IRC logs for #openttd on OFTC at 2020-07-25
⏴ go to previous day
00:23:22 <andythenorth> hmm maybe each reload should be added to browser history?
00:27:40 <andythenorth> hmm, option to save the page :P
00:44:13 <andythenorth> I want to comma-thousands format it, but that is traumatic for some Europeans
00:46:39 <andythenorth> I should do a score multiplier for each snowplough
01:56:23 <DorpsGek_III> [OpenTTD/OpenTTD] jostephd commented on pull request #8278: Feature: decitiles/day velocity unit https://git.io/JJ4Sq
03:20:13 *** Wormnest has joined #openttd
04:28:28 *** Wormnest has joined #openttd
06:55:12 *** ChanServ sets mode: +v tokai
07:06:56 *** ChanServ sets mode: +v tokai
07:11:52 *** tokai|noir has joined #openttd
07:11:52 *** ChanServ sets mode: +v tokai|noir
08:46:54 *** HerzogDeXtEr has joined #openttd
09:26:52 *** Progman has joined #openttd
09:41:44 *** sla_ro|master has joined #openttd
10:44:37 *** frosch123 has joined #openttd
13:34:50 <DorpsGek_III> [OpenTTD/OpenTTD] btzy updated pull request #7028: Feature: Option to group vehicle list by shared orders https://git.io/fhGxd
13:37:06 <DorpsGek_III> [OpenTTD/OpenTTD] btzy commented on pull request #7028: Feature: Option to group vehicle list by shared orders https://git.io/JJBtf
14:03:44 *** jottyfan has joined #openttd
14:07:42 <nielsm> looking at my modular airports notes again...
14:13:42 <nielsm> things an implementation would need: map storage format, definitions for tiles and modules, drawing code for the airport tiles, UI for placing modules, UI for viewing status/validity of the airport, storage and validation of the movement paths, converting existing airports to new format
14:13:53 <nielsm> and I'm sure there's more stuff
14:14:30 <nielsm> well actually making the vehicles use the movement paths and pick an appropriate entry point
14:15:25 <glx> isn't there already some kind of state machine for airports ?
14:15:42 <nielsm> there is, and I'm not sure a modular format would be easy to convert to it
14:15:56 <nielsm> at the very least it would probably need some extension
14:15:56 <glx> ah yes modular is another story
14:16:08 <Eddi|zuHause> it would probably be easier to convert the old format to a modular one
14:16:51 <glx> yes an entry/exit info per tile, with checks with neighbours
14:19:00 <Eddi|zuHause> the more modular it gets, the less likely you can ensure deadlock-free operation
14:19:54 <nielsm> I think some rules regarding planes moving to takeoff get priority, reserve their path, can make it pretty safe
14:20:19 <nielsm> so they should always be able to empty out eventually
14:20:38 <Eddi|zuHause> on radar: might allow planes to take off/land in quicker succession if present?
14:21:26 <nielsm> I think that'd be difficult
14:21:48 <nielsm> unless it's just that without radar there's an extra delay inserted on runways or similar
14:22:04 <nielsm> maybe only allow multiple runways to be reserved simultaneously if a radar is present
14:22:24 <nielsm> so without radar you can't have two planes land at the same time, or one land and one take off at the same time
14:24:21 <Eddi|zuHause> this all blends in the other issue that is occasionally requested: station buildings that have gameplay effects (warehouses, waiting areas, restaurants, etc.)
14:24:58 <nielsm> the movements graph for the airport would probably need to be stored in the savegame, to allow aircraft in the airport to refer to specific nodes and edges without depending on the graph being generated exactly the same after load
14:25:18 <nielsm> and keep reservation data properly
14:29:44 <Eddi|zuHause> and we need decorative station tiles that work for rail, road, dock and airport stations
14:30:28 <Eddi|zuHause> which causes a problem for "legacy" newgrf stations
14:32:55 <Eddi|zuHause> so you can easily make 3 separate projects out of this
14:33:36 <Eddi|zuHause> 1) modular movement patterns, 2) generalizing decorative tiles, 3) tiles that have side effects
14:36:26 <nielsm> hm yeah making old airports use new movements might be the first step
14:36:37 <nielsm> before trying to construct movements from tile layouts
14:48:45 <b_jonas> so as a drawback, the wind direction would vary, and you'd have to construct runways in at least two directions so that airplanes can always take off and land?
14:49:29 <b_jonas> nielsm: "without radar you can't have two planes land at the same time" => what stops you from just having two airports then?
14:49:53 <nielsm> that's two stations, works entirely different
14:51:25 <glx> usually noise level won't allow you to build 2 airports
14:52:28 <b_jonas> oh, as for station buildings. I love the graphics of the Industrial stations renewal newgrf, it lets me build different stations for different cargo. And I found that it at least partly supports FIRS, in that it can display the engineering supplies with the graphics that was made for goods,
14:53:16 <b_jonas> but I'd like some more newgrfs for this, perhaps adding stations with FIRS-specific cargos, perhaps with just more variations of the same thing, so if you can recommend railway station newgrfs like that, I'd be happy.
14:53:33 <b_jonas> also for airports, I loaded the newgrf that has rotated versions of the standard airports
15:30:53 <Eddi|zuHause> frosch123: maybe we should think about deprecating the current station spec, and develop a more object-like one
15:32:24 <frosch123> that's a common pattern. when people do not understand something, i believe it would be easier if they rewrite it
15:33:34 <frosch123> maybe you should first try to add some features to objects, which people asked for years, and which stations support
15:33:48 <frosch123> like dragg-able any-size objects
15:33:57 <frosch123> partial demolition and more
15:55:58 *** andythenorth has joined #openttd
15:58:31 <andythenorth> hmm do logs work?
15:58:55 <andythenorth> I posted Train Whack updates this morning at 9AM UK time or so
15:59:13 <frosch123> yes, i pressed F5 about a hundred times
15:59:18 <frosch123> never got 2 snowploughs
15:59:32 <andythenorth> testing the game was quite repetitive
15:59:39 <andythenorth> testing randomised games always is :P
16:04:48 <frosch123> You scored 1031240! Snowplough bonus: 2 = 4x score multiplier <- yay, finally
16:05:06 <frosch123> how much bandwidth did that cost?
16:05:35 <frosch123> one of the snowploughs is even at the front
16:07:26 <andythenorth> wonder what my AWS bill will be
16:29:36 <michi_cc> andythenorth: The game is missing an "Share your result" link, how am I supposed to show of my awesome creation otherwise? :P
16:29:53 <andythenorth> I don't know how many params I can fit on the query string
16:29:56 <andythenorth> iirc there's a limit
16:30:09 <andythenorth> I could hash table *every* possible combination of vehicles?
16:31:54 <andythenorth> hmm, 2048 chars per url
16:32:06 <michi_cc> I see you use random(). If it is possible in JS, you'd just need to set/read the radmon seed.
16:32:28 <andythenorth> vehicle IDs are 0...9999 range, there are max 15 in a row, 10 rows
16:32:45 <andythenorth> so 600 chars, plus boilerplate separators
16:32:54 <andythenorth> might fit in a query string
16:33:05 <andythenorth> if I just did one row, that would be trivial
16:34:08 <andythenorth> hmm, I could do a service where you take a screenshot, then I read the pixels, and detect the IDs from them
16:34:27 <andythenorth> I suspect that's not very hard if an uncompressed PNG is used
16:34:34 <b_jonas> you can always just use a POST request if you want more parameters in the request than fit in an url
16:34:47 <andythenorth> you can post to JS on S3 now? :o
16:35:03 <andythenorth> probably have to use a Lambda :P
16:35:06 <andythenorth> this might get out of hand
16:36:00 <andythenorth> I like the hash table idea, there are between 10 and 150 vehicles displayed, and there are 496 vehicles to choose from
16:36:06 <andythenorth> how many hashes do I need? o_O
16:36:13 * andythenorth never did hashing
16:36:48 <andythenorth> this will be nCr or some other thing I should have paid more attention to
16:38:25 <andythenorth> hashing each row is probably simpler :P
16:38:47 <andythenorth> that's just 15 entries, with 496 possible values per entry
16:40:43 <DorpsGek> andythenorth: Error: Something in there wasn't a valid number.
16:40:47 <frosch123> just stick to 3 digits per vehicle
16:41:05 <frosch123> 10 query params, one for each row. each a string of 15x3 chars
16:41:20 <andythenorth> I am using the newgrf IDs currently which are widely spaced
16:41:24 <andythenorth> I could reduce it though
16:43:02 <frosch123> what's the hero unit?
16:43:27 <frosch123> i thought it was some cheat to win at train whack
16:44:00 <andythenorth> TBH, for rebuilding permutations, a different design might be better :)
16:44:08 <andythenorth> this is kind of top-down or something
16:44:17 <andythenorth> whichever kind of coding is wrong in comp. sci. class
16:44:24 <andythenorth> I failed the only comp. sci. class I did
16:44:41 <andythenorth> I got the result, but I did the wrong kind of coding, I didn't go to the lecture
16:46:23 <frosch123> top-down for new stuff. bottom-up for established stuff that you never want to deal with again
16:47:09 <andythenorth> it was some parse-data-and-draw-a-chart thing
16:47:19 <andythenorth> so I just parsed the data and drew the chart
16:47:37 <andythenorth> I was supposed to create some charting component or something, and a data parsing module
16:47:42 <andythenorth> then connect them together
16:48:15 <andythenorth> Comp. Sci. seemed moderately less actual science than the Political Sci. degree I actually finished
16:48:18 <andythenorth> and more fashion
16:49:05 <andythenorth> I wonder if I can create an ephemeral high score table with an AWS queue?
16:50:03 <andythenorth> "AWS SQS message retention period is 14 days"
16:50:42 <frosch123> did you add code for local storage, to make results worse the longer you play?
16:50:55 <andythenorth> no, that would require a whole load of PECR compliance crap
16:51:12 <andythenorth> I'd need to do a Data Privacy Impact Assessment, appoint a Data Protection Officer
16:51:17 <andythenorth> get a privacy policy etc
16:51:23 <frosch123> oh, sorry, probably should not have mentioned that
16:51:30 <andythenorth> oh jeez, I'm still collecting logs
16:51:37 <andythenorth> I must remember not to look at them
16:52:13 <andythenorth> I probably should set a retention period and add them to my data register too
16:53:27 <andythenorth> the internet was fun once :P
17:04:34 *** snail_UES_ has joined #openttd
17:29:10 *** Progman has joined #openttd
18:34:33 *** Wormnest has joined #openttd
19:45:45 <DorpsGek_III> - Update: Translations from eints (by translators)
20:25:09 <andythenorth> anyone won TW yet?
20:32:30 <andythenorth> billing forecast: $6
20:32:52 <andythenorth> it's *way* cheaper with cloudfront in front of s3
20:33:39 <frosch123> what's the exchange rate between $ and TW points?
20:34:03 <Eddi|zuHause> does it have blockchain?
20:34:06 <andythenorth> are they tradeable?
20:34:22 <andythenorth> "Personal AWS (502284999405) costs are up $1.52 (6,596%)"
20:34:28 <andythenorth> scary % increase
20:34:33 <frosch123> i got about a million in the best game, is that more than aws earned?
20:34:59 <Eddi|zuHause> there's no highscore, how would i keep track?
20:35:05 <andythenorth> has anyone cloned it yet?
20:35:09 <andythenorth> 2048 got a lot of copies
20:35:57 <frosch123> but 2048 was a copy of 3s
20:36:04 <frosch123> what is TW a copy of?
20:36:29 <andythenorth> I didn't invent it :P
20:36:47 <frosch123> do you have distribution rights?
20:37:08 <andythenorth> I no longer discuss licensing issues :D
20:37:22 <andythenorth> is TW more or less fun than OpenTTD?
20:38:26 <frosch123> today i player a lot more tw than ottd
20:38:50 <andythenorth> I played a lot of OpenTTD this week
20:39:01 <andythenorth> 70 years of game or so
20:39:01 <frosch123> is it easier to add more trains to tw than to ottd?
20:39:13 <andythenorth> frosch123 pretty much identical amount of effort, oddly
20:39:24 <andythenorth> milek7 you had to ask :(
20:39:54 <andythenorth> someone here actually gave me the idea, but I forget who
20:40:16 <andythenorth> someone was hitting refresh on them
20:41:10 <frosch123> btw. "boring mode" is perfectly labeled
20:41:22 <frosch123> i tried it, and it was pretty accurate
20:41:24 <andythenorth> I thought you might like that
20:42:06 <frosch123> can you disable breakdowns?
20:42:21 <andythenorth> in the web inspector yes
20:42:25 <andythenorth> you have to hack the source
20:42:30 <andythenorth> I don't know the license on that
20:42:37 <andythenorth> you might have to send me your modifications
20:43:05 <andythenorth> GPL has this weirdly vague handwaving away about "html etc aren't really code, so we'll just ignore them"
20:43:35 <frosch123> i thought everything is solved by "source is the favorite format for editing"
20:44:05 <andythenorth> anyway just don't distribute, you'll be fine :P
20:44:21 <frosch123> my favorite format is putting weird idea into someone else's head, and see what comes out
20:44:31 <Timberwolf> This is great until you're doing serious retrocomputing and either punching cards or toggling front panel switches.
20:44:35 <andythenorth> I think my sense of lolz is a bit distorted recently, what with COVID and PECR so forth
20:44:59 * andythenorth making simplified mode for Iron Horse, clearly a good sign?
20:45:28 <andythenorth> I should call it pikka mode :P
20:55:38 <Timberwolf> Is there any way using sprite stacks to set pixels of the target sprite back to transparent/index 0? I'm guessing not, the only thing I could think of that might actually write a transparent pixel is using a recolour map to map something else to 0x00, but looks like the colour remap takes place before the transparency check.
20:56:09 * Timberwolf was looking for a lazy way to resolve the "long vehicles glitch in tunnels" problem.
20:58:54 <frosch123> Timberwolf: no, you cannot recolour to transparent
20:58:56 <Timberwolf> enorth> in the web inspector yes
20:59:00 <frosch123> recolouring to color 0 is just black
21:00:09 <frosch123> ottd can clip sprite (as it does for ground tiles on halftile foundations, and in a few other places), but that is not available to newgrf
21:04:11 <Timberwolf> Heh. I accidentally held down shift instead of ctrl and got a cost estimate for setting the start date for my timetable :)
21:04:51 <b_jonas> that's a good thing. I want shift to cancel everything, even things where the game can't give me a cost estimate, for intuitive UI
21:05:26 <Timberwolf> Yeah, it's consistent and it was also correct. The estimate was £0, and it did indeed cost me £0.
21:06:20 <b_jonas> I wonder if it would make sense to have some way to do land raise/lower/flatten/dynamite tools in such a way that it refuses to do expensive things (for early game), as in destroying water or city property
21:06:43 <b_jonas> though for dynamite I'd really like some much more fine-grained control
21:06:49 <b_jonas> where I can tell what exactly to destroy and what not
21:07:02 <Timberwolf> Kind of like minimum bank balance for autorenew, except with terraforming?
21:07:30 <Timberwolf> Making it intuitive for new players would be hard, although I guess the error could mention how to override it.
21:07:50 <frosch123> b_jonas: but tycoons are mad and throw dynamite in all directions
21:08:25 <b_jonas> frosch123: yes, and I want that as an option too, for late game, an option to destroy transmitters and whole cities, even if it costs a lot of money
21:08:35 <b_jonas> but I don't want to misclick with the dynamite and lose all my money that way
21:09:05 <frosch123> nah, i like transmitters
21:09:15 <frosch123> there must be limitations, otherwise there is no game
21:09:49 <b_jonas> frosch123: there are too many transmitters, I hate the map generator for it, I hate it as a game mechanic because it makes you want to generate flat maps just to avoid transmitters
21:09:54 * andythenorth just plays with magic bulldozer on most of the time
21:10:04 <andythenorth> sometimes that causes new problems :P
21:10:06 <b_jonas> I do use magic bulldozer occasionally
21:10:09 <frosch123> i prefer permaban for people playing on flat maps
21:10:12 <b_jonas> I don't leave it on most of the time
21:10:42 <b_jonas> frosch123: I don't want a flat map either, and that's why I magic bulldozered most of the transmitters and lighthouses,
21:11:03 <andythenorth> I played 88 game years this week, there's not much I'd bother changing about OpentTTD currently
21:11:04 <b_jonas> so I think it's the map generation's fault
21:11:20 <Timberwolf> Awful gameplay idea: can only start/stop vehicles or change orders if they have line of sight to a transmitter.
21:11:23 <b_jonas> andythenorth: yes, it's certainly very playable, I like it
21:11:29 <andythenorth> vehicles should show speed at the start of their window destination string, not the end
21:11:33 <andythenorth> I tried to patch that, but failed
21:11:39 <andythenorth> otherwise game is kinda done
21:11:51 <Timberwolf> Additional awful gameplay idea: transmitter based fog of war.
21:12:18 <andythenorth> we should add a bool to each tile - draw/don't draw
21:12:24 <andythenorth> like Neverending Story
21:12:27 <b_jonas> Timberwolf: ah yes, and pylons to supply vehicles and infrastructure with psi
21:12:28 <andythenorth> then let GS turn them on
21:12:51 <andythenorth> anyway OpenTTD is kinda done, we should stop changing it
21:13:11 <andythenorth> and I should fix the Bananas skin but eh
21:13:19 <frosch123> maybe construction cost should scale with the distance to your hq
21:13:50 <Eddi|zuHause> i always wanted statues to be replaced by branch offices
21:13:59 <b_jonas> andythenorth: yes, I'm playing FIRS now, I was asking for station newgrfs similar to industrial station renewal but optimized for it
21:14:10 <andythenorth> and the object sets also
21:14:18 <Timberwolf> Start merging in features from Civ type games.
21:14:33 <b_jonas> Eddi|zuHause: oh, I'd like that, that would mean I can destroy my own branch offices
21:14:36 <andythenorth> I am trying to figure out a pollution type mechanic
21:14:39 <b_jonas> currently I am not allowed to destroy my statues
21:14:39 <Timberwolf> "Senington Heights has revolted due to low ratings! Senington Heights is now owned by Frufingwell Transport."
21:14:41 <andythenorth> like original SimCity
21:15:23 <andythenorth> TBH, I think the only decent feature we should add is the ability to launch a rocket
21:15:28 <frosch123> Timberwolf: i was worried you wanted to make it turn based
21:15:29 <andythenorth> I am quite serious
21:15:43 <andythenorth> complete arbitrary goals; launch rocket
21:15:53 <frosch123> andythenorth: add it to the website
21:15:55 <b_jonas> andythenorth: yes, that could work
21:16:02 <andythenorth> frosch123 the suggestions wiki?
21:16:04 <frosch123> you can copy the code from the web
21:16:19 <andythenorth> shall we just move the game into the browser?
21:16:23 <frosch123> andythenorth: no, add the rocket to the website
21:16:25 <andythenorth> then we can just edit the wasm live
21:16:45 <andythenorth> rocket.openttd.org
21:16:59 <frosch123> firs docs should have the same, with trains
21:17:02 <andythenorth> I used to have a stupid thing like that
21:17:23 <frosch123> click a vehicle, and it drops and crashes to the fllor
21:17:48 <andythenorth> pure CSS animations? o_O
21:17:56 <andythenorth> I can't be bothered to find the stupid web things I used to do
21:18:00 <andythenorth> they were so stupid
21:18:23 <frosch123> weren't they in flash? and noone has flash anymore
21:19:03 <frosch123> is there some permafrost storage for flash players somewhere? so future archaeologist can play flash games?
21:20:03 <frosch123> is there a rosetta stone with carvings in c, java and flash?
21:21:52 <frosch123> it has frames, must be old
21:22:44 <andythenorth> oh the days when we just used to make stuff and upload it to a server :P
21:23:43 <frosch123> seems to work, kind of surprising
21:23:51 <frosch123> i expected it would block all that positiojning stuff
21:23:53 <andythenorth> some of the window sizing isn't quite right
21:24:01 <andythenorth> also screens got a lot bigger
21:24:07 <andythenorth> that was probably made for 800x600 or so
21:24:34 <nielsm> <Eddi|zuHause> i always wanted statues to be replaced by branch offices <- it would perhaps make sense to replace the "build statue" town action with a statue object you just place
21:24:42 <nielsm> but that would probably break AIs :P
21:24:58 <nielsm> but you could then have different objects that have "statue effect"
21:25:32 <b_jonas> like statues in Settlers game, as in expensive buildings that increase the morale but do nothing more direct than that?
21:26:16 <Eddi|zuHause> you must have played different settlers games...
21:26:20 <andythenorth> very design-of-a-certain-era
21:26:22 <frosch123> b_jonas: are statures the new gold?
21:26:39 <frosch123> i s1 you deposited 2 gold for 3 soldiers in houses
21:26:40 <andythenorth> ha my old site is quite similar style to OpenTTD site :)
21:27:08 <b_jonas> frosch123: you need gold to built them
21:27:08 <andythenorth> orange and dark grey, with stripes, was a thing
21:27:45 <andythenorth> looks like I used to draw pixel crap too
21:27:50 <Eddi|zuHause> in settlers 2 you would deliver gold to houses, and when they are used, one soldier of each rank in the house would get promoted
21:28:08 <Eddi|zuHause> so it worked more efficiently in larger houses
21:28:09 <frosch123> andythenorth: i made a screenshot of the old-old ottd website, before it was replaced by the old website
21:28:25 <frosch123> it's probably somewhere on some disk
21:28:43 <andythenorth> wayback machine?
21:29:36 <frosch123> i like the false-coloured header
21:30:24 <frosch123> i think we won hup 3 years in a row
21:30:42 <andythenorth> when was that, 2006?
21:30:44 <frosch123> maybe we got disqualified
21:30:58 <frosch123> andythenorth: old homepage was started in 2008
21:31:06 <frosch123> so that page is 2006-2008
21:31:28 <andythenorth> looks like I joined forums about 2004
21:31:33 <andythenorth> but I didn't really understand the project
21:31:49 <frosch123> did you play ttdp then?
21:31:57 <andythenorth> nah, I was on a mac
21:32:31 <andythenorth> check the SSL :P
21:32:57 <frosch123> so, ottd started on 2004-03-06, and it had a website on 2004-04-23
21:33:32 <frosch123> considering the headers, i bet most of the text is still on current website :p
21:33:38 *** Progman has joined #openttd
21:33:47 <andythenorth> this is a pattern right? In all projects, the old guard do less, and browse the archives more? :P
21:34:06 <Eddi|zuHause> i don't think i've ever seen this version of the website
21:34:12 <Timberwolf> ISTR it was really quick to get from vanilla TTD to near feature parity with TTDP.
21:34:31 <Timberwolf> Quick as in a lot of work happened fast :)
21:34:39 <frosch123> Timberwolf: feature-parity was 2008
21:34:42 <andythenorth> not as fast as you draw trains Timberwolf :P
21:34:58 <andythenorth> unrelated: are forums safe to visit currently? I am sort of avoiding internets mostly
21:35:44 <frosch123> Timberwolf: considering that ttdp only really started in 2002, ottd started in 2004, and ttdp died in 2007. 2008 is not that fast
21:35:54 <andythenorth> "so with any luck we will have an OS/2 binary for 0.5.0 final"
21:36:22 <Timberwolf> Really, that late? For some reason I recall it being a lot quicker, maybe it was only the features I cared about :p
21:36:42 <andythenorth> there are those who say feature parity has never been achieved
21:36:51 <andythenorth> but they mostly say it in places I don't go in
21:37:09 <andythenorth> in languages ich spreche nichst so gut
21:37:33 <Timberwolf> The current realistic acceleration model took a while, I remember hacking in my own years ago.
21:38:00 * andythenorth wonders when TW will overtake OpenTTD in features
21:38:02 <Timberwolf> Mine was evil in that trains could fully stall on hills.
21:38:22 <andythenorth> did turning them round solve it?
21:38:34 <andythenorth> maybe TW needs a content layer, for modding
21:38:35 <Timberwolf> It solved that particular problem.
21:38:59 <Timberwolf> GL if you'd put a one way signal halfway up the hill.
21:39:19 <andythenorth> pre-PBS I used to get trains so stuck that crashing them was the only option
21:39:31 <frosch123> andythenorth: how about a "feature request" link, that links to the gh fork action?
21:39:44 <andythenorth> or just opens web inspector? :P
21:39:55 <andythenorth> I've not actually tried editing JS live in the dom
21:40:01 <andythenorth> not sure what it does
21:41:26 <Timberwolf> My main network problems pre-PBS came from living through The Age Of Roundabouts.
21:41:35 <_dp_> I recently had an idea to do fog of war with landscaping
21:41:43 <frosch123> hmm, i just noticed, while it says "transport tycoon", it missed the "deluxe"
21:41:47 <frosch123> so what does the D stand for?
21:41:49 <_dp_> when fog lifts it raises the land and funds industries and towns
21:42:11 <andythenorth> if we're making changes, we should make big silly changes :)
21:42:22 <andythenorth> that don't require nml patches or newgrf wiki docs :P
21:43:21 <Timberwolf> vanilla pathfinder was interesting to work around, too. ISTR it has situations where a train will always turn in the compass direction to their destionation, even if it's not a viable path.
21:44:24 <Timberwolf> I always ended up with at least one "pathfinder correction loop" somewhere in a network.
21:45:08 <frosch123> Timberwolf: the first "improved" train pathfinder (named ntp: new train pathfinder), solved exactly that issue by picking a random direction when lost
21:46:18 <Eddi|zuHause> i always built my crossings in a way that compass direction was the right one
21:46:41 <Timberwolf> Oh yes! I remember NTP. Usually swapping to it mid-game when whatever the first full pathfinder got beyond what my PC could cope with.
21:47:07 <Timberwolf> Eddi|zuHause: well, that would be the "good at the game" option :)
21:47:18 <andythenorth> ok brains trust: help me our with FIRS industry closure, it nearly tipped me into a rage quit 2 weeks ago :D
21:47:42 <andythenorth> FIRS 3 has closure for secondary industries if unserviced for some years (I never use it)
21:47:54 <Eddi|zuHause> are we quoting scrubs now?
21:47:58 <Timberwolf> I think it wasn't until 1998 or 1999 that I played a successful game which wasn't either extending Megarail, 1960 or just building bus and lorry services on West Country 90210.
21:48:05 <andythenorth> I think scrubs quoted something else :)
21:48:07 <andythenorth> FIRS 4 closure is broken due to new production mechanics etc
21:48:19 <Eddi|zuHause> quite possibly :)
21:48:34 <andythenorth> I wonder if (a) it's acceptable to say closure = use magic bulldozer
21:48:46 * _dp_ hates closing with passion
21:48:47 <andythenorth> or (b) not including closure removes some richness from the game
21:48:51 <andythenorth> or (c) something else
21:48:52 <Eddi|zuHause> but even if i had watched that other stuff, it would have been in german so the reference might have been lost
21:49:35 <nielsm> industries opening and closing during the game, and having unpredictable production changes, is one of the things I like in Transport Tycoon over Transport Fever
21:49:56 <andythenorth> FIRS is very linear
21:50:12 <andythenorth> production only changes when you want it too - I have ideas to vary that
21:50:30 <Eddi|zuHause> andythenorth: somehow i doubt that scrubs would have referenced a UK thing
21:51:23 <andythenorth> what's old is new again
21:51:29 <andythenorth> the world turns on it's axis
21:51:52 <andythenorth> one man struggles
21:51:56 <andythenorth> while another relaxes
21:52:54 <frosch123> anyway, extrapolation time. first website 2004-04 to 2005-07, 2nd 2005-08 to 2008, 3rd 2008 to 2019, 4th 2019 to ??
21:53:01 <andythenorth> where even is TrueBrain? Holidays?
21:53:03 <frosch123> when do we get the next website?
21:53:23 <frosch123> don't wake the dragon
21:53:24 <andythenorth> srs answer: not this year, but I would like to fix bananas skin
21:54:25 <andythenorth> I quite like the website, except the page structure is no longer good
21:54:39 <Eddi|zuHause> maybe you should peel the banana first?
21:57:20 <Eddi|zuHause> This TrueLight person must be a real proper idiot :p
21:57:20 <andythenorth> hmm, how much town storage is there?
21:57:32 * andythenorth wonders about more intelligent closing
21:57:43 <andythenorth> how about global storage for newgrf? :P
21:58:24 <andythenorth> hmm, no, please no more newgrf spec changes
21:59:59 <Eddi|zuHause> andythenorth: my thoughts on closing: 1) have a parameter for "no closing", 2) have areas that the player never touched "protected" so they aren't empty when the player finally manages to get there
22:00:43 <Eddi|zuHause> as a rough estimate for "never touched" you can maybe query the rating of nearby towns?
22:00:57 <andythenorth> I was wondering about minimum number of industries per quadrant of the map or something
22:01:24 <andythenorth> inevitably, this is all GS stuff :P
22:01:33 <Eddi|zuHause> that sounds like it would be failing for different density settings
22:01:51 <andythenorth> I might break those in FIRS soon anyway
22:02:04 <andythenorth> they're not very good
22:02:24 <andythenorth> can I just let OpenTTD close the industries?
22:02:35 <andythenorth> for secondary, I don't understand the basis for default closure
22:02:52 <andythenorth> primary uses prod_multiplier I think
22:05:44 <andythenorth> if ( (byte)(_cur_year - i->last_prod_year) >= 5 && Chance16(1, smooth_economy ? 180 : 2)) { closeit = true; }
22:06:02 <andythenorth> so it's just no prod for 5 years and random chances
22:06:29 <andythenorth> but I have to reimplement that myself I think
22:06:40 <Eddi|zuHause> yeah, and this 5 year grace period is ... weird
22:06:57 <andythenorth> I could just reimplement it and move on
22:07:01 <andythenorth> that's what I did last time :)
22:07:14 <andythenorth> frosch123 said improving closure in newgrf was impossible and I should stop wasting my life :)
22:08:36 <Eddi|zuHause> no. newgrf can probably improve closing, but it cannot work with global economy data
22:10:16 <andythenorth> I didn't try giving each industry some unique seed
22:10:25 <_dp_> remove all closing and let GS solve it :p
22:10:25 <andythenorth> the biggest problem is waves of closure
22:10:37 <andythenorth> _dp_ we should just make a GS bot
22:10:55 <andythenorth> GS is the new XML
22:10:58 <Eddi|zuHause> yes, the waves of closure come from the 5 year period
22:11:31 <Eddi|zuHause> it'll catch all industries you haven't touched in the first 5 years of gameplay. then the rest of the map will be empty
22:12:06 <andythenorth> the game solves that because it only visits each industry occasionally, unless smooth economy is on
22:12:20 <Eddi|zuHause> that's why my suggestion would be to start the 5 years only if the nearby town has a rating for at least one compan
22:12:28 <andythenorth> yes, that could be a mitigation
22:12:45 <andythenorth> other mitigations could be giving each industry a periodicity
22:13:01 <andythenorth> this has to run on the monthly cb, the occasional cb can't be trusted due to economy setting
22:13:39 <andythenorth> but if the check was mostly 'return nothing' and only actually attempted closure 1 in n times
22:14:07 *** gelignite has joined #openttd
22:15:23 <Eddi|zuHause> the problem there is that N needs to depend on map settings
22:16:40 <andythenorth> fundamentally this is an induction vs deduction problem
22:16:53 <andythenorth> I just wonder if I can plaster over it enough
22:17:26 <andythenorth> the downside is I'll never know because I never use the setting :)
22:18:27 <Eddi|zuHause> you'll never find the right N without massive playtesting between different settings and playstyles
22:18:38 <andythenorth> ...which I'm not going to do
22:18:46 <andythenorth> I might do something different
22:19:38 <andythenorth> this is another class of the problem of distributing industries
22:19:56 <andythenorth> I use newgame a lot because mapgen gives non-viable maps (not enough coal mines or something)
22:20:32 <andythenorth> distributing closure values for arbitrary density and map size is similar issue, but not discovered until you have played for 5 years
22:21:11 <andythenorth> there might be crude solutions
22:21:13 <Eddi|zuHause> i'd still try the town rating idea
22:21:25 <andythenorth> I forsee a side effect though :)
22:21:36 <andythenorth> just as you start doing well in a town, the industries start closing :)
22:21:43 <_dp_> since I started working on newgrf server I'm really curious whether I'll end up actually using any grfs or it will just be a single megaforkgrf of my own ...
22:22:16 <Eddi|zuHause> you need to store the date when the first town rating appeared, and then count the 5 years from there
22:22:39 <andythenorth> I have related idea about making industries volatile or stable
22:22:52 <andythenorth> which I might use also for primary production
22:23:23 <andythenorth> when an industry is built, I can count the type, and distribute volatile / stable on odd/even
22:23:43 <andythenorth> so 50% will be of each type, at least until closures and openings start
22:24:47 <Eddi|zuHause> just make sure the first one is always stable. for smaller maps that only have 1 of each type :p
22:25:36 <andythenorth> much as I hate newgrf spec change these days, 256 grf-local storage registers for an industry newgrf might be useful :P
22:25:50 *** jottyfan has joined #openttd
22:28:20 <Eddi|zuHause> that sounds... problematic
22:29:24 <andythenorth> it only solves closure, not needed for much else
22:29:56 <andythenorth> not even sure it can solve closure, due to way cb works :P
22:30:44 <Eddi|zuHause> i mean, it would introduce state into previously steteless operations, and i don't think you can restrict it to industries only
22:43:01 *** HerzogDeXtEr has joined #openttd
22:50:02 *** virtualrandomnumber has joined #openttd
23:14:22 *** virtualrandomnumber has quit IRC
23:25:08 *** WormnestAndroid has quit IRC
23:26:50 *** WormnestAndroid has joined #openttd
23:34:49 <b_jonas> I think I'll dismantle my food trains and rebuild entirely. With the higher amount of money that I have now, and faster trains for the quickly spoiling cargo, and new industries that keep opening, I can probably make it work after a redesign, but the current network is terrible.
23:51:01 *** JacobD88 has joined #openttd
23:51:33 <b_jonas> oh I see! FIRS industries differ in how much cargo they want to start gung-ho production levels:
23:52:35 <b_jonas> iron ore mine only wants 80 crates of engineering supplies in 3 months, but bulk terminal wants 640 tons of clay in three month
23:54:00 <b_jonas> luckily the high production is not too important for the bulk terminal
continue to next day ⏵