00:04:22 -!- Phantom_Hoover has quit (Read error: Connection reset by peer).
00:11:57 <boily> MDudello. the IOCCC has produced many mind-bending gems.
00:12:05 <boily> I think my favourite is the tiling program.
00:12:19 <boily> Tanelle. do you still boardgame?
00:12:53 <Taneb> Or however many cs and ss that word has
00:13:09 <Taneb> Haven't for a couple of weeks, though
00:13:45 <boily> ah, the joys of studenting...
00:14:22 <Taneb> Got two more this season, both next week
00:14:29 <Taneb> One on Tuesday about computability and complexity
00:14:38 <Taneb> And one on Thursday about groups, rings, and fields
00:14:42 <Taneb> Not too worried about any of them
00:15:42 <Taneb> I should go to bed now, though
00:15:49 <Taneb> Didn't get much sleep last night
00:54:49 <HackEgo> olist 987: shachaf oerjan Sgeo FireFly boily nortti
01:04:11 -!- variable has joined.
01:10:29 -!- variable has quit (Ping timeout: 245 seconds).
01:11:34 -!- heroux has quit (Ping timeout: 255 seconds).
01:12:23 -!- heroux has joined.
01:21:21 <shachaf> boily: what is a godsmoot twh
01:25:01 <oren> i had some triuble renemberisnf what froups tings and fields are
01:26:57 <boily> helloren. still touchtyping?
01:27:10 <boily> shachaf: you'd probably be better off asking the fungot hth
01:27:10 <fungot> boily: the comrades of washington projected this monument. their love inspired it. their fear betrays to the first faint rumours of this calamity pitt would give no adequate representation to moslem opinion. in bombay the moslems are fnord/ 4 fnord per cent.
01:27:37 <boily> shachaf: first google hit here: http://mrtehcyborg.tumblr.com/post/116757446518/nihhussa-oh-my-god
01:28:24 <boily> oren: http://www.keybr.com/#!game
01:28:24 -!- M_I_Wright has quit (Read error: Connection reset by peer).
01:28:25 -!- Prime has joined.
01:28:32 <HackEgo> Prime: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: <http://esolangs.org/>. (For the other kind of esoterica, try #esoteric on irc.dal.net.)
01:28:48 -!- Prime has changed nick to Guest77553.
01:31:09 -!- hilquias has joined.
01:32:01 -!- boily has quit (Quit: CARMINATIVE CHICKEN).
01:52:41 -!- adu has joined.
01:58:39 -!- adu has quit (Quit: adu).
02:22:32 <HackEgo> [wiki] [[Talk:XMLfuck]] http://esolangs.org/w/index.php?diff=43015&oldid=29733 * Zzo38 * (+258)
02:22:48 <oren> Dad: "Hey oren, see if you can spot the gap in this proof." Me: "Uhh, is the gap where it says 'obviously'?"
02:28:14 <oren> Apparently my dad has invented a sport of finding crappy papers in supposedly reputable journals
02:28:24 <HackEgo> [wiki] [[XMLfuck]] http://esolangs.org/w/index.php?diff=43016&oldid=15399 * Zzo38 * (+29)
02:31:26 -!- hilquias` has joined.
02:33:12 -!- hilquias has quit (Ping timeout: 256 seconds).
02:56:56 <oren> how do you pronounce häagen dazs?
03:06:51 -!- ZombieAlive has quit (Remote host closed the connection).
03:14:18 <zzo38> Do you know if there is any free/open-source software to create MOD/XM/S3M that can use a piano-roll editor? I want to know so that I can add it into the AmigaMML wiki comparison charts.
03:35:23 -!- GeekDude has quit (Quit: {{{}}{{{}}{{}}}{{}}} (www.adiirc.com)).
03:51:52 -!- copumpkin has quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…).
04:09:07 -!- Herbalist has quit (Ping timeout: 264 seconds).
04:35:21 -!- Nihilumbra has joined.
04:35:38 <Nihilumbra> http://i.imgur.com/69uwy8ql.jpg I drew a thing
04:43:15 <zzo38> What is it supposed to be a picture of though? Some kind of strange person?
04:44:15 <Nihilumbra> When I'm not working on coding projects for people
04:46:52 <Nihilumbra> You make the OK as if its like OK wow what a creep
04:48:21 <zzo38> No, I meant, OK you can make such stuff if you don't have the other stuff to do
04:49:55 -!- augur has quit (Quit: Leaving...).
04:51:57 -!- heroux has quit (Ping timeout: 240 seconds).
04:59:04 <oren> good mroing Nihilumbra. That looks similar to some of Frida Kahlo's more... esoteric, works
05:04:26 <oren> Frida Kahlo is an artist of 20th century mexico, the wife of Diego Rivera
05:07:57 <Sgeo> I think I need an Ubuntu livecd
05:08:04 <MDude> Looks more like something on a metal album cover.
05:08:28 <MDude> Wait no, that deer with a man head would fit on one too.
05:10:03 <oren> Yeah, the little deer is the one I was thinking it sort of looked like, the animal-human surrealist morphing
05:11:51 <oren> Also she painted this one where her organs ther heart and lungs are shown through her clothing which reminded me
05:12:55 <oren> Ubuntu livecd? Well you'll need a cd burner and a blank cd, which are less common nowadays
05:13:23 <oren> only my oldest laptop has a bourner
05:13:34 <Nihilumbra> I used to have a lot of works on my computer but it got wiped
05:13:55 <Nihilumbra> So I'm buying a new drawing tablet and re drawing some old stuff
05:15:21 <Sgeo> Ugh this website is so scummy
05:15:42 <Sgeo> A million download ads, and "Please note that SolMiRe does not allow the download of any uploaded midi files."
05:15:46 <Sgeo> in small print
05:17:51 -!- heroux has joined.
05:22:46 -!- MDude has changed nick to MDream.
05:29:57 -!- augur has joined.
05:44:17 -!- ais523 has joined.
05:58:13 -!- oerjan has joined.
05:59:10 <oren> helloerjan. good mroing ais523
06:02:43 -!- oerjan has set topic: <fungot> oerjan: i've gotten to the metacircular evaluation chapter? | The chanteau | https://dl.dropboxusercontent.com/u/2023808/wisdom.pdf http://codu.org/logs/_esoteric/ http://tunes.org/~nef/logs/esoteric/ | http://esolangs.org/.
06:02:45 -!- variable has joined.
06:11:48 <oren> `run echo 'O=8"*16hello a=11 N=8 aW1' | scrip7
06:12:37 * oerjan has no idea what tte misspells
06:12:51 <oren> `run echo 'O=8"*13hello a=11 N=8 aW1' | scrip7
06:13:25 <oren> whu dosnr this work?
06:17:01 <oren> `run echo 'O=8"*13hello a=3 N=8 aW1' | scrip7
06:19:25 <oren> so colors are with ^C
06:19:42 <oren> `run echo 'O=8"*13;12hello a=3 N=8 aW1' | scrip7
06:20:05 <oren> `run echo 'O=11"*13,12hello a=3 N=11 aW1' | scrip7
06:23:09 <oren> `run echo 'O=12"**13,12hello a(1=22 a=3 N=11 aW1' | scrip7
06:24:15 <oren> `run echo 'O=12"**13,12hello a(1=3 a=22 N=11 aW1' | scrip7
06:24:22 <oren> `run echo 'O=12"**13,12hello a(1=3 a=22 N=12 aW1' | scrip7
06:24:47 <oren> so bright colors can be used as background if you set reverse video
06:26:53 <oren> `run echo 'O=12"*76hello? a=3 N=12 aW1' | scrip7
06:27:11 <oren> `run echo 'O=9"*76hello? a=3 N=12 aW1' | scrip7
06:28:08 <oren> `run echo 'O=10"*208hello? a=3 N=12 aW1' | scrip7
06:28:25 <oren> `run echo 'O=10"*208hello? a=3 N=10 aW1' | scrip7
06:29:59 <oren> `run echo 'O=10"*99hello? a=3 N=10 aW1' | scrip7
06:31:39 <oren> So it's doing a %16 on it
06:32:17 -!- Tritonio has joined.
06:37:29 -!- newsham has quit (*.net *.split).
06:39:58 -!- hilquias` has quit (Ping timeout: 258 seconds).
06:41:23 <oerjan> @let cne 0 f x = modify(+1) >> return x; cne n f x = cne (n-1) e f x >>= f
06:41:25 <lambdabot> Couldn't match type ‘Expr’ with ‘Expr -> a0’
06:41:37 <oerjan> @let cne 0 f x = modify(+1) >> return x; cne n f x = cne (n-1) f x >>= f
06:42:29 <oerjan> > flip execState 0 $ cne 2 (cne 2) undefined
06:42:31 <lambdabot> Couldn't match type ‘a10 -> m0 a10’
06:42:31 <lambdabot> with ‘StateT s Identity (a10 -> m0 a10)’
06:42:31 <lambdabot> Expected type: (a10 -> m0 a10) -> StateT s Identity (a10 -> m0 a10)
06:42:41 <lambdabot> (Eq a, Num a, Num s, MonadState s m) => a -> (a1 -> m a1) -> a1 -> m a1
06:55:57 <oren> `run echo 'O={Maybe this will work?} a=3 aL'} aW1' | scrip7
06:55:58 <HackEgo> bash: -c: line 0: unexpected EOF while looking for matching `'' \ bash: -c: line 1: syntax error: unexpected end of file
06:56:11 <oren> `run echo "O={Maybe this will work?} a=3 aL'} aW1" | scrip7
06:56:24 <oren> `run echo "O={Maybe this will work?} aL'} aW1" | scrip7
07:09:22 -!- ChanServ has set channel mode: +o oerjan.
07:09:37 -!- oerjan has set channel mode: -b *!Frooxius@199-241-202-205.PUBLIC.monkeybrains.net$#fix_your_connection.
07:11:13 -!- oerjan has set channel mode: -o oerjan.
07:11:39 -!- Herbalist has joined.
07:16:38 -!- Tritonio has quit (Ping timeout: 272 seconds).
07:16:58 -!- variable has quit (Ping timeout: 255 seconds).
07:18:07 -!- newsham has joined.
07:51:15 <b_jonas> hello, ais523. have you figured out a fix for the StackFlow interpreter over M:tG yet?
07:51:28 <ais523> I've had a lot of other things to think about
08:12:19 <oerjan> <oren> how do you pronounce häagen dazs? <-- istr those words are made up.
08:12:55 <oerjan> so, "american", i think.
08:12:58 -!- AnotherTest has joined.
08:26:02 -!- evalj has joined.
08:29:04 -!- Patashu has joined.
08:37:41 -!- oerjan has quit (Quit: leaving).
08:40:51 -!- AnotherTest has quit (Ping timeout: 272 seconds).
08:48:18 -!- Nihilumbra has quit (Quit: Connection closed for inactivity).
08:57:01 -!- AnotherTest has joined.
09:11:12 -!- AnotherTest has quit (Ping timeout: 252 seconds).
09:43:35 <FireFly> Indeed they are. They're supposed to look scandinavian, but neither the äa nor the z is really scandinavian... it looks more german to me
09:44:04 <FireFly> Although I'm not sure if äa makes any sense in german either
09:53:03 <b_jonas> FireFly: no, I think it's supposed to look foreign for everyone
09:53:14 <b_jonas> so they can sell anywhere under the same name
09:53:32 <b_jonas> that's why the name looks so riddiculous
09:53:58 <fungot> mroman_: the reason, too, have received instructions for the part i have undertaken as plain and intelligible as i possibly can. i want to tell you just how to do it.
10:02:38 -!- Wright__ has joined.
10:02:39 -!- Wright has quit (Read error: Connection reset by peer).
10:10:20 <FireFly> b_jonas: apparently it was danish that they tried to approximate: https://en.wikipedia.org/wiki/Häagen-Dazs#Origin_of_brand_name
10:10:49 <FireFly> Admittedly, at least Häagen-Dazs doesn't look as ridiculous in swedish as Frusen glädjé does
10:11:47 <FireFly> fungot: roger that. Clear and intelligible instructions are very useful.
10:11:47 <fungot> FireFly: the causes of that great and enlightened city, a run on the bank of england; whose credit had often supported a tottering state, and the waves run high, that the influence of these feelings, m. d'angers, the sculptor, his part in the memoirs of marmontel. many others might be fnord or not, before it can be done by a succession of absolute monarchs, guaranteed by irresistible force against the fnord heresy. but, when i
10:12:12 <HackEgo> /home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: style: not found
10:12:14 <fungot> Available: agora alice c64 ct darwin discworld enron europarl ff7 fisher fungot homestuck ic irc iwcs jargon lovecraft nethack oots pa qwantz sms speeches* ss wp youtube
10:12:43 <fungot> Selected style: speeches (misc. speeches from Project Gutenberg)
10:12:57 <fizzie> fungot: Are you sure they used the word fnord that much?
10:12:58 <fungot> fizzie: what are the elements of true greatness. of mankind i fnord the landing on plymouth rock? theirs indeed, were not merely as probable, but as evanescent, and that
10:13:26 <mroman_> fungot is all about the fnord.
10:13:27 <fungot> mroman_: " so you can see fnord all about where she has strained herself trying to do a lot of them; and i am persuaded that we have generally had a great battle which arrested the armies of europe, and also if they choose; but the executive, of france, who was generally the mouthpiece of the administration did i place more confidence than in the provision and distribution of the public weal enjoy high consideration, and i was
10:13:29 <b_jonas> FireFly: what's "Frusen glädjé"?
10:14:11 <FireFly> b_jonas: some competitor founded the same year, apparently, mentioned in the aforementioned article
10:14:38 <FireFly> I've never heard of them apart from that
10:29:04 -!- boily has joined.
10:31:16 <HackEgo> [wiki] [[MiniMAX]] M http://esolangs.org/w/index.php?diff=43017&oldid=42973 * Ais523 * (+0) /* Example */ typo fix
10:32:55 <HackEgo> [wiki] [[MiniMAX]] M http://esolangs.org/w/index.php?diff=43018&oldid=43017 * Ais523 * (+8) /* Computational class */ clarification
10:36:22 <ais523> ä and é in the same word? ouch
10:39:07 <Taneb> Almost like naivete?
10:39:17 <Taneb> Except with more dots and swoops and stuff
10:39:36 <ais523> oh, but that's a diaeresis, not an umlaut
10:39:56 <ais523> that ä can't be a diaeresis because the preceding letter's a consonant
10:40:34 <ais523> or, well, I guess you could claim that l is a vowel, but people don't normally use diaereses with vowels as dubious as that
10:41:27 <ais523> can #esoteric help me feel better about mockingbirds, bt?
10:41:34 <ais523> I tried plugging some of them into my day job research
10:41:40 <ais523> and the implications are driving me mad
10:41:49 <ais523> they have a tendency to explode type systems
10:42:03 <Taneb> They're a kind of bird
10:42:51 <b_jonas> explode how exactly? do they violate occurs check in type unification, or straight up try to unify two unequal non-unifyable types, or some other way?
10:43:36 <b_jonas> ais523: can you use more specialized less powerful loop functions instead, ones that are well-typed?
10:43:49 <ais523> the occurs check is basically a hack that's designed to stop mockingbirds crashing the compiler
10:43:51 <b_jonas> fold and unfold and the like
10:43:55 <ais523> and if I feed them to my compiler atm, it crashes ;-)
10:43:59 <Taneb> What do you mean by mockingbird?
10:44:28 <ais523> hurts to look at, right?
10:44:36 <Taneb> :: (a = a -> a) => a?
10:44:45 <ais523> b_jonas: anyway, in my current type theory, a mockingbird is actually well-typed
10:44:47 <ais523> but the double mockingbird isn't
10:45:11 <ais523> and when I run through the type inference algo to find out why, bad things start happening
10:45:30 <b_jonas> what's the double mockingbird? I'm not good in ornithology
10:45:44 <ais523> e.g. it works by constructing a table of definitions, but gives two different definitions for the same thing, which nonetheless seem to converge
10:45:57 <ais523> b_jonas: (\x.x(x))(\x.x(x))
10:46:20 <ais523> it's basically the Henkin statement of untyped lambda calculus
10:46:38 <ais523> if you try to work out what type it has, about the best you can do is to determine that it has the same type as itself
10:47:48 <b_jonas> and can you debug the compiler to see how exactly it crashes?
10:48:04 <b_jonas> like, as in stack blowup or memory trashing or something?
10:48:21 <ais523> b_jonas: the current algo is stack blowup, it tries to generate infintely many type constraints
10:48:28 <ais523> I'm working on a new algo manually
10:48:42 <ais523> where I can notice if things blow up before they exhaust my text document ;-)
10:49:20 <Taneb> ais523, is it really bad if you can't type something which can't be typed?
10:49:37 <ais523> Taneb: no, you'd expect to not type something that can't be typed
10:49:51 <ais523> but you want to know why it doesn't type
10:50:07 <ais523> note that type inference for intersection types is equivalent to the halting problem
10:50:16 <ais523> and in about the most direct possible way, too: the term has a type if and only if it halts
10:51:42 <Taneb> Are all semidecidable problems equivalent to the halting problem?
10:53:11 <ais523> I don't see a reason why a semidecidable problem would necessarily be equivalent to the halting problem
10:53:17 <ais523> it's amusing how that one is, though
10:53:34 <ais523> problem equivalence is basically never that astonishingly exact
10:54:27 <b_jonas> as for semi-decidable stuff, there's this nice new algebraic topology result I've been reading: http://arxiv.org/abs/1302.2370v1
10:54:37 <b_jonas> Martin Cadek, Marek Krcal, Jiri Matousek, Lukas Vokrinek, Uli Wagner, "Extendability of continuous maps is undecidable"
10:55:15 <b_jonas> yeah, I don't understand it either, but (one of) the results it proves itself is easy enough to understand (without proof)
10:55:44 <ais523> aa((!((aa)(!))))*:*^!**^a*^a*aa*(*:*^!**^)*^
10:55:45 <Taneb> Oooh, I'm doing a module in topology next year
10:55:49 <ais523> the scariest line of Underload I've seen
10:55:55 <ais523> I still don't really understand it
10:56:06 <Taneb> It just looks like screaming, ais523
10:56:07 <ais523> (oerjan came up with it, somehow)
10:56:14 <ais523> Taneb: it's an implementation of ~ without using ~
10:56:32 <ais523> this violates my mental model of substructual logics
10:56:32 <b_jonas> some similar results have been known for long, like that it's RE but not recursive to decide which pairs of simplicial complexes are homotopic
10:56:53 <ais523> b_jonas: I like the way you can say "simplicial complexes" with the IRC version of a straight face
10:57:51 <ais523> hmm, we need more ad-hoc prove-this-interesting-language-TC contests
10:57:57 <ais523> the last time was resplicate, i think
10:57:58 <b_jonas> what? "simplicial complex" is just the easiest to understand finite representation of "nice" finite dimensional topological spaces up to homeomorphism
10:58:08 <ais523> b_jonas: I was thinking of the name
10:58:29 <b_jonas> it's like when you say "represented in binary" about integers
10:58:31 <ais523> it sounds completely absurd if you don't know what it means
10:58:52 <b_jonas> in computability yuo have to take care about how you represent stuff
10:58:58 <b_jonas> so you get this kind of thing all the time
10:59:17 <b_jonas> also, topology people work with really crazy spaces, but simplicial complexes are nice spaces
10:59:33 <ais523> a simplex is basically just a generalized tetrahedron, right?
11:01:06 <b_jonas> ais523: and a simplicial complex is a space given as a union of simplexes such that (1) each lower-dimensional side of each simplex is in the set and (2) any two non-disjoint simplexes in the set intersect in a simplex that's the side of both of those simplexes.
11:01:34 <b_jonas> so it's like a polyhedron of any finite dimension but without half-overlapping faces and star-shaped faces all those ugly stuff
11:01:46 <ais523> it's basically the way it can be simple and complex at the same time
11:01:49 <b_jonas> it's like fully triangulated
11:01:57 <b_jonas> oh, that's why it sounds funny?
11:02:06 <b_jonas> but you have "simple complex lie algebras" too
11:02:19 <b_jonas> mind you, that's a different sense of "complex"
11:03:22 <ais523> hmm, my favourite name for anything is still a macro from Perl: SV_CHECK_THINKFIRST_COW_DROP
11:03:39 <b_jonas> the easy way to get simplicial complexes (and you can get any, up to homeomorphism) is to take n vertices (where n is natural number) affine independent in an n-1 dimensional space, and then any set of simplexes over that.
11:03:46 -!- Herbalist has quit (Ping timeout: 255 seconds).
11:03:46 <b_jonas> then there can be no uglyness.
11:03:58 <boily> ais523: hellais523. what is it for?
11:04:09 <b_jonas> we just usually imagine simplicial complexes in lower-dimensional container spaces because they're easier to draw.
11:04:43 <ais523> boily: the Perl macro? it checks to see if a scalar has special properties that would make normal-seeming operations on it not work; if the only such problem is copy-on-writeness, it does the copy
11:04:48 <ais523> so that it isn't copy-on-write any more
11:05:12 <b_jonas> ais523: ok, but what does the CHECK part mean?
11:05:18 <ais523> it checks to see if it's thinkfirst
11:05:31 <b_jonas> oh, so does this like return a boolean?
11:05:34 <ais523> (thinkfirst being a property that means that you can't do weird things to it, sort-of like the opposite of Plain Old Data)
11:05:40 <ais523> and also drops the cow as a side effect
11:06:14 <boily> Scalar Variable Check Weird Properties Drop COW. makes sense.
11:06:48 <ais523> my brain still expands "SV" to "scalar container", because Perl makes so much more sense with that mental expansion
11:06:55 <ais523> even though the acronym doesn't fit then
11:07:05 <boily> perhaps it's a vontainer?
11:07:25 <ais523> this explains why $x = 4 doesn't change any SV (just the /contents/ of an SV)
11:08:03 <ais523> right, unless $x doesn't exist at the time
11:08:20 <b_jonas> or it exists and is magical or tied or something
11:08:23 <ais523> in which case a new SV is created, and placed in *x{SCALAR} (which is also an SV; hash elements are)
11:08:35 <ais523> oh yes, if it's magical anything could happen
11:09:01 <b_jonas> um, *x{SCALAR} isn't a hash element
11:09:05 <ais523> I'm not sure how much of the slowness of my memory profiler is due to the fact that it's doing profiling activities during the main loop, and how much is just a consequence of making every single scalar magical
11:09:13 <b_jonas> do you mean $somepackage::{x} as the hash element?
11:09:32 <ais523> ($somepackage::{x} /is/ *x, right?)
11:09:50 <ais523> actually, is it *x or \*x?
11:10:08 <ais523> counting the number of containers involved in something can be weird
11:10:19 <ais523> especially because you can put an array in an SV just fine; you're not meant to but it works
11:10:22 <b_jonas> it can also be one of two magical optimization values: a reference to a scalar or a reference to a sub, or something
11:10:36 <b_jonas> and it can just not exist yet
11:10:40 <ais523> oh, so if you avoid having two variables with the same name, the program is faster?
11:10:42 <b_jonas> (as in the pair doesn't exist int he hash)
11:10:45 <ais523> Perl optimizations always sound so weird
11:11:04 <ais523> because their purpose is to make Perl work vaguely like other languages, as opposed to what would normally be considered a variable
11:11:13 <b_jonas> ais523: I think the optimization applies for constants only, which somehow magically work as both a sub and a scalar
11:11:14 <ais523> b_jonas: well if you have both $x and @x, then *x will need to be an actual glob
11:11:24 <ais523> oh right, that'd make sense
11:11:32 <b_jonas> ais523: no, I think if you have $x as a plain package variable then it has to be an actual glob
11:11:38 <ais523> constants are subs internally, for most purposes
11:11:40 <b_jonas> nad if you have @x then it _definitely_ has to be an actual glob
11:11:47 <b_jonas> and possibly $x has to exist as well
11:11:49 <ais523> I thought the ruling was "currently they're subs but that might change in future"
11:12:03 <b_jonas> ais523: no, I mean you can access them as scalar
11:12:44 <ais523> both at a) that being old, and b) my reaction on realising 5.14 is old
11:12:48 <ais523> Perl really has been releasing a lot recently
11:13:26 <b_jonas> but just look at Linux, do you know what version number they're at? 5.0
11:13:29 <b_jonas> it doesn't even look right
11:13:58 <ais523> b_jonas: Linus announced a new policy of incrementing the major version number whenever he feels like it, without any particular significance
11:14:08 <ais523> because otherwise it seemed doomed to stick at 2 indefinitely
11:14:17 <b_jonas> when they bumped the version number to 3.0, some user processes or libraries balked because they expected the uname to be 2.6.*
11:14:43 <ais523> that's a good argument for doing it more often
11:14:59 <ais523> otherwise you end up with the Windows 10 issue
11:15:12 <b_jonas> I've seen such a problem at work
11:15:22 <b_jonas> but a moment let me try this perl stuff still
11:17:06 <b_jonas> anyway, I've created video files that are encoded with fake timestamps, and you had to convert between the real timestamp and the fake timestamp using an auxiliary file
11:17:58 <b_jonas> but the problem is, I chose the wrong frame rate for the fake timestamps, and at the point when the stuff started to work, the real timestamps were very uniform linearly distributed and at the exact same framerate as the fake timestamp.
11:18:07 <b_jonas> the conversions were wrong at some points but because of this we didn't notice.
11:18:22 <b_jonas> I chose the wrong framerate for the fake stuff because it was too correct.
11:18:38 <b_jonas> Later we got videos with higher real framerate, and the errors started to show.
11:21:52 <b_jonas> And I can even blame myself because there was a time when I should have foreseen that this would happen and could have changed the fake timestamp framerate.
11:31:27 -!- boily has quit (Quit: FLORAL CHICKEN).
11:37:04 <fizzie> I'm surprised that `perl -e worked, since due to ` it had the effect of perl '-e print $]' -- I guess Perl's just being very unpicky about arguments.
11:37:09 <HackEgo> Linux umlbox 3.13.0-umlbox #1 Wed Jan 29 12:56:45 UTC 2014 x86_64 GNU/Linux
11:37:23 <fizzie> (That would've been fine either way, of course, since it only has the one argument.)
11:37:55 <ais523> fizzie: Perl's generally fine with argument stacking
11:40:03 <fizzie> Yes, although perldoc perlrun synopsis doesn't really suggest that. It's got e.g. [ -Fpattern ] but for -e it has [ [-e|-E] 'command' ] which makes it look like it "should" be separate.
11:41:46 <fizzie> Guess the difference is that for e.g. -F it can't be separated, while -e is fine either way.
11:49:35 <b_jonas> fizzie: why would that not work? most programs can take the argument for a switch in the same command-line argument or a different command-line argument
11:49:41 <b_jonas> fizzie: eg. either perl -e foo or perl -efoo works
11:49:52 <b_jonas> a few programs are more picky, but most aren't
11:51:20 <fizzie> Maybe I've just run across the picky ones more often than is standard. Although I can't recall any particular examples.
11:51:45 <fizzie> The usual suspects (sed, dc) seem to be friendly, too.
11:52:44 <b_jonas> fizzie: or maybe you just usually used a separate arg
11:53:09 <b_jonas> anyway, perl is parsing its command-line arguments in an untypical way, mostly to make shebang magic easier, but this isn't an example for it
11:53:58 <fizzie> Well, to quote POSIX: "The Utility Syntax Guidelines in Utility Syntax Guidelines require that the option be a separate argument from its option-argument, but there are some exceptions in IEEE Std 1003.1-2001 to ensure continued operation of historical applications: --"
11:54:14 <fizzie> "If the SYNOPSIS of a standard utility shows a <space> between an option and option-argument (as with [ -c option_argument] in the example), a conforming application shall use separate arguments for that option and its option-argument."
11:55:14 <b_jonas> fizzie: ok... but still, at least gnu programs usually call the libc getopt_long function which behaves this way.
11:55:16 <ais523> POSIX's option syntax is basically not used by anything, though
11:55:22 <ais523> ayacc uses it, but that shouldn't really be surprising
11:55:38 <b_jonas> except of course gcc which has a more complicated syntax for historical reasons
11:55:40 <fizzie> It does explicitly allow the "normal" getopt way.
11:55:41 <ais523> possibly the only time I've found Getopt::Std to be useful
11:55:45 <fizzie> "A standard utility may also be implemented to operate correctly when the required separation into multiple arguments is violated by a non-conforming application."
11:58:07 <b_jonas> I never used Getopt::Std in perl. I used Getopt::Long many times, though its default settings are idiotic (accepts + as an option starter) so I always cargo-cult this from a previous program:
11:58:12 <b_jonas> Getopt::Long::Configure "bundling", "gnu_compat", "prefix_pattern=(--|-)";
11:59:15 <fizzie> I like those programs that accept -foo to enable foo, and +foo to disable foo. They are delightfully unintuitive.
12:00:04 <b_jonas> fizzie: I wish programs started to use -t- as the negation of -t
12:00:22 <b_jonas> I mean, the hyphen can't be used as an option letter because of -- anyway, so this seems like the obvious syntax
12:00:26 <b_jonas> but no-one I've seen is using it
12:00:37 <b_jonas> they're just using --no-foo as the negation of --foo
12:00:46 <b_jonas> or another letter or something
12:00:50 <b_jonas> like -H being the negation of -h
12:01:09 <fizzie> splint uses +foo for turning foo on, and -foo for turning foo off. I can't remember what does the opposite, but I clearly remember it.
12:01:25 <b_jonas> yes, I know some programs do
12:01:29 <b_jonas> even with single-letter options
12:01:51 <b_jonas> but it's dangerous because you expect + to start a normal non-option argument
12:25:13 -!- Phantom_Hoover has joined.
12:42:55 <HackEgo> [wiki] [[RLS]] N http://esolangs.org/w/index.php?oldid=43019 * EzoLang * (+2189) Created page with "'''rLS''' (revised/reduced Lambdastack) is a stack programming language based on [[Lambdastack]]. It removes most of the ugliness and several features from the old one, but al..."
12:45:02 <HackEgo> [wiki] [[User:EzoLang]] M http://esolangs.org/w/index.php?diff=43020&oldid=39324 * EzoLang * (-18) Add rLS to language list and reformat
12:46:12 <HackEgo> [wiki] [[Language list]] M http://esolangs.org/w/index.php?diff=43021&oldid=42981 * EzoLang * (+14) Add rLS
12:48:00 -!- callforjudgement has joined.
12:48:16 -!- ais523 has quit (Disconnected by services).
12:48:18 -!- callforjudgement has changed nick to ais523.
12:59:57 -!- Guest77553 has quit (Read error: Connection reset by peer).
13:00:03 -!- M_I_Wright has joined.
13:00:37 -!- Wright__ has changed nick to Wright.
13:24:58 <Taneb> If I have a ring and a and b in the ring such that neither are 0, and a*b = 0, does a uniquely determine b?
13:25:22 <ais523> my initial thought is "no"
13:25:34 <Taneb> Worked it out just after I wrote it
13:25:35 <ais523> but my intuitions might be off because I've spent so long working with /semi/rings
13:25:39 <ais523> and thus I'm missing a couple of axioms
13:25:50 <Taneb> Z/8Z, 2*4 = 0, and 4*4 = 0
13:26:13 <b_jonas> ais523: because b=0 is always a solution, and there are other solutions in some rings
13:26:15 <ais523> Taneb: that's a ring? what are the multiplicative inverses?
13:26:21 <ais523> b_jonas: "neither are 0"
13:26:38 <ais523> I'll buy it as a semiring, but not as a full ring
13:26:52 <Taneb> ais523, rings don't have multiplicative inverses
13:26:53 <b_jonas> ais523: um, it's a _ring_. it doesn't have to have multiplicative inverses
13:27:06 <b_jonas> ais523: if it had multiplicative inverses, then it was a division ring aka skew-field
13:27:06 <ais523> so a semiring is a 3/4field?
13:27:20 <ais523> or 1/4field? or whatever?
13:27:26 <ais523> terminology is weird sometimes
13:27:37 <b_jonas> no, there's no such thing as a "semi-field
13:27:54 <b_jonas> (what were the crazy french terms for these two?)
13:29:50 <b_jonas> (“anneau à division” and “corps gauche” apparently)
13:31:22 <b_jonas> Taneb: yes, though “gauche” is used in a different meaning
13:31:32 <Taneb> My French is not very good
13:31:41 <b_jonas> the “corps” for “field” makes sense, it's reusing “field” in English that doesn't
13:32:26 <b_jonas> “field” is used as two unrelated mathematical root words in English that's distinguished in other languages
13:32:35 <b_jonas> English mathematical terminology is sometimes crazy
13:32:55 <b_jonas> mind you, it's not more crazy than those in other languages
13:33:17 -!- `^_^v has joined.
13:33:18 <b_jonas> ais523: no, English in general is way more crazy than other languages
13:33:28 <ais523> trying to get to sleep recently, I was going over words that formed gerunds via -tion and via -ing
13:33:32 <ais523> and completely failed to spot a pattern
13:33:39 <ais523> (also some verbs didn't form gerunds either way)
13:33:49 <Taneb> ais523, action vs acting
13:34:09 <b_jonas> the mathematical terminology is actually less crazy than most of English, and probably not more crazy than in other languages
13:34:28 <ais523> Taneb: oh bleh, those are both gerunds of different senses of "act", aren't they?
13:34:47 <b_jonas> I think two mathematical root words co-inciding happens in other languages too, it definitely happens in Hungarian because there's too few people inventing good maths terms for Hungarian
13:34:58 <ais523> b_jonas: are you one of them?
13:35:15 <ais523> actually I've seen this happening in game semantics
13:35:31 <ais523> there's debate about the meaning of "play" and "position", the meanings are swapped in some papers
13:35:52 <ais523> which, IMO, is evidence that an extended metaphor that doesn't fit properly is an awful way to produce mathematical terminology
13:36:20 -!- Patashu has quit (Ping timeout: 256 seconds).
13:36:49 <mroman_> Mathematicians are always trying to put a ring on it
13:36:53 <b_jonas> there's like ten root words that mostly miss a Hungarian equvalent, in particular, one of "disk" and "circle" and one of "ball" and "sphere" is missing, even though we _would_ actually have suitable short words for them but people aren't using them in maths,
13:37:39 <mroman_> they should teach university level math in english
13:37:53 <b_jonas> and it's worth in algorithms: nobody has figured out good enough words for "deque" and "trie" etc (and "stack" and "hash" have only half-good translations too)
13:37:58 <mroman_> mainly because if you read english papers you have no idea what they are talking about
13:38:12 <Taneb> I think they should teach it in Greek and Latin
13:38:17 <Taneb> With maybe a little German
13:38:18 <b_jonas> mroman_: that doesn't work, because when people enter the university, they don't speak enough English yet, they learn it in the first three years
13:38:27 <ais523> b_jonas: well, "deque" is an abbreviation for "double ended queue", so I guess take the same words in Hungarian then abbreviate into whatever seems pronounceable
13:38:28 <mroman_> b_jonas: Maybe in your country
13:38:29 <b_jonas> mroman_: teaching in Hungarian in the first few years reduces the latency of that
13:38:36 <mroman_> english is tought to little 8 year old kids in switzerland
13:38:48 <b_jonas> ais523: but it's also a crazy pun on "deck" in English
13:39:00 <ais523> (or unpronounceable, IME Hungarians tend to get quite good at pronouncing random series of letters)
13:39:03 <b_jonas> mroman_: right, that works in Switzerland and Sweden
13:39:07 <Taneb> b_jonas, I was taught to pronounce it "dee-cue"
13:39:12 <ais523> I was unaware of the pun and don't thing it's particularly important
13:39:21 <ais523> OTOH, "trie" definitely is a pun, but not really one that's worth saving
13:39:29 <b_jonas> Taneb: doesn't Knuth prescribe to pronounce it the same as "deck"?
13:40:01 <Taneb> I guess it's like a deck of cards
13:40:14 <ais523> b_jonas: well Knuth commissioned a new version of C-INTERCAL semi-recently
13:40:21 <mroman_> A deck of cards is a deque.
13:40:32 <mroman_> with random access if you're good enough
13:40:36 <ais523> so it may just be a complex act of trolling
13:40:52 <ais523> b_jonas: you probably wouldn't be surprised at how quickly that version came out :-)
13:40:55 <b_jonas> ais523: but I mean, iirc he specifically said it's a pun on "deck" and pronounced like that and that a deck of cards is a deque
13:41:07 <b_jonas> ais523: and writes this in TAOCP I believe, which is a serious enough work
13:41:09 <ais523> how often do people draw from the bottom of the deck?
13:41:12 <b_jonas> sure, it has jokes but still
13:41:21 <mroman_> ais523: More than they are willing to admit.
13:41:22 <Taneb> ais523, could I get a source for the Knuth C-Intercal thing?
13:41:36 <b_jonas> ais523: rarely, but that's because of power issues rather than because of it technically being hard
13:42:00 <b_jonas> ais523: but it does happen:
13:42:06 <ais523> Taneb: it was an email from his secretary to ESR that I eventually got copied into, but ESR mentions its existence here: http://esr.ibiblio.org/?p=2491
13:42:33 <b_jonas> in some games, the trump gets determined by flipping the bottom card of the deck, and sometimes you can even access the card above that, though even then to only a limited depth so it's not _really_ a deque
13:42:53 <ais523> (when reading that page as a whole, it's worth remembering what you know about me and mentally reconciling it with what you see written there)
13:43:13 <ais523> it seems incredibly unlikely to me that ESR forged the email, so I'm pretty sure that it's true
13:43:55 <b_jonas> can't you decide that by verifying that the Knuth reward cheque you get is authentic or something?
13:44:17 <ais523> the reward check is for finding mistakes in TAOCP, I think
13:44:24 <ais523> not for updating INTERCAL impls
13:44:32 <b_jonas> ais523: and other books and programs etc, but yeah
13:44:41 <b_jonas> oh, it's a bugfix in your program
13:44:46 <b_jonas> then yes, there'd be no check
13:44:59 <b_jonas> if it was a new feature he requested then there might be (of course he's under no obligation)
13:45:09 <ais523> the version ESR had at the time was very old, the bug may well have been fixed independently since then (I don't know what the specific bug was)
13:45:26 <mroman_> Who pays the refunds when Knuth has died?
13:45:42 <b_jonas> mroman_: that won't be your biggest problem
13:45:56 <ais523> the funny thing is that this probably sets a new record for "famous companies/people asking me for help with INTERCAL"; my previous record was maintaining CADIE for Google
13:46:03 <b_jonas> someone will step up if he hasn't named a heir, anyway
13:46:27 <ais523> which reminds me, could someone with a Github account export https://code.google.com/p/cadie/ >
13:46:42 <b_jonas> mroman_: if Knuth dies before you, your problem will be who finishes his books
13:47:02 <ais523> mroman_: google code is shutting down, someone needs to do an export to preserve the projects on it elsewhere
13:47:06 <ais523> I'm not sure what the process is like
13:47:25 <ais523> anyway: https://code.google.com/p/cadie/people/list : the only people who commit to CADIE are me and CADIE herself
13:47:29 <b_jonas> ais523: is there anything besides the git repository that has to be preserved?
13:47:33 <mroman_> I'm exporting it right now
13:47:58 <ais523> although luckily she can take care of herself mostly
13:48:06 <ais523> b_jonas: I don't think so, the style guide's in the repo
13:48:10 <mroman_> https://github.com/FMNSSun/cadie
13:48:14 <ais523> it went up, then went down again, but then went back up in the repo
13:48:32 <ais523> anyway, CADIE was a teenager (or acting like one) back in 2009, she's grown up somewhat since
13:48:41 <ais523> if you're a teenager in March 2009, you're an adult in May 22
13:49:32 <ais523> it's an april fools joke
13:49:40 <ais523> so yes, just two files
13:49:48 <ais523> basically because nobody wanted to write anything large in INTERCAL
13:49:57 <ais523> for all I know it was generated with yapp
13:50:23 <ais523> yeah, that looks a lot like yapp output actually in retrospect
13:50:31 <ais523> so the only actual INTERCAL programming involved was done by me
13:51:09 <ais523> yapp has much better compression
13:53:26 <b_jonas> ais523: have you read http://www.madore.org/~david/weblog/d.2015-03-20.2284.html ?
13:53:39 <ais523> possibly, I recognise the author at least
13:53:40 <ais523> let me look at the page
13:53:46 <ais523> ah no, I haven't read it
13:57:03 <b_jonas> can parts of it be from yapp "linked" with hand-written parts?
13:58:29 <mroman_> People not considering learning Esperanto are suckers :p
13:59:05 <ais523> b_jonas: I don't think so, I compared it to yapp output, it's different enough that it would need a total rewrite
13:59:51 <ais523> it's much more like convickt output – almost visually identical – except that convickt can't actually generate that sort of program
14:00:09 <ais523> and if someone had extended it to do that, I'd have hoped they'd have contributed the patch back again :-(
14:00:24 <b_jonas> ok, then how about a convickt output "linked' together with something handwritten?
14:00:29 <ais523> mroman_: what about Lojban? there's at least one casual Lojban speaker here (tswett)
14:00:41 <mroman_> also eo.wikipedia has way more articles than most real languages have
14:00:42 <ais523> b_jonas: the problem is that convickt output produces the wrong numbers
14:00:53 <ais523> you'd need to do a running sum on them, or possibly a running difference
14:01:56 <b_jonas> but mostly I'm infuriated with its crazy eso-grammar that I'm still trying to figure out how it can be modified consistently, because it's _so_ much not LR(1) a grammar and some some crazy custom preprocessing to become LR-parsable that it's not funny
14:02:27 <b_jonas> ais523: I don't know what convickt is or what program you're mentioning or anything, I'm just asking
14:02:37 <mroman_> ais523: Esperanto looks nicer
14:02:42 <mroman_> and probably has more speakers
14:02:42 <ais523> b_jonas: convickt converts between character sets
14:02:52 <ais523> it's basically an INTERCAL-specific version of iconv
14:03:17 <b_jonas> but probably that's what I should expect from intercal stuff, yeah
14:04:01 <mroman_> That's more than twice those greek folks have
14:04:45 <b_jonas> wait, you don't mean "pages" in the sheet of paper of text sense?
14:05:27 <ais523> presumably in the software sense
14:05:44 <mroman_> b_jonas: wikipedia articles
14:05:47 <ais523> the number of legal "title=" parameters
14:07:09 <ais523> err, that don't give you redlinks
14:07:21 <ais523> otherwise it's 255^256
14:08:00 <b_jonas> I was wondering if it somehow meant pages in the sense of learner or something
14:08:15 <b_jonas> ais523: no way, there's way more legal title parameters because there's some normalization rules
14:08:30 <b_jonas> like, whitespace stripped from the end already gives tons
14:08:51 <b_jonas> ah right, mroman_ did mention wikipedia indeed
14:10:27 <mroman_> ais523: You could learn Tok Pisin .
14:11:39 -!- adu has joined.
14:11:54 <mroman_> and it's probably easier than esperanto
14:12:00 <mroman_> since it only has a few hundred words
14:12:30 <mroman_> What I don't understand is wikipedia in regional dialects
14:12:30 -!- adu has quit (Client Quit).
14:12:35 <mroman_> I mean sure.. I like swiss german
14:12:45 <mroman_> but there's no point in maintaining a seperate wikipedia
14:13:11 <mroman_> it's too much effort for what it offers
14:13:58 -!- solid_whiskey has joined.
14:14:10 <b_jonas> mroman_: you only have to maintain the one in the one true language everyone should use
14:14:25 <b_jonas> everyone agrees in that, they just don't agree which language that is
14:14:30 <mroman_> b_jonas: Well that's why I'm learning Esperanto
14:14:34 <ais523> b_jonas: I'm having fun guessing which language you think that is, but I suspect you aren't thinking of one in particualr
14:14:58 <mroman_> obviously the third world war will be about languages
14:15:34 <b_jonas> ais523: I'm not thinking of one in particular, because I'm lucky, I'm not one of those people who have to decide whether serbo-croatian is one, two, three, four, or five different languages, and follow which of the seven language codes for it map to which subsets of the five different ones.
14:15:35 <mroman_> either the third or the fourth
14:16:02 <b_jonas> ais523: there's actually a separate serbo-croatial language wikipedia
14:16:16 <b_jonas> for those who think it's one language
14:16:42 <mroman_> but I'd teach them Esperanto
14:17:28 <mroman_> although Esperanto lacks on official pronunciation I think
14:18:11 <mroman_> I hate "al la" combinations in Esperanto
14:18:33 <mroman_> all natural languages will eventually invent short forms for those.
14:22:46 <mroman_> b_jonas: you mean everbody agrees except those native english folks
14:23:02 <mroman_> I thought those were the only ones not wanting to learn a new language
14:23:25 <mroman_> because their children are already overwhelmed with learning all those subjects and can't be bothered with more stuff
14:23:37 <mroman_> where we bother our children with TWO foreign languages
14:24:07 <mroman_> although 95% of those at age of 24 will have unlearnt one of those foreign languages because nobody uses it
14:24:32 <mroman_> Everybody knows it's a useless thing but due to political reasons they have to learn it
14:24:33 -!- solid_whiskey has quit (Remote host closed the connection).
14:24:45 <mroman_> It's like one of those scenarios where you marry a princess of another country to maintain peace
14:25:12 <mroman_> we maintain peace between our countries region by promising to learn each others language in school and then forget them after school
14:25:59 <ais523> mroman_: are you Swiss, then? I have an unfortunate habit of failing to guess that people are Swiss on IRC
14:26:03 <b_jonas> mroman_: no. the native English folks do want that everyone use their one true form of language, and everything else is a travesty, and think that there's no such thing as international english, and that even if the whole world is using english they have no right to prescribe what english is supposed to be like, and that only they, the speakers of the one true language, determine it,
14:26:08 -!- hilquias has joined.
14:26:17 <b_jonas> but they haven't so far tried to suggest that English is more than one languages afaik.
14:26:44 <ais523> b_jonas: I've been known to give non-native-English speakers advice on the differences between US and UK English
14:26:45 <mroman_> ais523: Switzerland is the one country where you can't study chemistry without being good enough in French.
14:27:11 <ais523> mroman_: ever since IUPAC got involved, I thought most names of chemicals were basically identical in all languages?
14:27:22 <mroman_> It's like: So.. you have a 5 in chemistry, a 5 in math and a 4.5 in physics but a 2 in french? YOU CAN NOT STUDY ANYTHING!
14:27:34 <b_jonas> ais523: I don't much follow those differences, and write a crazy mixture with "color" and "behaviour" mostly (though I try to train myself to type "behavior" these days)
14:28:09 <mroman_> ais523: To be able to attend universities you need to pass exams in French ;)
14:28:22 <b_jonas> mroman_: do people have to learn both hochdeutch and swiss german?
14:28:25 <ais523> I know at the University I work at, you need to prove you can understand English well enough to understand the lectures
14:28:34 <ais523> the normal method is via a prior exam in English, but there are other ways
14:28:43 <mroman_> b_jonas: swiss german is what we speak in the swiss german part of switzerland.
14:28:51 <ais523> up to "taking a year of remedial English before starting the course" if you really want to study but don't know the language (you have to pay for it, though)
14:28:56 <mroman_> Hochdeutsch is what we learn in school.
14:29:03 <mroman_> besides French and Englisch.
14:29:34 <fizzie> I don't know what to write in emails at work, because I'm in UK, but I'm in no way British.
14:29:34 <mroman_> ais523: French is part of "general education"
14:29:50 <b_jonas> mroman_: yes, but I mean do people have to learn to speak both, and how difficult overhead is that over knowing just one?
14:30:04 <ais523> fizzie: people don't care much; I know my habit is to avoid salutations and valedictions, and use a very short custom sig
14:30:12 <mroman_> If you suck at French there's no way you can study anything.
14:30:30 <mroman_> Even if you are brilliant in math and those subjects
14:30:42 <ais523> hmm, I wonder how it compares to, say, English versus Scots (not Scottish Gaelic)
14:30:48 <ais523> it's similar enough to English that it's mostly intelligible
14:30:52 <Taneb> fizzie, write whatever you are most comfortable with, I guess
14:31:03 <mroman_> swiss german isn't a language
14:31:07 <ais523> but the words are spelled and pronounced differently and you get the occasional word that's completely different
14:31:17 <Taneb> ais523, I was under the impression that English vs Scots is like Norwegian vs Danish
14:31:35 <mroman_> If you live here you are expected to be able to at least understand swiss german
14:31:56 <ais523> Taneb: that's believable I guess? mostly because I don't know either Norwegian or Danish
14:31:57 <mroman_> I.e. if you're german and move to switzerland you should be able to understand swiss german
14:32:08 <mroman_> but that's just a "cultural requirement"
14:32:13 <ais523> thus it isn't a very high bar to get me to consider things about them to not be obviously false
14:32:15 <Taneb> As in, mutually intelligible if they both talk slowly and clearly
14:32:31 <Taneb> And not immediately clear where the border between the languages is
14:32:35 <mroman_> swiss german isn't part of school or anything
14:33:32 <mroman_> if you're a hungarian and move to switzerland
14:33:38 <mroman_> yes, you should learn swiss german and german
14:33:58 <b_jonas> Taneb: yeah, as in you can understand adults who know how to speak in a way you understand, but you don't understand children
14:34:05 <mroman_> but you only really need to learn german (for official matters)
14:34:30 <b_jonas> and I think there's even some assymetry between Swedish and Danish where the words in one is easier to guess from the words in the other than backwards
14:34:43 <mroman_> but since german is only a written language in switzerland you eventually should be able to use swiss german as well
14:35:23 <b_jonas> mroman_: but even then that means that native swiss german people living there learn both, doesn't it?
14:35:47 <mroman_> swiss german is what you learn as a baby from your mother/father etc.
14:35:54 <mroman_> and german is what they'll teach you in school.
14:36:13 <Taneb> So, it's a local non-standard dialect?
14:36:27 <b_jonas> so what I'm asking is, how different are those? how difficult is it to learn both as opposed to learning just one.
14:36:40 <Taneb> A bit like geordie, say?
14:37:00 <b_jonas> that doesn't help me, sorry.
14:37:03 -!- Weloxux has joined.
14:37:05 <mroman_> That depends on what swiss german dialect exactly
14:37:14 <mroman_> there are some minor grammatical differences, pronunciation differences
14:37:18 <mroman_> and vocabulary differences
14:37:24 <mroman_> but all in all they are very alike
14:37:54 <mroman_> Isch habb's'm schunn vazehld, awwa där hod ma's nid geglawd
14:38:04 <b_jonas> is it sort of like with Austrian english (where they pronounce zwei as "zwo")? or more different?
14:38:04 <mroman_> Ich habe es ihm schon erzählt, aber er hat es mir nicht geglaubt
14:38:17 <mroman_> I has em scho verzehlt, aber är häts mer nöd glaubt.
14:38:37 <ais523> b_jonas: I was under the impression that "zwo" was an invented word to not be confused with "drei" over the phone
14:38:43 <ais523> sort-of like "niner" in English
14:39:00 <b_jonas> ais523: maybe it's that too, but I think Austrian German consistently pronounes it like that usually
14:39:20 <ais523> perhaps it caught on unexpectedly well
14:39:35 <b_jonas> They use “kettő” and “hetes” for that purpose here sometimes, though “kettő” is a pre-existing word and is used for other things too.
14:40:11 -!- lleu has joined.
14:40:14 <b_jonas> could you really confuse "zwei" with "drei"?
14:40:25 <mroman_> "zwei", "zwo", "zwee", "zwyy"
14:40:29 <ais523> they're pretty similar over a noisy connection
14:41:05 <b_jonas> afterall, “két” and “hét” aren't confused only with each other, but even with “négy”
14:41:05 <mroman_> swiss german mostly lacks genitive case
14:41:16 <mroman_> although some swiss german dialects have a genitive case
14:41:25 <b_jonas> you must be careful with numbers
14:41:54 <b_jonas> fingers can help when you're not in telephone
14:42:00 <ais523> b_jonas: because the é is all you can really hear?
14:43:13 <mroman_> b_jonas: but to answer the question: The biggest difference is pronunciation
14:43:20 <mroman_> which is slightly different for every dialect
14:43:43 <mroman_> some sei "haben", others say "ham" others say "händ" others say "hend"
14:43:56 <mroman_> others say "habn" dropping the e
14:44:49 <mroman_> every dialect has it's own vocabulary though
14:45:13 <mroman_> like uhm lift <-> elevator in english
14:45:21 -!- solid_whiskey has joined.
14:46:04 -!- solid_whiskey has quit (Max SendQ exceeded).
14:46:06 <ais523> mroman_: that's UK english vs. US englsih
14:46:26 <b_jonas> "lift" and "elevator" isn't even close to the craziest
14:46:33 -!- solid_whiskey has joined.
14:46:40 <mroman_> http://www.forums9.ch/sprachen/Rosetta.htm
14:46:56 <ais523> there's the words which are valid in both languages with different meanings, but similar enough to be confused without a lot of additional context
14:47:00 <ais523> "petrol" is a good one
14:47:15 <ais523> US:petrol = UK:paraffin, UK:petrol=US:gasoline
14:47:31 <ais523> but paraffin and gasoline are sufficiently similar that if the context isn't just right, you can be confused for many sentences
14:47:47 <mroman_> some swiss german dialects are to german what scottish is to southern texas accents
14:48:07 <ais523> i.e. completely different?
14:48:07 <mroman_> same language, different pronunciation, some vocabulary differences
14:48:39 <mroman_> ais523: to a foreigner probably yes
14:49:24 <mroman_> Without prior knowledge you wouldn't know that swiss german and german are the same languages
14:49:26 <Gregor> <ais523> but paraffin and gasoline are sufficiently similar that if the context isn't just right, you can be confused for many sentences // I have NEVER heard anyone say "petrol" when they mean paraffin.
14:49:58 <ais523> you hear "gasoline" occasionally in the UK, too (although "petrol" is still more common)
14:50:15 <ais523> maybe people are trending towards less ambiguous words over time
14:50:39 <ais523> b_jonas: "gas pedal" is heard of in the UK, and generally using it as a metaphor for speed/acceleration
14:50:53 <Gregor> Muahaha American imperialism
14:51:12 <ais523> "gas" for the actual volatile liquid, not really
14:51:16 <mroman_> meaning they have a limes of some sort
14:51:27 <ais523> mroman_: limes as in the esolang logo?
14:52:02 <mroman_> We say "limes" for the lim x -> foo stuff
14:53:52 <Gregor> LimeSTONE is, I don't think "lime" is used inyeah
14:54:04 <ais523> there's also a chemical called lime
14:54:09 <ais523> which is pretty different from limestone
14:54:11 <mroman_> a lime lime was laying on lime lime.
14:54:25 <ais523> I think the colour is named after the fruit
14:54:27 <Gregor> And of course the color came from the fruit, like orange.
14:54:38 <mroman_> there's a tree called lime
14:54:49 <mroman_> a lime lime was laying on lime lime under a lime.
14:54:52 <Gregor> No, it's called a lime tree.
14:55:15 <mroman_> except that limes don't grow on lime trees
14:55:48 <ais523> (this is not something I'd expect most non-native speakers to know/guess, but it's true)
14:55:48 <mroman_> leo.org says "lime" - Linde
14:56:01 <mroman_> but limes (Limetten) don't grow on lime (Linde)
14:56:21 <Gregor> Although I know lime trees aren't trees that grow lime because UK logic, I didn't think anyone called the trees just "lime"
14:56:28 <b_jonas> what the heck are lime tree?
14:56:33 <ais523> right, people always call them "lime trees"
14:56:49 <ais523> they're famous for being eaten by aphids, who then excrete a sticky substance
14:56:59 <ais523> meaning that it's considered a bad idea to park under one because it takes ages to clean your car afterwards
14:57:07 <b_jonas> who invents these stupid words
14:57:16 <ais523> people who don't realise they're already used
14:57:26 <mroman_> "to smear with a sticky substance"
14:57:49 <ais523> that's probably related to what happens if you park under a lime tree
14:57:58 <Gregor> Buffalo buffalo Buffalo buffalo buffalo buffalo Buffalo buffalo Buffalo buffalo buffalo.
14:58:00 <mroman_> your car get's limed by a lime tree
14:58:59 <b_jonas> But then where do lime fruits grow? Supermarkets?
14:59:21 <b_jonas> I mean, I know they can't grow in Britain because the climate is wrong
14:59:43 <Gregor> From lime trees, but not lime trees.
15:00:10 <Gregor> I think people usually say the specific species of lime. Key lime trees, kaffir lime trees, etc.
15:01:05 <ais523> wait, key lime pies are made of key limes?
15:01:20 <Gregor> Unless they're a filthy lie.
15:01:41 <b_jonas> ais523: they're probably some sort of search trees with keys
15:02:53 -!- callforjudgement has joined.
15:04:00 -!- ais523 has quit (Disconnected by services).
15:04:02 -!- callforjudgement has changed nick to ais523.
15:05:09 -!- Herbalist has joined.
15:06:42 -!- solid_whiskey has quit (Ping timeout: 264 seconds).
15:19:19 -!- solid_whiskey has joined.
15:24:12 -!- solid_whiskey has quit (Read error: Connection reset by peer).
15:44:52 <Gregor> There's a species of lime called Kaffir limes.
15:44:59 <Gregor> I know the name from Thai food.
15:49:52 -!- MDream has changed nick to MDude.
15:52:29 <Phantom_Hoover> Gregor, i wonder if it actually is cognate with the south african slur
15:54:31 <quintopia> gregor: that buffalo sentence seems to be a tautology
15:55:15 <Gregor> quintopia: That particular formulation is tautological, yes.
15:57:26 <quintopia> only because of the last three buffalo
15:58:17 -!- variable has joined.
15:59:46 <quintopia> if you have a 2-place function f, what do you call the 2-place "inverse" functions g and h such that g x f x y = y and h y f x y = x? is there a naming scheme that generalizes this to more arguments?
16:00:16 <ais523> "left inverse" and "right inverse" are the normal names I see; I don't know of a generalization of the naming scheme
16:00:54 <b_jonas> ais523: left inverse and right inverse are different I think
16:01:01 <b_jonas> are they used for this too?
16:01:24 <ais523> I think they might have exactly two meanings (with the other one being composition-related)
16:03:18 <Phantom_Hoover> quintopia, well if f is curried the former is just the inverse of f x
16:06:33 -!- variable has changed nick to trout.
16:08:52 -!- hjulle has joined.
16:45:37 <tswett> quintopia: do you mean g x (f x y) and h y (f x y)?
16:46:23 <tswett> I'd call them the "left inverse with respect to the first argument" and the "left inverse with respect to the second argument".
16:47:15 <tswett> "Left inverse" because you never specified that f x (g x y) = y and f (h y x) y = x.
16:48:57 -!- atrapado has joined.
17:09:09 <oren> `run echo 'X=0 X(1=1 X(1_X _pX' | scrip7
17:09:09 <HackEgo> -1056589062271330492704679569833033213037694652072243044255921418053347805113449718948834511775314375789348789986514257357764695119005371074501077956925879153816773367998010168337463035352852882106048465816422376808296056585503123477676793797534072952979077161795475996672.000000 \ bash: line 1: 293 Done echo 'X=0 X(1=1 X(1_X
17:09:12 -!- GeekDude has joined.
17:09:48 -!- hilquias has quit (Remote host closed the connection).
17:09:54 <oren> `run echo 'X=0 Y>1 Y=1 Y_X _pX' | scrip7
17:09:57 <b_jonas> oren: I think he's printf "%f"-ing a large number
17:10:04 <oren> `run echo 'X=0 Y>1 Y=1 Y_X _pY' | scrip7
17:10:17 <b_jonas> `perl -eprintf"%f",sqrt(10)*1e90
17:10:22 <HackEgo> 3162277660168380149484908708480656937183007146037772679597241557288343883945532816210526208.000000
17:10:48 <oren> Yeah but the code wasn't supposed to give a large number
17:11:05 <oren> It was supposed to give infinity
17:11:18 <b_jonas> `` <<<'X=0 Y>1 Y=1 Y_X _pY' scrip7 # we have a proper bash here, don't we?
17:11:47 <oren> `run echo 'X=-1 Y>1 Y=1 Y_X _pY' | scrip7
17:11:48 -!- oerjan has joined.
17:12:27 <oren> the operator does multiply by the log, even though I don't remember making it do that
17:13:02 <b_jonas> is there an easy way to printf %g though?
17:13:18 <oren> I think the x operator
17:13:25 <b_jonas> wait, what do all those underscores even mean? underscore isn't even a variable, is it?
17:13:51 <oren> underscore, as a variable, is a null var which is always zero and does nothing when written to
17:14:02 <tswett> https://karpathy.github.io/2015/05/21/rnn-effectiveness/ - neural nets which are capable of generating random Wikipedia text.
17:14:03 <b_jonas> but in Y_X it's a command?
17:14:04 <oren> underscore as a operator, is log
17:14:09 <tswett> http://cs.stanford.edu/people/karpathy/char-rnn/wiki.txt - the generated text.
17:14:29 <tswett> "Naturalism and decision for the majority of Arab countries' capitalide was grounded by the Irish language by [[John Clair]], [[An Imperial Japanese Revolt]], associated with Guangzham's sovereignty."
17:14:38 <tswett> "'''See also''': [[List of ethical consent processing]]"
17:15:43 <tswett> It also generates random Linux source code.
17:16:29 <tswett> The syntax is almost always correct, as is the indentation. Variable names are almost never correct.
17:16:57 <tswett> The code is, of course, commented.
17:17:49 <tswett> /* Various new destinations in associate data */
17:18:17 <tswett> void arizona_set_at86rfb(struct arizona_hw *ah, u8 *period);
17:18:39 <tswett> /* note: skb_info struct templates have extra read buffers */
17:19:30 <tswett> /* Software socket driver stuff */
17:20:42 <ais523> to be fair, humans don't do a good job of producing working source code by looking at examples they don't understand either
17:21:42 <oerjan> <ais523> I still don't really understand it <-- i thought i'd explained it on the talk page?
17:21:47 <tswett> It occasionally generates a random address for the Free Software Foundation.
17:22:07 <tswett> "You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA"
17:22:18 <ais523> oerjan: let me read that
17:22:44 <oerjan> hm possibly i just gave a hint, but it's important
17:22:48 <tswett> I guess that is, in fact, a real address that presumably appears in the Linux source code somewhere.
17:22:53 <tswett> It just isn't the address of the FSF.
17:23:00 <ais523> oerjan: keymaker posted a stack trace but it doesn't help much
17:23:05 <ais523> tswett: it looks like a real address
17:23:34 <oerjan> ais523: the thing is that the whole is really constructed from that fragment i pointed out
17:23:52 <ais523> I might have another try later
17:24:18 <oerjan> ^ul (Y)aa((!(X)))*:*^!**^SS
17:37:12 -!- GeekDude has quit (Ping timeout: 264 seconds).
17:38:00 -!- Weloxux has quit (Quit: Leaving).
17:41:05 <HackEgo> [wiki] [[Talk:Underload]] http://esolangs.org/w/index.php?diff=43022&oldid=37706 * Ais523 * (+710) /* Why the reserved characters? */ some info about Overload
17:41:58 <HackEgo> [wiki] [[Talk:Underload]] M http://esolangs.org/w/index.php?diff=43023&oldid=43022 * Ais523 * (+116) fix formatting
17:44:21 <quintopia> tswett: that's kind of what i had been saying, but it seems so wordy
17:45:17 <pikhq> tswett: Whose is it then?
17:46:45 <tswett> Ah, I lied. That is a correct address for the FSF.
17:47:09 <pikhq> Actually, it is not the current correct address for the FSF.
17:47:24 <ais523> maybe we should have some other organization to write to if you don't have a copy of the GPL
17:47:28 <pikhq> I have no idea if mail to that address still reaches that location.
17:47:55 <pikhq> Erm, still reaches the FSF.
17:48:29 <zzo38> If it is wrong then they should update the program?
17:49:08 <b_jonas> ais523: these days they just include a http address in the standard short text
17:49:19 <ais523> zzo38: this is boilerplate that the FSF recommended including at the start of every GPL-licensed file
17:49:31 <ais523> they can't go and change everyone else's files to update the postal address
17:49:55 <b_jonas> but postal addresses can be chosen such that they work for a very long time in civilized countries
17:50:05 <ais523> I'm sort-of tempted to write to them to ask for a copy of the GPL just for fun
17:50:17 <b_jonas> you can use redirection, give the address of a university which rarely moves and the post will know its address even if it does move
17:50:53 <b_jonas> eg. you can probably contact me twenty years in the future in snail mail if you write to the university department. they'll probably have my contact even if I'm not working there.
17:51:08 <pikhq> ais523: You should also buy copy of GCC on tape.
17:51:24 <b_jonas> ais523: send a SASE to the ubuntu guys and ask for free stickers instead
17:51:47 <pikhq> *Aaaw*, they finally stopped that.
17:52:08 <b_jonas> stopped what? tape or stickers?
17:52:28 <pikhq> For a long time RMS did that to try and raise funds.
17:52:39 <pikhq> It was kinda hilarious circa 2000.
17:53:28 <b_jonas> buy gcc on floppies instead
17:53:44 <b_jonas> or a full linux distro on floppies rather
17:54:53 <zzo38> Presumably a kind of small distro that can fit on three floppy disks
17:55:34 <tswett> tomsrtbt fits on a small number of floppies.
17:57:24 <pikhq> It's kinda hard (though not impossible) to fit a modern Linux setup on floppies these days.
17:57:31 <pikhq> The kernel is likely to need its own disk.
18:11:12 -!- GeekDude has joined.
18:13:25 <oren> Apparently you can at least fit a bootloader on a floppy.
18:14:28 <oren> The question is, is it possible to bootload froma floppy and then download the kernel fromt eh internet
18:14:45 -!- trn has quit (Remote host closed the connection).
18:15:26 <ais523> I'd be surprised if it were impossible to write a working Linux installer inside GRUB 2, without a separate kernel
18:15:47 <ais523> i.e. you get it to download and run the kernel from online, and that bootstraps the rest of the installer
18:16:25 <pikhq> No reason you couldn't. Boot over a network is a thing that is actually done.
18:16:51 <pikhq> Though usually with BOOTP and TFTP rather than (as you'd probably prefer here) DHCP and HTTP.
18:16:58 <b_jonas> ais523: why would you do that? you can just fit the kernel and a small initrd to the floppy
18:17:13 <b_jonas> it'd have to be smaller than usual, but it's possible
18:17:21 <ais523> b_jonas: to save room on the floppy
18:17:33 <ais523> you only have 720K to work with
18:17:54 <ais523> (also, double density floppies are the oldest for which hardware to read them is still reasonably available)
18:18:33 <pikhq> More bizarre thought: EFI libc. You could totally do that.
18:18:51 <b_jonas> if it had to be 720K, then I'd put MS-DOS and a small DOS program on that floppy, that program copied the files from that floppy and the next one to six floppies to hard disk or ramdisk,
18:19:03 <b_jonas> and then ran loadlin to boot it
18:19:16 <b_jonas> can loadlin load a kernel and initrd from ramdisk? I never tried
18:19:23 <b_jonas> it definitely works if you use a hard disk
18:19:40 <b_jonas> though then you need some reboots because you may have to make a partition on the hard disk first
18:20:11 <ais523> pikhq: I thought it had been done already
18:20:20 -!- trn has joined.
18:20:22 <ais523> b_jonas: not "and the next one"
18:20:42 <b_jonas> ais523: oh, if it's just one 720K floppy then you're screwed
18:20:50 <pikhq> I don't see why it can't. It loads the kernel into RAM normally and only then shoves off DOS.
18:20:56 <b_jonas> I think even with two 720K floppies you're screwed
18:21:09 <pikhq> Huh, does the EFI toolkit have a halfway decent libc?
18:22:00 <b_jonas> the debian x86_64 tiny installer (which continues the install from network, loading most components of the installer from network actually) is actually 18 megabytes size these days
18:22:13 <b_jonas> of which 15 megabytes is the initrd and 3 megabytes is the kernel
18:22:47 <b_jonas> when I first heared that people have ran linux 1.* on machines with only 1 megabyte of ram, I didn't believe it
18:23:05 <b_jonas> at those days the common wisdom was that you need at least 2 megabytes of ram even for a minimal system
18:23:10 <b_jonas> these days it's more like 8 megabytes
18:23:45 <b_jonas> anyway, if you want to make it work on just most configurations, it might be possible from one floppy
18:23:59 <b_jonas> the installer is so large because it has to have drivers for unusual hardware too, especially for unusual network stuff
18:24:04 <pikhq> Totally *has* a libc, but it looks like crap.
18:24:23 <b_jonas> pikhq: libc always looks like crap. duh.
18:24:37 <b_jonas> "decent libc" is a contradiction
18:24:48 <b_jonas> libc has to do the dirty work, it must be ugly
18:25:08 <pikhq> "Decent" here would mean "implements as much of POSIX as is practical given the environment limitations"
18:25:15 <b_jonas> for reasons like compatibility and shared libraries and stuff like that
18:25:40 <pikhq> Also, writing a libc with not-crap code is a lot more reasonable than you'd think.
18:25:42 <b_jonas> but wouldn't that make it too large? you don't want a multi-megabyte sized efi bios, right?
18:26:01 <pikhq> Note that the libc would be linked into the binaries.
18:26:17 <pikhq> Also, a *full* POSIX libc is like 700k.
18:26:25 <pikhq> If it's not crap. :)
18:27:04 <b_jonas> yeah, but you need stuff other than the libc too
18:27:33 <b_jonas> oh, and the boot loader has to do work in like four different virtual cpu architectures and switch between them during the boot process, or maybe fewer for efi
18:27:48 <pikhq> (and the libc would not be provided by EFI, it's supposed to be linked into EFI binaries -- what EFI provides is basically a syscall layer.)
18:28:14 <b_jonas> yes, that makes more sense
18:28:21 <pikhq> (... and by "would be" I mean "is" -- there is a libc in the EFI dev kit. It's just not fully featured.)
18:30:02 <pikhq> Nothing preventing them from making it implement the subset of POSIX that isn't patently absurd to implement.
18:30:58 <oerjan> <ais523> this violates my mental model of substructual logics <-- the very first trick is the observation that (a)(b)*:* contains b followed by a inside it. the next step is getting rid of the surrounding junk.
18:31:51 <pikhq> (hint, fork() is probably patently absurd.)
18:32:13 <b_jonas> pikhq: but before you have a libc, you need like an execution environment (abi) and what the program is allowed to do
18:32:25 <b_jonas> and there's like four or more of those in a boot loader depending on what stage it is in
18:32:29 <b_jonas> and you'd need separate libc for each
18:32:43 <b_jonas> maybe it's a bit better with efi though
18:32:51 <pikhq> Um, I think you're not entirely understanding what EFI is like.
18:33:04 <b_jonas> not that I understand that either
18:33:13 <b_jonas> but know much more about it than about efi
18:33:25 <pikhq> EFI provides an x86 or x86_64 (depending on the system) execution environment with a syscall layer, filesystem, program loader, drivers, etc.
18:34:35 <pikhq> Yep. EFI boots and loads a program (well, one of a few possible programs) off of the EFI boot partition, which is a FAT filesystem.
18:35:27 <b_jonas> you can have limited read-only filesystem drivers in very small space actually
18:35:35 <b_jonas> I don't know how they do it
18:35:57 <b_jonas> supposedly it even does the filename hashing and b-tree search on modern filesystems, rather than just traversing the whole directory to find the filename
18:36:12 <b_jonas> I don't know if that's true but it's the rumour I heared about grub-l
18:36:21 <pikhq> It's a read-write FS implementation.
18:36:34 <b_jonas> yeah, on FAT that's possible
18:36:54 <b_jonas> on real filesystems I wouldn't dare to put it in a boot loader
18:37:07 <pikhq> Apple's implementation is somewhat more impressive -- it's HFS+.
18:37:24 <b_jonas> but I'm satisfied with using a separate boot partition that is required to be in the first two terrabyte of the disk or something
18:37:40 <pikhq> It isn't required to be in the first two terabytes.
18:37:58 <b_jonas> I mean for bios and bios boot loaders
18:38:01 <pikhq> Only if you're using an MBR partitioning scheme.
18:38:09 <pikhq> (EFI supports this)
18:38:23 <pikhq> (... you can also do BIOS boot with GPT partitions)
18:38:36 <b_jonas> because the bios doesn't touch the partitions
18:38:39 <b_jonas> it's the boot loader that doe
18:39:01 <b_jonas> for hard disks and floppy disks, the bios boot is very simple
18:39:07 <pikhq> Yeah, the BIOS's notion of the disk is "load 512 bytes".
18:39:25 <b_jonas> yes, it loads one sector of the boot loader to a fixed address in x86_16 real mode, and jumps to it
18:39:41 <b_jonas> and gives an interface to read or write sectors of the disk
18:39:49 <pikhq> After starting up the hardware to some extent and setting up some interrupts.
18:39:56 <b_jonas> for booting from cd/dvd and network it's much more complicated
18:40:30 <pikhq> Not for the bootloader. By the time you're in the bootloader the BIOS is faking the same interface.
18:40:31 <b_jonas> it gives an easy interface for the keyboard and the vga display and serial port, which is very useful for a boot loader
18:40:57 <pikhq> Though if you want to break out from the fake hard drive or floppy drive it's booting from (to, say, "load a kernel") you're in for a rough time. :)
18:41:17 <b_jonas> pikhq: no, I think booting from cd with bios actually has two modes, one is the floppy emulation that fakes the floppy, but the other actually loads more than one sector and doesn't emulate a floppy
18:41:40 <pikhq> The other is actually a hard drive emulation rather than a floppy emulation.
18:41:52 <zzo38> What you will really need is a true PC BIOS which is open-source and has a Forth environment built-in (and will execute even if there is no hard disk, external disk, network, or anything else other than the keyboard and monitor)
18:42:06 <b_jonas> I thought there was a mode where it loads the whole boot loader, multiple sectors of it, and lets it access the cd, though not in an easy way
18:42:13 -!- lleu has quit (Quit: That's what she said).
18:42:35 <b_jonas> zzo38: no, a forth environment wouldn't make it a true PC BIOS. it would have to be a BASIC environment.
18:42:44 <pikhq> Nope, still a boot sector, it just maps the CD as a hard drive. Apparently.
18:42:45 <b_jonas> called OpenROM-BASIC or some such
18:42:54 <zzo38> Yes, BASIC is how the PC originally did it.
18:43:37 <b_jonas> not the original BASIC that loads from casettes of course, but, say, a 32-bit one that lets you read sectors and fat file systems from hard disk and floppy disk
18:43:39 <zzo38> But as long as the BIOS function calls are proper, you could put Forth instead.
18:44:26 <b_jonas> I've written bootable grub cds but it's possible that that's what it does internally
18:44:48 <pikhq> b_jonas: There was also a 16 bit one that did that. It was called IIRC "ABASIC". Which used the ROM BASIC and patched it with extra stuff.
18:45:07 <zzo38> You can run Forth in unreal mode so that you can access the full memory and so on; the built-in BOOT command switches back to normal real mode
18:45:55 <pikhq> zzo38: You can also use 32-bit prefixes on the instructions and access full memory from normal real mode. :)
18:46:09 <b_jonas> hmm, let's put a doom in the bios
18:46:26 <b_jonas> you can use 32-bit prefixes, but that doesn't let you access full memory
18:46:26 <fizzie> b_jonas: Segment limits aren't reset when switching to 16-bit mode.
18:46:32 <b_jonas> you're limited to the first 1 megabyte
18:46:56 <fizzie> b_jonas: If you never load any values to the segment registers after switching back, you can keep using more than that; that's the "unreal mode".
18:47:12 <pikhq> b_jonas: Erm, right, yeah. Segments.
18:47:13 <b_jonas> fizzie: what? I though the cpu docs forbids that
18:47:36 <b_jonas> they do forbid some stuff, but I don't know about this in particular
18:47:46 <fizzie> b_jonas: It's presumably not an intended feature, at least. But it's a very well-known one.
18:47:52 <pikhq> I don't think it was supposed to be legal, but it was heavily used back in the day so it has to still work.
18:48:19 <fizzie> Also if you're running in dosbox, it doesn't enforce segment limits. :p
18:48:40 <b_jonas> people misusing the 386 that way...
18:49:14 <fizzie> http://wiki.osdev.org/Unreal_Mode has a bit more details than the wikipedia article.
18:49:27 <pikhq> This is x86. The A20 hack is still around. :)
18:50:30 <zzo38> The Forth environment can have other command too such as CMOS-WRITE to update CMOS settings and DISK-BOOT you can tell which disk to boot, and so on.
19:01:16 -!- trout has quit (Ping timeout: 276 seconds).
19:01:26 <b_jonas> zzo38: well sure, it has to have special functions useful for a boot loader of course
19:04:24 <zzo38> You would only to need to use such command if the normal boot-sequence is interrupted though; otherwise it will just boot normally and don't load the Forth environment.
19:07:43 <b_jonas> sure, but it needs them so you can actually use it as a recovery console when something goes wrong, as opposed to just as a calculator
19:07:57 <zzo38> Yes, that is what I meant!
19:08:16 <zzo38> You can use it as a recovery console.
19:10:18 <b_jonas> mind you, I think these days grub2 is starting to look like a whole operating system, you can use it as a recovery console too
19:10:30 <b_jonas> they're putting everything in it
19:11:12 <b_jonas> will they, like, run Doom straight from the boot loader?
19:11:39 <b_jonas> I started to hate this stuff when people started to put graphical splash screens in boot loaders
19:11:45 <b_jonas> seriously, graphical splash screens
19:12:01 <b_jonas> also in the early linux kernel
19:12:42 <zzo38> I don't like that either
19:12:46 <ais523> b_jonas: the penguin in the linux kernel is actually amazingly useful for debugging
19:13:05 <ais523> apparently when porting Linux to a new hardware, a static bitmap like that is much /easier/ to display than text
19:13:12 <ais523> so the first thing that people normally get working is the penguin
19:13:18 <b_jonas> ais523: yeah, on other architectures maybe
19:13:33 <b_jonas> mind you, the kernel also has a static bitmap font in it
19:14:07 <zzo38> It shouldn't need a static bitmap font on PC; on non-PC computers though it can be compiled in the PC font
19:14:14 <ais523> yes but you still have to do font rendering
19:14:22 <b_jonas> anyway, on pc, the bios gets the vga console working very early, and it has a very easy interface even if you access the vga memory directly rather than through the bios
19:14:30 <zzo38> You should use the PC character set.
19:14:31 <ais523> also, people don't normally struggle to port Linux to x86 nowadays
19:14:35 <ais523> it has a very good x86 port already
19:14:48 <b_jonas> if you switch to graphics mode, you have to handle the font rendering
19:14:55 <b_jonas> the bios can do that, but only in real mode
19:15:59 <b_jonas> but no, people want fancy graphical splash screens in the vga card, then in the bios, then in the boot loader, then in the early kernel, then in the initrd, then in early x11, then in the login procedure
19:16:10 <ais523> it's because people get scared by tesxt
19:16:13 <b_jonas> then when you start the program or something
19:16:26 <b_jonas> it's like eight different splash screens, each implemented differently
19:16:30 <ais523> although Plymouth at least (between initrd and x11 IIRC) just disappears if you press esc
19:16:44 <zzo38> I would like it never switch to graphics mode until X starts...
19:17:17 <b_jonas> (oh, and there may be a monitor splash screen too in modern tft monitors)
19:17:34 -!- hilquias has joined.
19:17:38 -!- hilquias has quit (Changing host).
19:17:38 -!- hilquias has joined.
19:17:59 <zzo38> My opinion is monitor spash screen should not be displayed unless there is no other picture available to display
19:18:25 <b_jonas> I for one have only the vga card and bios splash screens on my computer, and I think those are in text mode.
19:18:27 <zzo38> But if there is no other picture then it should display its own splash screen for perhaps one second, you can see how it can display a picture at least.
19:18:33 <b_jonas> I don't have any of the others enabled.
19:19:47 <b_jonas> zzo38: yes, and the on-screen menus should tell what the allowed hsync and vrefresh ranges and resolution are for vga signal, rather than just a no-information error message about invalid signal
19:20:10 <b_jonas> that would be trivial, but no
19:20:12 <zzo38> Yes, that too; I agree that too very much.
19:20:21 <b_jonas> it's not anywhere in the monitor's documentation or anything
19:20:23 <zzo38> The on screen menus ought to still work if there is no signal!
19:20:31 <b_jonas> it used to be stamped on the back of the screen, but it's not on the label these days
19:20:31 <zzo38> And to tell you such information if it is available.
19:20:42 <b_jonas> even if the menus work, they never tell that information
19:21:02 <b_jonas> I can get the menus to work because, like, standard vga modes should work, right?
19:21:11 <b_jonas> but the menus don't tell any of this useful stuff
19:22:00 <zzo38> After displaying the splash screen though it should enter power-saving mode, until either there is a signal or you push some of the other buttons on the monitor. If there is a signal it can display it should bypass the splash screen if possible (if the input is HDMI then maybe it takes some time anyways; I don't quite know)
19:22:25 <b_jonas> you have to search the internet, and even there such info is usually hard to find, or just test what works, because at least these days you can't destroy the hardware so easily by giving invalid video signal
19:22:29 <zzo38> I think it should still be written physically on the back of the monitor too though, in case it is not plugged in.
19:22:53 <b_jonas> but I'm saying the menus should display it because that wouldn't even cost them anything
19:22:57 <oren> Puppy linux is relatively devoid of splash screens
19:23:14 <b_jonas> they know the ranges (the nominal ranges at least, obviously the actual range might be very slightly larger) and resolution
19:23:24 <b_jonas> and they have lots of long text compared to this in menus
19:23:30 <b_jonas> localized to ten languages too
19:24:33 <zzo38> Yes I agree the menu should still display if the monitor is turned on regardless if a input signal is available or not.
19:24:59 <b_jonas> eight splash screens… it's just crazy
19:26:31 <oren> I just rebooted puppy to count. First bios, then a blinking cursor of 1 second, then a picture of a puppy with text detailing the boot process, then more text, then desktop
19:26:59 <oren> not having a login screen helps
19:27:07 <fizzie> b_jonas: It's in the EDID data hth
19:27:16 <b_jonas> oren: by the way, sometimes you can't see the vga card splash screen because the tft monitor takes too long to start up
19:27:52 <oren> yeah that is probably what is happening
19:27:52 <b_jonas> it's less funny when you don't see the bios splash screen which tells you which key to press to access the setup because the monitor starts up that slow
19:28:11 <b_jonas> not all vga cards have a separate splash screen of course
19:28:23 <oren> I usually just spam the F keys
19:28:42 <b_jonas> oren: yes, the f keys, escape, del, insert, tab, and combinations with shift and control and alt
19:29:09 <oren> tab? really? hm..
19:29:09 <b_jonas> but you probably only get like 16 keys to try before the buffer fills up
19:29:32 <b_jonas> oren: I think tab isn't used for setup, but for suppressing the bios splash screen to get actual messages on text from the bios
19:29:43 <b_jonas> but even that could be useful if that screen says "press f1 for bios" or something
19:30:16 <b_jonas> too late to suppress the splash screen by the time it's read
19:35:08 <fizzie> http://sprunge.us/aaiF Okay, you still need to actually get it *out* of the EDID. (FWIW, it's decoded properly in Xorg log. I blame nvidia.)
19:38:01 <b_jonas> fizzie: wow, that line has very few words
19:38:10 -!- trn has quit (Remote host closed the connection).
19:38:11 <b_jonas> like, it's all abbreviations and technical stuff
19:39:33 <b_jonas> yes, it's supposedly there
19:40:15 <b_jonas> also, these days I insist on using DVI instead of VGA connection, and computers have enough resources that I run X11 all the time, no text console stuff, so all of that is getting less relevant
19:40:59 -!- trn has joined.
19:41:16 <oren> I run X and then run a terminal und4er X and aften nothing else
19:42:10 <b_jonas> back in the old days that could cause problems because it tied up some significant memory and cpu
19:42:32 <b_jonas> these days it's somewhat less (though it still ties up 1/10 of the memory bandwidth on this old machine)
19:43:06 <b_jonas> computers get faster and faster very quickly
19:43:49 <oren> this machine is 9 years old...
19:44:11 <b_jonas> I'm not sure how old exactly this is
19:44:12 <oren> but that means like, 2006. so not even that old
19:44:25 -!- MoALTz has joined.
19:44:58 <fizzie> I gave up not running X when I switched away from Matrox graphics cards.
19:45:06 <oren> hmm... switching to the console and then running the terminal on that works fairly well
19:45:41 <oren> the terminal...I means GNU screen not the terminal
19:46:03 <b_jonas> it works more than fine on this machine too. it's worked already on the two previous computers before this.
19:46:34 <b_jonas> I'm not sure about the PII one that was two before this
19:46:41 <b_jonas> but definitely fine on the P4 system before this
19:47:02 <b_jonas> I really like this system though
19:47:13 <b_jonas> I'll have to reinstall the software
19:47:37 <zzo38> I think Digi-RGB is better than DVI and HDMI and VGA and so on. A Digi-RGB monitor ought to take two frames to start up (because the screen resolution won't be known until one entire frame has passed; it won't know where to start drawing either until one entire frame has passed).
19:48:01 <fizzie> Had at least a Mystique 220 and a G450, and they both were very good. mplayer had a 'mga' output driver that didn't need X.
19:48:31 <zzo38> Some graphical programs can run even without X
19:49:15 <b_jonas> zzo38: well, it's not really just the frames that limit the startup
19:49:35 <b_jonas> I think it's mostly the electronics in the monitor, like powering up the backlight
19:50:07 <zzo38> Maybe it can power up the backlight independently though?
19:50:44 <b_jonas> it can't just keep the backlight powered up at all time because that would waste energy. it has to start that when it notices there's video signal as you turn on the machine.
19:51:08 <b_jonas> it might be other stuff it has to do too, I don't know electronics
19:51:18 <b_jonas> what's this Digi-RGB stuff?
19:51:22 <oren> like have a backlight switch that is on the side, independent of everything
19:51:23 <zzo38> I mean independently from the displaying signal; I don't mean it would keep the backlight on all the time!
19:51:42 <zzo38> oren: Although maybe that might work too
19:53:03 <zzo38> b_jonas: The electrical specification is not written yet, but the rest is like this: All signals go from the computer to the display; only one direction. Signals are 4x red, 4x green, 4x blue, clock, sync, power, ground. You have two clock cycles per pixel, and sync is between frames. Aspect ratio must be eiter 4:3 or 16:9.
19:53:30 <zzo38> That's how it works.
19:53:33 <b_jonas> I don't really know which of the modern digital formats (DVI and HDMI and there's like two more I think) is better, I'm just claiming I want to use a digital video signal format rather than VGA (or other analog formats)
19:54:11 <zzo38> There is no limit to screen resolution or clock rate, although a minimum "base case" should be specified
19:54:17 <b_jonas> zzo38: oh, that seems a bit limited
19:54:35 <b_jonas> for one, I have a 16:10 monitor
19:55:04 <b_jonas> also, even if most use cases don't need it, sometimes it does make sense to have more than 8 bit depth per color component
19:55:26 <oren> 16:10 is good because then you can display a movie, with the player interface in the smallextra space
19:55:40 <fizzie> HDMI equals DVI + stuff (- other stuff), AIUI.
19:55:49 <b_jonas> it's just that most people use 16:9 because that's what the cheap monitors do
19:55:50 <zzo38> I don't want to complicate it
19:55:58 <b_jonas> there's also a lot of 5:4 monitors
19:56:04 <fizzie> Okay, that description covers everything. But still.
19:56:27 <b_jonas> fizzie: in particular, hdmi can transmit sound in the same wire, dvi can't
19:57:07 <fizzie> For the basic part of transferring an image, they're supposedly identical. Which is why you can get passive adapters for the conversion.
19:57:10 <b_jonas> also dvi has pins for analog signal so you can have a passive dvi-vga converter or something like that
19:57:17 <zzo38> oren: Yes I suppose that is true if you want to display the timecode and track number while the movie is playing perhaps, but usually I just don't need this.
19:57:46 <b_jonas> fizzie: no, I think you get passive adapters because the hdmi monitors and cards specifically have support for the converters
19:57:54 <zzo38> Digi-RGB-Plus can transmit analog stereo sound on the same cable as digital video; Digi-RGB doesn't though. It is designed though that a passive cable can convert between them with no compatibility issues.
19:57:56 <b_jonas> it's like the ps2 to usb converters for mous
19:58:10 <b_jonas> which works because the mouse has built-in support for usb and ps2 or something
19:58:19 <fizzie> "DVI and HDMI have the same electrical specifications for their TMDS and VESA/DDC links."
19:58:23 <b_jonas> support both, just one plug for physical reasons
19:59:09 <b_jonas> then there's I think some, uh, mini-hdmi with a smaller plug, just not called that, and some other similar digital format
19:59:19 <b_jonas> and dvi is actually more than one formats too
19:59:27 <zzo38> I have a computer monitor connected to a VCR/DVD combo, you can connect the HDMI out of the VCR to the DVI in of the monitor but then there is no sound. However, the speakers can be connected to the other audio out on the VCR and then the sound will work too.
19:59:31 <pikhq> Basically, HDMI is specifically designed to have its on-wire protocol an extension of DVI.
19:59:32 <fizzie> Yes, it's all very complicated.
20:00:17 <pikhq> It's not a matter of the monitors and cards having explicit support but rather that HDMI itself is inherently DVI-compatible.
20:00:39 <b_jonas> pikhq: but how is that possible? doesn't HDMI have too few pins in the connector for that?
20:01:06 <fizzie> It does drop the analog parts.
20:01:24 <b_jonas> maybe dvi doesn't actually use that many pins
20:01:36 <b_jonas> maybe it's like the two rs232 serial port connectors
20:01:42 <b_jonas> that seems riddiculous too
20:01:57 <pikhq> DVI has 24 pins, but some of them are not at all necessary.
20:01:58 <fizzie> For that matter, I think one of my outputs in the current graphics cards is DVI-D-only.
20:01:59 <b_jonas> they're exactly the same, the long one just has some unused or duplicated pins
20:02:46 <fizzie> 13w3 for all, I say -- it looks the funniest.
20:03:05 <pikhq> DVI has 5 TMDS lines while HDMI has 2, but DVI doesn't require all 5 of those to work.
20:03:20 <zzo38> Digi-RGB-Plus is Digi-RGB + analog stereo sound + control signal. The specification requires that it will function properly even if one or the other device does not support the control signal. (Also, the control signal is the reverse direction from the other signals.)
20:03:45 <b_jonas> hehe, Digi-thigy actually has analog sound
20:04:04 <pikhq> (DVI single link I believe only requires 2 TMDS lines)
20:04:09 <b_jonas> they really master naming stuff
20:04:38 <b_jonas> strange, I thought DVI actually required a lot of pins
20:04:42 <b_jonas> maybe not all of them, but a lot
20:05:05 <pikhq> Yes, but those same pins are on HDMI as well.
20:05:26 <pikhq> HDMI *just* has the set of pins required for DVI it looks like.
20:05:32 <fizzie> MHL is that one thing that can share a port with micro-USB and be "HDMI-compatible" in a very weak sense -- in the way that the other end can share a HDMI port, but both ends need to specially support MHL.
20:05:42 <b_jonas> how many pins is that actually?
20:05:55 <pikhq> And does higher quality video by clocking the lines faster rather than adding more lines.
20:06:33 <fizzie> Then there's the DisplayPort side, and I think they had a "MHL-equivalent".
20:06:41 <b_jonas> oh by the way, I believe USB-3 also works in such a way that it's "compatible" with USB-2 because the USB-3 host actually has a full USB-2 host built in it
20:07:15 <b_jonas> and the connector of USB-3 has all the pins of USB-2 so you can plug in an USB-2 cable physically
20:07:22 <b_jonas> but it just handles USB-2 separately
20:07:30 <pikhq> Yes, USB-3 works entirely over separate wires with USB-2 (in USB-3 mode) used for some initial protocol setup.
20:07:58 <zzo38> My opinion is DisplayPort and HDMI is complicated and has some kind of more problem; I invented Digi-RGB and Digi-RGB-Plus to be better systems.
20:09:01 <pikhq> Huh, weird -- there's nothing useful even missing from DB-25 on DE-9. I figured it was just obscure stuff.
20:09:02 <b_jonas> zzo38: ok, but still, supporting only 16:9 and 4:3 is limited
20:09:03 <zzo38> You don't need to pay or register to use it, but, if you impleemnt it wrong and claim it is Digi-RGB then it is a trademark violation.
20:09:17 <b_jonas> it works for most projectors but not for some monitors
20:09:38 <fizzie> USB 3.1 type C cables can carry a MHL 3.0 or a DisplayPort 1.3 signal, to further muddle things.
20:10:01 <b_jonas> fizzie: what's the relation of USB* to ESATA?
20:10:13 <zzo38> b_jonas: One problem is if too many aspect ratios are possible then it might become difficult to figure out what the aspect ratio is from the signal.
20:10:44 <pikhq> b_jonas: Utterly unrelated, but some ports ("eSATAp") are built to accept both eSATA and USB connectors.
20:10:57 <b_jonas> zzo38: can't there be a dummy row between frames, with a sync signal before and after it, or something?
20:11:06 <b_jonas> I mean, if you don't want to pay for a hsync signal
20:11:27 <pikhq> It's just something that you can shove both types of connectors into and it'll work.
20:11:43 <pikhq> Or, slightly more usefully, you can have an eSATA device that draws power from that USB port too.
20:11:45 <b_jonas> probably a notebook thing then
20:11:55 <fizzie> b_jonas: You can also carry PCI Express over SATA ("SATA Express" or something like that).
20:11:55 <b_jonas> well, mind you, the whole idea of esata is probably a notebook thing
20:12:07 <b_jonas> because we desktop people have space in the desktop to install four hard disks
20:12:13 <b_jonas> notebook people can install only two
20:12:43 <b_jonas> that sucks, how can you use a notebook as your main computer when you can't have multiple hard disks for redundancy? that's something I never understood
20:12:51 <b_jonas> sure, they use usb external disks and stuff, but still
20:13:01 <pikhq> If you're feeling really fancy you can shove a port multiplier on there and run, like, all the hard drives.
20:13:08 <b_jonas> fizzie: what, that sounds crazy
20:13:13 <fizzie> They use the cloud for redundancy, too.
20:13:31 <oren> I use SD cards for backups
20:13:40 <b_jonas> pikhq: the problem is that you physically can't fit the hard disks inside the notebook chasis, not that there isn't enough port
20:14:12 <pikhq> It'd have to be a *big fucking laptop* to do more than two.
20:14:14 <fizzie> b_jonas: I think it's mainly so that SSD manufacturers can make single controllers for both PCIE and "SATA" SSDs without being limited to SATA.
20:14:36 <b_jonas> also, even those notebook hard disks have to be small size (one of three small sizes actually, with different but passive convertible connnectors), and I think small size 2 terabyte hard disks are significantly more expensive than large size
20:14:42 <pikhq> *My* solution is to use my laptop as an SSH terminal and web browser.
20:15:17 <b_jonas> and it's getting even crazier when people want to put an ssd in their notebook, because that takes up a hard disk slot
20:15:29 <b_jonas> pikhq: sure, that does work, what I don't understand is using them as a main machine
20:15:30 <fizzie> The "SATA Express" port looks like two side-by-side SATA ports, and I think they also generally work as regular SATA if you want.
20:15:50 <pikhq> b_jonas: A lot of people also don't bother with, y'know, redundancy on drives in general.
20:16:16 <b_jonas> they also don't want to upgrade parts of their machine too
20:16:24 <b_jonas> they just buy a whole new machine
20:16:44 <fizzie> That reminded me of http://www.penny-arcade.com/comic/2005/08/10
20:18:35 <zzo38> Can you explain how you would do such dummy row and those stuff?
20:19:03 <b_jonas> zzo38: well, if you had an extra hsync signal, you could send a hsync signal at the end of each row, right?
20:19:25 <b_jonas> you could probably do that in the digital signal too if you made one color value special, but that's not practical with 8 bit depth
20:19:43 <b_jonas> (I think analog encodes the hsync signal with some out of range value or something, I'm not sure)
20:20:15 <fizzie> ("They " also have a new thing called "NVMe" to replace AHCI as the logical interface. Basically, http://en.m.wikipedia.org/wiki/SATA_Express#/media/File:SATA_Express_interface.svg
20:20:55 <b_jonas> if you don't want to do that, but you still have a vrefresh signal, you could do like this: vrefresh signal, first row, vrefresh signal, second row, third row, fourth row, ... 1200th row, start from beginning
20:21:15 <b_jonas> you decode that by checking three consecutive vsync signals,
20:21:52 <b_jonas> the distance of the first and third gives the full resolution (plus whatever overhead there is), the distance of the two vsync signals that are closer give the length of one row (plus possible overhead)
20:22:06 <b_jonas> (provided there's at least three rows)
20:22:21 <b_jonas> but I don't do electronics, so I don't know how practical that would actually be
20:22:23 <zzo38> Yes you could do that with a hsync signal and it is correct it won't work with 8-bit depths only. My idea was that there is no clocks during hsync, and that hsync may be of zero time or more. A CRT display is allowed to require a specific hsync time but LCD/LED displays shouldn't.
20:22:39 <zzo38> b_jonas: That seems an interesting idea; I don't quite know how practical it is either though.
20:22:58 <zzo38> I can try discussing it with others I work with though and see
20:23:22 <b_jonas> or, you know, you could send digital metadata (telling the resolution and pixel depth and format stuff) after the vsync signal, but you probably don't want that
20:24:24 <b_jonas> zzo38: is this supposed to work with only lcd monitors and projectors? or more than that?
20:25:15 <b_jonas> zzo38: you could also waste a few pixels or rows at vsync or hsync and send metadata at that time
20:25:18 <zzo38> Probably only LCD monitors and projectors. Someone can make a CRT to work with it too if they want to but it isn't really designed to work with CRT.
20:26:01 <b_jonas> zzo38: like, use a hsync signal instead of a vsync signal, waste at least one pixel data during hsync, and distinguish vsync from hsync by special values on the pixel pins
20:26:14 <b_jonas> (or by a longer sync, which is what vga does)
20:26:24 <b_jonas> but I don't know if wasting pixels at every _hsync_ is a good idea
20:26:42 <b_jonas> zzo38: how about those fancy low-res led matrices?
20:26:50 <b_jonas> not led backlit, but made of leds
20:26:58 <b_jonas> low resolution, one or two or three color channels
20:27:25 <zzo38> Those are some ideas too but I don't want to add metadata or require hsync in specific ways. I don't know how those fancy low-res led matrices work. If you can provide details then I can answer you.
20:27:47 <b_jonas> (oh by the way, besides more than 8 pixel depth, what I'd like is more than 3 color channels, for both cameras and monitors, and custom metadata that describes the spectrum of each color component)
20:28:05 <b_jonas> (but that's like wishlist category)
20:28:16 <b_jonas> (as in, I also "want a pony")
20:28:33 <zzo38> Like I said, I do not want to complicate it.
20:29:02 <b_jonas> zzo38: is the vsync zero pixel long, or longer? fixed or variable length?
20:30:22 <zzo38> Vsync is zero pixel long but the amount of time it takes can be more than zero.
20:30:49 <b_jonas> we just time stuff in pixels for video signals
20:30:59 <b_jonas> because the video card runs on a clock
20:31:12 <zzo38> There are no clock signals during vsync or hsync though, but the vsync signal will indicate vsync.
20:31:16 <b_jonas> (it gets crazier with the optional 9/8 multipler of vga text mode. I don't know how that works.)
20:31:35 <b_jonas> zzo38: how does the monitor know how long the vsync is?
20:32:28 <zzo38> If it is a LCD monitor then does it need to know, if it can just start right after vsync?
20:32:43 <tswett> I'm going to create the Wikipedia page "Israel with sea download".
20:33:11 <zzo38> You should probably add a delay for vsync and hsync in case the display wants it anyways though, but it is not a requirement.
20:33:44 <b_jonas> zzo38: yeah, but then how does it know when the previous frame ends?
20:34:39 <b_jonas> I mean, it has to know when the previous frame ends and when the current frame starts, right?
20:35:08 <zzo38> The sync signal is active after the current frame ends
20:35:28 <b_jonas> also, even if limited to 8 bit depth and those fixed ratios, is this intended to support monitors with high resolution and high frame rate, possibly higher than what you can buy today?
20:35:54 <b_jonas> zzo38: and when does the sync signal gets passive?
20:36:31 <zzo38> And then sync is inactive then you will start the picture.
20:36:36 -!- lleu has joined.
20:36:38 <b_jonas> zzo38: but then you can't have zero time sync
20:36:45 <zzo38> And yes it is intended to support monitors with high resolution and high frame rate, as well as low ones.
20:37:01 <b_jonas> shouldn't it be active for at least the last one pixel (a fixed number of pixels) in the previous frame?
20:37:15 <zzo38> b_jonas: Well, yes, it isn't really "zero time"; it should be at least enough time for half a pixel, probably more
20:38:28 <b_jonas> but if you want to support zero time, shouldn't you make it active for the last few pixels in the previous frame?
20:39:02 <b_jonas> (in fact, possibly even shifted back in time so it's passive a few pixels before the next frame)
20:39:18 <b_jonas> how can it be half a pixel? there's a clock signal, isn't there?
20:39:21 <zzo38> Well, I made a mistake about zero time, but thank you for your suggestion anyways it might be considered. Perhaps at least the clock fall of the last half-pixel should be, at least
20:39:35 <zzo38> b_jonas: It is two clocks per pixel, so one clock is half of a pixel
20:39:35 <b_jonas> not, like, implicit timing like some protocols do
20:39:51 <zzo38> (Otherwise there is too many pins)
20:39:54 <b_jonas> how is it represented in the clock wire by the way?
20:40:44 <zzo38> Clock signal is high during each half-pixel and then is low, and then you do next one
20:41:22 <zzo38> Like many other things are
20:44:41 <b_jonas> how are you even detecting between 9:16 and 4:3 ratio? are some sizes of those disallowed?
20:47:14 <b_jonas> I mean, any 9:16 resolution has a pixel count that's valid in some 4:3 ratio resolution, just usually an unusual one
20:47:23 <b_jonas> because there's not that many resolutions actually in use
20:47:38 <b_jonas> I don't know if there's any ambiguity among already used resolutions
20:48:15 <b_jonas> or would a monitor support only one of those ratios?
20:51:22 <b_jonas> wow, I don't remember having seen ais523 on irc for such a long interval continuously (with short breaks which I assume are connection problems)
20:52:54 <zzo38> Can you show example of the pixel count?
20:53:09 <zzo38> OK, perhaps I missed it
20:53:10 <b_jonas> zzo38: I mean as in horizontal resolution times vertical resolution
20:53:20 <zzo38> Because it isn't supposed to be
20:53:35 <zzo38> You are probably right though
20:53:40 <b_jonas> so for my 1920x1200 pixel monitor, that would be 1920*1200
21:01:03 -!- oerjan has quit (Quit: Nite).
21:05:26 <zzo38> How you figure out is try to calculate the square root of the number of pixels and figure out what factor is left over. If it is 1 then it is 16:9. If it is 3 then it is 4:3.
21:06:40 -!- GeekDude has quit (Ping timeout: 256 seconds).
21:07:16 -!- `^_^v has quit (Quit: This computer has gone to sleep).
21:10:00 <zzo38> You can also figure out if it is square during counting, because the square number is added up 1+3+5+7+9+11+13+...
21:11:09 <b_jonas> zzo38: no. there's always a lcm of any two rectangular screen ratios, so there's always a screen resolution where the pixel count is ambiguous
21:11:23 <b_jonas> zzo38: as in lcm(16*9, 4*3)
21:11:50 <b_jonas> that lcm could be very large so only large resolutions can be ambiguous, but not for these ratios, where it's 144
21:12:40 <b_jonas> so unless you use non-rectangular screens (eg. adding an extra pixel to the end of 4*3 screens so you can distinguish from the parity of the pixel count) you're screweed
21:14:31 <ais523> b_jonas: this laptop supports both 1360x768 and 1366x768 as resolutions
21:15:24 <b_jonas> maybe I'm wrong because I'm tired
21:15:40 <b_jonas> maybe there's no collision actually
21:18:49 <b_jonas> yep, no collision, you're right
21:20:41 <ais523> b_jonas: also lcm(16*9, 4*3) is 16*9, for what I hope are obvious reasons
21:21:04 <b_jonas> isn't that the value I said?
21:21:08 <b_jonas> but it turns out it's not the lcm that matters
21:38:35 <zzo38> Do you have any more feature-suggestion/complaints about AmigaMML today?
21:39:14 <ais523> I did answer a question you asked in 2011, though: http://esolangs.org/wiki/Talk:Underload#Why_the_reserved_characters.3F
21:39:17 <ais523> not sure if you're still interested in the answer
21:40:04 <zzo38> I did look at that answer; thank you for that
21:46:11 <fizzie> Oh, since this is a bit weird, maybe I'll mention here: I've got a mouse (just a plain Logitech M500-or-something-like-that), and a USB 2.0 hub (just your basic cheap unpowered 4-port thing), and on this desktop, the mouse stops working generally after 1-60 minutes of use; but it works fine if (a) the mouse is plugged into the machine, not the hub, or (b) the hub is plugged into the laptop, ...
21:46:17 <fizzie> ... not the desktop. Oh, and the original setup used to work fine earlier, but now reliably fails.
21:46:28 -!- atrapado has quit (Quit: Leaving).
21:46:52 <fizzie> Oh, and unplugging + replugging makes it start working for a while again.
21:46:58 <fizzie> Debugging tips? As I recall, dmesg doesn't contain any particularly insightful messages when it stops working; it just... stops.
21:47:38 <fizzie> (Maybe I should just get another cheap hub.)
21:47:47 <ais523> fizzie: check the X log, just in case?
21:48:41 <fizzie> Hmm. I guess I should. I'll plug it into the hub and wait.
21:51:44 <fizzie> Well, it hung up. Nothing in dmesg; no 'disconnected' message or anything.
21:52:58 <fizzie> Nothing in Xorg.0.log either.
21:53:07 <fizzie> Although I think the cursor just disappeared.
21:53:22 <fizzie> Or maybe I just forgot where it was.
21:53:36 <fizzie> Oh, it's moved to the other monitor. Hmm.
21:53:45 <fizzie> Don't know what's up with that either.
21:54:38 <fizzie> It's one of those infrared dealies, so I can't even see whether it's emitting light.
21:55:37 <ais523> now I'm wondering what the easiest way to see in infrared is
21:55:46 <ais523> you can't rely on flourescence like you can with UV
21:55:54 <fizzie> The generally accepted answer to that is I think "cheap webcam".
21:56:00 <fizzie> For near infrared, that is.
21:56:05 <ais523> fizzie: oh, that's too cheap to filter the IR out?
21:56:08 <ais523> that's actually pretty hilarious
21:56:22 <fizzie> Yes, with a lot of them you can see at least a TV remote IR led.
21:56:55 <fizzie> And some of them have a "night vision" mode which turns on an IR led in the camera.
21:58:25 <ais523> fizzie: hmm, I remember many years ago
21:58:46 <ais523> reading a New Scientist article which was talking about this new innovation that made photodiodes much more efficient, but as a side effect made them see into the infrared
21:58:52 <ais523> presumably it's been widely adopted since
21:59:02 <ais523> (they were unclear on whether this was an advantage or disadvantage)
22:02:36 -!- Wright_ has joined.
22:02:36 -!- Wright has quit (Read error: Connection reset by peer).
22:05:37 <fizzie> Mmmaybe. I am at least under the impression that "most" CCD sensors are IR-sensitive enough to require an explicit IR filter. (People post instructions on how to dismantle some DSLRs to remove the sensor, so that you can do IR photography for artistic purposes.)
22:05:50 <fizzie> s/remove the sensor/remove the filter/
22:06:03 <fizzie> Although I'm sure removing the sensor would be a valid artistic choice too.
22:06:16 -!- idris-bot has quit (Quit: Terminated).
22:08:26 <ais523> fizzie: aren't there easier ways to get an entirely black image? :-)
22:08:31 <b_jonas> ais523: telephone cameras are also easy to check ir with
22:08:47 <b_jonas> I'm not sure if that's because the ir filter is expensive or thick
22:08:53 <fizzie> I tried my phone camera, and it didn't show anything. Of course I don't know anything about what sort of light comes out of the mouse.
22:09:06 -!- idris-bot has joined.
22:09:07 <b_jonas> my phone camera definitely shows some IR lights
22:09:15 <b_jonas> it might depend on the frequency of course
22:09:49 <fizzie> Yes. Don't see anything here; not that I know how bright it is, either, or if it's easy to trigger.
22:10:05 <fizzie> The MS IntelliMouse I had was this ridiculously bright (visible) red.
22:10:16 <fizzie> You pretty much didn't need any lights in the room.
22:11:11 <ais523> fizzie: there's a computer lab at the university that's full of those mice
22:11:33 <ais523> even when they turn the lights off in the room at night, there's still this angry red glow that's visible at a huge distance
22:11:51 <fizzie> The mice are plotting the doom of the human race there.
22:12:11 <pikhq> I don't think that's what they're plotting.
22:13:05 <zzo38> I downloaded the new version of SoX but now it says there is no default audio device configured.
22:13:09 <b_jonas> like http://www.xkcd.com/251/ ?
22:14:57 <zzo38> Specifying "-t waveaudio default" instead of "-d" works, but how to make it so that "-d" will also work?
22:15:58 <fizzie> I think it reads some environment variables for the default?
22:16:14 <zzo38> What environment variables are these?
22:16:22 <fizzie> (I might be wrong here.)
22:17:29 <fizzie> AUDIODRIVER and AUDIODEV are mentioned in the man page. Although it's curious that a (presumably) Windows binary wouldn't have the reasonable default.
22:17:29 <zzo38> O, I think I found it
22:20:34 -!- variable has joined.
22:31:27 <zzo38> I have the "libsox-3.dll" but how to link it into a C program with GCC?
22:36:40 <fizzie> I've done a little bit of that with MinGW, but I've forgotten the process. I think it involved the use of the MinGW 'dlltool'. (I'm not sure about Cygwin and such.)
22:37:24 -!- evalj has quit (Remote host closed the connection).
22:39:05 <fizzie> I have a vague feeling it was complicated, since the normal way was to start from the source code of the .dll.
22:39:55 -!- Patashu has joined.
22:40:06 <zzo38> I want to work with UNIX too not only on Windows
22:41:14 <fizzie> I believe the build steps will need to be platform-specific. Unless something like libtool can abstract that away.
22:41:22 <zzo38> Will specifying the .dll as an additional source file work?
22:42:19 <fizzie> Hm, maybe this has changed.
22:42:33 <fizzie> When I was doing this, you definitely needed to generate the special "import library" to link against.
22:42:53 <fizzie> But now http://www.mingw.org/wiki/createimportlibraries claims that: "Usually (read: for all DLLs created with MinGW and also a few others) MinGW links fine against a DLL. No special import library is necessary (see sampleDLL)."
22:42:56 <zzo38> Creating DLLs with MinGW works easily and I have done it perfectly fine; the only issue is that if you want to call any functions in the DLL from Visual Basic then you need to write "_stdcall" in front of those function definitions.
22:43:21 <ais523> having the separate import library helps to avoid circular dependencies though
22:43:38 <zzo38> (Typing _stdcall isn't necessary in any other cases, it seems)
22:44:24 <pikhq> Essentially what it does when you pass a dll to it is generate the import library automatically.
22:45:28 <pikhq> (if the symbols are exported by name, of course)
22:45:35 <ais523> pikhq: I thought it generated more efficient code than the import library based version
22:45:48 <ais523> presumably it inlines the import library, or something like that?
22:46:05 <pikhq> Nah, that's what __declspec(dllimport) does.
22:46:40 <pikhq> Such inlining cannot possibly be done at link time.
22:49:51 -!- copumpkin has joined.
22:51:44 <zzo38> They do seem to be exported by name; if I look in the dependency viewer, all of the names are listed there.
22:52:33 <zzo38> Can the program be used on Linux too if libsox is also available on Linux?
22:53:30 <pikhq> Likely. libsox *is* on Linux.
22:53:32 -!- contrapumpkin has joined.
22:53:35 <tswett> 13:52:33 --- quit: zies- (*.net *.spit)
22:54:29 <tswett> Neural net still has things to learn.
22:54:33 <zzo38> But, what changes are then needed in the source-file of the program (if any)?
22:54:53 -!- copumpkin has quit (Remote host closed the connection).
22:56:10 <zzo38> Also how to find the header file for libsox?
23:00:28 -!- copumpkin has joined.
23:01:58 -!- contrapumpkin has quit (Remote host closed the connection).
23:02:54 <Taneb> Mad Max: Fury Road was a film
23:07:22 -!- M_I_Wright has quit (Ping timeout: 256 seconds).
23:08:32 -!- contrapumpkin has joined.
23:10:00 -!- copumpkin has quit (Remote host closed the connection).
23:13:59 <zzo38> I found the sox.h file and now it compiles but it says "The procedure entry point GOMP_parallel could not be located in the dynamic link library libgomp-1.dll."
23:14:21 <zzo38> That file is in the path though
23:15:15 -!- contrapumpkin has quit (Remote host closed the connection).
23:15:52 <zzo38> Dependency Walker says that GOMP_parallel is exported from libgomp-1.dll too
23:17:41 <zzo38> The program works when the current directory contains libgomp-1.dll
23:21:16 -!- supay has quit (Ping timeout: 276 seconds).
23:23:45 -!- hilquias has quit (Remote host closed the connection).
23:27:44 -!- supay has joined.
23:40:37 <zzo38> I figured out the mistake is that I had a different version of libgomp-1.dll in a different directory and it found the wrong one.
23:40:57 <zzo38> Changing the order of the path partially fixed it.
23:41:24 <HackEgo> [wiki] [[Groovy]] N http://esolangs.org/w/index.php?oldid=43024 * 189.8.69.39 * (+15) Created page with "is a java thing"
23:42:54 -!- GeekDude has joined.
23:48:18 <zzo38> Now I fixed it more
23:48:28 <oren> Hold on I thought Groovy was that port of python to JRE?
23:48:47 <oren> No wait that's Jython
23:53:17 <oren> Well it's not an esoteric language in any case
23:54:45 <zzo38> And the article (which just says "is a java thing") is worthless, in any case.