IRC logs for #openttd on OFTC at 2015-05-25
⏴ go to previous day
00:51:57 *** zeknurn has joined #openttd
01:01:26 *** sim-al2 has joined #openttd
01:26:55 <glx> some routers prevent servers from being accessed from local ip with outside ip
01:33:48 <sim-al2> I did get that. It seems to work if I use dedicated mode now
02:08:45 *** Ketsuban has joined #openttd
02:24:39 *** Biolunar has joined #openttd
02:57:04 <sim-al2> Ok update: I think that the server advertises for a minute then stops. I am using the dedicated mode and a player was able to find the server and join from the list but now the server does not appear.
03:03:48 <Eddi|zuHause> that might mean you have a fishy connection
03:04:15 <Eddi|zuHause> the masterserver uses UDP to test your server. if lots of packages get lost, it might not get a reply
03:35:07 *** DarkenMoon has joined #openttd
03:43:51 *** shirish has joined #openttd
03:57:37 *** supermop has joined #openttd
04:08:40 *** Stimrol has joined #openttd
04:56:24 *** Biolunar_ has joined #openttd
05:18:20 *** CompuDesktop has joined #openttd
05:29:17 *** smoke_fumus has joined #openttd
06:56:17 *** Eddi|zuHause has joined #openttd
07:41:51 *** sla_ro|master has joined #openttd
07:47:55 *** Alberth has joined #openttd
07:47:55 *** ChanServ sets mode: +o Alberth
08:01:03 *** andythenorth has joined #openttd
08:32:48 *** andythenorth has joined #openttd
08:33:05 * andythenorth considers a ‘casual’ supplies mechanic
08:33:12 <andythenorth> for the player who can’t really be bothered :)
08:33:16 <andythenorth> Alberth: ^ ? o_O
08:35:46 <Alberth> and while flattered, tbh I am not sure you should do that, most players play firs for exactly the supplies, as I understand it
08:36:49 <Alberth> so as game mechanic it's a good concept
08:37:09 *** shirish has joined #openttd
08:38:22 <Alberth> obviously, I am aware I am dropping a large part of the game play by not handling supplies :)
08:40:34 <V453000> XD yeah how does it work
08:40:40 <Alberth> oh, I do use cdist for supplies, just drop supplies at a point 'nearby', and let a different train run a circle delivering; cdist does some form of splitting the delivered cargo :)
08:42:26 <andythenorth> how to code “you have made a token effort at delivering supplies, here is a small reward”
08:43:26 <Alberth> with BB, it's fun to deliver 25t coal to the other side of the map
08:43:46 <Alberth> if I have to deliver supplies, I can just as well do it properly
08:44:12 <Alberth> it's basically not much more than buying a longer train to move the stuff
08:47:05 <Alberth> perhaps my main problem with supplies is the precise timing of it
08:47:34 <Alberth> and lots of players seem to have troubles delivering the precise amount too
08:48:29 <Alberth> it's a bit like feeding tropic towns in the desert, they need food+water in the same month
08:48:49 <andythenorth> it’s a very similar mechanic
08:48:49 <Alberth> it's hopeless if you try that with independent supply chains
08:49:42 <Alberth> supplies is better in the sense that you have 1 cargo, and a time window, instead of a calendar month
08:50:14 * andythenorth wonders about quarterly amounts
08:50:35 <Alberth> to me, it's too much detailed messing about
08:50:47 <andythenorth> and more stable production
08:50:49 <Alberth> on the other hand, lots of people really seem to like the idea
08:51:21 <andythenorth> I am +/-0 to the whole concept
08:51:42 <andythenorth> I don’t like the original TTD mechanic for increasing production
08:51:48 <Alberth> timetables, worried about the bearings of russian wagons, all very much detail
08:51:55 <andythenorth> but I don’t think FIRS is categorically better
08:52:24 <Alberth> I don't think you can generalize such notions
08:52:46 <Alberth> the best you can do is decide for yourself
08:53:04 <Alberth> and even then, there is normally not a single winner
08:53:37 <Alberth> I like the increasing production, but it's my sytle of playing
08:55:24 <Alberth> if you make changes, make them additions, and keep the original behavior
08:55:25 <argoneus> good morning train friends
08:55:45 <Alberth> people can try different variations, and decide what they like best
08:58:32 <andythenorth> I am keeping current version
08:58:50 <andythenorth> the question is whether there are 1 or 2 alternatives that would be fun
09:02:58 *** zeknurn has joined #openttd
09:08:39 *** Pokka is now known as Pikka
09:41:23 *** shirish_ has joined #openttd
10:03:24 *** Progman has joined #openttd
10:21:19 *** andythenorth has joined #openttd
10:22:20 <andythenorth> Pikka bob: industry production that increases by year. You canned all that?
10:24:37 <Alberth> what are you trying to solve?
10:24:39 <V453000> that sounds simple nicely, but I think it would be great if the player still gets the higher production only if it is serviced
10:24:53 <Alberth> change the mechanism, or add one as alternative, or improve it?
10:24:56 <V453000> ie, just capping original industries at ~1000 units of cargo per month max would be fine
10:27:48 <Alberth> if you want to reduce the 3-months stress, you could simply store unused supplies rather than consume it without using it
10:28:20 <Alberth> perhaps with a small degradation, not sure you need that
10:29:11 <Alberth> an alternative can be to used delivery-based instead of time-based
10:30:08 <Alberth> if you want a more extreme variant, you could require a minimal amount of deliveries before producing stuff
10:42:13 <andythenorth> “Alberth: what are you trying to solve?” <- I wonder that too
10:42:22 <andythenorth> this is more of the experimental thinking
10:43:55 <Alberth> towards simplifying for casual players, towards more advanced players, something else?
10:44:22 <andythenorth> dunno, towards inventing variety in gameplay I think
10:44:24 <Alberth> making it more complicated or more challenging?
10:44:32 <andythenorth> challenging / different
10:44:38 <andythenorth> PBI does that well
10:44:59 <andythenorth> I can’t play PBI anymore, the stockpiling mechanic bores me horrid now, but it was fun for some time
10:45:32 <Alberth> stockpiling in itself isn't a problem, it's the capping that breaks things, imho
10:45:40 <andythenorth> eh standard game progression for a player is modelled as: (1) learning the game (2) mastering the game (3) boredom
10:45:50 <andythenorth> that’s your basic ludic theory
10:46:03 <andythenorth> people write books that are much longer, but say the same :P
10:46:05 <Alberth> I think that holds universal
10:46:40 <Alberth> I have great respect for people that manage to eg run a shop for a decade
10:47:01 <Alberth> I would be bored beyond belief
10:47:03 <andythenorth> so I am looking for new options for (1) a new mechanic to learn (2) a mechanic that is fun, not tedious
10:47:26 <andythenorth> doesn’t have to be supplies
10:47:39 <andythenorth> problem is, every “interesting” idea I think of, would be better in GS
10:48:38 <Alberth> at local industry level, I think you can only reward steady deliveries as 'new'
10:49:36 <Alberth> perhaps getting more output for the same output, if you keep input at the same level for a longer time?
10:49:54 <Alberth> ie the industry 'optimizes' on the input level
10:50:58 <andythenorth> I think a ‘steady’ or ‘gradual’ mechanic is worth doing, probably along the lines Eddi suggested
10:51:28 <andythenorth> anything else is beyond FIRS
10:51:28 <Alberth> the more optimized, the narrower the band of allowed deviations
10:52:24 <Pikka> although something more deterministic and controllable than "random" for sure.
10:52:25 <Alberth> I think it's all experimental anyway, you have to code and try it, to decide if it's useful
10:52:49 <Pikka> by year makes assumptions about increasing carriage capacity over time
10:53:05 <Pikka> and also non-interactive
10:53:08 <andythenorth> Pikka: what will your supplies do?
10:53:14 <andythenorth> also how do Yetis work?
10:53:20 <Alberth> maybe have different mechanics for different industry types?
10:53:23 * andythenorth could just steal someone else’s idea
10:53:36 <Alberth> afaik, yetis work very hard :)
10:53:47 <Pikka> my supplies will either temporarily increase production, progress the industry towards an upgrade, or both. haven't really worked out any mechanics yet :)
10:54:26 <andythenorth> (1) delivering gives a temp (30 day) boost to next level
10:54:40 <andythenorth> (2) delivering consistently for 12 months makes that level permanent
10:54:50 <andythenorth> this is like old FIRS + new FIRS
10:55:28 <andythenorth> is there a ‘level down’?
10:55:59 <Pikka> nope, I've got some idea about industries going bust though. if you level them up you have to maintain high service or they'll close down on you, or something.
10:56:31 * andythenorth ponders designing for small maps
10:56:38 <andythenorth> more micro-management
10:57:05 <Pikka> test everything on 256*512 imo :P
10:57:17 <Pikka> bigger maps are fundementally silly.
10:58:22 <Alberth> ECS closes down on you, Pikka
10:58:51 <andythenorth> bigger maps introduce all kinds of problems
10:59:11 <Alberth> couldn't you just 'punish' by lowering production for some time?
10:59:18 <andythenorth> eh, also micro-management is too easily dismissed as ‘micro-management is bad, mkay'
10:59:46 <andythenorth> more micro-management imo, mediated by GS
11:00:08 <Pikka> I think most of the specific mechanics will be switchonoffable
11:00:19 <Pikka> since no-one can agree on what makes a good industry grf :P
11:00:35 <andythenorth> I thought it was understood that ECS has nailed it?
11:00:54 <Pikka> in general, I quite like the idea of industries closing and opening during gameplay, both primary and secondary.
11:00:54 <Alberth> by definition, people are never happy with what you give them
11:01:09 <Pikka> makes the game world more alive, makes network building less fire-and-forget
11:01:16 <Pikka> more micromanagement ;)
11:01:21 <Pikka> or middle-management, perhaps
11:01:28 <andythenorth> let’s extend NoGo so it can handle the industry monthly and random cbs
11:01:35 <andythenorth> then we get production level and closure
11:01:40 <Pikka> it needs to happen in an explicable and predictable way though
11:03:05 * Pikka never played ECS... at least not for many years.
11:03:09 <Alberth> I would prefer a saner interface than n*2 times
11:03:58 <andythenorth> that’s how ‘old’ FIRS did it
11:04:20 <andythenorth> that cb is feature-complete imo
11:04:27 <andythenorth> it does everything a GS would need
11:04:45 <andythenorth> and it’s only called monthly per industry, so not CPU intensive
11:05:06 <andythenorth> or there’s a random equivalent, which varies by smooth / non-smooth economy iirc
11:05:49 <andythenorth> “locusts eat your grain, your grain network is in chaos"
11:06:11 <andythenorth> “3 years of good weather means bumper harvest"
11:06:29 <Alberth> cb can be used once / month only?
11:06:48 * Pikka sometimes wishes for a global random seed accessable to newgrf
11:06:57 <andythenorth> I thought there was one?
11:06:57 <Pikka> so we could have a bumper harvest, across all farms.
11:07:09 <Pikka> I don't think so? I've been asking for one for a while...
11:07:14 <andythenorth> I worked out a hack once by counting towns
11:07:31 <andythenorth> oh you can read town registers
11:07:38 <andythenorth> strictly they could be used for message passing
11:07:49 <andythenorth> was advised it was a Bad Idea
11:46:47 *** tokai|mdlx has joined #openttd
11:55:29 *** Pensacola has joined #openttd
11:58:04 *** tycoondemon has joined #openttd
11:58:52 *** CompuDesktop is now known as Compu
12:02:06 *** tycoondemon has joined #openttd
12:14:26 *** Ketsuban has joined #openttd
12:19:10 *** tycoondemon has joined #openttd
12:20:20 <supermop> is it wrong to use this huge 'centennial' diesel to pull passenggers?
12:28:30 <Flygon> Heavy Harry was designed for pax, afterall
12:28:45 <Flygon> And it was Australia's singular most powerful locomotive from the 1930s through to 1995
12:39:43 <Eddi|zuHause> you don't really have a lot of choice anyway
12:44:49 *** frosch123 has joined #openttd
12:53:56 <supermop> the fp40 or whatever looks boring doubleheaded
12:54:32 <Alberth> ugh, palette of newgrfs is not saved :(
12:55:11 <Eddi|zuHause> no, but where would you save it?
12:58:16 *** Klanticus has joined #openttd
13:01:49 <Alberth> in openttd.cfg probably
13:06:16 <Eddi|zuHause> i suppose nobody bothered to implement it
13:06:28 <Eddi|zuHause> it's only needed if your ancient grfs have mixed palettes
13:06:45 <Eddi|zuHause> if all have the same palette, just use the default switch
13:36:40 *** HerzogDeXtEr has joined #openttd
13:36:46 *** andythenorth has joined #openttd
13:40:35 <andythenorth> Pikka: wanna make an industry set together? o_O
13:40:52 <Pikka> sure. when do you want to start? :D
13:41:02 <andythenorth> about 4 years ago
13:41:38 <Pikka> I'll write some pineappleindustry documentation some time soon
13:41:42 <Pikka> maybe after the kickstarter
13:41:59 <Pikka> definitely after tonight, I have about another thousand words to write in the next 2 hours. :)
13:42:05 <andythenorth> I like GarryG, he has sent me some PMs
13:42:18 <andythenorth> someone should kick Pikka
13:42:23 <Eddi|zuHause> i'm having a hard time distinguishing which of these lines are ironic :p
13:42:38 <Eddi|zuHause> (if in doubt: all of them.)
13:42:43 <andythenorth> Eddi|zuHause: sometimes even English people struggle with that
13:42:50 <andythenorth> sometimes even the speaker [writer]
13:54:26 * andythenorth wonders what this ‘div’ and ‘multi’ do in FIRS clustering, and what they should be set to
13:56:05 <andythenorth> I want slightly more clusters than I’m getting
13:57:51 <Alberth> it's a work around for the obsessiveness of newgrf for integer numbers :)
13:58:07 <Alberth> ie you cannot multiply with 0.6
13:58:26 <andythenorth> as far as I can see, if mult is 1, and div is 2
13:58:27 <Alberth> so you multiply with 3, and then divide by 5
13:58:31 <andythenorth> I should get 2 clusters per map
13:59:23 <andythenorth> a better algorithm would limit the number of industries per cluster, then go build more :P
13:59:40 <Alberth> mult and div just means ( (real)mult/div ) as multiplication factor
14:00:06 <Alberth> but in the integer domain, so you get an integer number as result
14:01:43 <Alberth> hmm, yeah, T3rkhen looked into that one at some time, but it appeared you cannot express "I want N industries at most" (for N=3)
14:02:35 <andythenorth> so if mult = 1 and div = 3
14:02:52 <andythenorth> how many clusters? :P
14:05:25 <Alberth> Any idea what "industry_clusters" is at line 84?
14:05:49 <andythenorth> map scaling factor
14:05:52 <andythenorth> it’s a parameter
14:06:19 <Alberth> oh, wrong macro problably?
14:07:46 <andythenorth> the comment is wrong on L6 I think
14:07:56 <andythenorth> closures / clusters /s
14:08:05 <Alberth> industry_count(related) * mult / div / industry_clusters != 0 <-- this counts the related industries, multiplies it with 1/3 (in your case), and checks it's less than 1
14:08:39 <Alberth> which of course fails if you have 3 or more industry_count(related)
14:08:54 <Alberth> (assuming the parameter is 1)
14:10:08 <andythenorth> < 1 evaluates same as !=0 :o
14:10:16 <andythenorth> integer maths, floor?
14:13:05 <Alberth> yes, unless they are negative
14:13:47 <Alberth> but that's about how you truncate a floating point number
14:28:04 *** zeknurn has joined #openttd
14:36:59 <supermop> can i re-order my orders, but in the same order?
14:38:17 <supermop> like right now the order list is "CDAB", can i easily make it "ABCD" without messing up my timetable?
14:39:42 <supermop> for some definitions of 'easily"
14:40:21 <V453000> heyoo you arent asleep yet :) sending email in 1 minute :P
14:41:00 <andythenorth> so coal mine has mult = 2 and div = 3
14:41:13 <andythenorth> so 1 cluster per 256*256 of map size
14:45:44 * andythenorth wonders if fractional amounts are worth bothering with
14:45:51 <andythenorth> and also, FIRS doesn’t work at 256*256
14:46:14 <andythenorth> maybe just specify number of clusters
14:57:59 *** Celestar has joined #openttd
15:09:03 <andythenorth> that’s plausible hax :P
15:10:40 <andythenorth> I can’t see any code to enforce distance _between_ clusters
15:11:20 <andythenorth> until specified number of clusters, there is no check for distance to related industry
15:12:15 <andythenorth> so clusters might easily co-locate
15:12:21 <andythenorth> and I suspect that actually they do this
15:31:27 <andythenorth> hmm farms map probability was set insanely high
16:01:38 <Alberth> '&&' means 'and', ie 'both must hold'
16:01:59 <Alberth> '||' means 'or', ie 'at least one must hold'
16:04:21 <andythenorth> L84, I need to add another check
16:04:41 <andythenorth> need to check “industry_count(related) * mult / div / industry_clusters == 0"
16:04:57 <andythenorth> and also enforce a min. distance to another industry
16:05:12 <andythenorth> would be easier done by splitting switches tbh, more readable
16:05:21 <andythenorth> also if there were fewer \ chars :P
16:09:28 <andythenorth> Alberth: got a game to hand? :P
16:09:31 <andythenorth> nvm if you’re busy
16:10:11 <Alberth> I have 3 conversations going currently :)
16:11:11 <andythenorth> I need a second opinion on last FIRS commit
16:11:36 <andythenorth> basically means running map gen a few times and giving opinion on primary industry distribution
16:17:26 *** Pensacola has joined #openttd
16:24:08 *** sla_ro|master has joined #openttd
16:26:33 <Alberth> some industries are missing in 1950, that is intended I assume?
16:32:15 <andythenorth> you saw the hg branch is ‘snakebite’ btw, not default? o_O
16:32:40 <Alberth> I assumed that was intended :)
16:32:57 <andythenorth> ah, the default map is interesting
16:33:06 <andythenorth> FIRS ‘consumes’ quite a lot of industries for towns
16:33:21 <andythenorth> so relatively fewer primary + secondary
16:33:25 *** HerzogDeXtEr1 has joined #openttd
16:33:47 <andythenorth> we talked once about letting the industry grf specify the number of industries to generate, over-riding TTD
16:34:18 <andythenorth> the map in your screenie loses about 23 industries to towns
16:35:53 * andythenorth never thought about the implications of that
16:58:13 *** Celestar has joined #openttd
17:24:25 <Alberth> How can GarryG merge TAI and FIRS? I thought you used all cargoes?
17:29:53 <andythenorth> he’s either getting lucky with overlapping cargos
17:30:04 <andythenorth> or it’s broken and he hasn’t found that yet (he’s still testing)
17:30:32 <andythenorth> TAI is mostly default cargos
17:30:43 <andythenorth> and Full FIRS provides most of those
17:35:50 <V453000> yapf.rail_doubleslip_penalty = 100 what is this?
17:37:21 <Alberth> ///< penalty for passing a double slip switch <-- thats the comment
17:38:02 <andythenorth> it really bothers about double slips?
17:38:16 <V453000> and yapf.rail_pbs_cross_penalty = 300 ?
17:39:16 <andythenorth> on 2-rail DC electric trains, wiring one is a bit fiddly
17:39:29 <andythenorth> but that’s my long-ago past :P
17:39:33 * andythenorth is not train kid anymore
17:39:40 <andythenorth> now I wire up pythons :P
17:39:48 <V453000> so the pahtfinder knows if my tracks are like |X| ?
17:39:56 <V453000> like the typical X at the front of terminus?
17:40:13 <V453000> and the PBS cross is then what? :D same with PBS?
17:42:30 <andythenorth> eh, I’ll let you finish train things :P
17:43:43 <Alberth> I have no clue about train things, compared to most other people here :)
17:44:10 <andythenorth> but keeps it quiet :P
17:44:24 * andythenorth probably is most-train-thing of anyone here :P
17:45:23 <Alberth> well, you do find very nice train and industry pictures, I very much enjoy those
17:46:31 <andythenorth> so, eh, should farms cluster at all?
17:47:20 <Alberth> it makes them different from other industries, which I think is good
17:47:58 <Alberth> truck or tram feeders are nicely feasible with them
17:48:24 <andythenorth> it leaves them unevenly distributed across the map
17:48:39 <andythenorth> maybe I just need to figure out how to get more clusters
17:49:09 <Alberth> reduce mult/div factor?
17:49:45 <Alberth> then it takes longer until you reach 1
17:50:14 <Alberth> ie add something to 'div'
17:50:44 <Alberth> make it a big number as a test :)
17:52:06 <andythenorth> faster FIRS compiles => more FIRS development
18:03:51 <andythenorth> ok so 1/4 works better
18:03:55 <andythenorth> on a 512x512 map
18:04:00 <andythenorth> for limited number of tests :P
18:07:37 *** FLHerne has joined #openttd
18:43:50 * andythenorth will ignore those
18:50:36 <andythenorth> why would I even care where industries get built on a 2048x2048 map?
18:50:50 <andythenorth> I should stick em all within 128 tiles of the NE map edge :P
18:51:31 <V453000> good feature, I like it
18:51:59 <andythenorth> maybe that’s the best placement rule for _all_ map sizes?
18:52:03 <andythenorth> what could go wrong?
18:52:07 <andythenorth> all processors on NE edge
18:52:12 <andythenorth> all primaries on SW
18:52:12 <Alberth> do it in the top-left 1024, and it works for all maps :p
18:52:54 <Alberth> I do hope I don't have a huge sea at the SW :p
18:55:48 * andythenorth generates more test maps
18:56:14 <andythenorth> so a dairy that is > 128 tiles from a farm producing milk
18:56:25 <andythenorth> what’s the point? o_O
18:56:51 <andythenorth> there are 5 other dairies between this one and the nearest milk farms
18:57:03 <andythenorth> this kind of map gen bugs me
18:57:14 <andythenorth> especially when I get GS orders to provide cargo to it :)
18:57:54 <Eddi|zuHause> i don't see the problem
19:04:31 <andythenorth> it’s one of the cases where realism is a valid concern
19:04:54 <andythenorth> makes the map gen seem…daft
19:08:29 <Eddi|zuHause> the problem with "realism" as an argument is that people selectively choose the parts of "realism" that suit them and ignore the other parts
19:08:53 <Eddi|zuHause> which stops it from being "realistic" in the first place
19:19:09 <andythenorth> why would the game locate an industry you’re never going to serve?
19:21:31 <Alberth> interesting point, but how do you fix that?
19:22:41 <andythenorth> enforce max distance to source industry?
19:23:46 <andythenorth> every source-destination pair would need to do it
19:24:23 <Alberth> unless max distance is equal to size of the industry, I don't it would work
19:27:37 <Alberth> wouldn't it require a form of analysis to see if there are same industries "on the path" ?
19:28:22 <andythenorth> could be done by FIRS at compile time
19:28:29 <andythenorth> but yeah, unpleasant
19:29:36 <Alberth> it fails on the first splash of water at the wrong place :p
19:30:09 <andythenorth> allow industries to raise land :P
19:30:29 <Alberth> neat, islands in the sea :)
19:30:40 <andythenorth> I wanted to do nitrate mine that way
19:30:43 <andythenorth> but can’t raise land :P
19:33:25 * andythenorth invents complicated location rules
19:35:22 *** roidal_ has joined #openttd
19:38:10 <andythenorth> those industries bug me htough
19:51:51 <Eddi|zuHause> i think you're overengineering this
19:54:25 <Eddi|zuHause> andythenorth: whatever placement rule you can come up with, it will fail when the order of industry construction is changed
19:55:03 <Eddi|zuHause> you cannot check for industries that will be constructed later
20:06:23 *** JacobD88 has joined #openttd
20:09:23 <andythenorth> that’s why I’m not going to fix it
20:09:37 <andythenorth> it’s just going to bug me every time I see it :P
20:10:04 <andythenorth> any code I write will likely have unwanted side effects, at minimum it will increase map gen time
20:10:16 <andythenorth> at worst, pathological case, entire chains might go missing :P
20:29:30 * andythenorth must to motorways
20:50:57 <Eddi|zuHause> do motorways have to bathe now as well?
22:37:31 *** Celestar has joined #openttd
22:47:49 *** Progman_ has joined #openttd
22:53:39 *** Progman_ is now known as Progman
23:05:25 *** andythenorth has joined #openttd
23:09:50 <andythenorth> motorways complete
23:33:03 *** Biolunar has joined #openttd
23:45:46 *** andythenorth has left #openttd
continue to next day ⏵
