oerjan, I think it should require exactly correct syntax
15:27:28 <AnMaster> oerjan, and only give:
15:27:33 <AnMaster> SYNTAX ERROR
15:27:40 <AnMaster> and not even line number
15:27:45 <AnMaster> in case of an error
15:28:23 <oerjan> i think that is fairly common as esolangs go.
15:28:33 <oerjan> if even that
15:29:14 <AnMaster> oerjan, maybe very very verbose ones that are still, somehow, not helpful at all?
15:29:32 <oerjan> maybe
15:29:47 <oerjan> it could say precisely what the error is, but not where it is
15:30:12 <oerjan> oh and try to do it without mentioning any actual text used
15:30:31 <Deewiant> "you fail because you are an idiot"
15:30:50 <oerjan> that's not precise
15:31:12 <Deewiant> well, it can't know why you're an idiot ;-0
15:31:15 <Deewiant> s/0/)/
15:31:16 <oerjan> "you fail because of increased dopamine level in the hippocampus region of your brain"
15:31:30 <Deewiant> "PEBKAC"
15:31:35 <ehird> "Fuck you and die, all you ever do is lie. You say 'this is a program' but it's _not_ - it may look like one, but it is not. You always lie to me. You never just tell the truth, maybe say 'this is not a program'. I'm going."
15:31:39 <ehird> then it removes the compiler
15:31:46 <oklopol> perhaps it should try to correct your error, and add errors elsewhere?
15:31:51 <ehird> (that happens after 10 syntax errors (spread out over different program runs))
15:32:14 <oklopol> so if you have say a syntax error somewhere, you might get weird overflows somewhere else.
15:32:51 <oerjan> ehird: after the first errors you need to add "SORRY ABOUT THAT" and increasingly profuse apologies to get the next ones to compile
15:32:59 <ehird> :-D
15:33:04 <AnMaster> hehe
15:33:16 <ehird> maybe have it ask you to stop beating it (beating = running it with an invalid program)
15:33:20 <ehird> and plead and plead
15:33:26 <ehird> until it runs away (deletes itself) sobbing
15:34:05 <oerjan> when installing the compiler, it may have to choose a random personality
15:34:14 <AnMaster> oerjan, hahah
15:34:19 <AnMaster> one should be...
15:34:28 <AnMaster> genuine people personality
15:34:28 <AnMaster> :D
15:34:35 <oerjan> "sane" should not be an option
15:37:55 <oerjan> think Sirius Corporation
15:38:53 <AnMaster> yes
15:41:09 <oerjan> oh wait that was what you were referring to
15:41:11 <AnMaster> Oct 19 12:55:04 tux hpijs: WARNING: black pen has low ink
15:41:11 <AnMaster> Oct 19 12:55:04 tux hpijs: STATE: marker-supply-low-warning
15:41:17 <AnMaster> odd, that has been going on for weeks
15:41:20 <AnMaster> and still printing works fine
15:41:29 * oerjan didn't notice until he read it on the wiki page
15:41:39 <AnMaster> oerjan, of course it was...
15:42:31 <oerjan> that's odd? i thought that was how all printers did it
15:43:04 <AnMaster> oerjan, reporting the level was low when there are still more than 100 pages left on "normal quality"?
15:43:10 <oerjan> hint: they want to sell marker
15:43:17 <oerjan> er wait
15:43:24 <oerjan> ink, whatever
15:43:28 <AnMaster> yep ink
15:43:36 <AnMaster> no idea why hp call it marker
15:44:09 <oerjan> AnMaster: mind you with some corporate procedures you may have to order new ink that far in advance :D
15:44:37 <AnMaster> oerjan, heh? I just went to the local shop a few days ago. But not going to replace it until I actually run out of black ink
15:44:58 -!- Slereah_ has quit (Read error: 113 (No route to host)).
15:45:02 <oerjan> i mean, Dilbert-style corporations
15:45:10 <AnMaster> what really hurts are the colour ones... Because you usually run out of one colour long before the other ones
15:45:45 <AnMaster> oerjan, don't have any experience of those
15:45:54 <oerjan> ooh, you could have a company policy that said you had to use all colors balancedly...
15:46:18 <oerjan> neither have i but i do read Dilbert
15:46:34 <AnMaster> oerjan, maybe I should start doing that
15:52:33 -!- hando has quit ("Leaving").
15:54:17 <ehird> http://impressive.net/people/gerald/1999/01/http-archive/ This is very clever.
15:54:31 -!- Judofyr has joined.
15:54:34 <ehird> I think I shall write a firefox plugin to do that, maybe a greasemonkey script for simplicity.
15:55:02 <ehird> But... I visit an awful lot of pages per day, probably bordering on 10,000.
15:55:07 <ehird> Might take too much space.
15:56:01 <AnMaster> ehird, 10 000 on a day? That many *static* pages? Since the system doesn't seem to do dynamic ones yet
15:56:12 <ehird> AnMaster: Um, that archives all pages.
15:56:20 <AnMaster> "This setup doesn't quite archive enough stuff because Squid doesn't cache dynamic resources etc.; I plan to install Jigsaw and make it archive a copy of everything it retrieves. (or,"
15:56:36 <ehird> It's impossible to know what a dynamic resource is via HTTP.
15:56:39 <ehird> It's just a resource.
15:56:40 -!- Corun has joined.
15:56:42 <ehird> Besides, I said I'd write my own.
15:56:46 <AnMaster> ah yes
15:56:47 <ehird> As that was made in 1999, last updated 2004.
15:57:03 <ehird> But yes, my browsing style is rapid.
15:57:09 <ehird> Due to obsessive link-clickery. ;-)
15:57:14 <AnMaster> also the issue of different pages, same url
15:57:17 <AnMaster> for example ajax
15:57:26 <AnMaster> how to handle something like gmail?
15:57:38 <ehird> AnMaster: well, you can't.
15:57:39 <ehird> Well.
15:57:40 <ehird> Yes you can.
15:57:43 <AnMaster> partly yes
15:57:48 <ehird> gmail changes the hash-portion of the URI as a hack for the back button/history.
15:57:52 <ehird> /bookmarks
15:57:54 <AnMaster> ah
15:57:56 <ehird> So you could hack that in, somehow.
15:57:59 <ehird> Generally, though?
15:58:04 <AnMaster> generally impossible
15:58:04 <ehird> Ajaxy pages should not change the very nature of the page.
15:58:11 <ehird> They should just update content/submit stuff, etc. :-P
15:58:23 <AnMaster> hm
15:58:41 <ehird> Anything else is bad style.
15:58:41 <AnMaster> ehird, what about plugins? Such as when watching youtube
15:58:43 <ehird> Well.
15:58:44 <oerjan> fungot: What is the meaning of life?
15:58:44 <fungot> oerjan: yeah yeah that's
15:58:45 <ehird> Unless you are making a site specialized for the iPhone.
15:59:02 <ehird> Because it's nicer to have a seamless view there (e.g. mimic the OS transitions) by doing everything with Ajax.
15:59:03 <oerjan> fungot: are you still stuck on phone conversations?
15:59:03 <fungot> oerjan: ( ( oh okay oh
15:59:05 <ehird> But that's a corner case.
15:59:11 <ehird> AnMaster: What's that got to do with ajax?
15:59:23 <Deewiant> fungot: oh?
15:59:24 <fungot> Deewiant: i'm sure they do they come and serve you something it's not fnord for whatever bizarre reason
15:59:25 <AnMaster> ehird, nothing, it was just another potential issue
15:59:37 <Deewiant> fungot: yes, being fnord would make much more sense.
15:59:38 <fungot> Deewiant: i guess so
15:59:38 <ehird> AnMaster: Well, it'd fail at mirroring youtube pages, yeah.
15:59:46 <AnMaster> fungot, are you still on the data from those phone calls?
15:59:46 <oerjan> fungot: that's because fnord is so hard to get hold of
15:59:47 <fungot> AnMaster: ( ( laughter mn))
15:59:47 <fungot> oerjan: ( ( laughter oh i don't know i'm
15:59:49 <ehird> I could specialcase it though.
15:59:51 <AnMaster> yeah I guess so
16:00:22 <ehird> OK, now for a directory structure...
16:00:24 <AnMaster> ehird, and then pages using POST data with same url. And what not
16:00:30 <ehird> AnMaster: that's not a problem
16:00:40 <AnMaster> or pages with same data but different ? arguments
16:00:48 <ehird> that's not a problem.
16:00:55 <AnMaster> well see how much you get during a day would be interesting.
16:01:29 <AnMaster> and then there is the issue of pages marked as no-cache heh, I guess you'll just ignore that
16:01:37 <ehird> <archive>/http/example.com/:80/i/love/astronauts/they/are/cool/?cool=1/#hello/2008/10/19/16/01/1
16:01:43 <ehird> Wait, need the method in there.
16:01:54 <ehird> <archive>/http/example.com/:80/i/love/astronauts/they/are/cool/?cool=1/#hello/POST/2008/10/19/16/01/1
16:01:57 <ehird> It's really verbose, but complete.
16:02:21 <ehird> Or.
16:02:24 <ehird> I could just not cache POSTs.
16:02:27 <ehird> As you're not meant to.
16:05:24 <ehird> Hmm...
16:05:37 <ehird> I would, ideally, modify my browser so that it does not use a regular cache anymore, but instead fetches from the archive.
16:05:40 <ehird> Otherwise, that's some duplication.
16:05:53 <fizzie> Heh, the Bruce Schneier facts randomizator gave me: "Bruce Schneier is able to read every Unlambda program."
16:06:00 <oerjan> cache not the POST, lest ye be cached
16:06:02 <ehird> Also, actually viewing archived pages will be difficult - they will have to be preprocessed to resolve links to other pages from the archive at the right time.
16:06:08 <ehird> fizzie: *g*
16:07:46 <fizzie> fungot: Go back to IRC-speak, it's less boring.
16:07:47 <fungot> fizzie: what makes you think i need to recompile all the strands or whatever you wanna call it
16:08:12 <ehird> fizzie: feed it the brown corpus
16:08:15 <ehird> it's hilariously bad
16:09:09 <AnMaster> ehird, you need to consider cache control directives
16:09:21 <AnMaster> if you replace the cache
16:09:24 <ehird> Yes.
16:09:35 <ehird> I should probably have a metadata system - 1/meta and 1/content
16:09:49 <ehird> Of course, the more usable this system gets the less liklely I am to implement it out of complexity.
16:09:55 <AnMaster> ehird, you could use file attributes if OS X have that
16:10:00 <ehird> Still, it'd be nice never to see a 404 after visiting a page again.
16:10:06 <ehird> AnMaster: I'd prefer something more mungible.
16:10:23 <AnMaster> "No definitions were found for mungible."
16:10:35 <AnMaster> Did you mean: define:fungible
16:10:44 <ehird> AnMaster: It's one of those words that you can interpret just by reading them.
16:10:54 <AnMaster> ehird, well maybe a native speaker can
16:10:55 <AnMaster> I can't
16:11:11 <ehird> AnMaster: It's in the phonetics, kinda. :-P
16:11:19 <ehird> I can't really explain it.
16:11:25 <AnMaster> ehird, so what does it mean?
16:11:37 <ehird> AnMaster: Hard to explain. Kind of... manipulatable, except... more gloopy...
16:11:53 <AnMaster> ...... No definitions were found for gloopy.
16:11:55 <AnMaster> :(
16:12:01 <ehird> Gooey.
16:12:02 <ehird> Globby.
16:12:04 <ehird> :-P
16:12:29 <AnMaster> Definitions of gooey on the Web: <-- some really really strange ones before the sane ones
16:12:35 <AnMaster> "Graphical User Interface. Esentially, it is using pictures (or graphics) instead of words to give commands or exchange information with the computer.
16:12:35 <AnMaster> homepages.vvm.com/~jhunt/compupedia/comp_glos/g_h.htm"
16:12:52 <AnMaster> "of, or related to goo; soft, sticky and viscous " seems sane
16:13:05 <AnMaster> ehird, as for globby: "No definitions were found for globby."
16:13:17 <ehird> AnMaster: Oh fer gods sake, just pronounce the words in your head for a bit. :-P
16:13:34 <AnMaster> ehird, "like a glob"? Wildcard glob?
16:13:54 <ehird> Globby, goey, like... a gloopy slimeball...monster...thing.
16:14:01 <ehird> (Also, the first gooey definition is from the pronounciation of GUI)
16:14:35 <AnMaster> you want something that is a manipulatable slime ball in your file system :D
16:14:36 <AnMaster> ?
16:14:50 <ehird> no
16:14:50 <ehird> XD
16:14:54 <Deewiant> AnMaster: like tar balls?
16:14:58 <ehird> mungible...sort of...
16:15:03 <fizzie> ehird: I could feed it the LDC Gigaword corpus -- that's 12 gigabytes of text, mainly news articles, gzipped. But my script would need some serious fixation before processing that stuff would be even remotely possible.
16:15:04 <ehird> Deewiant: do you understand what i mean by 'mungible'?
16:15:13 <Deewiant> ehird: I don't even know what you're talking about.
16:15:13 <ehird> fizzie: the brown corpus is just a meg or two
16:15:21 <Deewiant> I might, if I knew what you were talking about.
16:15:22 <ehird> Deewiant: the word.
16:15:23 <ehird> mungible.
16:15:32 <ehird> i'm talking about a metadata file, AnMaster suggested using attributes on the actual file
16:15:37 <ehird> I said i'd prefer an actual file because it's more mungible
16:15:46 <Deewiant> yeah, you can munge it
16:16:11 <Deewiant> more easily, anyway
16:16:33 <ehird> Deewiant: you're bluffing
16:16:34 <ehird> :-P
16:17:11 <Deewiant> no, I don't think I am
16:17:21 <Deewiant> I think AnMaster got it to an extent with "manipulatable"
16:17:30 <Deewiant> ehird: but, you want to tell me why this says "mong"? http://www.doomworld.com/idgames/?search
16:17:44 <Deewiant> you can ponder it while I eat
16:17:47 <ehird> I have absolutely no idea.
16:23:00 <AnMaster> hm?
16:23:20 <AnMaster> what?
16:23:25 <ehird> WHAT
16:23:28 <ehird> *what
16:23:57 <AnMaster> I'm being highlighted on around 5 channels, if if is something important, please repeat it on one line, or I can read it later.
16:24:21 <fizzie> ehird: I used the Europarl corpus (European parliament speeches, the English translations, for authors with names beginning with A-F) because I happened to have it here locally already. So fungot's now a politician.
16:24:21 <fungot> fizzie: mr president, i support your positions on this.
16:24:29 <ehird> LMAO
16:24:35 <ehird> fungot: lie
16:24:35 <fungot> ehird: mr president, clearly we are going to continue to take place. we cannot ask for this in the near future
16:24:55 <ehird> ok fungot
16:24:56 <fungot> ehird: madam president, when discussing the issue. i should perhaps make a few specific positions. as i have said, that corruption comes from within. a decision in nice to the effect that the existing accommodation for such instruments within the european parliament at this time as well.
16:25:04 <ehird> fungot: stop bullshitting
16:25:05 <fungot> ehird: mr president, commissioner, ensure that the impact of the directive, and a conference held in new york is a very important time, precisely because the area is the second recommendation. a joint declaration on free and fair trade bananas, but about the method. as this is indeed a breakthrough for the right of public access to administrative procedures.
16:25:13 <ehird> fungot: AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
16:25:14 <fungot> ehird: to avoid any cuts in relation to your reference to this, and with each day that the french should be " fnord'.
16:25:54 -!- zbrown has joined.
16:33:24 -!- omniscient_idiot has quit (Connection timed out).
16:33:25 -!- Jiminy_Cricket has quit (Connection timed out).
16:34:20 <oerjan> free and fair trade bananas, indeed
16:35:25 <oerjan> fungot: so i can count on your vote, then?
16:35:26 <fungot> oerjan: mr president, honourable members, it may be said that particularly here, parliament has the power to assess the impact of this agreement which should contribute to the international agreements on limiting co2 emissions and reducing environmental pollution and intellectual property in third countries.
16:37:04 <oerjan> fungot: yeah reduce their intellectual property already
16:37:05 <fungot> oerjan: madam president, i thank mr watson for the information on spain. sorry, i have no information about, but nothing has happened here. i am slowly getting fnord up by these fnord arguments. we listened to our fnord impulses again.
16:37:34 <oerjan> fungot: yeah i can see that
16:37:35 <fungot> oerjan: madam president, mr schulz: we did indeed restrict ourselves to condemning the atrocities of the taliban protection of mr bin laden has a london bank account. in the interests of major fnord multinational monopolies which want to join in when the action really does start; and, when eventually slaughtered, of meat and bonemeal amounts to less than 2 in the form of a draft regulation that will ensure the economic viabilit
16:38:14 <GregorR> WTF
16:38:53 <Deewiant> fungot: WTF
16:38:53 <fungot> Deewiant: mr president, you were wonderful!' i am sorry to say that the work has been completed they will be included in the list of members in the lower chamber in the netherlands, which has saved europe from dictators in the past.
16:38:59 <oerjan> clearly people should listen more to parliament procedures. this is horrible.
16:39:09 <AnMaster> hehe
16:39:13 <GregorR> WTF
16:39:24 <oerjan> thank you, dutch
16:39:29 <Deewiant> fungot: WTF
16:39:29 <fungot> Deewiant: the impressive number of projects in third countries. if the transatlantic relationship, but the parliament put forward a proposal on voluntary agreements aimed at considering how these can be overcome without secondary legislation, we must jointly take this approach.
16:39:32 <AnMaster> GregorR, fizzie fed som mad text into fungot's dict
16:39:42 <Deewiant> fungot: WTF
16:39:48 <Deewiant> :-(
16:39:50 <oerjan> oh noes
16:39:53 <Deewiant> fungot: dammit
16:39:55 <AnMaster> fungot, there?
16:39:58 <AnMaster> ah
16:39:59 <AnMaster> fizzie, !
16:40:02 <AnMaster> crash I fear
16:40:04 <Deewiant> ah well
16:40:05 <fizzie> Yes.
16:40:07 -!- fungot has quit (Read error: 131 (Connection reset by peer)).
16:40:09 <AnMaster> fizzie, debugging?
16:40:26 <fizzie> Later. :p
16:40:34 -!- fungot has joined.
16:40:41 <fizzie> There you have our little politician again.
16:40:42 <AnMaster> fizzie, with cfunge you could put it in trace mode, then redirect trace to a file
16:40:52 <AnMaster> fizzie, trace is by default to stderr
16:40:55 <fizzie> There's far too much trace information.
16:41:02 <fizzie> RC/Funge has got a trace mode too.
16:41:05 <AnMaster> fizzie, you just want the last bit
16:41:08 <GregorR> Hm, I wonder if you get more points for killing everyone slowly and horribly, or quickly and dramatically ...
16:41:13 <ehird> fizzie: just tail it :-P
16:41:19 <oerjan> fungot: what about the crisis?
16:41:20 <fungot> oerjan: mr president, because i believe that these tasks will have to hold your committee meeting without me. i worry about the conditions in which all these qualities will continue to be monitored but it must then be prepared to move into the city. that is very important that we support in particular paragraphs 11 and 13 refer would be too many control centres is nonsense. last year 390 000 people applied for asylum in member
16:41:39 <GregorR> fungot: How much wood would a woodchuck chuck if a woodchuck would chuck wood?
16:41:39 <fungot> GregorR: mr president, that we currently hear from the commission to adopt them, because of the pressure that the european research area. the role of a central disciplinary system which will ensure that the provisions of the treaty.
16:41:53 <AnMaster> fizzie, cfunge can just dump current instruction, so one char per executed instruction, or "x,y, integer value of instruction, instruction, thread id"
16:41:59 <AnMaster> or even dump stack every time
16:42:04 <AnMaster> no actual debugger though
16:42:15 <Deewiant> AnMaster: RC/Funge-98 has a trace mode as well as a debugger.
16:42:20 <AnMaster> Deewiant, ah
16:42:26 <AnMaster> and ccbi lacks a trace mode
16:42:33 <Deewiant> yep
16:42:44 <AnMaster> efunge lacks both unless you uncomment a few lines in the source
16:42:47 <AnMaster> currenly that is
16:42:51 <Asztal> "i am slowly getting fnord up by these fnord arguments" <- I like that
16:42:58 <AnMaster> I plan to make tracing available in a *VERY ENTERPRISY WAY*
16:42:59 -!- Jiminy_Cricket has joined.
16:43:10 <Deewiant> AnMaster: XML output? :-P
16:43:12 <AnMaster> using the SASL error logger (System Architecture Support Libraries)
16:43:12 -!- omniscient_idiot has joined.
16:43:26 <AnMaster> it would look like:
16:43:41 <AnMaster> =PROGRESS REPORT==== 19-Oct-2008::17:43:35 ===
16:43:46 <AnMaster> application: efunge
16:43:55 <AnMaster> started_at: foo@node
16:44:10 <AnMaster> and then the rest of the data
16:45:10 <Deewiant> one of those for every tick, right?
16:45:15 <AnMaster> wait no
16:45:19 <AnMaster> it would be a different message
16:45:20 <AnMaster> a sec
16:45:53 <fizzie> Yes, there is a trace mode in RC/Funge. Assuming tail is clever enough to throw stuff away (should be) I guess I could just tail the last ten thousand lines or so.
16:46:01 <AnMaster> =INFO REPORT==== 19-Oct-2008::17:45:56 ===
16:46:01 <AnMaster> Executing instruction 43 at {42,43} in thread 0 IP 0
16:46:02 <AnMaster> like that
16:46:04 <AnMaster> every tick
16:46:05 <AnMaster> :D
16:46:14 <AnMaster> 1> error_logger:info_msg("Executing instruction ~p at ~p in thread ~p IP ~p~n", [$+, {42,43}, 0, 0]).
16:46:17 <AnMaster> was how I generated it
16:46:51 <AnMaster> Deewiant, however a bit mad for tracing
16:47:02 <Asztal> very good, but it should be XML with XSL
16:47:22 <AnMaster> Asztal, you could install a different logger callback I believe
16:47:26 -!- ENKI-][ has quit (Read error: 104 (Connection reset by peer)).
16:47:30 <AnMaster> not sure how, but pretty sure it is possible
16:47:45 <AnMaster> and you could make it log to rotating log files
16:47:48 <Deewiant> fizzie: no, tail reads all the input into a red-black tree indexed by line number, then when it hits EOF it repeatedly gets the lowest key, checks whether it's in the requested range to be printed, and prints it if so
16:47:50 <AnMaster> that rotate when they reach some size
16:48:02 <AnMaster> by a few lines of config
16:49:38 <AnMaster> Asztal, I think there is also some true tracing functionality
16:49:53 <AnMaster> sys:trace seems to be for system messages to a thread only, but pretty sure there are stuff for user tracing too
16:50:01 <fizzie> Usually those just rename the file, start a new one, and signal the logging thing to reopen the log file so it starts writing in the new one. I don't think RC/Funge is clever enough to do the reopening.
16:50:38 <AnMaster> in fact erlang got several tracing frontends
16:51:01 <AnMaster> since the low level trace functionality is too powerful and flexible, using it directly is pretty hard
16:51:18 <oerjan> Deewiant: you've gotta be kidding O_O
16:51:41 <ehird> oerjan: he is...
16:51:53 <oerjan> whew
16:51:53 <AnMaster> oerjan, I thought you if anyone would have humor
16:52:04 <ehird> AnMaster: Life imitates art.
16:52:05 <oerjan> i do have humor
16:52:17 <AnMaster> I found Deewiant pretty funny
16:52:31 <oerjan> but that doesn't mean i can recognize when the real world is insane :D
16:53:38 <AnMaster> yes of course, everyone know that tail uses a self-balanced B-tree, not a red-black tree.
16:53:44 <AnMaster> Of course he was kidding...
16:53:59 -!- ENKI-][ has joined.
16:54:07 <fizzie> Thanks, people, you made me check out tail sources just to be sure.
16:54:12 <oerjan> ha ha
16:54:40 <ehird> AnMaster: wow, an actual funny joke
16:54:40 <ehird> :O
16:54:55 <AnMaster> ehird, well as I said, humor is subjective
16:55:04 <ehird> AN ACTUAL FUNNY JOKE
16:55:10 <AnMaster> ehird, unlike yours
16:55:15 <oerjan> and ehird doesn't like to be subjected to humor
16:55:17 <ehird> no, like mine
16:55:20 <ehird> oerjan: it burns
16:55:22 <AnMaster> fizzie, and the erlang support for rotating log files is done in erlang
16:57:33 <AnMaster> fizzie, http://rafb.net/p/8VgYea83.html
16:57:46 <fizzie> Maybe I'll restart fungot under the tracing+tail now. Although I'm not sure how much it'll help. Certainly I'll find out where it ends up looping around, but if it's something data-dependent (like it'll probably be) most likely the infinite-looping has pushed the interesting file offsets and such out of the log. Maybe I'd have some space to add a ":." sequence there where it talks about file offsets.
16:57:46 <fungot> fizzie: mr president, ladies and gentlemen, is it basically poor management or is it due to the refusal of this house to enable us to take environmental criteria into consideration in a subsequent review. the draft framework decision on organised crime.
16:57:47 <Deewiant> it makes you think, though: why the hell is 'tail' 1700 lines of code
16:58:03 <AnMaster> Deewiant, the GNU one maybe
16:58:08 <AnMaster> I'm pretty sure the BSD one is shorter
16:58:13 <AnMaster> let me check (freebsd 6.3)
16:58:27 <Deewiant> 'echo' is a sweet 256 lines
16:58:38 <Deewiant> and 'cat' is 785
16:58:42 <AnMaster> $ wc -l /usr/src/usr.bin/tail/tail.c
16:58:42 <AnMaster> 338 /usr/src/usr.bin/tail/tail.c
16:58:43 <AnMaster> there
16:58:47 <fizzie> I didn't know it does files and pipes differently, although I guess that makes a lot of sense.
16:59:03 <AnMaster> $ wc -l /usr/src/bin/cat/cat.c
16:59:03 <AnMaster> 314 /usr/src/bin/cat/cat.c
16:59:13 <ehird> cat should not be 314
16:59:14 <ehird> lines
16:59:16 <ehird> it should be 30
16:59:17 <ehird> seriously
16:59:21 <AnMaster> $ wc -l /usr/src/bin/echo/echo.c
16:59:22 <AnMaster> 137 /usr/src/bin/echo/echo.c
16:59:22 <ehird> fucking gnu software :\
16:59:23 <AnMaster> ehird, agreed
16:59:27 <Deewiant> 'ls' is a whopping 4542 lines
16:59:27 <AnMaster> ehird, this is FreeBSD one
16:59:28 <AnMaster> ....
16:59:32 <AnMaster> ehird, so not gnu at all
16:59:33 <ehird> AnMaster: fucking bsd :-P
16:59:39 <ehird> Deewiant: you are fucking kidding me
16:59:39 <Deewiant> that's over half the size of CCBI
16:59:40 * Corun read that as "fucking gnu showers"
16:59:42 <ehird> 4000 lines?
16:59:44 <ehird> jesus
16:59:47 <Deewiant> ehird: 4542.
16:59:56 <AnMaster> $ wc -l /usr/src/bin/ls/ls.c
16:59:56 <AnMaster> 873 /usr/src/bin/ls/ls.c
16:59:59 <Deewiant> ehird: the last 180 being the --help text.
17:00:00 <ehird> Corun: god, you just made me imagine rms having sex in a shower... how did you do that? how can anything make me imagine anything so unspeakably awful
17:00:01 <ehird> i hate you
17:00:07 <ehird> i hope you die in a fire tonight >:(
17:00:12 <ehird> Deewiant: LOL
17:00:22 <AnMaster> ehird, there are multiple flags for cat, even posix defines one
17:00:23 <oerjan> LOL
17:00:26 <Deewiant> AnMaster: how long is ls --help
17:00:27 <AnMaster> -u Write bytes from the input file to the standard output without delay as each is read.
17:00:37 <AnMaster> $ ls --help
17:00:37 <AnMaster> ls: illegal option -- -
17:00:37 <AnMaster> usage: ls [-ABCFGHILPRSTUWZabcdfghiklmnopqrstuwx1] [file ...]
17:00:38 <AnMaster> :P
17:00:45 <AnMaster> Deewiant, see man page instead
17:00:49 <ehird> what is that program that has 'has even more flags than ls(1)' in BUGS>
17:01:05 <AnMaster> Deewiant, since *bsd tool doesn't use -- flags
17:01:11 <AnMaster> only single letter ones
17:01:15 <Deewiant> AnMaster: ls -h?
17:01:16 <AnMaster> and -G is the colours one
17:01:28 <AnMaster> /usr/src $ ls -h
17:01:28 <AnMaster> COPYRIGHT Makefile.inc1 bin games lib sbin tools
17:01:28 <AnMaster> LOCKS ObsoleteFiles.inc contrib gnu libexec secure usr.bin
17:01:28 <AnMaster> MAINTAINERS README crypto include release share usr.sbin
17:01:29 <AnMaster> Makefile UPDATING etc kerberos5 rescue sys
17:01:38 <Deewiant> you really didn't need to paste that :-P
17:01:41 <AnMaster> Deewiant, there is *no help except short usage*
17:01:43 <AnMaster> and man page
17:01:44 <fizzie> ls.c is the longest source file in coreutils; sort at a bit over 3k lines and pr at a bit less than 3k come next.
17:01:49 <AnMaster> Deewiant, on *BSD you sue the *man page*
17:01:54 <fizzie> Heh, sue.
17:01:55 <AnMaster> not some funky --help
17:02:08 <fizzie> BSD is so litigation-obsessed, you even sue the man pages!
17:02:19 <AnMaster> use*
17:02:19 <Deewiant> I've always preferred --help and such
17:02:20 <AnMaster> typo
17:02:21 <AnMaster> duh
17:02:35 <Deewiant> ideally a program is just a single executable with no other data files
17:02:54 <AnMaster> the *BSD sort seems to just use the GNU one
17:03:03 <AnMaster> FreeBSD that is
17:03:15 <AnMaster> same for grep
17:03:19 <AnMaster> but their own awk
17:03:49 <AnMaster> there is also something called tsort
17:03:51 * AnMaster reads man page
17:03:56 -!- fungot has quit ("restarting with tracing...").
17:04:02 <AnMaster> TSORT(1) FreeBSD General Commands Manual TSORT(1)
17:04:02 <AnMaster> NAME
17:04:02 <AnMaster> tsort -- topological sort of a directed graph
17:04:03 <AnMaster> *blink*
17:04:24 <AnMaster> fizzie, also cfunge wouldn't know to reopen the file for trace since it just use plain stderr
17:04:30 <AnMaster> adding a signal handler should be simple
17:04:31 -!- Judofyr has quit (Read error: 104 (Connection reset by peer)).
17:05:03 <Deewiant> hmm, I don't quite get tsort
17:05:08 -!- Judofyr has joined.
17:05:10 <Deewiant> "input contains an odd number of tokens"?
17:05:14 <AnMaster> Deewiant, as for "<Deewiant> ideally a program is just a single executable with no other data files"
17:05:16 <AnMaster> well they are
17:05:19 <AnMaster> file is an exception
17:05:23 <AnMaster> it needs a definition file
17:05:30 <AnMaster> most other simple core tools are one file
17:05:32 -!- fungot has joined.
17:05:43 <Deewiant> AnMaster: on BSD they aren't, because you don't know what to do with them without the man file, which is external.
17:05:47 <oerjan> AnMaster: a topological sort finds an order of the vertices such that all source vertices come before the corresponding target vertices
17:05:50 <oerjan> iirc
17:05:54 <Deewiant> yes
17:05:59 <Deewiant> but I don't get tsort.
17:05:59 <AnMaster> Deewiant, don't bloat binary with manual
17:06:09 <Deewiant> AnMaster: don't bloat package with extra file
17:06:26 <AnMaster> Deewiant, you have man pages on linux too
17:06:39 <AnMaster> Deewiant, except the man pages on *bsd are way way higher quality
17:06:41 <Deewiant> AnMaster: yes, I know, I wasn't arguing for linux in particular
17:06:44 <AnMaster> not just some "see info page"
17:06:47 <AnMaster> but proper man pages
17:06:59 <Deewiant> at least on linux you don't usually need the man/info page
17:07:02 <Deewiant> --help is sufficient
17:07:23 <fizzie> Deewiant: It works so that on alternating lines you have source/destination for the graph edges.
17:07:56 <fizzie> Deewiant: So you input "foo", "bar", "foo", "baz", "bar", "quux", "baz", "quux" and it outputs either "foo", "bar", "baz", "quux" or "foo", "baz", "bar", "quux".
17:08:00 <oerjan> Deewiant: "`tsort' reads its input as pairs of strings, separated by blanks,
17:08:01 <oerjan> indicating a partial ordering. The output is a total ordering that
17:08:04 <oerjan> corresponds to the given partial ordering.
17:08:13 <Deewiant> yeah, I looked at the info page. I tried "foo bar baz" but it seems it wants "foo bar foo baz"
17:08:16 <fizzie> Okay, so any sort of blank works, not just newline.
17:08:31 <AnMaster> http://rafb.net/p/1iIMoG85.html
17:08:34 <AnMaster> that is linux and freebsd
17:08:35 <oerjan> Deewiant: note that each edge connects a pair of vertices
17:08:36 <fizzie> Deewiant: I think the "odd number of tokens" is quite a clear clue that it expects pairs.
17:08:38 <AnMaster> man pages for tsort
17:08:39 <AnMaster> Deewiant, ^
17:08:46 <AnMaster> Deewiant, which one is most helpful?
17:08:54 <Deewiant> fizzie: should have been obvious I suppose, yes :-)
17:09:11 <Deewiant> AnMaster: BSD. what's your point? My point was that that kind of stuff should be in --help.
17:09:20 <Deewiant> or otherwise embedded in the executable.
17:09:27 <AnMaster> Deewiant, isn't in tsort --help on my linux
17:09:37 <Deewiant> AnMaster: What's your point? My point was that that kind of stuff should be in --help.
17:09:43 <AnMaster> well it isn't
17:09:53 <AnMaster> on gnu it is all in horrible info pages
17:09:57 <Deewiant> AnMaster: My point was that that kind of stuff ******should***** be in --help.
17:10:05 <AnMaster> Deewiant, yes, but show me a system where it is
17:10:09 <Deewiant> AnMaster: CCBI.
17:10:11 <Deewiant> for instance.
17:10:23 <AnMaster> Deewiant, I don't want to have to pipe --help into less
17:10:25 <Deewiant> I wasn't talking about operating systems, I was talking about programs.
17:10:50 <Deewiant> AnMaster: you'd rather pipe a man page into less?
17:10:58 <AnMaster> Deewiant, well yes
17:11:06 <AnMaster> Deewiant, since that is what man does
17:11:13 <Deewiant> oerjan: Yeah, I was hoping that one could give a vertex followed by a list of vertices it connects to
17:11:15 <AnMaster> man tsort
17:11:15 <AnMaster> or
17:11:19 <AnMaster> tsort --help | less
17:11:22 <AnMaster> or whatever
17:11:24 <Deewiant> AnMaster: Yes, and that's my point. It's the _same thing_.
17:11:29 <AnMaster> I think the first is shorter
17:11:30 <AnMaster> :)
17:11:35 <AnMaster> that was my point
17:11:38 <AnMaster> it is shorter to write
17:11:39 <Deewiant> Then have a short --help, and a --help-long.
17:11:45 <Deewiant> Whatever.
17:11:50 <AnMaster> Deewiant, ah like portage
17:11:54 <AnMaster> --help and --help --verbose
17:12:04 <AnMaster> -v/--verbose affects a lot of other stuff too
17:12:44 <AnMaster> Deewiant, now I think a good example of how to not do it: nmap --help
17:12:46 <AnMaster> it is just a clutter
17:12:49 <AnMaster> white nc -h
17:12:50 <AnMaster> is easy
17:13:11 <AnMaster> http://rafb.net/p/LUULdC31.html
17:13:22 <AnMaster> "SEE THE MAN PAGE FOR MANY MORE OPTIONS, DESCRIPTIONS, AND EXAMPLES" is how the nmap help ends
17:13:48 <Deewiant> nmap --help is pretty good IMO, could use more whitespace
17:13:55 <AnMaster> Deewiant, the nc one?
17:14:08 <AnMaster> I strongly prefer the nc -h
17:14:18 <Deewiant> AnMaster: unfair comparison, nmap supports hundreds of options, nc supports, looks like 20 or so
17:14:30 <Deewiant> well okay, probably not hundreds either
17:14:34 <AnMaster> Deewiant, also nmap says "see man page for more options"
17:14:34 <Deewiant> but lots more anyhoo
17:16:05 <AnMaster> Deewiant, anyway yes --help is good, but with too many options it breaks badly
17:16:18 <Deewiant> yes, hence --help --verbose
17:16:46 <AnMaster> Deewiant, I think the erlang solution would really piss you off
17:16:50 <AnMaster> erl -man erlc
17:16:53 <AnMaster> erl -man lists
17:17:06 <AnMaster> opens man with the relevant documentation page
17:17:26 <Deewiant> In general I guess best would be to have --help have the shortopts in the usual format and at most 20 lines of description of most important options
17:17:29 <AnMaster> oh and they are stored in a special place, not in standard man path, since most of them are for erlang modules, not tools
17:17:46 <AnMaster> Deewiant, erlang tools seem to like ignoring unknown options
17:17:56 <AnMaster> so you get erlang shell if you try --help
17:18:11 <AnMaster> also some command line parameters use + or such
17:18:16 <AnMaster> like +K true or +A 128
17:18:24 <Deewiant> But, you know, when you're talking about things like compilers + libraries it does make sense to have separate manuals because they can be hundreds of pages long.
17:18:53 <AnMaster> (+K = kernel pool, like epoll/kqueue and such, no idea why default is false; +A = Async threads for IO to prevent blocking erlang itself)
17:19:15 <AnMaster> Deewiant, yes... tried man zsh?
17:19:29 <AnMaster> "Because zsh contains many features, the zsh manual has been split into a number of sections"
17:19:50 <AnMaster> 17 man pages it seems here
17:20:05 <Deewiant> yep
17:20:18 <Deewiant> that's a good example, but the --help is terrible
17:20:29 <Deewiant> too much whitespace :-P
17:20:49 <AnMaster> (Använd '-v --help' för att visa kommandoradsflaggor för barnprocesser)
17:20:50 <AnMaster> for gcc
17:21:01 <AnMaster> which is where all the interesting options are
17:21:15 <AnMaster> $ gcc --help -v 2>&1 | wc -l
17:21:15 <AnMaster> 1487
17:21:16 <AnMaster> wow
17:21:23 <Deewiant> ew, non-english locale
17:21:37 <AnMaster> Deewiant, well yeah I want UTF-8, and that breaks in C locale
17:21:46 <AnMaster> there is no C.UTF-8
17:22:08 <Deewiant> just use en_GB
17:22:24 <AnMaster> Deewiant, with sv_SE.UTF-8 for some like date and so?
17:22:44 <AnMaster> LC_MONETARY, LC_PAPER, LC_MEASUREMENT and so on
17:22:54 <AnMaster> (type locale for the full list of variables)
17:23:20 <AnMaster> Deewiant, actually:
17:23:26 <AnMaster> export LC_MESSAGES=C
17:23:28 <AnMaster> issue solved
17:24:24 <AnMaster> Deewiant, what do you think of cfunge's -h?
17:24:55 <AnMaster> Deewiant, http://rafb.net/p/J2zlDX34.html
17:25:11 <Deewiant> fine apart from the "see README" bit ;-)
17:25:24 <AnMaster> Deewiant, well that is where I put most docs
17:25:44 <AnMaster> for efunge I should do something like:
17:25:47 <AnMaster> The options are a lie
17:25:52 <AnMaster> since it doesn't have any options
17:25:55 <Deewiant> :-P
17:26:00 <AnMaster> just want file name and script args
17:27:18 <AnMaster> actually a better idea:
17:27:36 <AnMaster> Usage: efunge [OPTIONS] [FILE] [PROGRAM OPTIONS]
17:27:45 <AnMaster> -h Show this help
17:27:51 <AnMaster> That's all :D
17:28:01 <AnMaster> Deewiant, is that what you want?
17:29:26 <Deewiant> Yeah, I guess :-P
17:30:05 <AnMaster> Deewiant, agree it seems slightly insane though?
17:30:14 <Deewiant> better than having no help
17:31:09 <AnMaster> Deewiant, anyway the recommended way to start efunge is different if you encounter a bug, then you should start it from inside erlang
17:31:24 <AnMaster> hm... maybe adding that to the wrapper script would be a good idea
17:31:38 <AnMaster> but then I'd get command line parameters
17:31:48 <AnMaster> ah I know...
17:31:50 <AnMaster> hehehehe
17:32:54 <AnMaster> EFUNGE_OPTS="[option1, option2]" ./efunge
17:33:00 <AnMaster> Deewiant, what do you think of that idea?
17:33:07 <AnMaster> it has to be done in erlang term format too
17:33:24 <Deewiant> I think it's terrible
17:33:42 <AnMaster> Deewiant, well it follows established tradition:
17:33:44 <Deewiant> for one thing because you can't do that in windows
17:33:45 <AnMaster> ERL_COMPILER_OPTIONS='[inline,native,{hipe,[o3]}]' make
17:33:47 <AnMaster> for example
17:33:59 <AnMaster> Deewiant, you can use set MYENVVAR="FOO" right?
17:34:02 <AnMaster> or something like that
17:34:08 <AnMaster> don't remember syntax
17:34:39 <AnMaster> Deewiant, also can't windows powershell do it?
17:34:44 <AnMaster> or is that so powerless?
17:34:51 <Deewiant> no, it can't, I just tried
17:34:57 <Deewiant> at least not with that syntax
17:35:01 <AnMaster> heh
17:35:02 <Deewiant> it could have it in some weird syntax
17:35:31 <AnMaster> Deewiant, well anyway the wrapper script is a shell script
17:35:35 <AnMaster> using portable /bin/sh
17:35:40 <AnMaster> it even works with busybox
17:35:49 <AnMaster> Deewiant, you would need a custom wrapper script on windows anyway
17:35:54 <Deewiant> so it requires posix and is thus non-portable :-P
17:35:59 <AnMaster> Deewiant, not at all
17:36:06 <AnMaster> you just have to start it in an erlang shell
17:36:11 <AnMaster> or write a *.bat or whatever
17:36:15 <AnMaster> Deewiant, just I can't write *.bat
17:36:20 <Deewiant> I meant the /bin/sh
17:36:46 <AnMaster> Deewiant, well... funge is unportable, it requires the concept of "environment variables"
17:37:16 <AnMaster> which is in the os module in Erlang, and nothing in that module is guaranteed to be portable
17:37:53 <Deewiant> funge doesn't require it, you can just push an empty list if they don't exist
17:38:06 <AnMaster> Deewiant, http://rafb.net/p/ftSSjH10.html Notice how extremely simple it is
17:38:15 <AnMaster> just some sanity checking and setting up some paths
17:38:27 <Deewiant> you can't write that in C?
17:38:42 <AnMaster> Deewiant, well no, would need fork and execv
17:38:48 <AnMaster> which would be less portable
17:38:53 <Deewiant> system()
17:38:56 <AnMaster> also using C for a wrapper script would be silly
17:39:22 <AnMaster> Deewiant, and /bin/sh exists on Windows. Wasn't it POSIX okish or such? ;P
17:39:44 <AnMaster> and then there is msys, cygwin and various other ones
17:39:50 <Deewiant> /bin/sh cannot exist on windows because "/" is not valid in a path
17:40:08 <AnMaster> Deewiant, well yet windows passed POSIX exam iirc? Or at least didn't fail it
17:40:15 <AnMaster> I believe it has to exist then
17:40:17 <Deewiant> only with extensions
17:40:23 <Deewiant> and hacks to workaround stuff like that
17:41:14 <Asztal> well, / works in paths, but it's not a valid root path without something like Services For Unix or msys
17:41:24 <Asztal> C:/windows/system32/notepad.exe works
17:41:24 <Deewiant> no, / doesn't work in paths
17:41:33 <Deewiant> yeah, but the path isn't that
17:41:36 <AnMaster> yes it does in cmd.exe
17:41:36 <Deewiant> that's just translated
17:41:44 <Deewiant> no it doesn't
17:41:48 <Deewiant> it specifically doesn't work in cmd.exe
17:41:55 -!- Judofyr_ has joined.
17:42:06 -!- Judofyr has quit (Read error: 104 (Connection reset by peer)).
17:42:20 <AnMaster> Deewiant, anyway 1) you can write a *.bat file I guess. or 2) You could just do it from inside erlang as the README describes
17:42:23 <Deewiant> "cd /" will not work
17:42:40 <Deewiant> yes, it's /possible/
17:42:44 -!- Judofyr_ has changed nick to Judofyr.
17:42:45 <Asztal> cd / works
17:42:58 <Deewiant> not here it doesn't
17:43:01 <AnMaster> Deewiant, I'm pretty sure cd / works in cmd.exe under windows xp
17:43:05 <Asztal> (I didn't expect it to, but it did)
17:43:07 <AnMaster> I'm 99% certain I seen it working
17:43:13 <Deewiant> I'm pretty sure it doesn't since I just did it and it didn't
17:43:15 <AnMaster> using Swedish windows xp home
17:43:25 <Deewiant> "cd /" does absolutely nothing, doesn't error out either though
17:43:29 <AnMaster> was a year or so ago
17:43:35 <Asztal> both the notepad thing and cd / worked for me just now
17:43:44 <AnMaster> Asztal, windows version?
17:43:46 <Deewiant> "cd foo/bar" does work
17:43:48 <Asztal> this is the evil vista
17:43:51 <AnMaster> ah
17:43:57 <Deewiant> ah, it's a bug then :-P
17:44:11 <AnMaster> Well, there is no spec to check against is there Deewiant?
17:44:14 <Asztal> though, I was sure XP supported it :)
17:44:21 <AnMaster> anyway...
17:44:27 <Deewiant> AnMaster: there may be docs
17:44:53 <AnMaster> Deewiant, doing this on windows would be non-trivial anyway
17:45:00 <AnMaster> since usually erlang isn't in path there
17:45:05 <AnMaster> and it called werl.exe
17:45:37 <Deewiant> interesting that 'erl' starts an Eshell then
17:45:54 <AnMaster> Deewiant, hm it does? Doesn't match what I was able to find in docs
17:46:08 <AnMaster> Deewiant, what version does it state that it is btw?
17:46:13 <Deewiant> 5.5.5
17:46:21 <Asztal> (I suspect cmd.exe is doing something to explicitly support '/' as a filesystem root, though, since "ls.exe /Users" doesn't work)
17:46:27 <AnMaster> Deewiant, I don't think efunge will work on that, Only tested on 5.6.4
17:46:32 <Deewiant> hmm
17:46:33 <Deewiant> Eshell V5.5.5 (abort with ^G)
17:46:33 <Deewiant> 1> ^G
17:46:33 <Deewiant> Eshell V5.5.5 (abort with ^G)
17:46:33 <Deewiant> 1>
17:46:40 <AnMaster> Deewiant, XD
17:46:43 <AnMaster> Deewiant, q().
17:46:45 <AnMaster> may work
17:46:49 <AnMaster> takes a second or so
17:46:51 <Deewiant> ^C worked
17:46:53 <AnMaster> before it actually quits
17:46:56 <Deewiant> but, hoorays for ^G
17:47:00 <AnMaster> Deewiant, well, I ^G works for me
17:47:06 <AnMaster> 1>
17:47:06 <AnMaster> User switch command
17:47:06 <AnMaster> -->
17:47:17 <AnMaster> ^G drops me into that mode
17:47:45 <AnMaster> where you can start another shell on a local or remote node, jump between existing shells and so on
17:47:55 <AnMaster> and quit erlang
17:48:04 <AnMaster> Deewiant, no clue why it doesn't work on windows.
17:48:25 <AnMaster> Asztal, there is ls.exe?
17:48:27 <AnMaster> huh
17:49:02 <AnMaster> Deewiant, anyway I'm quite sure erlang will need at least BEAM 5.6.x
17:49:10 <AnMaster> probably even 5.6.4
17:49:14 <Deewiant> s/erlang/efunge/
17:49:15 <Deewiant> why
17:49:16 <AnMaster> (which is last version as of writing this)
17:49:26 <AnMaster> Deewiant, because it uses some very very useful recent features
17:49:52 <Deewiant> like what
17:50:11 <AnMaster> -spec and -type notification are experimentally supported in 5.6.4 at least. Allows better static analysis and easier to find bugs in advance
17:50:22 <AnMaster> sed could remove those if needed I guess
17:50:33 <AnMaster> Deewiant, apart from that it may or may not use other bits I don't know about
17:50:47 <AnMaster> but I think Erlang 5.5 is a different major release
17:50:48 <AnMaster> R11B
17:50:54 <AnMaster> while 5.6 is R12B
17:51:04 <AnMaster> and those can introduce major new features
17:51:23 <AnMaster> Deewiant, so it wouldn't surprise me if it broke on the legacy R11B
17:51:41 <AnMaster> and my response would be: SOL
17:51:47 <AnMaster> since that is outdated by years
17:52:30 <oerjan> statute of limitations? :D
17:52:34 <AnMaster> oerjan, ?
17:52:36 <Deewiant> well, by one year
17:52:42 <AnMaster> and a few months
17:52:46 * AnMaster just went and checked
17:52:53 <oerjan> AnMaster: SOL
17:53:03 <AnMaster> oerjan, S*** out of luck
17:53:19 <AnMaster> where * count doesn't match real word
17:53:33 <AnMaster> I think
17:53:37 <AnMaster> actually it does
17:53:39 <AnMaster> oh well
17:54:06 <oerjan> i think there are various options there
17:54:11 <Deewiant> I've always read it as "so outta luck"
17:54:21 <AnMaster> Deewiant, that works too
17:54:27 <Deewiant> but I guess "shittily" is the rude form and "sadly" the more common polite form
17:54:32 <AnMaster> anyway: you are free to try efunge on R11B, but I won't fix it breaks, recommending R12B
17:54:57 <Deewiant> evidently "shit" is the common form, although it doesn't make grammatical sense
17:54:57 <AnMaster> when R13B is released I will probably keep R12B compat for a few months or a bit more at least
17:55:02 <Deewiant> but then, expletives rarely do
17:57:17 <fizzie> Come to think of it, I could just run fungot's babble-generation locally repeatedly to find that bug faster than just waiting for it to crash.
17:57:19 <fungot> fizzie: madam president, i should briefly like to focus on the most fundamental human rights.
17:57:55 <fizzie> fungot: Er, certainly, I don't think we have anything against human rights here.
17:58:05 <fungot> fizzie: first and foremost, stability and security, which already give the commission something to consider, when commissioner palacio brings forward her nuclear package this autumn, that the poorest regions is not exacerbated by another divide linked to the activities of the european union
17:58:24 <fizzie> (It's quite a lot slower with that trace mode enabled.)
17:58:27 <Deewiant> fungot: Right, exactly.
17:58:39 <Deewiant> maybe I should try and reproduce the bug
17:58:43 <fungot> Deewiant: mr president, mr poms ruiz, to whom they really do, as the elected representative of a battle of principles, you will find it more difficult to resolve, but it is essential to finding a solution to this particular matter and my views were clearly laid out. nevertheless, i hope that, by taking part in what was said at the first reading.
17:58:45 <Deewiant> fungot: WTF
17:59:09 <fungot> Deewiant: madam president, as an indirect consequence, holding us prisoner on other subjects that are excluded from so many contradictory standpoints. it would be too expensive. in this respect, i should like to remind her and the house' s work just a few days ago here to introduce a uniform control system which is more fnord and export a certain model of society that suits them. it would be very interesting for economists to s
17:59:16 <Deewiant> fungot: WTF
17:59:25 <fizzie> It's funny how it alternates between madam/mr all the time.
17:59:30 <Deewiant> fizzie: hmm, maybe you could make it cutoff messages in a smarter way
17:59:32 <oerjan> "clearly laid out" my ass
17:59:35 <Deewiant> fizzie: i.e. not mid-word
17:59:45 <fungot> Deewiant: mr president, i would say to those who need to disclose one's inventions before filing for a patent, its aim is that, in 2004, especially in mr poignant' s report, certain elements of the tens, but we cannot ignore. the us government has threatened to move its entire development department and corporate management to another country if it cannot put a stop to. secondly, the establishment of a european parliament and t
17:59:48 <Deewiant> fungot: WTF
17:59:49 <fungot> Deewiant: mr president, the discussion is so interesting and there are a number of people.
17:59:54 <Deewiant> meh
17:59:59 <Deewiant> fungot: bug out already, will ya
18:00:19 <fizzie> Deewiant: That's just the IRC message length. There's also a limit to the number of words, which should be small enough not to hit the IRC limits.
18:00:44 <fizzie> Maybe it's all those long words.
18:00:48 <Deewiant> fizzie: Yes, I know what it is, but as you can see it keeps getting cut off :-)
18:00:53 <fizzie> I don't remember what I set the word limit to.
18:01:22 <Deewiant> count the message length and cut off at a word boundary where the wanted message length is exceeded
18:01:28 <Deewiant> fizzie: hmm, did I kill it?
18:01:44 <oerjan> fungot: you alive?
18:01:47 <fungot> oerjan: as i demonstrated during the debate. as regards priority 1, namely broadening the scope and structure of the directorate-general for transport essentially deals with this very difficult issue. a solution must be sought initially on a voluntary basis. it is a pity.
18:02:15 <oerjan> in a manner of speaking
18:02:19 <Deewiant> fungot: WTF
18:02:22 <fungot> Deewiant: mr president, many pensioners paint or create works of art and antiques'. according to a report by the independent opinion of the committee, issues on labour law and free and fair way, in spite of what mr bouwman said. in fact, with the game of musical chairs that is being done.
18:02:33 <Deewiant> fungot: I'm not a president
18:03:02 <fizzie> Actually it should be stopping quite early if I'm reading that source right. (Except that it will only stop when the context is suitable for ending a sentence.)
18:03:05 <fungot> Deewiant: mr president, ladies and gentlemen, first of all that is bad comes from elsewhere; this is why i am against increasing taxation on employment, under the present legislation. both treaties are based on different understandings of human health and the environment, but also for animal welfare where many millions of new jobs as it did not do is to base programming on a political solution to end spam. let me also address t
18:03:16 <Deewiant> fungot: I'm still not a president
18:03:23 <fungot> Deewiant: question no 30 by ( fnord/ 2001. i can allow the exemption for mortgages are all adopted, and comparable objectives have been set, but whose idea of political responsibility observed during the process of 2004.
18:03:38 <oerjan> Deewiant: that's just what you want us to believe, Ms. Halonen
18:03:39 <Deewiant> fungot: Who's fnord
18:03:43 <fungot> Deewiant: it is easy to see that there is a job we will have to be addressed: greece, spain, greece and spain.
18:04:06 <Deewiant> oerjan: oh noes, you caught me
18:04:14 <fizzie> Haha: "Greece, Spain, Greece and Spain -- a joh we will have to address!"
18:04:20 <Deewiant> fungot: WTF
18:04:49 <oerjan> fizzie: it's counting commas as tokens?
18:05:01 <Deewiant> fungot: WTF
18:05:02 <fizzie> oerjan: Yes, punctuation is separate tokens.
18:05:21 <fizzie> So it's just got a context of ", spain," when it generated that second 'greece'.
18:05:52 <oerjan> so i thought
18:06:00 <fizzie> fungot: Did you just ignore Deewiant, or crash?
18:06:23 <fungot> fizzie: mr president, the pensioners of the different legal frameworks would probably make it very clear that the eu must develop a process, namely a fnord victory for president eyadema. on a financial basis, nor to care about their neighbour. the socialists say yes. allow me to outline the way in which you will accept these amendments, nos 1, 2, 5 and 6 i can support the draft act which is before us here today, of her excellen
18:06:29 <Deewiant> waa
18:06:35 <Deewiant> I don't like the politician-fungot
18:06:36 <fizzie> Deewiant: Yours was a fnord victory.
18:06:38 <fungot> Deewiant: i wish to ask you, mr president-in-office, you did hear me say that i fully endorse what mr salafranca has just said.
18:06:49 <Deewiant> fungot: now who the hell is mr salafranca
18:06:57 <fungot> Deewiant: mr president, i was in fact our motion for a resolution calling for a committee of inquiry into the current cohesion policy. it is my opinion that the controls that are applied to economies whose needs remain very different.
18:07:13 <Deewiant> fungot: you didn't answer my question
18:07:14 <fungot> Deewiant: mr president, i am sorry, in particular candidate countries.
18:07:16 <fizzie> Deewiant: That would be José Ignacio Salafranca Sánchez-Neyra.
18:07:38 <oerjan> not in wikipedia
18:07:45 <Deewiant> oerjan: http://en.wikipedia.org/wiki/Jos%C3%A9_Salafranca_S%C3%A1nchez-Neira
18:07:45 <fizzie> http://en.wikipedia.org/wiki/José_Salafranca_Sánchez-Neira
18:07:49 <oerjan> (english one)
18:08:04 <fizzie> The page title doesn't include the "Ignacio" part.
18:08:05 <Deewiant> fungot is sorry; but only in particular candidate countries
18:08:10 <oerjan> oh wait
18:08:21 <fungot> Deewiant: mr president, mr president, is extremely limited. the commission feels that to increase the number of subjects, the european council decided at its meeting on 30 april, or even this very budget assuming the administrative costs. moreover, i do not believe it would have been preferable and more effective fnord and monitoring powers. your support for our approach as set out in the end to violence, subjected to terrible
18:08:24 <oerjan> wp only searches for starting parts?
18:08:43 <Deewiant> fungot: Yes, I think it would have been less effective fnord
18:09:41 <fizzie> I think you're being ignored again. Okay, I don't think it makes sense to do more of this on-channel; I'll test it elsewhere.
18:10:08 <Deewiant> fizzie: I suspect you can reconstruct, from your logs, everything that was said to it before it crashed
18:10:18 <fizzie> Yes, but it shouldn't matter.
18:10:22 <Deewiant> so just keep throwing that at it
18:10:23 <fizzie> Since it doesn't look at the input at all.
18:10:27 <Deewiant> ah.
18:10:43 <oerjan> Eyadema was an interesting guy too
18:11:46 <oerjan> Gnassingbé is such a household name
18:16:51 <AnMaster> Deewiant, haven't you read fungot source?
18:16:59 <AnMaster> http://zem.fi/~fis/fungot.b98.txt
18:17:18 <AnMaster> Deewiant, being a befunge expert (author of mycology) I assume you will understand it right away :P
18:17:32 <AnMaster> (like you see to expect me to do with mycology)
18:17:42 <AnMaster> bbl food
18:17:46 <fizzie> Should probably make a real web page for fungot.
18:17:46 <fungot> AnMaster: mr president, commissioner, for this legislature. we have member states who so wish, to reduce pollution by 60% over the last 20 years this industry has one of the questions, so i shall not go into the details of your speeches, ladies and gentlemen, on behalf of the commission to ensure that the principle of mutual recognition of decisions to expel third-country citizens. i congratulate joe fnord and fnord and catalon
18:18:48 <fizzie> I didn't know there was a Joe Fnord in the Parliament. And I wonder if it crashed; it's certainly taking a long time to answer.
18:18:53 <fungot> AnMaster: mr president, i want to stress the extremely high social expectations, we are being ideological. they are told off for living too long and that it is only natural, when already in a europe that is not industry which determines the courts that involve an interpretation of the new policies, such as tag replacement, without adding any ifs or buts. the report reveals a motley fnord of organisations, which are far from mat
18:18:58 <fizzie> No, it's just the slow.
18:19:20 <fungot> fizzie: mr president, with the participation of the president-in-office. of course, under the existing substance regulations of 1993. the final point i am addressing our fellow parliamentarians from macedonia should be able to have olaf's report so that chapters for the individual sectors. secondly, we attach great importance to us that there could be this consensus and that the meat is from animals which were tested and whethe
18:20:06 <oerjan> ^echo are we here yet?
18:20:07 <fungot> are we here yet? are we here yet?
18:20:38 <Asztal> fungot, your prolixity is overly quixotic >:(
18:20:53 <fungot> Asztal: madam president, mr posselt, let me say that we should take the decision we need to look at the simple trend, i have one difficulty which is not covered by specific community legislation, so ias can only be achieved by applying lower rates of growth, employment and the unwieldiness of the procedures, favouring a more competitive and makes his company more competitive. everyone agrees on that, are dying as a result of ad
18:21:59 <oerjan> presumably http://en.wikipedia.org/wiki/Bernd_Posselt
18:23:33 <oerjan> also an interesting guy it seems...
18:26:45 <fizzie> The bot does sound quite much like a politician, but maybe it's not overly interesting babbling.
18:27:26 <AnMaster> fizzie, you want low-overhead tracing?
18:27:49 * AnMaster considers that
18:28:10 <AnMaster> interesting idea
18:28:38 <AnMaster> you want to append them to a buffer or such, and then let another thread format the messages
18:28:57 <AnMaster> (this would be a bad idea in erlang, because every thread have a separate heap
18:29:41 <AnMaster> (so a lot of data copying)
18:32:58 <fizzie> Heh, nice output: "let us not talk of positive discrimination more effective than for other types of intensive animal husbandry, the commission is willing to launch pre-emptive nuclear strikes"
18:33:09 <fizzie> Pre-emptive nuclear strikes does not sound like the right solution to me.
18:33:18 <fizzie> fungot: Are you sure about that?
18:33:19 <fungot> fizzie: without being afraid of the difficulties currently facing the union.
18:33:34 <AnMaster> Ouch that bad
18:33:40 <AnMaster> fungot, ...
18:33:41 <fungot> AnMaster: madam president, with reference to mr pronk on his report which is evidently to establish the existence of the internal market.
18:33:47 <fizzie> Hopefully no-one's going to put fungot in charge.
18:33:52 <AnMaster> fungot, you got my gender wrong
18:33:55 <fungot> fizzie: mr president, i welcome the resolve that was shown here today by the college of commissioners has already decided, was literally betrayed at the last moment, i woke up, i am happy to support, veiled criticism and open criticism. this evening's debate is certainly the case for amendments nos 3 and 4, so that we can all see the work we do with small-scale fishermen. i await mr henderson's answer on burma was rather meanin
18:34:02 <fungot> AnMaster: mr president, i am aware that this is all the more necessary to seek equal treatment for agency workers in ireland.
18:34:05 <AnMaster> better
18:34:39 <oerjan> who is mr pronk?
18:34:48 <AnMaster> fizzie, where did you get this data from did you say?
18:35:12 <AnMaster> oerjan, what about mr henderson?
18:35:32 <oerjan> ah, maybe Jan Pronk
18:35:39 <fizzie> oerjan: Bartho Pronk, actually.
18:35:47 <fizzie> (PPE-DE)
18:35:50 <AnMaster> eh?
18:36:11 <fizzie> AnMaster: It's the European Parliament speeches. Those are translated to all N official languages; I just used the English version.
18:36:12 <oerjan> hm not in wp
18:36:16 <AnMaster> fizzie, I would prefer telephone or discworld over this
18:36:45 <fizzie> Years 1996-2006, apparently.
18:37:04 <AnMaster> fizzie, do they speak to presidents all the time?
18:37:07 <fizzie> And only those speakers whose names start with a letter from A to F.
18:37:26 <fizzie> It's the President of the Commission, I think.
18:37:42 <AnMaster> ah
18:38:00 <oerjan> or parliament?
18:38:14 <fizzie> Yes, actually.
18:38:29 <fizzie> I just grepped the data a bit and saw "AFFILIATION="President of the Commission"", but that was just one of the speakers.
18:40:03 <oerjan> henderson is probably too common a name
18:40:22 <fizzie> Just about every speaker starts with "Mr President" or some-such, and since we grouped those so that one "paragraph" (which is what fungot outputs) is one complete "speech"/statement, that's what fungot does too.
18:40:26 <fungot> fizzie: mr president, i wish to put on record that our starting point and encouraging that a solution is not very effective.
18:41:16 <oerjan> that's relevant, sort of
18:41:29 <fizzie> fungot: You're such a complainer.
18:41:41 <fungot> fizzie: mr president, mr camisn asensio, things are really up and running for 8 years, we have to realise that this is our parliament, a citizen many of which are very sensible. that very fact illustrates the european vagueness of the report.
18:42:25 <fizzie> Yes, the politician-fungot is not very funny at all. But at least it sounds like it's speaking and not like it's reading a book.
18:42:34 <fungot> fizzie: mr president, i will begin with the nine technical amendments. amendments nos 12, 65, 66, 67, 69, 70 and 71. the possible effects on other community policies, new proposals will have to submit every three years.
18:43:32 <oerjan> but he cannot count
18:44:13 -!- KingOfKarlsruhe has joined.
18:46:48 <fizzie> I've been looking at the Gutenberg authors I happen to have here, but none of them sound like they'd make an especially interesting data source.
18:51:25 <Deewiant> fizzie: try lewis carroll
18:51:39 <Deewiant> and james joyce
18:52:59 <AnMaster> lewis carroll should be fun
18:53:40 <fizzie> fungot: Well, you heard them. Do Lewis Carroll now.
18:54:09 <fizzie> Or are you ignoring me or something?
18:54:59 <fizzie> ^rot13 fgvyy jbexf
18:55:22 <fungot> still works
18:55:28 <fizzie> Ignoring, I guess.
18:55:34 <Deewiant> fungot: Bah.
18:55:42 <fizzie> Oh, wait!
18:55:47 <fizzie> No read permissions to those files.
18:55:52 <fizzie> Forgot to give g+r to them.
18:56:14 <fizzie> fungot: Can you read it now?
18:56:15 <fungot> fizzie: the witnesses proved, without fnord,' he went on, " the conclusion is fallacious."
18:57:20 <fizzie> fungot: What's that mean?
18:57:22 <fungot> fizzie: he saw them drooping here and there among them, though i don't want to get somewhere else, you must run at least twice as fast as they put them upon the table, merrily running round and round after her own shawl, which was still patiently fnord its fnord!"
18:58:38 -!- Judofyr has quit (Read error: 104 (Connection reset by peer)).
18:59:14 <oerjan> the fnord solution may not work very well with an author reknown for inventing words...
18:59:15 -!- Judofyr has joined.
19:00:01 <fizzie> oerjan: It mapped 5703 words to "fnord".
19:00:33 <oerjan> is that much?
19:00:44 <AnMaster> Deewiant, by the way to show that I wouldn't oppose MVRS done correctly and properly specced my funge-space changes for ATHR are designed so the change for adding multiple funge spaces should be simple
19:00:57 <AnMaster> I'm designing the API with that in mind
19:01:42 <AnMaster> fizzie, what books?
19:01:56 <oerjan> fungot: say 'cheese'
19:01:57 <fungot> oerjan: there was no " one, two! and through and through her head till she felt quite certain he would only refer to his fnord code, and read it " some x are y"?
19:02:24 <fizzie> oerjan: It's 3.4 % of the tokens in the data. For comparison purposes, in the irclogs it mapped only 1.2 % of the tokens to fnord. And irclogs have lots of unique flutzpahs too.
19:02:30 <AnMaster> Only books by Lewis Carroll I read are Alice in wonderland, and Allice through the looking glass
19:02:34 <AnMaster> both in Swedish translations too
19:02:37 <oerjan> "some x are y"? are you including his math? :D
19:02:41 <AnMaster> so. I can't identify anything
19:03:18 <fizzie> The books were: "Alice's Adventures in Wonderland", "Phantasmagoria and Other Poems", "Sylvie and Bruno", "The Game of Logic", "The Hunting of the Snark" and "Through the Looking-Glass"
19:03:32 <AnMaster> "The Hunting of the Snark"?
19:03:37 <oerjan> the game of logic looks suspicious
19:03:42 <fizzie> The "some x are y" is from the game of logic, yes.
19:03:51 <AnMaster> sounds like hunting of the fnord...
19:04:30 <Deewiant> fizzie: remind me what the fnord-mapping was all about
19:04:58 <fizzie> Deewiant: All words that occur only once are mapped to a special token "UNKNOWN", and fungot converts that to text as "fnord".
19:04:59 <fungot> fizzie: " i will call on the earl.
19:05:20 <Deewiant> fizzie: Why?
19:05:22 <oerjan> snark surely occurs more than once though
19:05:40 <fizzie> And Snarks have been seen at least in a short story by Niven, it's somewhat famous.
19:06:01 <fizzie> Deewiant: To cut down the size of the lexicon, mostly. Also to make the output more silly.
19:06:37 <Deewiant> meh
19:06:40 <oerjan> snarks? not bandersnatchi?
19:07:38 <fizzie> oerjan: Those are everywhere in the Known Space books, but I think there was a Snark too.
19:07:57 <fizzie> I'm not absolutely certain, though.
19:09:46 <fizzie> oerjan: At least in the story "Like Banquo's Ghost" (from the Convergent Series collection) there's a space ship called "Snarkhunter #3".
19:10:19 <AnMaster> huh. fizzie you make *less* sense than fungot now..
19:10:20 <fungot> AnMaster: all this i saw from the open window of the warden's breakfast-saloon, looking across the garden, examining the fastenings of the drawing-room window.
19:10:31 <fizzie> No other Snarks, though; I've misremembered, it seems.
19:10:52 <AnMaster> fizzie, which books are that from ^
19:11:14 <AnMaster> is*
19:12:32 <fizzie> AnMaster: Seems to be mostly from that "Sylvie and Bruno" book. Haven't read it.
19:13:54 <fizzie> Still, whenever I feed it books the output also sounds like a book, not like someone talking.
19:16:21 <fizzie> I should feed it some chatting, but I have a smaller supply of that available.
19:16:27 <fizzie> Maybe I should try Wikipedia "talk" pages. :p
19:17:27 <AnMaster> fizzie, no... that would be suicidal amount of "consensus version"
19:17:32 <AnMaster> be a*
19:18:10 <fizzie> Youtube video comments, then; I hear there's a lot of thoughtful debate there.
19:18:40 * oerjan gasps
19:21:18 <AnMaster> fizzie, that is *worse*
19:21:37 <AnMaster> oh and even worse: /b/ of four chan
19:21:38 <AnMaster> I guess
19:21:45 <fizzie> Or maybe spam! I think I gzipped some ten thousand messages somewhere just in case I need a sample spam mail sometimes.
19:22:01 <AnMaster> fizzie, what department were you at?
19:22:06 <AnMaster> also spam changes over time
19:22:07 <AnMaster> a lot
19:22:29 * oerjan wonders what fizzie is taking
19:22:31 <AnMaster> fizzie, and we don't want all the obscene stuff
19:22:33 <fizzie> Yes, well, not for effective spam *filtering*, just generally.
19:23:08 <fizzie> Don't you ever get a craving for some spam?
19:23:25 <oerjan> SPAM SPAM WONDERFUL SPAM
19:23:33 <AnMaster> fizzie, no
19:26:16 <ehird> lol, a /b/ markov chain would be indistinguishable from the real thing.
19:27:16 <AnMaster> ehird, yeah
19:28:09 <ehird> I've always thought that the best way to run a markov chain is to analyze a forum/mailing list, make it post on that mailing list with its results, and start up about 10 instances.
19:28:16 <ehird> After a while they just start spewing out nonsense based on their own failures
19:28:32 <fizzie> Okay, I'll try one more author, let's see if it (a) works and (b) you recognize it. The source text at least should be pretty recognizable.
19:28:38 <fizzie> fungot: Let's see what you've got.
19:28:40 <fungot> fizzie: fnord with a fnord negroid mouth, pulled forth a dirty, crumpled paper and handed it to me, for there are elements i had not independently suspected before, though we had no cause to think the regions beyond the range of modern human knowledge.
19:29:49 <fizzie> fungot: Give us a second example too, okay?
19:29:50 <fungot> fizzie: miss tilton, comparing all possible fnord and if we be fnord and besides, was not beyond normal credibility.
19:30:01 <fizzie> Well, that wasn't very useful.
19:30:08 <fizzie> Pretty fnordy text this time.
19:35:13 <oerjan> fungot: huh?
19:35:14 <fungot> oerjan: even the small piece refused to grow cool, it soon had the college in a state of
19:35:27 <oerjan> fungot: good grief
19:35:30 <fungot> oerjan: by h. p. lovecraft and c. m... eddy jr.
19:35:37 <fizzie> fungot: Hey, you gave it away!
19:35:42 <fungot> fizzie: at my violent start the speaker paused a moment before there had been a gorgeous sunset, and judged it was a million or ten million or fifty million years ago and that of only thirty million years old.
19:35:53 <oerjan> and i was _just_ guessing lovecraft too :(
19:36:18 <oerjan> (although i haven't read any)
19:36:20 <fizzie> Should've stripped those parts out of them books, but couldn't be bothered.
19:42:05 -!- Mony has quit ("Ne regrettons pas le passé, ne nous préoccupons pas de l'avenir : le sage vit dans le présent.").
19:47:48 -!- Slereah_ has joined.
19:59:32 -!- Corun has quit ("Leaving").
20:09:10 -!- Slereah_ has quit (Read error: 54 (Connection reset by peer)).
20:09:13 -!- Slereah_ has joined.
20:14:33 -!- Slereah has joined.
20:14:33 -!- Slereah_ has quit (Read error: 104 (Connection reset by peer)).
20:18:13 -!- Slereah has quit (Read error: 104 (Connection reset by peer)).
20:21:11 -!- Slereah has joined.
20:42:48 -!- atrapado has joined.
20:57:10 -!- Linus` has joined.
21:00:14 -!- Slereah_ has joined.
21:00:14 -!- Slereah has quit (Read error: 104 (Connection reset by peer)).
21:07:08 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | <CTCP>PING<CTCP>.
21:08:49 <oerjan> ah the wiki is back
21:10:36 <psygnisfive> hey oerjan
21:11:32 -!- LinuS has quit (Connection timed out).
21:13:06 <oerjan> hi psygnisfive
21:13:44 <oklopol> hi oerjan
21:13:46 <oklopol> hi psygnisfive
21:13:55 <psygnisfive> oklopol <d
21:13:57 <psygnisfive> <3
21:14:33 <oerjan> hi oklopol
21:17:59 <oklopol> was wondering about turning boolean expressions into maxterm normal form
21:18:04 <oklopol> after ages of thinking and pondering
21:18:10 <oerjan> maxterm?
21:18:20 <oklopol> i realized the general case just amounts to enumerating the whole truth table :|
21:18:38 <oklopol> anding ors
21:19:43 <oklopol> maxterms are when you and or clauses, kinda taking the maximum of the truth values of each clause
21:19:50 <oklopol> minterms are when you and them
21:20:03 <oklopol> the two normal forms
21:20:18 <oklopol> i need the maxterm one so i can convert to 3-sat
21:21:51 <oklopol> the problem with making unimplementable turing tarpits is even *i* feel i'm wasting my time
21:21:54 <oklopol> only slightly though
21:22:09 <psygnisfive> hey oklopol, oerjan, didnt you want to learn something about like.. syntax??
21:22:14 <oerjan> well the conversion can blow up exponentially
21:23:03 <oklopol> btw is there anything new and interesting on the wiki?
21:23:07 <oklopol> i haven't read it for ages
21:23:16 <oklopol> only recently got reinterested in esolangs
21:23:54 <oklopol> well, actively reinterested, i'm always *interested*, but i'm not always using my brain for esolang-related ponderings
21:23:59 <oklopol> psygnisfive: NO
21:24:14 <psygnisfive> i remember you and oerjan saying something about itd be interesting to learn such and such. :|
21:24:48 <oklopol> all is interesting to learn that is abstract and useless.
21:24:51 * oerjan has a fever
21:25:21 <psygnisfive> lots of people think syntax is useless :P
21:25:24 <oklopol> i don't recall saying anything like that, i have only asked about *phonetics*
21:25:37 <oklopol> do they really? then perhaps i should try it
21:27:17 <oklopol> oerjan: do you have insights on the 3-sat conversion?
21:28:12 <oerjan> well you introduce extra variables, i think
21:28:25 <oklopol> hmm
21:28:40 <oklopol> (a v b)
21:28:41 <oklopol> =>
21:28:52 <oklopol> (a v b v c) ^ (a v b v C)
21:29:02 <oklopol> (using noprob negation)
21:29:18 <oklopol> and C can actually be used as the temp of any amount of clauses
21:29:25 <oklopol> because why couldn't it
21:29:40 <oerjan> er i'm not sure of _that_
21:29:45 <oerjan> oh wait
21:29:52 <oerjan> in the case you show there, it can
21:30:17 <oerjan> but not for handling splitting
21:30:31 <oklopol> yeah
21:30:34 <oklopol> at splitting
21:30:36 <oklopol> like
21:30:42 <oklopol> (a v b v c v d)
21:30:47 <oklopol> you'd to something like
21:30:48 <oerjan> (a v b v c v d) => (a v b v e) ^ (c v d v E)
21:30:56 <oklopol> err
21:31:01 <oklopol> oh
21:31:10 <oklopol> i didn't think of that
21:31:10 <oerjan> not sure if i got that right
21:31:28 <oklopol> that looks right
21:31:37 <psygnisfive> ah yes phonetics
21:31:39 <oklopol> but i can't put into words why
21:31:39 <psygnisfive> thats what it was
21:31:52 <psygnisfive> also, oklopol
21:32:00 <oerjan> if e is true, then c or d must be true
21:32:09 <oerjan> if e is false, then a or b must be true
21:32:09 <psygnisfive> theres a theory of syntax from a guy at UPenn, Aravind Joshi, called Tree Adjoining Grammar
21:32:14 <psygnisfive> its all about tree rewriting :D
21:32:31 <oerjan> and e can be chosen arbitrarily, but here it is important that it is used nowhere else
21:32:37 <oklopol> yes
21:32:51 <oklopol> psygnisfive: tree rewriting!!
21:32:57 <psygnisfive> yeah its cool :o
21:33:06 <oklopol> pretty cool yeah
21:33:10 <psygnisfive> it lets you do all sorts of crazy shit with non-local dependencies
21:33:12 <oklopol> but how about REWRITING HYPERGRAPHS?!?
21:33:18 <psygnisfive> well
21:33:30 <psygnisfive> i do know that some models of dependency grammar are based on /multigraphs/
21:33:38 <psygnisfive> i dont know what a hypergraph is tho
21:33:44 <oklopol> oerjan: seems you had quite a lot of insight, thanks
21:33:57 <oklopol> i probably wouldn't have come up with that without giving it tons of thought
21:34:18 <psygnisfive> oh i see what a hypergraph is
21:34:19 <psygnisfive> interesting
21:35:21 <psygnisfive> hm
21:35:31 <psygnisfive> i think i might be using hypergraphs in my syntactic formalism
21:36:01 <psygnisfive> maybe not.
21:36:02 <oklopol> (a v b v c v d v e v f v g), i should just split this as evenly as possible, right? (a v b v c v d v temp1) (e v f v g v Temp1)
21:36:03 <psygnisfive> i dont know. D:
21:36:12 <oklopol> and then recurse
21:36:18 <psygnisfive> oklopol whatchu doin?
21:36:38 <oklopol> psygnisfive: i'm doing small-scale research so i can start implementing noprob
21:36:49 <psygnisfive> what on
21:38:17 <oerjan> oklopol: well i've seen the reduction SAT -> 3SAT at some time...
21:38:23 <oklopol> (a v b v c v d v temp1) ^ (e v f v g v Temp1) ==> (a v b v temp4) ^ (c v temp2 v Temp4) ^ (d v temp1 v Temp2) ^ (e v f v temp3) ^ (g v Temp1 v Temp 3)
21:38:48 <oklopol> basically just recursing on the resulting subclauses
21:39:07 <oklopol> psygnisfive: what on what?
21:39:17 <psygnisfive> what are you researching
21:39:46 <oklopol> i'm researching the subject of <the thing i'm asking oerjan about>
21:39:52 <psygnisfive> :P
21:39:57 <oerjan> oklopol: looks good, of course there may be some more elegant way
21:40:03 -!- ryanakca has quit (Read error: 104 (Connection reset by peer)).
21:40:48 <oklopol> oerjan: well it's a logarithmic growth in size
21:40:55 <oklopol> so i'd say it's enough
21:41:13 <oklopol> as i'm already doing exponential work when creating the sat...
21:42:12 <oklopol> umm
21:42:16 <oklopol> or is it logarithmic...
21:42:47 <oklopol> the reduction is O(lg n) deep, because i always split the clause
21:42:59 <oklopol> but i'm also doubling the work each time
21:43:15 <oklopol> because there are twice as many clauses on the next level
21:43:43 <oklopol> so O(lg n) steps, exponential growth in size of one step, that would make O(n) in layman's math
21:43:48 <oerjan> when creating the sat, do you mean reducing from general boolean expression to conjunctive normal form?
21:43:58 <oklopol> yes
21:44:03 <oerjan> because i think some of the same tricks can be used there
21:44:54 <oerjan> introducing variables to split things cheaply
21:45:12 <oklopol> well i was just thinking, if you have just a few clauses of disjunctive normal form, the conjunctive normal form will have an exponential number of clauses
21:45:50 <oklopol> because the numbers of clauses are 2^|variables|'s complements if i'm not mistaken
21:45:54 <oklopol> oh
21:46:02 <oklopol> introducing variables.
21:46:05 <oklopol> goddamnit
21:46:11 <oklopol> why didn't i think of that
21:46:16 <oklopol> :P
21:46:55 <oerjan> in fact i think you can start this by doing the same thing to the large disjunction
21:47:13 <oklopol> err what?
21:47:14 <oerjan> then you end up with only relatively small ones
21:47:30 <oklopol> what large disjunction
21:47:50 <oerjan> if you start with a disjunctive normal form
21:47:57 <oklopol> ah
21:47:58 <oklopol> yeah
21:48:32 -!- atrapado has quit ("Abandonando").
21:49:10 <oklopol> my initial approach was to try and use boolean algebra to reduce things to the normal form
21:49:25 <oklopol> but that's quite limited
21:49:51 <oerjan> yeah because it blows up
21:49:54 <oklopol> yes
21:50:03 * ehird realises that his semantic wiki project thing is already realised by RDF, and he just needs to hack up an interface to viewing/editing rdf
21:50:05 <ehird> yay
21:50:15 <oerjan> after all, if you could convert things to conj. form then by duality you could convert it to disj. form
21:50:27 <oklopol> yes
21:50:47 <oerjan> and if you could do the latter without blowing up, satisfiability becomes easy to solve
21:51:08 <oklopol> it does?
21:51:33 <oklopol> is satisfiability trivial in disjunctive normal form?
21:51:39 <oerjan> yes. because checking a disjunctive normal form for satisfiability is just checking each clause
21:51:39 <oklopol> hmph, i don't know shit.
21:51:58 <oerjan> if any of them is satisfiable, then the whole is
21:52:01 <oerjan> if not, not
21:52:05 <oklopol> right
21:52:56 <oklopol> how does this not make it impossible to do the conversion with newly introduced variables?
21:53:18 <oerjan> ehird: ah, Reality Distortion Fields
21:53:44 <ehird> oerjan: no, http://www.w3.org/RDF/ :-
21:53:45 <ehird> P
21:54:05 <ehird> (Note: It isn't always XML, there's a non-eyeball-renching plaintext serialization of it too... thank god.)
21:54:19 * oerjan loves picking the wrong de-acronym
21:54:27 <ehird> :-D
21:55:07 <oerjan> oklopol: the thing here is that when introducing variables while keeping satisfiability you always introduce new _conjunctions_ outermost
21:55:38 <oerjan> so it gets worse, not better, as far as actually solving it goes
21:55:53 <oklopol> and you cannot get to DNF by adding new vars?
21:56:06 <oklopol> you have to realize i don't actually see how any of this is realized :P
21:56:12 <oklopol> realizeeeee
21:56:29 <oerjan> if you tried to do the dual i guess you would find you preserve the dual of satisfiability instead
21:56:42 -!- kar8nga has left (?).
21:56:58 <oerjan> and _that_ is hard to solve on disjunctive normal form, but easy on conjunctive
21:57:22 <oklopol> ah, find something that does *not* satisfy it
21:57:30 <oklopol> okay, it's all becoming clear now.
21:57:39 <oerjan> yeah
21:58:11 <oklopol> but
21:58:23 <oklopol> so err
21:58:46 <oklopol> (a ^ b) v (c ^ d), how would i turn this upside down into CNF?
21:59:13 <oerjan> i think once you have it down to just two terms you'll have to use boolean algebra a bit
21:59:22 <oerjan> so, distributive law
21:59:26 <oklopol> umm okay
21:59:27 <oklopol> let's see
21:59:46 <oklopol> (A v B)' v (C v D)'
21:59:49 <oklopol> and
22:00:00 <oklopol> ((A v B)' ^ (C v D))'
22:00:02 <oklopol> umm
22:00:04 <oklopol> ((A v B' ^ (C v D))'
22:00:08 <oklopol> but that's no use
22:00:09 <oklopol> and fuck
22:00:13 <oklopol> ((A v B) ^ (C v D))'
22:00:16 <oklopol> .
22:00:20 <oklopol> hmmhmm
22:00:57 <oerjan> um it's: (a ^ b) v (c ^ d) => (a v c) ^ (a v d) ^ (b v c) ^ (b v d)
22:01:54 <oklopol> right, so to not it, i have to list all the *other* possibilities?
22:02:18 <oerjan> not?
22:02:24 <oklopol> not as a verb
22:02:25 <oklopol> negate
22:02:33 <oklopol> complement
22:02:36 <oklopol> dunnnnnno
22:02:40 <oerjan> to negate a boolean expression you use deMorgan's laws
22:03:20 <oerjan> switch ^ and v, and negate the arguments recursively
22:03:50 <oklopol> why does that add new clauses
22:03:56 <oerjan> it doesn't
22:04:14 <oerjan> are we trying to do the same thing...
22:04:21 <oklopol> :D
22:04:55 <oerjan> basically, you usually use deMorgan's law to get all negation down to just the variables before doing anything else
22:04:56 <oklopol> what exactly did you negate in (a ^ b) v (c ^ d) to get that next thing?
22:05:21 <oerjan> i didn't negate, i used the distributive law for v over ^
22:05:29 <oklopol> well that makes more sense
22:05:48 <oerjan> (they're dual, so distribution goes both ways)
22:05:55 <oklopol> okay, yeah, i see it now
22:06:33 <oklopol> how about something more complex then, say i simply have (a ^ b ^ c) v (d ^ e ^ f)
22:06:39 <oklopol> well
22:06:53 <oklopol> i guess i can use the same law...
22:06:59 <oerjan> i got to thinking that this might be a bit wrong if you have deeply nested things
22:07:25 <oerjan> a better way may be to introduce variables _standing_ for the subterms
22:07:48 <oerjan> i think that's how the reduction NP -> SAT usually goes anyway
22:07:51 <oklopol> yes, that's what i tried initially
22:08:01 <oklopol> and NP means?
22:08:08 <oklopol> naughty proposition?
22:08:10 <oerjan> any NP problem
22:08:15 <oklopol> ahh
22:08:59 <oerjan> so then, what we want is a clause that implies abc == a ^ b ^ c
22:09:11 <oklopol> yeah
22:09:23 <oklopol> we somehow add some structure somewhere
22:09:38 <oklopol> that makes (a ^ b ^ c) true exactly when abc is true
22:09:40 <oerjan> this is easy since equivalence is a boolean relation
22:09:47 <oklopol> and then use abc as that subterm
22:09:54 <oklopol> oh
22:10:08 <oklopol> right, i used that reduction in my noprob examples
22:10:31 <oklopol> (a <=> b) <=> ((a ^ b) v (A ^ B))
22:10:39 <oklopol> both true or both false
22:10:48 <oerjan> except you want it conjunctive, presumably
22:10:51 <oklopol> right
22:10:52 <oklopol> so
22:11:03 <oklopol> ((a v B) ^ (A v b))
22:11:14 <oklopol> which i cannot really explain.
22:11:30 <oklopol> errrr is that even right
22:11:47 <oklopol> ah
22:12:03 <oklopol> if the vars were different
22:12:11 <oklopol> then one of those would definitely be false
22:12:15 <oklopol> so okay, yeah, that's it
22:12:27 <oklopol> so
22:12:29 <oerjan> you can get it with distributivity, then all the a ^ A terms disappear because they're inconsistent
22:12:43 <oklopol> yup
22:13:00 <oklopol> but i like to explain things rather than prove them, because i'm a softie
22:13:07 <oklopol> okay, so
22:13:29 <oklopol> if i have the subterm (a ^ b ^ c)
22:13:33 <oklopol> in a larger thingie
22:14:21 <oklopol> err... okay now i'm thinking i'd add an "^ (abc <=> (a ^ b ^ c))" on the toplevel and substitute abc for all (a ^ b ^ c)'s
22:14:29 <oerjan> yeah
22:14:56 <oklopol> the thing i ripped abc out of is now definitely smaller
22:16:09 <oklopol> but the problem is, when i've removed all the substructures
22:16:24 <oklopol> i have ands on the toplevel, anding up all the equivalences
22:16:36 <oerjan> yep, and that's good
22:16:43 <oklopol> ...
22:16:50 <oklopol> yes
22:16:53 <oklopol> of course it is
22:17:04 <oklopol> it's just i keep flipping ands and ors together.
22:17:23 <oklopol> but hey
22:17:30 <oklopol> umm
22:17:42 <oklopol> (abc <=> (a ^ b ^ c))
22:17:53 <oklopol> guide me through this
22:17:54 <oklopol> basically
22:18:02 <oklopol> i first do the flippedy
22:18:04 <oklopol> and get like
22:18:26 <oerjan> note if (a ^ b ^ c) were something larger you could always split it up more
22:18:30 <oklopol> (abc v (a ^ b ^ c)') ^ (Abc v (a ^ b ^ c))
22:18:41 <oklopol> ohh, i can distribute
22:19:03 <oklopol> err
22:19:31 <oklopol> (abc v (a ^ b ^ c)') ^ (Abc v a) ^ (Abc v b) ^ (Abc v c)
22:19:36 <oklopol> rright?
22:19:42 <oklopol> but the leftmost one
22:19:57 <oerjan> the left half needs a deMorgan
22:20:02 <oklopol> (abc v A v B v C) ^ (Abc v a) ^ (Abc v b) ^ (Abc v c) ?
22:20:07 <oerjan> yeah
22:20:10 <oklopol> yay
22:20:30 <oklopol> okay, i'm pretty sure i could do the conversion manually now
22:20:55 <oklopol> so i can probably automatize it with a bit of further consideration
22:21:10 -!- cathyal has joined.
22:21:27 <oklopol> thanks, this has been eye-opening
22:22:05 <oerjan> i think this is about half the proof that SAT _is_ NP-complete. the other half is turning a Turing machine into a boolean circuit
22:22:52 <oklopol> umm, with a finite playground?
22:22:55 <cathyal> anyone worked on a brainfuck compiler
22:23:08 <oerjan> polynomial size in the input
22:23:11 <oklopol> oh oh oh
22:23:30 <oklopol> that's how it's proven that SAT is np-complete without reducing it to anything
22:23:38 <oklopol> turing machine
22:23:51 <fizzie> Still trying to get fungot to crash, but I think he's losing hope:
22:23:51 <cathyal> what are you guys talking about
22:23:52 <fizzie> 00:22:01 <fizzie> fungot: Crash!
22:23:53 <fizzie> 00:22:02 <fungot> fizzie: my feelings toward these shelves cannot be described there is no hope. then,
22:23:53 <fungot> fizzie: " bragging rights" out of " scary dead grandma made us fake the stamp" stickers
22:23:54 <fungot> fizzie: the environment.' environments ( better referred to as ' characters.'
22:23:54 <fungot> fizzie: i'll give you mine if you want
22:23:55 <cathyal> turing complete machines?
22:24:04 <oklopol> my algo book just said something reeeally vague
22:24:36 <oerjan> cathyal: oklopol is trying to invent a language based on an NP-complete problem
22:24:44 <cathyal> oh please
22:24:44 <cathyal> NOT
22:24:45 <cathyal> lol
22:24:54 <oklopol> trying to *implement*, the language is ready
22:25:03 <oerjan> it's practically unimplementable
22:25:14 <oklopol> yes, that's the beauty of implementing it
22:25:18 <oerjan> but that has never stopped oklopol
22:25:37 <oklopol> :P
22:26:01 <oklopol> cise's parsing is most likely np-complete
22:26:12 <oklopol> i should try proving that
22:26:31 <oerjan> fizzie: are you not succeeding? maybe something other than just chatting is required
22:27:05 <fizzie> So far it just chatting has been enough, but it only occurs rarely.
22:27:09 <oklopol> two exams tomorrow, and not about 3-sat, perhaps i should sleep a bit ->
22:27:25 -!- zbrown has quit (Remote closed the connection).
22:27:29 -!- zbrown has joined.
22:27:44 <fizzie> I'm sure I'd get it to crash pretty quickly if I just feed "fungot\nfungot\nfungot\n" as the input without involving an IRC server, but that would feel like cheating.
22:28:12 <oerjan> night oklopol
22:28:13 <cathyal> so
22:28:16 <fizzie> So I've just been talking with it.
22:28:17 <cathyal> whose worked on haskell
22:28:25 <cathyal> or implemented symbolic languages
22:28:27 <oerjan> haskell haskell haskell
22:29:04 <oerjan> i have half of a brainfuck interpreter in haskell laying around
22:29:23 <oerjan> my usual vaporware
22:30:27 <fizzie> I've got half of a Befunge interpreter in Haskell, it's my default "testing a new language" program.
22:30:27 <oerjan> fungot: CRASH
22:30:36 -!- KingOfKarlsruhe has quit (Remote closed the connection).
22:30:38 <fungot> oerjan: ive got balls of steel?
22:30:43 <Linus`> is there an esoteric OS?
22:30:47 <Linus`> would be fun :o
22:30:57 <oerjan> fungot: DAMN YOU
22:30:58 <fungot> oerjan: undefined local variable or two
22:31:18 <fizzie> (It's back to irclogs from the politician-talk, if you didn't guess.)
22:31:22 <oerjan> Linus`: the idea crops up frequently
22:32:34 <oerjan> cathyal: somehow the languages implemented here tend not to be symbolic.
22:33:20 <oerjan> fizzie: i don't know, that _could_ be Bush saying that, don't you think?
22:33:28 <Asztal> except in the sense that most of the instructions end up being symbols in the other sense of the word :)
22:34:01 <oerjan> actually what is a symbolic language, precisely?
22:34:03 <fizzie> oerjan: Not the "undefined local variable" one.
22:34:14 <oerjan> fizzie: maybe not.
22:34:17 <cathyal> oerjan: nice
22:36:13 <oerjan> i recall the Reader[T] monads are considered nice for environments. i think there are some tutorials.
22:36:39 <fizzie> Hey, I think I crasheded it finally.
22:36:58 <oerjan> also, #haskell is a very friendly channel too, although i haven't been there in a while
22:37:29 <oerjan> fungot: you dead?
22:37:52 <ehird> hi cathyal.
22:39:00 <cathyal> hi ehird
22:39:03 -!- fungot has quit (Read error: 131 (Connection reset by peer)).
22:39:27 -!- fungot has joined.
22:40:34 <fizzie> There's a two-megabyte log for me to look at now. Later.
22:40:42 <oerjan> now we wait for the results of the autopsy
22:42:46 <fizzie> There's a lot of negative numbers in the stack, and it's somewhere there in the tokens-to-text code. Have to sleep now, the corpse will still be there tomorrow too.
22:52:54 <Asztal> no respect for the dead :(
22:53:21 <oerjan> but but ... it's for science!
22:53:29 <oerjan> right fungot?
22:53:29 <fungot> oerjan: is it possible to do all the design :)... when some of the implementations.
22:53:33 <Asztal> and did you get fungot's permission?
22:53:33 <fungot> Asztal: what's a metasyntactic variable there.
22:56:39 <oerjan> also, fungot is not dead.
22:56:40 <fungot> oerjan: are you a number of much higher-level concurrency abstractions and some other
22:57:07 <oerjan> fungot: i couldn't say
22:57:08 <fungot> oerjan: helsinki.fi cs entrance exam is coding instead of surfing?
22:57:25 <oerjan> fungot: shocking, i know
22:57:26 <fungot> oerjan: which then got reddit'd or something.
23:03:58 <psygnisfive> asztal, are you hungarian?
23:05:02 -!- Judofyr has quit (Read error: 104 (Connection reset by peer)).
23:05:31 -!- Judofyr has joined.
23:06:47 <Asztal> psygnisfive: no, but I am learning that language
23:08:38 <psygnisfive> ok
23:09:20 <Asztal> how did you know? do you know this word somehow?
23:09:59 * oerjan recalls assuming Asztal was hungarian before
23:10:43 <oklopol> thought occurred:
23:10:52 <oklopol> is there some kinda time transition thingie
23:11:03 <oklopol> somewhere around these times
23:11:27 <ehird> what
23:11:40 <oklopol> you know, like you turn the knob of the clock
23:11:43 <oklopol> around and around
23:11:48 <oklopol> happens a few times per year
23:12:03 <oerjan> next week
23:12:10 <oklopol> okay.
23:12:11 <oklopol> thanks.
23:12:13 <oklopol> sleep.
23:12:14 <oklopol> ->
23:12:31 <ehird> lol
23:17:55 <psygnisfive> night oklopol
23:18:06 <psygnisfive> asztal: the sz spelling is very hungarian
23:18:24 <psygnisfive> any combination of s and z is hungarian looking
23:19:14 <Asztal> ah, I see
23:19:37 <Asztal> I usually find that it's polish when I see sz :)
23:19:40 <psygnisfive> its like how you can identify finnish by its excessive double vowels, double consonants, and umlauts EVERYWHERE
23:19:49 <Asztal> gy is very hungarian imo
23:20:03 <psygnisfive> or dutch by its vowel pairs, ij, and short works
23:20:24 <psygnisfive> words*
23:20:35 <psygnisfive> gy looks hungarian indeed
23:20:37 <psygnisfive> or japanese
←2008-10-18 2008-10-19 2008-10-20→ ↑2008 ↑all