show different age for tree objects

[Frank]

show different age for tree objects

old:

[age] months old

new:

age 0 months = 0 months old
age = 1 month = 1 month old
age 2 -11 months = x months old
age 12 - 23 months = 1 year old
age > 23 months = x years old

changed translate string "Monate alt" to "months old"
add translate strings "month old", "year old" and "years old"

different_show_age_treeobjects.diff

[VS]

Good idea.



Actually... we might as well consider using gettext if we're going to dive into plurals. Czech has three forms (1, 2-4, 5+) and rules for some other languages are a lot more complicated.

[jamespetts]

I do like this idea...

[prissi]

Not sure if this is such a good idea, given the speed of many translations. Apart from the plural issues.

[MHD]

At least the idea with years instead months is better than now. The numbers are big and noone will calculate the years from it (except he feels very very bored)

[skreyola]

Yeah. Tree ages are generally understood in years (if not centuries) IRL.

[Fabio]

IMO, 0 months old is a bit lame... "just planted" would sound much better, wouldn't it?

[Isaac Eiland-Hall]

It would, unless it causes too much havoc in the code. Except that also, was it planted? Or did it just germinate naturally?

[jamespetts]

And is it growing on chalk or clay soil?

[VS]

An useful addition would be one comprehensive info window for all trees on tile, instead of a score of these... Of course that is entirely different matter.

[prissi]

I was also thinking of merging trees in a single structure to save memory (like TTD does), but then the "pure" OOP approach let me abstain from it.

However, it seems that tree age in years is a majority vote, so it will be done.

[skreyola]

I think trees under 1month old should still say 1 month... perhaps tree ages should start counting at 1... or display in this fashion: 1 month (for 0-1) ... 12 months (for 11-12), then 1 year, etc.
I don't think it should ever give the age as 0 months... Maybe 'sapling' or 'new' or 'days old' (without specifying the number).

[prissi]

I would think a string like "%i years %i months" will be added. And honestly, you will almost never ever find a zero aged tree, unless you planted it deliberately at zero age.

[Spike]

I was also thinking of merging trees in a single structure to save memory (like TTD does), but then the "pure" OOP approach let me abstain from it.

A "flyweight" design pattern might help with this. And it's OOP.

[prissi]

Ok, the new translation has been added.

[Spike]

Also: http://en.wikipedia.org/wiki/Flyweight_pattern

[prissi]

Thiis is slowing going towards offtopic but:
Trees are already only 12bit references and the position and so one is needed, as a tree in simutrans is allowed to be anywhere on a tile.
The only way out I see would be forest tiles with a predefined pattern for the trees (maximum four or so) with individual type and age but fixed position and so on. But I am not sure how to handle loading, or I rather thing this is a lot of work with not so much gain, as memory is usually not the limit nowadays.

[Dwachs]

Thiis is slowing going towards offtopic but:
Trees are already only 12bit references and the position and so one is needed, as a tree in simutrans is allowed to be anywhere on a tile.
The only way out I see would be forest tiles with a predefined pattern for the trees (maximum four or so) with individual type and age but fixed position and so on. But I am not sure how to handle loading, or I rather thing this is a lot of work with not so much gain, as memory is usually not the limit nowadays.

Also the displaying of multi-tree tiles will become a problem, if multiple trees are organized in one instance.

One possible change could be to remove the position from the ding_t class, since many derived classes only need to know their position on the map if their image is to be calculated (almost everything in dings/*). Only vehicles/grounds/maybe ways need to know their position always. This could save 5 bytes per object.

[Spike]

I just thought, since a few times it was suggested to players to turn treees off, since they make up for 30% of mein memory apparently, that there must be quite an overhead. I didn't knwo that trees had been optimized this much already.

For me, the trees are something important, particularly that they age, seed new trees, basicially that they are a micro ecological simulation in the Simutrans world.

[prissi]

One could also pass the position as step() method, then the tree could do without. And we can remove x offset too. Hmm.

[z9999+]

[off-topic]
Rate of wood on rectangle map is very bad.
If we start a new game with rectangle map like 1024x256 or 256x1024, more than 80% of land is wood. I feel this is too much.

[Dwachs]

yes, the create forest function do not scale right with map size.