Topic: Trees and Terrains

Because the issue of trees and terrains are coming up from time to time i think it is time for extra thread to this.

Just a few words about the current implementation:

• There are 3 values related to each terrain which are called "terrain affnitys" (temperature, humidity, fertility)
• Therea are 4 values related to each tree (preferred_temperature, preferred_humidity, preferred_fertility, pickiness)

Depending on this values a calculation is made if a tree grows very good or a tree grows not as good on a specific terrain. And this values are needed form a forester to plant a tree which is one of the best for a specific terrain.

The problems of current implementation, from my point of view, are described in this bug where the main problem is related to the mapmaker:

1.a It is not clear, which terrain affect treegrowing in which way. He may place f.e. mushroom trees (from blackland) on desert and while testing the map all the trees will probably dy before they get old

1.b. He may mix terrains from different worlds in one place nearby and treegrowing isn't as suspected.

Palm trees are another issue, because with the current implementation they are planted by foresters on meadow, which looks annoying (to some people ).

Some ideas to solve the current problems:

• king_of_nowhere proposed an additional value to the current 4 values
• Tibor suggested a complete new implementation of "biome" types

My personal (current) favorit implementation would be a "mapmakers decision" thing. If a mapmaker decide to plant whatever tree on whatever terrain, this is the rule. Automatic seeding of new trees depends then on the trees which are already in this area, independent form the terrain type (except "dry" and "dead" terrains). A problem with this idea are foresters who wants to plant trees where formerly no trees are growing (which sort of tree should they plant there? -> maybe one of the trees which are already found in this map?).

If you thinking about another solution please keep in mind:

backwards compatibility: We'd like to be able to still load maps from b18 and earlier

Top

Quote

I think that current model (humidity, fertility, temperature, pickness) is not the perfect one, but it is good enough to make it better. How? There are few possibilities:

• Tweak values of humidity, fertility, temperature, pickness (lots of values)
• Change a bit equations (now they are basicly expontents with linear argument)
• any other method you can think

I was one of the designer of current model and I will vote for upgrading current model instead of complete change.

First values (fertility, ...) were given "from hand" (no specific simulation were created before putting to widelands), and they were wrong. Also they changed behaviour of trees everywhere. Some maps were very hard to play (build18 and before). So SirVer proposed and created simple program that tweaked the values to current state. This program changed the values to current state: trees are growing as they were growing before, plus palm trees like ice

There is a possibility to change the values again using the same program, but different initial data. That will (hopefully) change the palm trees on ice. But it will not solve the problem of mix terrain types. Why? Because I know that most of people don't like (or even understand) equations which will make or make not some trees growing on different types of ground. Also mixing is strange too.

So I was testing (with help!) current state of trees-soil dependances, and I will try to make equations that will help people to understand how everything is working, but it is not enough.

If I will solve last problems (I will give some equations which will tell how specific tree will grow on specific terrain type), I can make simple www-calculator. How can it work? You pick some soils you want to mix (or only one which you want to use) and the site will tell you which tree types will grow the best. I know that this is not the best solution, but it should be enough for map makers for now. Also this will cause some "more work to do" and some (lazy) people will not like that

The idea of biomes or tree resources are good for developers, but for players it is not enough. Why? Because players will not know what resources are where. That is the main problem for Widelands itself.

einstein13
calculations & maps packages: http://wuatek.no-ip.org/~rak/widelands/
backup website files: http://kartezjusz.ddns.net/upload/widelands/

Top

Quote

einstein13 wrote:

The idea of biomes or tree resources are good for developers, but for players it is not enough. Why? Because players will not know what resources are where. That is the main problem for Widelands itself.

With "resources" you meant "good place for planting trees" ? There is a Geologist who could may do that.

Top

Quote

I imagine intersections would use the average of the six triangles around them. Don't want trees on the edge of lava? No problem; set them to -1000 fertility (assuming fertility is normally 0-100).

Top

Quote

I have prepared a table about my proposal on biomes. Basically it is a step back but will some customization. I created 9 biomes (see columns headers) and grouped trees to 5 groups (rows). The % is probability of growth of such group of trees in this biome. 0% means never. 100% means the biome is ideal for this group of trees.



The model is quite trivial and makes sure there would be no surprises like palms in arctic biome.

Top

Quote