IRC logs for #openttd on OFTC at 2012-07-04
⏴ go to previous day
00:19:26 *** kkimlabs_ has joined #openttd
00:19:51 *** Arendtse1 has joined #openttd
00:20:58 *** Zeknurn` has joined #openttd
00:21:28 *** TomyLobo2 has joined #openttd
00:22:34 *** TrueBrain_ has joined #openttd
00:23:53 *** Rienzilha has joined #openttd
00:24:18 *** dihedral has joined #openttd
00:24:27 *** Zeknurn` is now known as Zeknurn
00:24:27 *** TomyLobo2 is now known as TomyLobo
00:24:51 *** TrueBrain has joined #openttd
00:48:11 *** mahmoud has joined #openttd
00:48:42 *** mahmoud has joined #openttd
02:20:13 *** kkimlabs_ has joined #openttd
03:37:27 *** plastics has joined #openttd
04:39:17 *** tokai|noir has joined #openttd
04:39:17 *** ChanServ sets mode: +v tokai|noir
04:43:19 *** kkimlabs_ has joined #openttd
04:44:02 *** Supercheese has joined #openttd
04:56:17 *** Eddi|zuHause has joined #openttd
04:59:38 *** andythenorth has joined #openttd
05:25:10 *** andythenorth has joined #openttd
05:30:44 *** kkimlabs_ has joined #openttd
05:33:18 *** andythenorth has joined #openttd
05:43:07 *** andythenorth is now known as Guest2112
05:43:08 *** andythenorth has joined #openttd
05:48:54 <Supercheese> Heh, almost 11 PM here
06:07:08 *** sla_ro|master has joined #openttd
06:22:36 *** peter1138 has joined #openttd
06:22:36 *** ChanServ sets mode: +o peter1138
07:13:07 *** Zeknurn has joined #openttd
07:37:37 <peter1138> Indeed, happy Wednesday.
07:38:10 <Supercheese> Independence Day!
07:38:32 <Supercheese> Hmm, I wonder if more Americans default to "Fourth of July" or "Independence Day", my guess is the former
07:39:15 *** telanus has joined #openttd
07:44:27 *** andythenorth has joined #openttd
07:49:10 *** Eddi|zuHause has joined #openttd
08:16:06 <Hirundo> Ladies and gentlemen, we've got him! (Higgs particle)
08:17:38 <peter1138> I'd've thought Americans would default to "July four", as that's how the film trailers say dates...
08:19:38 <andythenorth> why are the internets fucked?
08:19:45 <andythenorth> are the internets fucked by the higgs?
08:20:00 <andythenorth> nothing works, not here in the office, not at home, not on my phone
08:20:13 <andythenorth> "nothing" ~= Twitter
08:20:16 <Supercheese> If you're in the Eastern US, there's apparently power outages all over
08:20:30 <peter1138> Twitter's working great.
08:20:31 <andythenorth> probably I'm visiting sites who are depending on Amazon Web Services
08:24:45 <planetmaker> what about those lang file errors with fish?
08:25:03 <andythenorth> not sure what causes those
08:25:11 <andythenorth> looks like it's mangling nml / nfo
08:25:36 <planetmaker> no, it's a wrong default lang file. Or rather insufficient header info
08:26:31 <planetmaker> or possibly a mix-up between english.lng and default.lng. Both exist
08:27:21 <planetmaker> and I've the feeling that default.lng is the offending one
08:27:24 <andythenorth> it's got "STR_LOG_RAFT_LARGE"
08:27:32 <andythenorth> which doesn't exist anywhere in the nml version, only in nfo
08:27:40 <planetmaker> hm... though... maybe it's some local issue...
08:27:40 <Hirundo> Both english.lng and default.lng is probably bad... IIRC
08:28:13 <planetmaker> I just checked that default.lng is in my ignored files... deleting it solves it :-)
08:28:31 <planetmaker> good that we talked about it ;-)
08:28:46 <andythenorth> so does it build for you?
08:29:52 <planetmaker> mind that the build script installs into andy's home dir. Which for curious reasons doesn't exist locally here ;-)
08:29:55 <andythenorth> is that with my dirty shell script?
08:30:22 <andythenorth> it's only got one vehicle, so time is not very representative :P
08:30:31 <planetmaker> ok, should I try to adapt the (new) makefile with it?
08:30:36 <andythenorth> please, if you have time
08:30:37 <planetmaker> and commit, if it works
08:30:51 <planetmaker> I may. Or I may not. Yesterday I worked like 17 hours... Maybe better today
08:31:04 <andythenorth> the main deps would be the .cfg file and the pngs
08:31:21 <andythenorth> although...I guess if the build script or templates change, they are also deps :P
08:31:30 <andythenorth> probably everything is a dep :P
08:31:35 <planetmaker> I'd not use the build script...
08:31:46 <andythenorth> not the shell script
08:31:48 <planetmaker> well... the py script I do
08:32:01 <andythenorth> the shell script is just a dirty hack for me when developing
08:32:14 <planetmaker> though it might be good to split, if possible into lang and nml generation. Feasible?
08:32:35 <andythenorth> but might be tmwftlb
08:59:09 *** sla_ro|master has joined #openttd
09:01:32 *** TGYoshi has joined #openttd
09:10:50 <planetmaker> what should I employ: I want to add a fixed number to the 3rd column of a csv file. sed, awk, ...?
09:25:54 <andythenorth> planetmaker: python :P
09:26:15 *** FLHerne has joined #openttd
09:28:45 <Terkhen> I used awk when I was formatting the spanish town names csv data; I dont remember much of it though
09:30:36 *** Arafangion has joined #openttd
09:32:45 <planetmaker> andythenorth: excel is no option. I have to do it dozens of times. With different values (which are read from the filename itself)
09:33:09 <planetmaker> and I need the data in the original format unspoiled except the changed numbers
09:33:26 <andythenorth> I'd use python, but that's because it's the only tool I know
09:34:05 <Arafangion> andythenorth: One trick pony!
09:34:16 * Arafangion would also use python... Despite not knowing the problem you guys are trying to solve.
09:34:45 <andythenorth> sounds like a .split() .join() kind of problem
09:35:28 <planetmaker> I want to add a fixed number to the 3rd column of a csv
09:35:45 <Arafangion> I'd use the CSV parser, and writer.
09:36:07 <planetmaker> and this task done dozens of times :-)
09:36:08 <Arafangion> The only limitation I've ever had with the python CSV implementation is that it doesn't natively support file encodings.
09:36:33 <planetmaker> file format needs remain unspoiled
09:36:45 <planetmaker> for further post-processing
09:37:00 <andythenorth> I'd do import codecs
09:37:03 <Arafangion> You can specify the type of CSV you wish to use, as well.
09:37:06 <andythenorth> then just read the raw text
09:37:11 <andythenorth> python csv processing seems to smell
09:37:28 <andythenorth> ','.split' worked well for me when I last did this
09:37:42 <Arafangion> Until the day you get an element that itself contains ','.
09:37:54 <andythenorth> in that case, you have a problem :)
09:39:05 <Arafangion> Which is why I suggest using the csv module. :)
09:40:32 <planetmaker> actually... "awk '{ $4 = $4 + 100; print $0 }' test.txt" might do the trick
09:41:07 <Arafangion> It naturally splits on commas?
09:41:20 <planetmaker> but that's what I need :-)
09:41:33 <Arafangion> So you didn't need CSV!
09:41:46 <planetmaker> well. Same thing with different separation char
09:41:54 <Eddi|zuHause> <Arafangion> Until the day you get an element that itself contains ',' <-- which is why i used tsv for CETS
09:42:02 <Arafangion> Eddi|zuHause: Heh.
09:42:19 <Arafangion> That said, CSV tends to imply *wacked* escaping rules.
09:42:29 <Arafangion> tsv and just about anything else is far more sane.
09:42:52 <Arafangion> Although understandably, people criticise tsv for using 'syntactically significant whitespace'.
09:43:41 <Eddi|zuHause> i'd love to see someone who enters the tab character into an excel or googledocs field :)
09:43:51 <blathijs> Arafangion: It seems to me that standard CSV has pretty ok escaping? How do you escape a tab character in tsv, then?
09:45:31 <Arafangion> blathijs: Typically with the literal \t
09:45:40 <Arafangion> (ie, two-characters)
09:46:17 <blathijs> and do people actually escape \ as \\, then?
09:46:41 <blathijs> Probably easier than the quoted escaping by CSV, yes
09:46:42 <Arafangion> Which does lead to monstrosities... But at least it's a *consistent* monstrosity.
09:47:18 *** TomyLobo has joined #openttd
09:50:26 <Arafangion> Not even Microsoft Excel interprets CSV consistently. :(
09:51:37 <Eddi|zuHause> did microsoft ever do something "consistently"?
09:52:44 <Arafangion> They still only support that standard. :)
10:00:44 <Hirundo> In my experience, excel doesn't understand that a csv file with hundreds of ',' but no ';' might be comma- instead of semicolon-separated
10:02:56 <Arafangion> Apparently Excel sometimes insists on ;'s rather than ,'s for CSV files.
10:05:49 <Eddi|zuHause> wait... doesn't importing a csv file ask what the separator is?
10:07:20 <Arafangion> Eddi|zuHause: Not sure, I think that depends on the reader, I know that the CSV reader has changed a few times in Excel.
10:11:46 <peter1138> Last time I tried to import a CSV into Excel, it needed to be tab separated.
10:24:57 *** Devroush has joined #openttd
11:58:54 *** TrueBrain_ is now known as TrueBrain
12:24:42 *** cmircea has joined #openttd
12:24:51 *** plastics has joined #openttd
13:06:29 *** sla_ro|master has joined #openttd
13:08:48 <NGC3982> Arafangion: i have never seen that happend.
13:09:45 <NGC3982> afaik, excel versions from "2007" and forward doesnt need tab separation to import CSV completely
13:32:23 <nido> question; is the libtimidity music backend supposed to work (as in, generate audible music) on non-psp platforms foor version 1.2.1?
13:34:48 <Eddi|zuHause> nido: i'd lean towards "no"... libtimidity is some very outdated fork of timidity
13:36:00 <Eddi|zuHause> i don't think anybody properly maintained it in the last 5 years
13:36:29 <Eddi|zuHause> use timidity, not libtimidity
13:36:43 <nido> in that case I shall take out libtimidity support in the package I make
13:37:08 <nido> would you be interested in the resulting centos6/fedora17 compatible specfile?
13:37:56 <Eddi|zuHause> specfiles are Ammler's business
13:48:44 *** kkimlabs_ has joined #openttd
13:56:32 *** HerzogDeXtEr1 has joined #openttd
14:05:14 *** TWerkhoven has joined #openttd
14:09:29 *** andythenorth has joined #openttd
14:14:14 <DorpsGek> Eddi|zuHause: 18047.8333333
14:15:11 <NGC3982> the ods of things can be quite perculiar
14:15:34 <NGC3982> that number is one digit away from a number i just recieved in a document.
14:15:43 <NGC3982> seconds before opening irc, reading it.
14:19:15 <__ln__> and you think that's a mere coincidence
14:20:35 <Ammler> the spec in the openttd vcs is quite outdated,
14:23:31 <NGC3982> __ln__: i cant say im a man of superstition.
14:29:04 <Ammler> nido: you find there also the specs for all the other openttd related projects
14:29:27 <Ammler> easiest would be, if you branch it and submit the diff
14:29:55 <Ammler> but hardly you can make better specs :-P
14:32:01 <Ammler> about sound, I added that to the opensfx package only
14:33:17 *** Biolunar has joined #openttd
14:33:28 <__ln__> NGC3982: please be a man of apostrohpes though.
14:37:44 <__ln__> they look like this: '
14:40:46 *** andythenorth has joined #openttd
14:44:05 <andythenorth> TrueBrain: internets says 'no' to openttd in JS
14:44:11 <andythenorth> apparently a very bad and silly idea :P
14:44:20 <andythenorth> at least according to forum
14:44:56 <TrueBrain> I am just really really happy 'those' people weren't around when Ludde wrote OpenTTD 1.0.0
14:45:05 <TrueBrain> as by their arguments, it was worse than TT
14:45:09 <TrueBrain> so nothing impressive
14:45:18 <TrueBrain> it btw highlights the attitude problem on tt-forums
14:45:39 <TrueBrain> it also made me hit "Sign Out" button on the forum btw
14:46:05 <andythenorth> the attitude problem is what?
14:46:14 * andythenorth thinks it's "wtf has the fun gone"
14:46:50 <TrueBrain> people fail to stimulate others to do better ;)
14:46:59 <andythenorth> openttd in JS is brilliant
14:47:07 <andythenorth> the stupidity of it makes it brilliant
14:47:16 <TrueBrain> the process behind it is even more briliant
14:47:20 <TrueBrain> I am still surprised it exists :P
14:47:57 <TrueBrain> always makes me smile when people do the "impossible"
14:48:05 <andythenorth> because I started web coding in 1999, I refuse to believe anything works in JS at all
14:48:09 <TrueBrain> like a MIT professor once told us: the impossible only takes a little bit longer
14:48:11 <andythenorth> I'm amazed by a simple popup
14:48:22 <TrueBrain> hehe; Javascript is one of the youngest languages we have
14:48:42 <andythenorth> everytime I use jQuery, and it works *even in IE6* I am amazed
14:48:52 <TrueBrain> do you still hav eIE6?
14:49:13 <andythenorth> our clients use it
14:49:24 <TrueBrain> they should be punished :P
14:51:41 <TrueBrain> but with a clearly working C -> Javascript, I am unsure what to expect from JS next
14:52:47 <TrueBrain> a full OS written in JS, I guess
14:53:49 <andythenorth> I once figured out you could write an OS in macromedia director
14:54:01 <TrueBrain> lol; that is scary :P
14:54:08 <andythenorth> as long as you had an OS to run it in :P
14:55:03 <andythenorth> JS proves to me that I am now a technological dinosaur
14:55:16 <andythenorth> I've gone from 'bright young thing' to 'old guy who moans and talks about the past'
14:56:03 * FLHerne made/wrote a paper-tape-fed 4-bit VM with an NXT :P
14:56:07 <andythenorth> I shall have to get a beard
14:56:17 <FLHerne> It was a bit hopeless :-(
14:56:33 <TrueBrain> andythenorth: I already have a beard :D MWHAHAHAHA :P
14:56:47 <TrueBrain> and yes, you know you are old if you have seen a Language been created, grown up, and do amasing shit :)
14:57:08 <TrueBrain> FLHerne: back to 1960s much? :P
14:58:26 * andythenorth actually knows nothing about PHP really, but knows about PHP *developers*
14:58:29 <TrueBrain> Python is a better example I guess :)
14:58:55 <FLHerne> TrueBrain: Not enough sensor inputs for a keyboard (binary keyboard, maybe, but that would be silly :P )
14:59:31 <FLHerne> "Hello World" needed 3m of tape though - LEGO sensors are 'opeless :-(
14:59:36 <TrueBrain> hmm, all this talk, I need a new hobby ...
15:01:49 <andythenorth> TrueBrain: write some code!
15:01:58 <andythenorth> [any child will do]
15:02:05 <andythenorth> buy a fast car, then break it, then fix it
15:02:39 <FLHerne> TrueBrain: Get a rat :P Just as attention-seeking as a child, but sleeps less :D
15:03:21 <TrueBrain> lol .. "get a child"
15:03:25 * TrueBrain walks in a busy store
15:03:28 <TrueBrain> pick up a random child
15:03:59 <TrueBrain> glx: love that URL:
15:04:01 <TrueBrain> For the same reason, "6" == " 6", "4.2" == "4.20", and "133" == "0133". But note that 133 != 0133, because 0133 is octal. But "0x10" == "16" and "1e3" == "1000"!
15:06:18 <andythenorth> wtf is 'mysql_real_escape_string'
15:06:52 <TrueBrain> it escapes a stirng to be MySQL safe
15:06:58 <TrueBrain> you used to have mysql_escape_string
15:07:01 <TrueBrain> but something was wrong with it I guess?
15:07:05 <TrueBrain> and now you have REAL escape :P
15:07:16 <TrueBrain> Namespace sin PHP 5.3 are hilarious
15:07:30 <TrueBrain> It's like: \Name\Space\Class::method
15:07:41 <TrueBrain> and no, the \ doesn't escape .. it is namespace seperator
15:09:39 <Kylie> "\name\space\class::method"
15:10:08 <TrueBrain> you can do: $f = "\name\space\class"; $i = $f();
15:10:16 <TrueBrain> but I hav eno clue if that works
15:10:21 <TrueBrain> or if you have to do: \\name\\space\\class
15:10:25 <TrueBrain> which would be amusing at best
15:10:43 <TrueBrain> it is DOS with it's \ all ove ragain; I hoped everyone learnt that leason
15:10:48 <TrueBrain> PHP clearly didn't. So here we just go again :)
15:11:29 <glx> they had to do it differently ;)
15:11:40 <TrueBrain> "php_uname tells you about the current OS. Unless PHP can’t tell what it’s running on; then it tells you about the OS it was built on. It doesn’t tell you if this has happened."
15:11:49 <Kylie> TrueBrain: lowercase should break and not work
15:12:13 <andythenorth> "133" == "0133" but 133 != 0133
15:12:22 <TrueBrain> dynamic casting going wrong :D
15:12:41 <TrueBrain> it is why I dislike weak typed languages .. I rather work in strong typed :P
15:12:52 <TrueBrain> (and yet I program in Python .. hmm ...)
15:13:13 <Kylie> TrueBrain: reword "and yet I..." please
15:13:16 <andythenorth> I have never ever had a type problem in the languages I use :o
15:13:48 <andythenorth> PHP leaves land mines for the unwary developer :O
15:13:56 <TrueBrain> andythenorth: like any language tbh
15:14:10 <TrueBrain> it is only 'recent' GCC warns for statements like if (a = 1)
15:15:21 <andythenorth> null < -1 and null == 0
15:15:53 <TrueBrain> again, dynamic typecasting at work :)
15:16:43 <TrueBrain> this article makes it sound much worse than it is btw .. you just have to act a bit more stupid when programming PHP than when you would program Javascript or C :P
15:17:12 <TrueBrain> don't think: owh, array_pop changes the variable I give to him, so array_reverse would too
15:17:15 <TrueBrain> no, just open php.net
15:17:17 <TrueBrain> look the function up
15:17:19 <TrueBrain> and see what it does
15:17:22 <TrueBrain> MAKE NO ASSUMPTIONS :P
15:17:39 <andythenorth> TrueBrain: when is it time for new bananananaas?
15:17:55 <TrueBrain> weren't you going on a holiday or what-ever? :P
15:21:09 <andythenorth> should I get our python 2.4 app ported to something that isn't dead and buried? :P
15:23:18 <TrueBrain> I wonder if I want to write a Minecraft mod or not ... or maybe an Erlang MMO .. sounds fun too ...
15:32:52 <Warod> TrueBrain: There is Erlang MMO already
15:32:59 <Warod> TrueBrain: Something EvE-ish...
15:33:02 <TrueBrain> there is also a Minecraft mod already
15:33:30 <Warod> wtf was it called... I even registered once but then was too lazy to download the client. :>
15:33:43 <TrueBrain> so yeah .. what can I say
15:33:57 <Warod> that's a challenge for a programmer like you!
15:34:00 <TrueBrain> poems are for boring people :P
15:34:57 <TrueBrain> damn, I need to poo .. bbl :)
15:35:37 *** andythenorth has joined #openttd
15:38:47 *** Prof_Frink has joined #openttd
15:39:10 <Arafangion> Getting rid of shit is always good.
15:42:38 <TrueBrain> and good it was ....
15:42:44 *** kkimlabs__ has joined #openttd
15:49:22 *** frosch123 has joined #openttd
15:51:22 *** valhallasw has joined #openttd
15:56:23 *** valhallasw has joined #openttd
16:40:16 <Eddi|zuHause> <TrueBrain> (and yet I program in Python .. hmm ...) <- python is strongly typed. just dynamically typed. (not a contradiction)
16:40:45 *** kkimlabs_ has joined #openttd
16:43:05 <Eddi|zuHause> C is both strongly and statically typed, but there is only one type.
16:45:06 <frosch123> what's the name of that type?
16:46:03 *** plastics has joined #openttd
16:46:27 <Rubidium> TrueBrain: implement Javascript in OpenTTD, using logic with trains. Then run OpenTTD in OpenTTD ;)
16:51:30 <frosch123> hmm, so what stuff is stored in the map array only for caching, i.e. could be recomputed on load?
16:52:35 <Rubidium> you just need the inital settings, seeds and DoCommand log ;)
16:53:16 <frosch123> well, it should not depend on a certain version of ottd
16:53:38 <Rubidium> what are you trying to achieve?
16:53:41 <frosch123> resp. know the behaviour of all past versions of ottd
16:53:58 <frosch123> Rubidium: foobar asks for townzone for rail tiles
16:54:08 <frosch123> the usual argument is: there is not enough space in the map array
16:54:23 <frosch123> but since we increased the memory usage by factor 64 (?) in the last release
16:54:31 <frosch123> that point does not sound so nice anymore :p
16:55:03 <frosch123> so, i was wondering about adding an extra townid for every tile on the map, while only saving it for town roads
16:55:13 <Rubidium> m6 bits 7..6 : Possibility of a bridge above, in the direction specified
16:56:11 <Rubidium> m2: Index into the array of towns (owning town for town roads; closest town otherwise, INVALID_TOWN if there is no town or we are creating a town) <- seems like it, but isn't really (towns growing into other towns)
16:56:49 <frosch123> yeah, does not work for roads and houses
16:58:08 <frosch123> single tile stuff which is also inside pools would not need storing the id either
16:58:38 *** kkimlabs_ has joined #openttd
16:58:56 <Rubidium> m7 bits 7..6: present road types (for road tiles only)
16:59:06 <Rubidium> the rest doesn't seem to be computable
16:59:26 <Rubidium> not sure whether the slope is cached somewhere though
16:59:38 <frosch123> it isn't, but could be added
17:00:13 <frosch123> idea would be to increase memory usage for caches, but not increase savegame size / network transfer time
17:00:40 <CIA-4> OpenTTD: rubidium * r24374 /trunk/src/lang/korean.txt: -Fix: cargo lists can't have genders (mostly because it's very unclear what gender it would have)
17:00:57 * andythenorth earlier made a mistake in a web form, multiplied value of sales by 100. Made our sales look good for a bit :P
17:01:03 <Rubidium> can random bits on the map be re-randomised?
17:01:38 <Rubidium> then in theory all those random bits can be cached as well
17:02:11 <Rubidium> just need to store the seed and give every tile a byte or so of pseudo random numbers that won't change during the game
17:02:17 <frosch123> err, sorry, i inverted your question, so you need to invert my answer :)
17:02:36 <frosch123> random bits cannot be recomputed
17:02:50 <frosch123> even if they cannot be randomised, they depend on the build date
17:03:07 <Rubidium> frosch123: now they do, but we could change that
17:03:34 <Rubidium> so the map just has a m8 with pseudo random values initialised during map generation with some seed (game seed?)
17:03:52 <Rubidium> but that only works when the random bits aren't changed by anything
17:04:02 <frosch123> this would only affect objects currenlty
17:04:11 <frosch123> and it was already asked to rerandomise object bits
17:04:26 <frosch123> trees and rails already do not store random bits
17:04:37 <frosch123> oh, canals cannot be rerandomised either
17:05:29 <frosch123> i am wondering about the hedge flags of clear land
17:05:47 <Rubidium> those are 'grown', just like water floods
17:06:24 <frosch123> hmm, so far the bridge flags have the biggest impact
17:06:29 <Rubidium> anyhow, trees use pseudorandom based on tile x/y to determine which of the N tree sprites to choose for a particular tree type + amount + age
17:07:10 <frosch123> yeah, we might improve the pseudorandomness for trees, houses and railtypes by adding a random array
17:08:34 *** KylieBrooks has joined #openttd
17:08:42 *** Progman has joined #openttd
17:31:45 *** Alberth has joined #openttd
17:31:45 *** ChanServ sets mode: +o Alberth
17:35:39 *** kkimlabs has joined #openttd
17:36:43 <Wolf01> hell... wait for it... o
17:37:23 <Alberth> so you evacuated into the house?
17:37:50 <Wolf01> and turned on my air conditioner
17:40:06 *** sla_ro|master has joined #openttd
17:40:56 <Rubidium> frosch123: you could also cache the slope (with foundation applied)
17:41:10 <TrueBrain> Rubidium: will you take lead on that project? :D
17:42:15 <frosch123> meh, foundations involve newgrf callbacks
17:42:37 <Rubidium> TrueBrain: hell no ;)
17:42:41 <frosch123> but, ok, it could be cached for the majority of tiles
17:42:51 <TrueBrain> Rubidium: I doubt it is possible in a 2kx2k map, sad :P
17:43:35 <frosch123> TrueBrain: extent them in the depth :p
17:43:53 <TrueBrain> frosch123: will you take lead on that project? :D
17:43:57 <frosch123> why would you need a 8kx8k map, if you could have 64 levels of subway?
17:44:15 <frosch123> what does leading imply? that i tell others what to do?
17:45:28 <CIA-4> OpenTTD: translators * r24375 /trunk/src/lang/japanese.txt:
17:45:28 <CIA-4> OpenTTD: -Update from WebTranslator v3.0:
17:45:28 <CIA-4> OpenTTD: japanese - 34 changes by TzakShrike
18:11:10 <Belugas> that is what i call a truck!
18:11:28 <Belugas> no need to wonder about speed limit, though hehehe
18:16:11 <andythenorth> there is one in HEQS, but it's a bit smaller than that
18:16:53 <Rubidium> the top one is a truck
18:17:10 <Rubidium> the bottom one is/was apparantly the best rally car
18:18:46 <andythenorth> in a straight line, HP wins?
18:19:02 <Rubidium> 220 km/h with a truck?
18:19:06 <andythenorth> even if you're driving something the shape of a piece of cheese, blunt end first :P
18:20:42 <Rubidium> that truck won Dakar by a day
18:23:21 <Rubidium> also, I guess the truck you showed has more HP, but would lose
18:25:13 <Mazur> What has been changed, that suddenly on stable everyone seems to get too long map processing?
18:25:37 <Mazur> No, that should be local.
18:25:58 <Rubidium> suddenly as in since when?
18:31:17 *** kkimlabs_ has joined #openttd
18:32:36 <Alberth> wasn't there a network setting added to get better control over map download?
18:32:57 <Mazur> Anything I can set that'd help me?
18:34:49 <Alberth> I don't know what your problem is tbh, you only gave a very global description
18:40:12 <Alberth> what's map processing?
18:49:20 <Mazur> <Stablean> *** Mazur has left the game (processing map took too long)
18:49:21 <Mazur> <Stablean> *** Mazur has left the game (connection lost)
18:49:50 <Mazur> And that 10 times in a row, or after 1-5 minutes if I do get in.
18:53:03 <Alberth> it's only in the FAQ ;)
18:58:22 *** plastics has joined #openttd
19:01:08 <frosch123> why is it actually called FAQ, if everyone wants the FAA?
19:01:50 <CornishPasty> frosch123: I like to read it as frequently answered questions :P
19:03:04 <Alberth> or the FGA (frequently given answers) ?
19:04:03 <frosch123> well, maybe the FGA/FAA would just be flaming about reading the FAQ :p
19:40:44 <andythenorth> is newgrf smoke done yet? :D
19:41:29 <TomyLobo> burn the midnight oil
19:41:54 <andythenorth> I will be awoken by children at ungodly hours
19:42:16 <Alberth> beds are burning already
19:42:17 <andythenorth> is underground building done yet?
19:42:40 <Alberth> yes, and it is all nicely covered up again, you can't see a thing of it
19:42:45 <frosch123> yes, you just can't see it cause its underground
19:42:45 * andythenorth is plotting a particle accelerator railtype
19:43:35 <andythenorth> no, the trains are protons(?)
19:44:35 <TomyLobo> no, they're photons, can have shared orders
19:45:05 <Eddi|zuHause> frosch123: would it not be possible to calculate the town zone in the tileloop (fence/pavement-like), and not have the townid available?
19:45:20 <andythenorth> TomyLobo: are they entangled?
19:45:29 <andythenorth> but travelling in opposite directions?
19:45:37 <andythenorth> one unloads when the other arrives in a station?
19:45:51 <frosch123> Eddi|zuHause: and what would be the use of that?
19:45:51 <andythenorth> or checking how much cargo they have causes the cargo in the other one to change?
19:46:05 <TomyLobo> cant have both in the same state no?
19:46:10 <TomyLobo> that'd be a train crash
19:46:26 <andythenorth> reality actually works like TTD right?
19:46:44 <andythenorth> it's not 'spooky action at a distance', it's just the game loop tidying things up
19:46:44 <Alberth> reality is a bit flawed
19:46:54 <Eddi|zuHause> frosch123: checking in varaction chain (e.g. determining type of fence. noise-wall in urban areas)
19:46:57 <TomyLobo> no one riding a train or plane knows where they're going until the train stops and kicks them out
19:47:13 <frosch123> Eddi|zuHause: the expensive part of finding the townzone is finding the town
19:47:23 <frosch123> but the nearest town is actually const during the game
19:47:29 <TomyLobo> k-d trees to the rescue, frosch123
19:47:31 <frosch123> so, it is cheap to do it just once
19:47:38 <frosch123> instead of some weird updating in the tileloop
19:48:15 <TomyLobo> just have a k-d tree of town zones (i'm guessing those are static, right?)
19:48:16 <Eddi|zuHause> so, you take something like a delaunay triangulation, and store the result somewhere?
19:48:24 <frosch123> they change all the time
19:48:36 <frosch123> but if you have the nearest town, the townzone is easy to compute
19:48:39 <TomyLobo> no static size? ah then k-d trees might not be the solution :)
19:48:41 <frosch123> and the nearest town is constant
19:48:48 <TomyLobo> ah you just need the nearest?
19:48:53 <TomyLobo> k-d tree to the rescue!
19:49:09 <Eddi|zuHause> the delaunay triangulation is storage size O(#towns), not O(#tiles)
19:49:20 <TomyLobo> O(logn) nearest-point lookup... not sure if it's exact though :D
19:49:21 <andythenorth> can we found towns currently?
19:49:42 <TomyLobo> founding/removing a town => tree needs to be redone
19:50:19 <Eddi|zuHause> recalculation on town funding is not such a big deal, as it's a rare thing, so performance is a minor issue
19:51:11 <Eddi|zuHause> actually, with a delaunay triangulation, it's not even an expensive operation, because only the direct neighbours of the new town are affected
19:51:14 <TomyLobo> i dont know if it's worth investing time in a sub-linear algo
19:51:17 <Eddi|zuHause> it's a local operation
19:51:32 <TomyLobo> how many towns are there, at most?
19:51:57 <Eddi|zuHause> 6000-ish on huge maps
19:52:11 <TomyLobo> hmm ok, that's something else then
19:52:54 <Eddi|zuHause> my gamescript delaunay triangulation took quite a while on those maps
19:53:12 <TomyLobo> i didnt follow from the start... what's your aim?
19:53:31 <Eddi|zuHause> TomyLobo: for any tile, find the nearest town. fast.
19:54:08 <TomyLobo> can you afford, memory-wise, to have one set of towns for each 16x16 cluster?
19:54:12 <Eddi|zuHause> frosch123's initial suggestion was: store the townid for each tile
19:54:16 <frosch123> we will likely just add some cache for the whole map size, without any fancy algorithm :p
19:54:46 <Eddi|zuHause> it's simple, but rather inefficient, both memory and time wise
19:54:56 <Eddi|zuHause> (initialisation time)
19:55:04 <Eddi|zuHause> but very fast on access
19:55:22 <frosch123> TomyLobo: we always denied adding more stuff per tile in the past to save space. but since the introduction of 32bpp and extra zoom increased the memory usage by factor 64, we can no longer argue about saving 8MB for tile->townid caches
19:55:56 <Eddi|zuHause> can we make the tiles 16 bytes while we're at it?
19:56:13 <Eddi|zuHause> so we won't have that discussion 20 times over again :p
19:56:15 <frosch123> the new arguemtn is: we can use more memory locally to cache stuff, as long as it does not go to the savegame and thus does not affect join times
19:56:25 <TomyLobo> so for each 16x16 cluster, you could determine the towns that can be closest to a tile in there
19:57:03 <TomyLobo> you could even re-use the lists in these clusters, if they (and that's one likely case) only contain one city or a very common set of 2 cities or something
19:57:16 <Eddi|zuHause> frosch123: the question there is: does the download size/time decrease warrant the time increase for recalculating the cache?
19:57:54 <TomyLobo> unless it's a phone, yes
19:58:35 <Eddi|zuHause> TomyLobo: the problem with such ideas is that they need a new data structure, while the idea to just save it for every tile can reuse the existing (slightly enlarged) datastructure of the "map array"
19:58:46 <Alberth> TomyLobo: looked at stored save game sizes lately?
19:59:05 <frosch123> Eddi|zuHause: that argument is not valid
19:59:23 <TomyLobo> Alberth noe, but doubling a large part of it is bad either way
19:59:29 <frosch123> it makes no difference to store an array compared to an addtional tile member
19:59:40 <Eddi|zuHause> TomyLobo: afair there's something like a 20 tile minimum distance between towns
19:59:48 <frosch123> and we would not change the tile structure to something not-power-of-two-ish
20:00:22 <TomyLobo> Eddi|zuHause could even use a wider grid then
20:01:06 <frosch123> and if you want to optimise the computation you can first check the towns on a rough grid
20:01:23 <TomyLobo> all the towns in or close to a 64x64 chunk vs all towns on the whole map would still be a lot less to check
20:01:39 <Eddi|zuHause> TomyLobo: but you can have many nearest towns in a cluster, when all the towns are several clusters away
20:01:54 <frosch123> if all edges of a square have the same nearest town, then the whole square have it
20:02:46 <TomyLobo> to fill in the grid, you could use a bleeding algo like notch uses for minecraft's lighting
20:02:46 <Eddi|zuHause> that can't possibly behave better than a delaunay triangulation
20:03:33 <frosch123> are you sure delaunay is the right think to do on an integer-grid?
20:03:53 <Eddi|zuHause> no, but it's the closest thing i know
20:04:29 <Eddi|zuHause> integer-versions of matching algebraic problems tend to be hard. (as in NP-hard)
20:04:40 <TomyLobo> i.e. put all towns into their cells, then have them extend until you hit something that's closer
20:05:24 <Eddi|zuHause> TomyLobo: you mean some kind of flood fill?
20:05:49 <TomyLobo> with 1 overlap of course
20:06:05 <frosch123> that will be weird to do for euclidian distances
20:06:07 <Eddi|zuHause> overlap of what?
20:06:20 <TomyLobo> frosch123 oh i thought manhattan distance
20:06:25 <frosch123> breadth-first works better for manhattan or maximum distance
20:06:45 <Eddi|zuHause> town zones are euclidean, but why would "nearest town" need to be?
20:07:20 <TomyLobo> uh, ok what's a town zone?
20:07:28 <TomyLobo> not the "local authority" thing, right?
20:08:07 <frosch123> CalcClosestTownFromTile uses manhattan
20:08:08 <Eddi|zuHause> TomyLobo: the rings around the town that determine whether pavement, lights or trees are planted on a road and determins which houses can be placed
20:08:23 <TomyLobo> yeah i know what you mean
20:08:41 <TomyLobo> yeah bleeding could be bad for that
20:08:47 <frosch123> then breathfirst might be an option
20:08:59 <Eddi|zuHause> TomyLobo: but that wasn't the problem that was to be solved
20:09:02 <TomyLobo> yeah, breath first :D
20:09:16 <TomyLobo> Eddi|zuHause what else was it?
20:09:22 <Eddi|zuHause> "breathe in" "breathe out" "breathe in"
20:09:48 <Eddi|zuHause> (is that joke known to anyone outside of germany?)
20:10:01 <TomyLobo> i dont get it and i'm german
20:10:17 <frosch123> sounds like a joke abuot blondes
20:10:30 <TomyLobo> you are now breathing manually!
20:11:38 <Eddi|zuHause> goes something like this: "a blonde is getting told to take out her earphones, afterwards she collapses. the other person picks up the earphones, and hears 'breathe in, breathe out, breathe in, ...'"
20:12:27 <Eddi|zuHause> IRC is a bad medium for telling jokes, though :p
20:12:53 <CornishPasty> Why did the chicken cross the road?
20:12:57 <Eddi|zuHause> anyway, the problem was not to determine the town zone, but to determine which town to calculate the town zone from
20:13:24 <Eddi|zuHause> CornishPasty: i never understood why that's supposed to be a joke
20:13:33 <CornishPasty> Eddi|zuHause: it's a double meaning
20:14:07 <CornishPasty> "the other side" refers to the other side of the road, and in a more spiritual sense, the afterlife
20:14:24 <CornishPasty> It's a play on words, but not very funny :P
20:14:32 <Eddi|zuHause> and which part of that makes it a joke?
20:15:25 <Alberth> the double meaning, for people that think that human language is unambigious
20:15:28 <CornishPasty> It's an antijoke apparently :P
20:15:31 <frosch123> ah, i see, the joke is that people write wiki articles about that
20:15:56 <Eddi|zuHause> if at all, i'd like my chicken "kross", though
20:20:11 *** kkimlabs_ has joined #openttd
20:36:07 *** Chris_Booth has joined #openttd
20:37:35 *** Chris_Booth is now known as SmAlL_mEeCh
20:37:40 *** andythenorth has left #openttd
20:45:03 *** SmAlL_mEeCh is now known as Chris_Booth
22:26:22 *** Rhamphoryncus has joined #openttd
22:43:21 *** kkimlabs__ has joined #openttd
22:58:34 *** CptHindsight has joined #openttd
23:30:38 *** Eddi|zuHause has joined #openttd
23:58:16 <Mazur> Anyway, that's next years physics Nobel prize sorted.
continue to next day ⏵