Map Generator Model

[Richard Bruns]

It seems that no one has been working on this model for over eight months. There is already a lot of overlap between this and the ecology model, so do you all think I should start working on this model? I think I could make things run more smoothly if I integrated this with the ecology model.

But there seems to be code already in existence for this model, and Mark mentioned earlier that there were things that had been decided on yet were not listed in this forum. Could you send me all of the information about this model that already has been decided on or coded?

[Mark_Everson]

Richard:

I think I have already mentioned the only thing that is really locked in by the map AI, which is that the map should be cylindrical. MCA, who did the coding for the previous model, disappeared a long time ago, and the links to his site with the code are no longer functional. It is possible that Paul Crocker or Henrique Duarte have copies of that.

Actually, given that you are interested in learning Java, I think going ahead learning how to program, and getting the technology model going is vastly more important to the project than anything you could possibly do in this area. IMO this model could lay dormant until well after demo 5 is out and it would not hurt the project one bit. Whereas getting the tech model basically up is one of our vital issues.

[Richard Bruns]

Okay, thanks for telling me that. I'll put all of my work on the ecology model on hold and start concentrating on technology. I can get back to this and the ecology model later.

BYW, Mark:

quote: For these reasons, most of the stuff that you described is already planned (and some of it implemented in code) but there is no mention of it here. Some of that there is no distinct mention of anywhere except in the code.

That was the stuff I wanted to know about. But if the code is gone I'll just have to start from scratch anyway, so I guess I don't need to know it.
[This message has been edited by Richard Bruns (edited September 16, 2000).]

[Mark_Everson]

I will write Henrique. I think he has it.

[Richard Bruns]

I got the code; it will eventually be integrated into the Ecology model. I was planning on doing this myself sometime in the future, but it looks as if F_Smith will end up coding all of this before I learn how to code.

[Lord God Jinnai]

Hey Mark heard anything?

[Mark_Everson]

Yeah I got the code from Henrique and forwarded it to Richard about a week ago.

[Rüdiger]

Hi!

I recently joined the Clash Project and volunteered to code the Map Generator. I chose it, because it is not so tightly integrated into the project and I am not so experienced in Java, yet.

Since I agreed with Mark to start the whole Map Generator anew, I think I will begin with coding a heightmap generator, using the model depicted at the beginning of this thread, or something similiar. After having calculated the height of each map tile one could set the "water height", resulting in every tile with lesser height beeing covered with water. I might do it so, that you preset the percentage of the surface of the world, which should be covered with water, and the model derives therefrom the necessary water height.
Determining the ecological properties of the map tiles would follow thereafter.

I do not know yet if I can do it that way, but what do you think about it?

--------------------------

Programming today is a race between software engineers striving to build bigger and better idiot-proof programs, and the universe trying to build bigger and better idiots. So far, the universe is winning.

Robert Cringley

[Mark_Everson]

Great to have you getting into this, and reviving this long-dormant topic! Your approach sounds good to me.

So when can we expect a demo no pressure...

[LDiCesare]

Your idea sounds good, but I want to point out the fact that some lands are under sea level. That's probably complicated to start with, though, but I think it could be done if you use seeds for ocean/lakes (for instance only fill with water lands below level X, then flood up to level X + 3).

[Rüdiger]

I have looked at the old code map generation code and the results it produces. Well, I don't like it very much. The continents are looking too geometric, not quite natural. The mountains are not nicely spread out and the many tiny islands produced aren't that appealing either.
Maybe we should use some more heuristic algorithm, one not based on real geological evolution, but one just optimised for good looking results? I haven't made up my mind yet, but I think the model as it is now may not quite be what we need.
Please comment what you think of that. It is a completely different approach and may not be what you like.

LDiCesare:
Yes, good point, hadn't thought about that. But maybe it wouldn't hurt too much leaving it for a later version.

[Mark_Everson]

Hi Rüdiger:

Well, the person doing the work gets the most say in how things get developed. So why don't we think about this for a bit, and then we can decide what the best direction to go in is. I'm going to try and get into contact with Paul Crocker, and see if I can interest him in contributing to this discussion.

On the continents looking too geometric, I raised that issue before with MCA, but he assured me that that was due to a bug (or perhaps lack of refinement in the code). I think another problem may be with initialization. By starting everything as a big continent, fracturing it, and only letting it evolve a bit from there, you will tend to get compact things like are shown in MCA's examples. It may be much better done with a cylindrical map with a few large starting continents.

So I advise giving this a few days to talk over before we make any decisions. But you could certainly investigate other possibilities during that time. You could also try some simple modifications to MCA's code along the lines of what I suggested, or let it evolve longer and see if it doesn't look better.

-Mark

[Lord God Jinnai]

My concern is that whatever we use for generating the maps should work resoanbly well whether the player wants a closed map, cylindral map (like civ), or a globe (my proposal).

[Mark_Everson]

Rüdiger:

I heard from Paul on the subject I posted above. He's really busy so he thought it might take a bit for him to get back to the forum. But he said that mca's code output has many limitations that are simply the result of not enough iterations and other factors. So pending Paul's re-emergence I'll again suggest that playing with the old demo code with more iterations or different initial landmass distributions seems valuable to me. BTW, It looks from the results that the map may not be cylindrical either. That would be another thing to look into changing, since plates might tend to spread out more given the extra freedom.

What do you think?

Hi LGJ:

While I agree it would be valuable to have a map generation routine that can make a sphere, I don't think we can call it essential. Till we've got all the models doing Something reasonable, I don't think we can make big detours in our algorithms etc. based on options it might be nice to have. Hopefully you'll get your wish, but I don't think we should make the map generation stuff much more difficult (and so take potentially a lot longer) just to allow for spherical maps.

[Rüdiger]

Ok, ok, I'll do what you say Mark!

I did not have much time the last days, so I haven't looked into the whole thing again. I'll rewrite the old code I have and try to make it better.
If cylindrical map means going off the map to the left is entering it on the right, then the code already handles cylindrical maps.
[This message has been edited by Rüdiger (edited February 28, 2001).]