The International Simutrans Forum

 

Author Topic: forest generation  (Read 6997 times)

0 Members and 1 Guest are viewing this topic.

Offline uci

  • *
  • Posts: 120
  • Without owner
  • Languages: EN, RO
forest generation
« on: April 28, 2011, 11:34:38 AM »
In spite of the settings

Code: [Select]
# climates with no trees at all (desert and arctic at the moment)
no_tree_climates = 130

forests can be generated/added on every climate. With the same settings pak64 behaves correctly.


Offline VS

  • Senior Plumber (Devotee)
  • Devotee
  • *
  • Posts: 4855
  • Vladimír Slávik
    • VS's Simutrans site
  • Languages: CS,EN
Re: forest generation
« Reply #1 on: April 28, 2011, 03:38:21 PM »
With the same settings pak64 behaves correctly.
You mean, if you add this to pak64 simuconf, it works correctly?

Offline uci

  • *
  • Posts: 120
  • Without owner
  • Languages: EN, RO
Re: forest generation
« Reply #2 on: April 28, 2011, 04:31:49 PM »
You mean, if you add this to pak64 simuconf, it works correctly?
Yes, in fact this is the default. You think it is not pak128 related?

Offline DirrrtyDirk

  • Devotees (Inactive)
  • *
  • Posts: 1253
  • JR 700 Series Shinkansen
  • Languages: EN,DE
Re: forest generation
« Reply #3 on: April 28, 2011, 06:16:20 PM »
I cannot see why it should not work in pak128... ??? The entry is the same as in pak64 - so for some reason it works there but not here.

Offline VS

  • Senior Plumber (Devotee)
  • Devotee
  • *
  • Posts: 4855
  • Vladimír Slávik
    • VS's Simutrans site
  • Languages: CS,EN
Re: forest generation
« Reply #4 on: April 28, 2011, 08:21:55 PM »
This is actually true, for a rather weird reason... No idea why, but pak128 has explicitly set no_trees=0. Removing that option fixes this. So fixing that is simple :)

However, there is some bug or undocumented logic in program, which makes this happen... I don't feel like debugging it though, there is nothing obvious... :-/

Offline DirrrtyDirk

  • Devotees (Inactive)
  • *
  • Posts: 1253
  • JR 700 Series Shinkansen
  • Languages: EN,DE
Re: forest generation
« Reply #5 on: April 28, 2011, 08:41:43 PM »
Removing this setting does not change anything for me - it's still not working properly in pak128, while pak64 runs just fine...

Offline VS

  • Senior Plumber (Devotee)
  • Devotee
  • *
  • Posts: 4855
  • Vladimír Slávik
    • VS's Simutrans site
  • Languages: CS,EN
Re: forest generation
« Reply #6 on: April 28, 2011, 09:48:08 PM »
Well, it turns out that pak64 has no trees for desert. So one oops for me... Another program bug, it seems...

It would be hilarious if my previous finding of "nothing obvious" could mean there is nothing at all :P

edit: No code, just thinking out loud.. Could it actually be that this entry has been made obsolete by tree climate data? Because there is little point in duplicating that functionality.

edit2: Moving to program bugs, please give it back if there is no program bug :)
« Last Edit: April 28, 2011, 09:56:20 PM by VS »

Offline Dwachs

  • DevTeam, Coder/patcher
  • Administrator
  • *
  • Posts: 4531
  • Languages: EN, DE, AT
Re: forest generation
« Reply #7 on: April 29, 2011, 06:11:40 AM »
This setting is only used to distribute those single trees across the map, not for forest generation.

Offline uci

  • *
  • Posts: 120
  • Without owner
  • Languages: EN, RO
Re: forest generation
« Reply #8 on: April 29, 2011, 08:34:47 AM »
This setting is only used to distribute those single trees across the map, not for forest generation.
Well, it turns out that pak64 has no trees for desert...
These facts explain the behaviour.

Let me suggest some changes for pak128:
1. Trees should not be allowed in arctic climate (always covered by snow).
2. At least palm trees should not be allowed in desert climate but in water climate (like pak64). This creates a nice effect along coastal lines.

Edit: Since we are still talking about trees, I want to point out here a visual bug (white pixel) in cactus000 tree (image).
« Last Edit: April 29, 2011, 08:54:44 AM by uci »

Offline DirrrtyDirk

  • Devotees (Inactive)
  • *
  • Posts: 1253
  • JR 700 Series Shinkansen
  • Languages: EN,DE
Re: forest generation
« Reply #9 on: April 29, 2011, 12:59:03 PM »
This setting is only used to distribute those single trees across the map, not for forest generation.

Hmm wouldn't it make more sense the other way round? So that single a single cactus may be placed in the desert once in a while, but not dense forests of them?

Offline VS

  • Senior Plumber (Devotee)
  • Devotee
  • *
  • Posts: 4855
  • Vladimír Slávik
    • VS's Simutrans site
  • Languages: CS,EN
Re: forest generation
« Reply #10 on: April 29, 2011, 01:54:49 PM »
I think some types of desert plants actually do grow in almost-forests... they're low, though, so it doesn't look that way. And it might be a question of density for climate zones, too. Shame on me, but I don't remember how dense is vegetation outside Simutrans forests. If we analyze this seriously, we'll turn Simutrans into a botanical simulator :D

Perhaps forbidding single both trees and forests could be configured that way.

Anyway, what uci says is a good starting approximation - forbid everything except cacti in desert, forbid everything in arctic.
« Last Edit: April 29, 2011, 02:01:49 PM by VS »

Offline DirrrtyDirk

  • Devotees (Inactive)
  • *
  • Posts: 1253
  • JR 700 Series Shinkansen
  • Languages: EN,DE
Re: forest generation
« Reply #11 on: April 29, 2011, 02:01:26 PM »
Yes, but for a game, the stereotype is enough for me. I require no more realism for desert vegetation than that.  ;)

Or, as a work-around, we could change the cactus plants from trees into ground objects - that would solve the desert problem, I think...

Offline VS

  • Senior Plumber (Devotee)
  • Devotee
  • *
  • Posts: 4855
  • Vladimír Slávik
    • VS's Simutrans site
  • Languages: CS,EN
Re: forest generation
« Reply #12 on: April 29, 2011, 07:55:25 PM »
Then they would stay the same all the time, no? Is that a problem? Could we say that cacti are simply long-lived and thus don't grow during game? Questions, questions...

Fixed the cactus problem and a few other stray tree pixels...

Also... should this thread stay here or go back to pak128? edit: yes, I think.

edit2: current climate report - http://vs.simutrans.com/remote_images/trees/128-tree-climates.png
« Last Edit: April 29, 2011, 10:00:46 PM by VS »

Offline uci

  • *
  • Posts: 120
  • Without owner
  • Languages: EN, RO
Re: forest generation
« Reply #13 on: April 30, 2011, 08:09:07 AM »
Nice overview. I think arctic climate should be avoided anyway.

Edit: Changing the settings only for palms does not produce significant differences (image), but allowing some trees on water is not bad.
« Last Edit: April 30, 2011, 09:47:21 AM by uci »

Offline VS

  • Senior Plumber (Devotee)
  • Devotee
  • *
  • Posts: 4855
  • Vladimír Slávik
    • VS's Simutrans site
  • Languages: CS,EN
Re: forest generation
« Reply #14 on: April 30, 2011, 12:30:16 PM »

Offline uci

  • *
  • Posts: 120
  • Without owner
  • Languages: EN, RO
Re: forest generation
« Reply #15 on: May 02, 2011, 01:31:58 PM »
Just a remark:
There is an implicit assumption in Simutrans that waters are hot. Suppose a nordic island with tundra climate. We get palms along beaches. This is not pak128 related.

Edit: of course, only tundra and alpine climates are enabled in landscape settings.
« Last Edit: May 03, 2011, 12:13:18 PM by uci »

Offline Václav

  • Devotee
  • *
  • Posts: 3230
  • formerly VaclavMacurek
  • Languages: CZ, EN
Re: forest generation
« Reply #16 on: May 03, 2011, 09:40:37 AM »
uci, in game start dialogue you can set climates - but on some maps it can be problem.

Offline VS

  • Senior Plumber (Devotee)
  • Devotee
  • *
  • Posts: 4855
  • Vladimír Slávik
    • VS's Simutrans site
  • Languages: CS,EN
Re: forest generation
« Reply #17 on: May 23, 2011, 08:58:36 PM »
Today I made cacti groundobjs, which means they are distributed randomly at start of the game... and that's it. No settings tweaked so far, but to me the results seem reasonable.

Reactions?

(changes finished with r512)