3d & Spherical maps discussion

[VetLegion]

forgot the source: http://www.golemlabs.com, defcon, an interesting computer game in development.

[Leland]

Well, since no one else said anything about the tile size, I suppose I can chip in my two cents worth...

First, I think 50km tiles would be fine. I finally realized that the more tiles there are, the more memory it consumes, and the more stuff has to be sent over the network each turn. We can and will strive to use the most efficient compression algorithms, of course, but the size of each package could still be in the megabytes... that's kind of steep for PBeM games, as well as fast-paced online games. Of course, if you're on-line it's probably possible to device a way to subdivide the package into smaller parts so that you could see the results of migration when still waiting for the builds... or something like that. But anyway, a map with 30km tile diamater is going to take almost three times as much space as a map with 50km tiles. I think that's somewhat undeniable.

(Just for the record: this is not an excuse to stop programming the map... the map diamater thingy mostly affects designs, since it's only about the granularity and resolution of the models.)

Also, it should be noted that

• I can't help repeating it: Britain is small! Take a look at world map and you'll get the idea how my examples would fit in. It's also possible to "cheat" by making the British isles larger in comparison than the real world, just like Civs have always done.
• The small tiles would mostly come in handy in the early game, whereas in the late game they make less difference.
• The cons of smaller tiles in addition to the memory consumption are that it could make drawing the map slower... I think that we should make the map all the way rendered, so for instance when you're looking at it zoomed back to show the whole Europe for instance there would be a lot of rendering to do... this probably affects the smoothness of how fast the map can be scrolled or something like that, and may not be much of a problem.
• It's possible and not even very difficult (I think) to smoothen the corners of the hexes so that you wouldn't get sharp corners and edges on the map. So the final map would visually look much more realistic than the above images.

About Defcon's map, I do think that it's rather too detailed for our purposes! I'm seeing nightmares about that level of complexity (you see the individual tanks and airplanes) on a global scale...

[S. Kroeze]

In this thread Leland made several maps of the British Isles.
According to his own words, in the first map the distance between the centre of individual hexes is about 50 km -I did not verify this.
I would be happy with such a map, but opinions did vary.
I still agree with Leland's arguments.

[Jon Shafer]

From what I can tell just by briefly looking over things, I agree with you and Leland Kroeze. The size of Britain in the 2nd map looks acceptable too.

But as mentioned earlier, if each hex is 40-50 km across, then the world would include 700,000 hexes. Let's say that 200,000 of those hexes are land, and populated. Running a fairly complex function on 200,000 hexes to keep track of population migration, properties, etc. would be far too much to handle for most machines. I'm sure my PIII 550 couldn't handle that. However, I'll suggest my proposition in a bit, my brain isn't in a thinking mood right now.

[Leland]

Yup, you're right that the hardware requirements will be a bit hefty. And the gameplay will probably consist of you thinking about your turn for fifteen minutes (considering that this is multiplayer), and then waiting for five or so minutes when the host computer crunches the numbers. Another thing that could be problematic is the network bandwidth... it looks like modem users are going to get shafted.

[Jon Shafer]

Just a discussion with Kroeze and Leland, trying to show that triangles cannot form perfect shapes with hexes.

[Azail]

Have you guys decided to go with a 3d map yet, or is it still open for discussion? I'd be interested in programming this game if it used a more traditional map (isometric, hex or square); though you could still use 3d just not a spherical map. The reason I would only work on this with a traditional map is that in my opinion, a spherical map would be a huge waste of computational resources in a game like this, especially with the depth of the models you plan to implement.

[Leland]

The map is definitely not going to be spherical, that much is certain. We're using hexagonal tiles on a flat, cylindrical map.

[Azail]

Alright then, not spherical (but still 3d - from the looks of Trips post above about triangles) that's good with me. I'm going to start a new thread introducing myself & what I bring to the group as well as voice my opinions about a few things I feel you have done wrong thus far.

[Nath]

Nice to see some activity here.

voice my opinions about a few things I feel you have done wrong thus far.

We've done something wrong? I didn't know we'd done anything!

[Azail]

That's exactly what I meant, Nath , at least as far as demos go.

[aeturneus]

One of the things I've fantasized about for a long time is a spherical world that is completely 3d, where you can fly out to planetary perspective or go right down to the ground... I know you aren't going to be able to do this in GGS, but has anyone else got this same dream?

I like to dream...

[RonHiler]

Originally posted by aeturneus
has anyone else got this same dream?

Yes.

https://www.rjcyberware.com

Ron