Chill's patchpack v14_7

[Eddi]

please provide a savegame and/or steps to deterministically reproduce this output.

[welshdragon]

using hca7f5804 speed signals seem to prevent a PBS reservation through them, thus meaning if i have:

Code: Select all

===PBS===SpeedSignal===Block Signal

my trains are stuck at red and I have to Pass the Signal At Danger.

Screenshot attached, save here

[ChillCore]

I made a map 32768 x 64 and flew an aircraft along it.
The game printed the following in my Linux console, about 60 times.

dbg: [misc] Drawing string using newlines with DrawString instead of DrawStringMultiLine. Please notify the developers of this: []

v14, Linux x86_64

Thank you for reporting.
I will have a looksie in a bit. (most likely tomorrow)

As eddi mentions a savegame would be very helpful as on my 64bit system I have already fixed a few things in regards of 64bit binaries ... the savegame would allow me to make a debug build of my current version and simply check if the notificatons are still there (read: instead of doing the same with my older version)

However, the "bug" seems to be something I simply forgot to do while bumping to trunk some time ago and most likely still needs to be fixed.

using hca7f5804 speed signals seem to prevent a PBS reservation through them, thus meaning if i have:
Code:

===PBS===SpeedSignal===Block Signal

my trains are stuck at red and I have to Pass the Signal At Danger.

Thank you for the bugreport too.
I was already aware of path reservation issues with speedsignals but it was shown to me with a much more complex setup.
Hopefully your setup will allow me to fix the issue more easily.

I see you found the bugtracker too on my repo page. Thank you for reporting there too so I do not forget about it.



I will soon-ish add a wiki page to my repo too (adding the content of the first posts and a bit more) and update the bugtracker with all known issues too.

[BruceLei]

First of all thanks to ChillCore for this amazing patch pack.

I've a question about the towngrowth_needs_goods patch. Basically i try to set up a configuration/set of patches that allows me to play for decades without getting immensley rich. I changed the base costs for a lot of things and with property maintenance it worked excellent so far, the only problem i have is very fast growing cities.
With my settings i have to build passenger busses first an then wait a few years before i can do anything else. Problem is that my main city (biggest served, 1-2 stations) population just explodes, although once the towngrowth patch kicks in (set to very hard or even above, town growth in advanced settings is on slow) growth should be at 1-2%.
Nevertheless, sometimes i see '85% growth last month' which should be impossible, cuz i am only at 5-10% in each of the pass/mail/goods categories and it would grow even without a mail or goods service.

So my city is at 10,000 after just two or three years and my budget is still way to small for trains.. Everything else works great and i'm looking forward to my truely epic ottd game once i/we solved this.

I'm using the binary for Win32 (running XP) from dec. 11th, and the modified_basecosts_v2.grf

Cheers

Bruce

[ChillCore]

Thank you for the kind words BruceLei.

Your town is not growing too fast; in fact 85% is still 15% slower than normal.
Basically the growth rate is calculated normally (still dependant on slow/fast settings) and then multiplied by the pecentage of requested/delivered goods.

Normal town growth (without boosters; eg. excellent station service or funding buildigs) is 100% and you reach maximum town growth speed when you have 320%.
The difficulty is in the number of demanded 'goods' only at the moment but if you want I could reduce the growth also on the two hardest setting even further.
?devided by two and four perhaps?

Also as soon as you service 5 or more stations regulary your city will grow fast; Do you want me to make it ten stations on the 'drive me nuts' setting? Note that this will not change much because once you have those ten stations you will see growth like you have nowm; hmm ... on second thought ... ?I could perhaps increase that number according to population numbers too?

You are aware that you can make the town growth patch kick in as soon as the town reaches 500 inhabitants in the advanced settings? (seperate settings for towns and cities)


Feel free to ask more questions and comment on the above but as I am not at home right now (fixing someones computer) I might not reply before tomorrow.

[BruceLei]

Thanks Chill, you were able to clear up a few things.

Also as soon as you service 5 or more stations regulary your city will grow fast; Do you want me to make it ten stations on the 'drive me nuts' setting? Note that this will not change much because once you have those ten stations you will see growth like you have nowm; hmm ... on second thought ... ?I could perhaps increase that number according to population numbers too?

I think linking the number of stations to the amount of population sounds great.


After reading your post, i think the feature i would like to see is a factor, that reduces all towngrowth by a factor (1/x where x is ranging from 1 to lets say 20) that is multiplied with the final towngrowth factor. So that ultimately you could slow down the overall towngrowth 20 times.

Since i want to set up a game that can be played for centuries, while profits are much smaller and costs higher so i can have a very long lasting financial challenge, the game span changes drastically and cities just grow too fast.
For example, i would like to see cities that are provided with everything they need only grow by 100 people a year.
My train depot construction costs are multiplied by 64 and station tile costs by 32, so i have to wait more than 10 years until i can build my first trains and by then the city is just to big to find a good spot for a station.

I hope this is a feature easy and worthy to implement. I understand you put a lot of work in this without getting paid at all, so if adding this factor to your pack is out of question or would take more than a few weeks i'd also appreciate any hint on how and where to change it in the code. My programming skills are a bit rusty but i would be able to figure out a way eventually.

Good luck with your friends computer!

[Eddi]

scaling the required stations by town population sounds like a great idea. something like 1 station for every 1000 population. that makes small towns grow faster which don't warrant 5 stations, and large cities give an incentive to increase coverage of your (cargodist) network, besides catching more passengers.

[ChillCore]

For example, i would like to see cities that are provided with everything they need only grow by 100 people a year.

I do not think that I can control it in that way ... as population depends on the houses that are built.
What I can try to do is adjust the speed setting for towns to grow (slow <-> fast) or add a extremely slow setting.

I hope this is a feature easy and worthy to implement.

Sure it is. I disable town growth completely from time to time as I am a slow player. Read: I like to tweak and evaluate (often in Fast Forward mode) sections of my network and by the time I am done I am out of room elsewhere ...
which then requires to make room and upset the towns.

i'd also appreciate any hint on how and where to change it in the code. My programming skills are a bit rusty but i would be able to figure out a way eventually.

town.h line 260 and following has the town growth percentage that is displayed in the gui.
town_cmd.cpp line 2919 until end of the function has the actual code that calculates demand/delivered goods percentage.
town_gui.cpp line 384 and following has the code for displaying numbers in the gui.

All of the above is in regards of the town growth patch and most is documented.
The setting for influencing the speed of growth I mentioned above is located somewhere else. (I'd have to search for it myself)
Feel free to tweak and mess up the code as you wish. If you post a patch I will look at it but might still make adjustments before implementing (no offence; I have not yet seen code that you wrote )


Here is what I have planned for v14_2 ... might take a few days to do all of it but I hope to have it done by the weekend.

Code: Select all

- Update german translation (pETe)
- Chunnel entries configurable from shore (2-20) + performance warning + serverside only in multiplayer (romazoon)
- Improve speedsignals speedlimit display (hthhs, suggested by welshdragon)
- Move changelog and buglist to bin
- Tweak default settings
- Fix debug message => aircraft multistring (FLHerne)
- Make demanded stations for town growth variable dependant on town population (Brucelei)
- Tweak town growth speed setting to allow slower setting (Brucelei)
- Fix 64 bit compiler warnings
- Backport from trunk -> YAIM
  - r23443 
  - r23593
  - r23676
  - r23686

- Maybe adjust ingame version string to include version of patchpack ...
  if only I knew how exactly without hacking it in a non desired way.

if anyone notices something else that should be fixed now ... please let me know.
I will look at the path reservation with speedsignals too but I can make no promises; the above list are things that I am sure that I can fix/update.



[/offtopic]

Good luck with your friends computer!

Thank you I will need it.
She had a genuine Vista installation (HP budget PC) she was happy with but she let someone mess with it after her computer started to misbehave (after installing lots of crap); he found no better solution then to wipe the disk and install a cracked Windows 7. (Duh how stupid can you be really?).
Luckily I myself made the recovery discs (3 DVDs) shortly after she bought it two years ago so I should be able to restore the system to almost default.

Yesterday I cleaned out the dust (I could not even see the grill underneath the processor fan anymore ) and we started backing up her personal files.
Tonight I am going to put in a new HDD and will try to restore the original (and genuine) system on that.
afterwards am going to attach the "dirty" HDD to my linux machine and try to recover her program preferences and files not stored in the personal directory.

Anyone happens to know where I can find stored passwords? she has auto login enabled everywhere (email, msn, facebook and everything else) but she can not remember half of them.
Any clue would be greatly appreciated as I'd rather concentrate on restoring the system than to go and search half of the internet to find relevant info ... M$ and HP are great at talking a lot without giving crucial info as I have experienced before.
I have two locked disks, second hand HP computers that claimed that I was not the owner and went click,
I can not even connect them in linux as they are no longer recognised as being HDDs... not that I care much since I have replaced them by bigger ones; still they could be of great use to mess about with various linux Distros.
[/oftopic]


edit: spelling and mishave is not a word

[Eddi]

Anyone happens to know where I can find stored paswords? she has auto login enabled everywhere (email, msn, facebook and everything else) but she can not remember half of them.

i'd start to search in appdata->identity->{XXX-YYY-ZZZZZ} or something like that.

[ChillCore]

^^^ thank you for the hint Eddi, I will check there first.

Also ... to get back on topic ...
I think 1 station per 1000 inhabitants seems be a bit much ... I have a city of 50000 and I do not really see myself placing 50 stations in it.
Would/could you agree to me trying 1 per 2500 or something like that first? I can tweak it afterwards still.

[Eddi]

i just checked my YACD game (12MB screenshot), and it had

• 19 stations in a 20k city
• 24 stations in a 27k city
• 32 stations in a 47k city (not covering the whole area)

so 1000 might actually be a pretty good estimate.

good goals are ones that you just barely reach...

[ChillCore]

okidoki ... 1500 (for the average setting) it is then as I will make the number depend on difficulty (but not on daylength factor) also like all the rest of the demands

I agree that good goals are the ones you need to work for.

[Eddi]

maybe you can have a multiplier per station type, e.g. a bus station has less effect than a train station, a city airport, an intercontinental airport, ...

[BruceLei]

Thanks for your detailed reply, i'll look through the code sections as soon as i have time. For now, an extremly slow towngrowth setting should do the job.

I agree with 1500 inhabitants per station.

Unfortunately i have no clue what to do about Vista and those HDDs. Hope you don't have to spend too much time on it.

[Leanden]

A question about save game compatibility. Are savegames from earlier versions of your patchpack compatible with later version, for example v13 savegame usable in v14?

And if so, will this be something guarenteed to remain true in future releases of the patch pack?

[Eddi]

A question about save game compatibility. Are savegames from earlier versions of your patchpack compatible with later version, for example v13 savegame usable in v14?

yes

And if so, will this be something guarenteed to remain true in future releases of the patch pack?

no.

[schack]

I think 1 station per 1000 inhabitants seems be a bit much ... I have a city of 50000 and I do not really see myself placing 50 stations in it.
Would/could you agree to me trying 1 per 2500 or something like that first? I can tweak it afterwards still.

I don't think I've ever taken the opportunity to thank you for making this patchpack which for me includes things I can never imagine playing without again (mainly CargoDist, Automated Timetables, Larger Maps, Daylength and probably more that I don't notice I'm using).

So thank you a million times for spending so much time on making all these fantastic patches work together, and thanks a lot to all the people who made them in the first place. I spend a lot of time playing your patch pack in multiplayer with some friends, and it's really flawless and adds a whole new level of depth to the game.

But about number of stations vs. inhabitants, I think you're trying to simplify a complex problem. The main problem is that the population density increases as cities increase. You'll always see the city center being more dense, and as the city expands geographically, the higher density city centre will also increase geographically.
So I don't believe you can make a linear relation between population and required stations. The most natural thing would be to make the requirements a percentage of "total catchment area/city tiles", where city tiles is the number of tiles the city has built on.
This also due to the fact that you don't increase the number of stations in a city when the population goes up, you increase it when you need to cover a larger area. In order to facilitate population increase, you instead use different types of transport (more dense busses, a metro network or light rail).

Now, there's a good chance it's too computationally intensive to calculate your total coverage area in a city as well as total geographical area of a city, but that still doesn't mean a linear equation is the right thing to use because of the non-linear relationship between population and amount of stations. One option in that case would be to get a larger sample size of how many stations you use if you want to cover x% of a city at various population sizes and then run a non-linear regression on it to figure out the equation needed to provide the desired result.

But overall, I don't think number of stations is the best route to go since it's always been based on coverage, and a large amount of city planning in real life is spent throttling which services are available to a given part of a city or region, which plays a large part in determining whether or not it grows, and at which rate. Obviously, the only real services we can provide in OTTD are various goods and the ability to use transportation. This isn't SimCity, so we don't fiddle about with schools or medical coverage etc.

I hope that makes sense.

[ChillCore]

I don't think I've ever taken the opportunity to thank you for making this patchpack which for me includes things I can never imagine playing without again (mainly CargoDist, Automated Timetables, Larger Maps, Daylength and probably more that I don't notice I'm using).

Since that was your first post on the forums ... I guess not.
Welcome to the forums.

So thank you a million times for spending so much time on making all these fantastic patches work together, and thanks a lot to all the people who made them in the first place. I spend a lot of time playing your patch pack in multiplayer with some friends, and it's really flawless and adds a whole new level of depth to the game.

Thank you for the kind words, much appreciated.

But about number of stations vs. inhabitants, I think you're trying to simplify a complex problem.

I like complexity by simplicity.

The main problem is that the population density increases as cities increase. You'll always see the city center being more dense, and as the city expands geographically, the higher density city centre will also increase geographically.
So I don't believe you can make a linear relation between population and required stations. The most natural thing would be to make the requirements a percentage of "total catchment area/city tiles", where city tiles is the number of tiles the city has built on.
This also due to the fact that you don't increase the number of stations in a city when the population goes up, you increase it when you need to cover a larger area. In order to facilitate population increase, you instead use different types of transport (more dense busses, a metro network or light rail).

I agree that the densest population concentration is near the city center ... however the city also grows outwards so you will need to place stations anyway to cover them new houses.
¿Demanding more stations would perhaps discourage people extending the catchment area by distant joining stations far away from eachother? (don't get me wrong I do like distant joining stations for better looks but covering an entire city with just one station is IHMO silly.)

Now, there's a good chance it's too computationally intensive to calculate your total coverage area in a city as well as total geographical area of a city, but that still doesn't mean a linear equation is the right thing to use because of the non-linear relationship between population and amount of stations. One option in that case would be to get a larger sample size of how many stations you use if you want to cover x% of a city at various population sizes and then run a non-linear regression on it to figure out the equation needed to provide the desired result.

- I do not know if such a function is already in trunk but if it is not I think it would indeed be too CPU intensive to calculate the total geographical area of a city ... remember we are talking lots of cities on the map that change their size all the time ... requiring to recalculate every so much ticks over and over again.
- Coverage areas of stations are already calculated and as they do not increase in size that often that is not so much of a problem (CPU wise that is)
- City population is also already calculated and available for use without extra cost.
- City growth is already influenced by how well you serve those stations ... simply plonking down the number required and not using them will not do.

It is not that I do not like what you are saying; I just do not know (yet) how I am going to to implement this.

But overall, I don't think number of stations is the best route to go since it's always been based on coverage, and a large amount of city planning in real life is spent throttling which services are available to a given part of a city or region, which plays a large part in determining whether or not it grows, and at which rate.

Ofcourse the more of a town that you cover the more goods and people you can ship in and out.
However ... those 5 stations trunk cares about do not need to cover a big area ... they just need to be regulary visited for towns to be happy about them. I'm not sure if changing that to city-coverage percentage is a good idea as explained above.

Obviously, the only real services we can provide in OTTD are various goods and the ability to use transportation. This isn't SimCity, so we don't fiddle about with schools or medical coverage etc.

Indeed.

I hope that makes sense.

Sure it does.
If someone tells me that the cities footprints are already calculated somewhere in the code (or if doing so will not eat CPU much and not cause lag) ... I will consider making city coverage part of the calculation.

Feel free to experiment with the code and/or provide me a formula that I can use.
I mentioned this before ... stupid young me did not finish shool.



Also before I forget:
BruceLei, welcome to the forums to you too ... I just noticed that those were your first few posts too.

[schack]

I just discovered this was my first post too, which I find hard to believe somehow. I've been reading these forums for a number of years, and I'm pretty sure I've even posted once or twice, but that must've been some other account I've had way back when, attached to some old email address.

Anyways, back the point. I'd forgotten about distant joins, as that's something I shun. Regardless, I'm not sure how big a weight I'd want to attribute to it, as that's something you can abuse too if you don't play with town growth.

I agree that the simplest solution is still amount of stations per inhabitant, but I'll still maintain that a linear approach is not a good idea, because as population density increases, the number of stations needed to cover them decreases.

So say you have 500 inhabitants (medium sized town, slightly disjointed), you might need 2 stations for 80% coverage. As that then expands to 2000 inhabitants and you have a larger population density in the city centre, you could still get away with 4-5 stations. As it expands to 5000, you might need 8 and let's for argument's sake say you need about 20 to cover a city with a population of 3,0000.
That's not a linear progression. Obviously it's also just an example, and probably not a super accurate one. On top of that, it's been a while since I went to school, so I've forgotten how to do non-linear regression, but I'm pretty certain what you'd want is a second degree polynomial that crosses through (0,0) on a graph. Now what the values for a and b would be... good question.
But I think the way to find that would be to set up some test scenarios with varying cities with varying number of inhabitants, decide on a factor you should cover, put down stations and see how many that amounts to - then plot it all in to a non-linear regression thingy that I'm sure someone can figure out. That should make for a pretty solid case.

The reason I'm quite insistant on this, is that, as you say, you have to increase it from 1 station per 1000 inhabitants to 1 station per 2500 inhabitants. That's quite a big increase. And that's to accomodate for the fact that you have a city of size 50,000. But what then when you get a city of size 100,000? Then you'd need 40 stations, but while the city has doubled it's population, it won't have doubled it's geographical area, and you'd need to adjust your factor again to accomodate for that (meaning "easier" gameplay at lower population levels). With the 1 station per 2500 inhabitants, you can serve a city with a population of 5000 with just two stations, and that's not really a lot. Small cities are quite large geographically due to the lower population density.

I know it's mainly what I was arguing before, I just hope this elaborates a bit on why I think it's not necessarily a good idea to use a linear formula to determine requirements.

[Eddi]

- I do not know if such a function is already in trunk but if it is not I think it would indeed be too CPU intensive to calculate the total geographical area of a city ... remember we are talking lots of cities on the map that change their size all the time ... requiring to recalculate every so much ticks over and over again.

i think YACD somewhat keeps track of which houses are covered by stations. And i believe the subsidy rewrite reused that part of the code. i'm not sure if that was already included in trunk...