IRC logs for #openttd on OFTC at 2012-02-01
⏴ go to previous day
00:04:21 *** Strid__ has joined #openttd
02:31:46 *** kkb110_ has joined #openttd
03:06:20 *** Rhamphoryncus has joined #openttd
04:19:33 <xiong> For "Show time in minutes rather than days" is this purely visual or is there an effect on play? Is this a simple substitution of 1 minute for 1 day?
04:52:27 *** kkb110_ has joined #openttd
05:03:10 *** Sergio is now known as Guest1184
05:07:40 *** tokai|noir has joined #openttd
05:07:40 *** ChanServ sets mode: +v tokai|noir
05:19:07 *** tokai|mdlx has joined #openttd
05:56:18 *** Eddi|zuHause has joined #openttd
06:57:17 *** Cybertinus has joined #openttd
07:01:40 *** JVassie has joined #openttd
07:10:55 *** sla_ro|master has joined #openttd
07:11:08 *** glevans2 has joined #openttd
07:15:45 <Eddi|zuHause> austrian steam engines are a huge mess...
07:17:43 * Rhamphoryncus swears at tropic refurbishment for finally giving him the highest speed box car.. but splitting it into 3 types, one of which being slower than the others and thus dictating the speed of his entire line
07:19:30 <Rhamphoryncus> Clearly such things require a very different style of play, one I have not discovered yet
07:20:19 <Elukka> i think it requires accepting that sometimes trains have to stop at signals :P
07:20:40 <Elukka> it would be nice if trains knew to slow down when needed so they wouldn't have to come to a complete stop
07:20:58 <Rhamphoryncus> I have a high volume trunk line. They *need* to be in harmony to move effectively
07:23:10 <Rhamphoryncus> I was thinking that. Check the speed of the train in the block ahead (actual speed, not max speed), distance to our signal, etc, and slow down a bit
07:24:34 <Rhamphoryncus> Path signals need some concept of caravaning too, so they'll use that and stick to their line normally, but switch to the other lane when it's nice and open
07:24:50 <Rhamphoryncus> Or a way to put speed rules on lanes :P
07:27:02 <Rhamphoryncus> Actually, multiple rail types could do that. Expensive high-speed rail could get a mild penalty for slow-speed vehicles
07:32:30 <Elukka> well, they already know to pick an empty line between two equal choices (except that one has a train somewhere down the line)
07:32:49 <Elukka> there was the speed signs patch a long time ago... i think it died
07:34:50 *** Progman has joined #openttd
07:36:40 *** eQualizer has joined #openttd
07:37:35 <Rhamphoryncus> Picking an empty line with path signals doesn't work. They switch lines *far* too often. I do have a priority merge by abusing presignals, but I need to redesign it to move them faster
07:46:17 <planetmaker> don't build line switches. Just build your hubs that each lane can go each other on branches
07:49:21 <Rhamphoryncus> So transfer at stations? Or don't transfer, just pass through?
07:49:38 <Eddi|zuHause> planetmaker: i think we need to split the early austrian railway into several, just "kkStB" doesn't cut it
07:49:51 <Eddi|zuHause> planetmaker: i'm thinking StEG, KFNB and SB for starters
07:51:31 *** Twofish has joined #openttd
07:56:39 *** sla_ro|master has joined #openttd
07:57:23 <Rhamphoryncus> Interesting idea.. rather than a rolling merge do a stopping merge. I could also just have a 6-wide waiting area. The point is to suck up the backlog in an efficient way, rather than letting it propagate back through the line
07:59:27 <Eddi|zuHause> Rhamphoryncus: i think the openttdcoop people called that a "tight packer"
08:28:23 <Rhamphoryncus> Heh. Just did a manual autorefit
08:36:35 <Elukka> i am so happy cargodist binaries are again available
08:42:35 *** HerzogDeXtEr has joined #openttd
08:48:46 *** DayDreamer has joined #openttd
09:13:35 <planetmaker> Eddi|zuHause, sure it's needed? And... what are those new abbreviations?
09:15:16 <Eddi|zuHause> planetmaker: the problem is that the development in austria is kinda reverse to the development in germany. the first railways were state-owned and later privatize
09:15:56 <Eddi|zuHause> so around 1900 you get lots of similar engine types
09:16:52 <Eddi|zuHause> the abbreviations are: SB: Südbahn, StEG: Staats-Eisenbahn Gesellschaft and KFNB: Kaiser Ferdinand Nordbahn (i think)
09:17:16 <Eddi|zuHause> where it's my impression that those are the three biggest
09:18:21 <Eddi|zuHause> possibly also the ÖNWB (österreichische Nordwestbahn)
09:19:42 <Eddi|zuHause> where the north and northwest of austria actually means modern czechia :)
09:22:40 *** mahmoud has joined #openttd
09:43:02 <planetmaker> I see. Well. I'm not familiar with the evolution of railways in that area really. If the diversity or amount of engines justifies a split - why not. On the other hand one could make a unified core (as now) and just allow three different expansion settings for that
09:43:54 <Rhamphoryncus> I found a use for break vans!
09:44:29 <Rhamphoryncus> When a UFO lands nearby.. you buy one, 'cuz they're free, and leave it in the depot. This avoids having to fix orders that involve said depot.
09:46:58 <Eddi|zuHause> doesn't really matter whether it's free. as long as it's stopped, you don't pay anything and get the full price back (afaik)
09:47:48 *** andythenorth has joined #openttd
09:48:30 *** TGYoshi has joined #openttd
09:51:27 *** sla_ro|master has joined #openttd
09:53:01 <Rhamphoryncus> It doesn't depreciate when stopped?
10:07:14 *** theholyduck has joined #openttd
10:26:21 <Elukka> the painting in the background looks a tad out of place...
10:34:47 <Ammler> heffer: how does fedora solve the license issue with opensfx?
10:35:57 <Ammler> I got requests to submit openttd to the standard repos
10:36:09 <Ammler> but there opensfx wouldn't be allowed
10:52:34 *** glevans2 has joined #openttd
11:16:45 <planetmaker> Ammler, providing OpenSFX is not necessary really... important is really only OpenGFX as base graphics to make the game start at all
11:24:23 *** Brianetta has joined #openttd
11:44:50 <Eddi|zuHause> dear cat. you are now standing right in front of my screen. would you please bother to move away from there?
11:46:11 <Arafangion> Eddi|zuHause: Why should it? ;)
11:55:43 *** tokai|noir has joined #openttd
11:55:43 *** ChanServ sets mode: +v tokai|noir
12:20:45 *** Progman has joined #openttd
12:36:37 *** roboboy has joined #openttd
12:39:10 *** cmircea has joined #openttd
13:13:27 *** creeper has joined #openttd
13:17:10 *** theholyduck has joined #openttd
13:29:30 *** tokai|mdlx has joined #openttd
13:39:30 *** tokai|noir has joined #openttd
13:39:30 *** ChanServ sets mode: +v tokai|noir
13:44:33 <Rhamphoryncus> "I'll just replace these old oilwell trams with some trains. Sure, they'll cost a lot more, but they'll amuse me, and I'm sure they'll so outstrip demand that I can do amusing but inefficient layout!"
13:45:04 <Rhamphoryncus> said wells have been operating for quite some time and are at 2.4 million litres/month x_x
14:22:49 *** theholyduck has joined #openttd
14:23:33 * andythenorth found a very sane way to do globals in python templating
14:30:25 *** tensai_cirno has joined #openttd
14:30:46 <tensai_cirno> is there developer channel?
14:31:14 <tensai_cirno> I'm playing around with alpha-stage port of OpenTTD for Blackberry PlayBook tablet
14:36:20 <Rubidium> there's no specific openttd port/patch developer channel; most of that just happens in here
14:37:32 <tensai_cirno> Rubidium, okay. Looks like I figured out my little hurdle.
14:48:23 *** TheBrick has joined #openttd
14:49:50 <TheBrick> I've got about 20 hours to spend contributing to this project. got anything to do for me?
14:50:01 <TheBrick> I've got basic programming skills
14:50:47 <TheBrick> or is there a todo list somewhere that I can take a look at?
14:51:11 <Rubidium> the "to do" items are all pretty complex
14:51:18 <Rubidium> though... do you have Mac OS X?
14:51:44 <TheBrick> I could fetch a virtual machine maybe
14:52:31 <TheBrick> if there's brainless grinding work to be done, I'd be happy to do it
14:53:24 <Rubidium> the only brainless stuff to do is fixing some of all the warnings/errors Doxygen generates
14:53:32 <Rubidium> i.e. document/fix documentation
14:54:24 <TheBrick> could I spend 20 hours on that and make significant headway?
14:55:40 <Rubidium> I guess you could write a significant amount of documentation, though start with the simple stuff unless you want to figure out how the complex functions really work
14:56:19 <Rubidium> the advantage is that you can usually stop somewhere midway with documentation and it's not unfinished, just some stuff isn't documented yet
14:56:50 <Rubidium> whereas trying to write some feature or fix a bug within 20 hours might be take too much to get it to completion within that time
14:56:53 <TheBrick> sounds like a job. is it separate documentation, or do I have to put comments into the code?
14:57:20 <TheBrick> fixing bugs is something I could likely do
14:57:53 <TheBrick> I'm just not very familiar with C++'s specific features and depths of hell
15:01:08 <TinoDidriksen> What is the need for OS X stuff?
15:01:41 <andythenorth> OS X has lots of open bugs
15:02:00 <andythenorth> enough that dropping OS X support has been considered
15:03:16 <TinoDidriksen> Nobody has OS X to work on it?
15:07:28 <andythenorth> 'some people' have OS X
15:07:39 <andythenorth> some of those people don't fix bugs
15:07:47 <glx> TinoDidriksen: the main problem is you have to redo everything when a new OSX version is out
15:10:39 <TinoDidriksen> Why? There's no project files for Mac?
15:12:40 <glx> big parts of API get deprecated when a new version is out
15:12:50 <TinoDidriksen> Oh OS X version..
15:16:53 *** TWerkhoven[l] has joined #openttd
15:18:04 <planetmaker> TinoDidriksen, for OSX you should (also) be familiar with Objective C(++) and the specific OSX APIs
15:18:19 <TinoDidriksen> ...or just use Qt
15:19:27 <planetmaker> how's that a solution?
15:19:46 <planetmaker> one could at the same time - and simpler - demand to use SDL
15:20:00 <planetmaker> and spend time in making that work nicer on OSX
15:20:18 <planetmaker> but all that won't help really. Or at least not quickly
15:20:32 <planetmaker> one of the main issues is that OSX is quite hard to virtualize
15:20:41 <planetmaker> thus you cannot well create test environments.
15:20:59 <planetmaker> And that the API deprecates quickly
15:21:22 <planetmaker> thus you're forced to re-write stuff which works fine on version X, but ceases to compile on version X+2
15:21:25 <TinoDidriksen> Qt's QGraphicsView could do the graphics OpenGL accelerated, and Qt on OS X has a very nifty deploy script that bundles all required libs into a .dmg. No need to fiddle with platform specific APIs.
15:21:56 <planetmaker> good. Then fiddling with apple apis is replaced by fiddling with qt apis
15:22:20 <planetmaker> what is gained then?
15:22:26 <planetmaker> Given that qt is used nowhere?
15:23:02 <planetmaker> and usually given that native APIs tend to be faster?
15:23:10 <tensai_cirno> Which OS-level API's are used? Threading, files, network stack.
15:23:24 <tensai_cirno> SDL used for graphics, sound and input
15:23:31 <planetmaker> SDL is not used on osx
15:24:23 <glx> because it often fails maybe
15:24:28 <planetmaker> it's osx support is worse than the native api
15:24:30 <__ln__> because it has various bugs of its own.
15:24:49 <planetmaker> and the osx SDL version is kinda buggy
15:25:23 <TinoDidriksen> Well, making a Qt project would mean one project that would work on WinXP+, Linux, OS X. I might take a stab at that in 2-3 months time...going to invest in a real Mac instead of my current VirtualBox one that I use for building my Qt apps.
15:25:53 <glx> sdl is not used for windows port either
15:26:30 <planetmaker> TinoDidriksen, adding a Qt backend certainly is a project on its own. Maybe it's worth it. But maybe also not
15:26:40 <planetmaker> I'm not sure how well it really supports the platform specifics
15:27:47 <tensai_cirno> Could anyone point me, where code for save/load dialogs located? Application crashes, when I'm trying to leave sandbox.
15:27:49 <TinoDidriksen> So far, all the Qt stuff I've done as "just worked" on all platforms.
15:28:19 <tensai_cirno> TinoDidriksen, there are lots of platforms, where Qt doesn't work
15:28:22 <glx> tensai_cirno: should be in a _gui.cpp file
15:28:31 <tensai_cirno> including toasters powered by NetBSD
15:31:46 <planetmaker> tensai_cirno, what is used on OSX is ApplicationServices, CoreServices, Cocoa (for video, input and music), QuickDraw (optional video)
15:31:59 <tensai_cirno> planetmaker, ouch
15:32:09 <tensai_cirno> i guess it's pretty hard to maintain
15:32:18 <planetmaker> thus the usual frameworks which *should* be used on OSX
15:32:22 <planetmaker> but ... they change :S
15:33:08 <planetmaker> and worse, there's quite some reports which I can't reproduce as they seem to depend on certain hardware arrangements
15:35:41 <glx> like blitters planetmaker ?
15:35:57 <planetmaker> that's tricky terrain, yes
15:36:24 <planetmaker> and it's not like I'm going to buy several macs just for test purposes ;-)
15:40:53 *** spongie has joined #openttd
15:41:38 <spongie> Hi guys. Is there an image or info on the optimal 2 track balancer. I have two tracks in the same direction, so far all the "sollutions" online present different problem.
15:42:50 <spongie> when I use pre signals i have the problem that a train might start moving towards the switch and then get blocked by a speedy train on the other track
15:42:52 <planetmaker> I dare say there is no 'optimal' as the "best" solution depends on the circumstances and the definition of "best"
15:43:19 <spongie> when I use path signals they mysteriously halt the trains sometimes even though a clear path is obviously available.
15:43:36 <spongie> almost as if it takes time to calculate the reservation
15:45:08 <spongie> both the coop site and the regular openttd wiki examples of this result in this behavior since both are using either pre signals or path signals
15:47:16 <spongie> planetmaker: what would you consider the best solution?
15:47:19 *** bondmain_ has joined #openttd
15:51:27 <TheBrick> I'm mostly interested in the sorting issues
15:54:25 <TheBrick> or if there's a place in the forums for these kind of topics, I'd be happily referred to that
16:04:11 <Polsie> i've downloaded the patch for big maps...
16:04:26 <__ln__> that's not a question, that's a statement.
16:04:27 <Polsie> but were must i place that patch? or how do I run that one?
16:06:23 <polsie> so have anyone a solution?
16:07:11 <polsie> because i can't solve it
16:08:54 *** bondmain_ has joined #openttd
16:10:16 <polsie> I didn't understand it al...
16:12:16 <Elukka> it's probably the fourth post you should look at
16:12:28 <Elukka> though i remember trying that one time and it didn't work for whatever reason...
16:17:08 <planetmaker> spongie, as said: there's no general answer. It really depends on the exact circumstances. I'd build a good system with prios where side lines merge and maybe a shift-mainline system
16:18:44 <planetmaker> TheBrick, it's a bit difficult. Not all info are available prior to join. But a solution making use of the existing info should not re-invent the wheel but use the existing sorting framework
16:19:55 <TheBrick> alright. I'm familiar with sorting algorithms, so it sounds like a good idea to try?
16:21:43 <planetmaker> TheBrick, there are other parts which might be nice to fix. Like if you use the OpenGFX+ BigGUI NewGRF you'll notice that there are several places where the sprite size is not properly taken care of in the GUI
16:22:11 <planetmaker> like the x close button on the top left of some windows
16:22:18 <planetmaker> or the <- or -> arrows in some places
16:24:08 <TheBrick> alright, is there a buglist entry for this?
16:25:12 *** George is now known as Guest1227
16:26:45 *** cmircea has joined #openttd
16:28:29 <TheBrick> if there is, I cannot find it. could you do me a favor and open one for this issue? That makes it clear cut for me and a nice place for discussion
16:28:49 *** sla_ro|master has joined #openttd
16:30:03 <planetmaker> how does it make it clearer? :-)
16:30:20 <planetmaker> and yes, one can also just open reports with patches attached
16:33:21 <spongie> planetmaker: its the shift mainline thingy im wondering about
16:34:19 <planetmaker> spongie, it will make sure that the lines are used absolutely unbalanced :-)
16:35:44 <spongie> ok. basically, I have an extra track in the same direction as the mainline and im looking to take some load of the first track
16:36:12 <spongie> my mainlines are usually LLRR
16:36:52 <spongie> and although it sorts of balances with either the pbs or pre signal solutions, theres still a fair amount of erronous shifting and blocking
16:37:34 <TheBrick> I'll see what I can do, planetmaker, but I'm unfamiliar with everything here (project, game, codebase, forums, etc) so I'm looking for directions on how to get started with all this. I have to reference my work here, so an existing bug that I work on sounds like the best approach
16:47:35 *** bondmain_ has joined #openttd
16:51:14 *** theholyduck has joined #openttd
16:56:12 *** tokai|mdlx has joined #openttd
16:59:07 <planetmaker> TheBrick, alternatively you could of course try to improve an AI or write a nice goal script for OpenTTD
16:59:24 <planetmaker> Or write for example a bot logging the commands via the admin port
16:59:48 <planetmaker> those might be projects which are possibly easier contained and don't require you to dig that deeply into OpenTTD's vast source code base
17:00:52 <planetmaker> AIs and goal scripts are written in squirrel, OpenTTD provides a more or less extensive API. And there are libraries available
17:01:19 <planetmaker> For such logging script you might use whatever you like. There exist libraries to access the admin port in both python and java
17:01:34 <planetmaker> To my knowledge at least. Other non-published stuff might exist
17:04:08 <TheBrick> thanks for the info, planetmaker. are you familiar with an AI that could use some work? AI is one of my favorite subjects, and I know a lot of the basics (probability, all kinds of search algorithms, machine learning)
17:05:51 <planetmaker> Well... Most AI luckily are GPL-licensed, so one can start with any. "Need" work is relative... I'd suggest to talk to Zuu. He's quite much into writing AI and scripts and libraries for them
17:06:06 <planetmaker> Hi might also give you more detailed ideas and might welcome contributions to one of his AIs
17:07:25 <planetmaker> Hm, though Zuu's might not yet be there :-)
17:08:00 <TheBrick> is he online on IRC here somewhere?
17:08:05 <TheBrick> I don't see his name
17:08:19 <planetmaker> then he's currently not online
17:08:42 <planetmaker> relatively regularily, yes
17:08:48 <planetmaker> European evening hours
17:08:56 <TheBrick> alright good. I hope to talk to him tonight
17:09:23 <planetmaker> if / when you're registered to tt-forums.net, you could write him a message there
17:10:30 <planetmaker> which also gives you the names of his AIs
17:10:35 <planetmaker> in the description
17:10:49 <planetmaker> And as it's a scripting language... downloading the AI will give you their source, too
17:11:40 *** frosch123 has joined #openttd
17:14:16 <TheBrick> I can't download the SuperLib repository for some reason
17:14:36 <TheBrick> sigh I'm such a noob at this
17:15:01 <TheBrick> trying to checkout the svn base
17:18:38 *** chester has joined #openttd
17:21:57 *** Prof_Frink has joined #openttd
17:28:29 <TheBrick> I'll be back after dinner and see if I can fix these things
17:28:35 <TheBrick> bye for now, and thanks.
17:50:59 *** TheMask96 has joined #openttd
18:03:20 *** bondmain_ has joined #openttd
18:06:55 *** sla_ro|master has joined #openttd
18:32:09 *** Devroush has joined #openttd
18:35:29 *** valhallasw has joined #openttd
18:44:12 *** bondmain_ has joined #openttd
18:45:38 <CIA-1> OpenTTD: translators * r23874 /trunk/src/lang/ (french.txt lithuanian.txt):
18:45:38 <CIA-1> OpenTTD: -Update from WebTranslator v3.0:
18:45:38 <CIA-1> OpenTTD: french - 1 changes by Snail_
18:45:38 <CIA-1> OpenTTD: lithuanian - 11 changes by Stabilitronas
18:53:38 *** Alberth has joined #openttd
18:53:38 *** ChanServ sets mode: +o Alberth
18:54:53 <frosch123> he, citydomination is not even by zuu :o
18:55:12 <frosch123> why did i search for zuu yesterday? :p
18:58:12 <Alberth> the one and only king of the squirrel scripts? ;)
18:58:56 <frosch123> yeah, i just assumed that any script must be by him :)
18:59:13 <frosch123> then i was confused that someone else replied to my bug report :p
19:08:04 *** HerzogDeXtEr1 has joined #openttd
19:09:11 *** andythenorth has joined #openttd
19:11:53 *** Chris_Booth has joined #openttd
19:12:11 <tensai_cirno> right mouse button used only for scrolling map?
19:12:53 <tensai_cirno> i want to implement scrolling with double-finger gesture
19:13:08 <frosch123> rmb can also be used to show tooltips
19:13:28 <frosch123> the depotview also has a speical ctrl+right click tooltip
19:13:36 <frosch123> minimap is also scrolled using rmb
19:13:36 <tensai_cirno> tooltips for what?
19:13:57 <frosch123> those thingies which pop up after some seconds
19:14:11 <frosch123> instead of the delay than can also be shown via rmb
19:14:21 <frosch123> there is an advancted setting for that
19:14:33 <frosch123> tooltips via rmb is the traditional ttd method
19:14:39 <frosch123> delay is more modern
19:15:16 <frosch123> but i think in the train depot the rmb tooltip always works, since it is not suitable to do that kind of tooltip via an hoover delay
19:15:38 <tensai_cirno> usual tooltips could be sacrificed.
19:15:49 <tensai_cirno> rmb+click on trains in train depot?
19:16:46 <tensai_cirno> first time noticed them
19:17:15 <tensai_cirno> Well, playbook can detect up to 4 fingers, I think I'll work around that somehow :)
19:17:41 <frosch123> even ctrl+rmb+click
19:23:34 *** JVassie has joined #openttd
19:38:43 *** tensai_cirno has joined #openttd
19:42:52 *** peteris has joined #openttd
19:44:29 *** theholyduck has joined #openttd
19:54:00 *** tensai_cirno has joined #openttd
19:55:10 *** Biolunar has joined #openttd
19:55:49 *** Chris_Booth has joined #openttd
20:02:04 <V453000> hi, when I need to load a newGRF on a dedicated server, how do I download the content which is missing for that savegame?
20:02:15 <V453000> so that the server loads it correctly
20:03:02 <planetmaker> V453000, via rcon
20:03:11 <Chris_Booth> V453000: you have to make sure that the newgrf is on the server
20:03:13 <planetmaker> but on our servers use !content and it should do the rest for you
20:03:29 <V453000> yeah but it isnt the current content
20:03:35 <V453000> there is an older version of a bananas newgrf
20:06:46 <planetmaker> which NewGRF (incl. md5sum) do you miss?
20:10:37 <V453000> I am already sending those grfs to them
20:10:51 <V453000> but for exaple basecosts 3.2 is in the save, but on bananas is 3.2.1 as latest
20:13:08 * andythenorth starts a not-quite-a-guide
20:17:02 *** kkb110_ has joined #openttd
20:17:05 *** tensai_cirno has joined #openttd
20:23:14 <V453000> [2012-02-01 21:18:06] dbg: [net] Loading game failed, so a new (random) game will be started! [2012-02-01 21:18:07] dbg: [net] Generating map, please wait... this could be caused only by incorrect version right?
20:23:22 <V453000> I am sure I have the version correctly
20:29:50 <Chris_Booth> might be you are loading an unsafe grf config
20:29:54 <Chris_Booth> or have grf errors
20:34:36 <Terkhen> fine, I just finished moving to a new room :)
20:44:58 *** zsoltbp has joined #openttd
20:46:02 <__ln__> i'm not sure it's an entirely positive thing that there are only three HDD manufacturers left in the world...
20:46:57 <TheBrick> partly positive is a greater demand for SSD => cheaper SSDs
20:51:18 <valhallasw> TheBrick: in general higher demand leads to more expensive products
20:51:29 <valhallasw> however, a high demand will also stimulate competition
20:51:53 <TheBrick> I mean there is a high demand for low priced SSDs
20:52:04 <TheBrick> so yes, competition for low prices
20:58:30 *** valhallasw has joined #openttd
21:08:35 *** devilsadvocate has quit IRC
21:16:17 <andythenorth> Eddi|zuHause: is it fair to say your generator is a pure code generator?
21:16:35 <andythenorth> rather than some horrible frankenstein of code generator + substitution-templating
21:16:56 <andythenorth> I'm overlooking the CPP, just referring to the python
21:19:19 <andythenorth> I want to draw a line somewhere that says "either choose simple templating, or learn how compilers work, but don't mix them, it's hell"
21:20:58 <andythenorth> (in a templating guide)
21:21:20 <frosch123> oh, "grf" is short for "graf dracula"?
21:31:29 <andythenorth> otherwise I move onto the more interesting bit. that's all pre-amble
21:31:46 * andythenorth fears lack of sleep = tpyos
21:43:53 *** devilsadvocate has joined #openttd
21:54:55 *** devilsadvocate has quit IRC
22:09:42 <cypher> I just thought of a way to compute matrix determinant with On^64 complexity. May I get my Nobel prize please?
22:10:41 <andythenorth> does it compile nml faster?
22:10:52 <__ln__> since when is ordo notation used without parentheses?
22:11:31 <andythenorth> is there any harm in duplicating about 95% of the code in a newgrf several times?
22:11:43 <andythenorth> it might be a brute force way to do things like FIRS economies
22:11:57 <cypher> __ln__ : since I'm inventing things.
22:12:14 <cypher> andythenorth : Probably not.
22:12:51 <andythenorth> it's probably a huge waste of file size anyway
22:13:13 <andythenorth> and nmlc would barf on duplicate identifiers, so they'd have to be made unique
22:14:15 <cypher> I still thing it's a good idea. Think about it. For 4x4 matrix it's 340282366920938463463374607431768211456 operations, so for 2048*2048 TTD map, it's probably somewhat more.
22:16:36 <TheBrick> I'd say that's an impressive feat
22:16:47 <TheBrick> coming up with such an order of complexity
22:16:56 <TheBrick> are you sure it would actually work?
22:17:48 <frosch123> hmm, i think i only learned to compute them with n!, but wiki says there are n^3 algorithms
22:18:17 <cypher> TheBrick : I'm almost positive.
22:19:49 <cypher> I only thought of it while studying for linear algebra exam, I didn't actually write it. See, I have a cheap HP laptop and no fire alarm whatsoever...
22:19:51 <frosch123> oh, apparently i never learned how to compute a determinant effectively... it's so obvious when i read it...
22:20:48 <cypher> frosch123 : how uneffectively did you do it? Was it anywhere near say... n^5?
22:21:02 <cypher> I think i'm gonna turn that into a competition.
22:21:32 <frosch123> no idea how to do it in n^5. i can only do it in n! or n^3
22:21:44 <frosch123> if you limit the band width, it becomes linear though
22:21:56 <TheBrick> just do bogosort with 128 bit numbers
22:22:22 <cypher> Anyways, I'll rather plug myself off.
22:22:57 <Rubidium> isn't O(n!) > O(n^c)
22:23:42 <frosch123> O(n!) > O(c^n) > O(n^c)
22:23:59 <frosch123> except for certain small values of c :)
22:26:34 <TheBrick> bogosort is O(infinite), isn't it? so I'd win
22:27:14 <frosch123> ottd 1.2-beta3 uses bogosort :)
22:27:15 <Rubidium> just do heapsort with choochoo comparator
22:32:56 *** Chris_Booth has joined #openttd
22:59:35 *** Stimrol_ has joined #openttd
23:03:10 *** David is now known as Guest1264
23:18:44 *** andythenorth has left #openttd
23:27:55 <Rhamphoryncus> TheBrick: bogosort is only O(∞) worst case. Average case isn't quite as bad.
23:28:45 <TheBrick> while (); is that on average though
23:28:46 <__ln__> worst case is not denoted by O()
23:29:00 <TheBrick> well, type me an omega
23:30:14 <TheBrick> eh, well I'm not fluent in this shit anymore, it's been quite a whie
23:30:18 <TheBrick> which one is the right one?
23:30:29 <TheBrick> theta was average case
23:30:55 <TheBrick> what bound is O then?
23:31:00 <TheBrick> I thought it was worst case
23:31:10 <TheBrick> time to look that shit up
23:32:17 <TheBrick> where's wikipedia for the mathematically declined?
23:34:07 <Rhamphoryncus> We're programmers, we don't need to use it "correctly". It's like the english language: we stole the word for ourselves.
23:34:51 <__ln__> i'll check the Book... to see if it possibly is the worst case
23:35:12 <TheBrick> I think I get it again though. O is worst case, omega is best case, theta is when both have equal complexities
23:35:39 <TheBrick> that's probably the undergraduate level understanding of it
23:36:50 <TheBrick> omega is apparently not that clear cut
23:42:20 <__ln__> i have the definition right here
23:43:08 <__ln__> O(g(n)) = { f(n) : there exist positive constants c and n0 such that 0 <= f(n) <= cg(n) for all n >= n0 }.
23:44:45 <TheBrick> this is not the time for maths
23:44:57 <TheBrick> it's time for... sleep
23:45:01 <__ln__> O() is an asymptotic upper bound, the Book says. Ω is asymptotic lower bound.
23:45:18 <TheBrick> so it's worse than the worst case?
23:45:53 <TheBrick> if you consider infinities not to exist
23:46:45 <__ln__> worse than worse sounds bad
23:47:14 <TheBrick> well if you're an optimist; it's never quite as bad as it says
continue to next day ⏵
