Spherical Map

[Vince278]

I don't mind the idea of dropping the tile system entirely and replacing them with something like a proximity-based system but we should have little problem using hex-based tiles on a sphere (the game would then kinda resemble board wargames with hex maps). The world would look similar (but not exactly) to a football (FIFA, not NFL ).

[Dauphin]

Edit . Nevermind that idea.

[geeslaka]

Tell me what your idea was!

[Vince278]

Probably something about maps being wrapped around other things.

[Dauphin]

I was just thinking about all the weird and wonderful maps you could create and was trying to visualise particular maps in real space.

As such trying to play civ in/on a Klein bottle would probably not be a good idea.

[Vince278]

It would be really cool to do a map that's non-euclidean.

[Leland]

I have an idea for a pentagonal map... that is, a map whose tiles are regular pentagons that fill an arbitrarily large area.

First, start off with a dodecahedron (consisting of 12 regular pentagons) and think of it as a sphere with a circumference of 1. Now, imagine another sphere inside it, which is only slightly smaller, and that there's a circular "edge" which connects these two spheres so that if you walk over the edge you always either drop from the outer sphere to the inner one, or climb from the inner one to the outer one in a continuous manner. On a computer where we think of the surface of the sphere (i.e. the actual map we see) as flat, the spheres can be exactly the same size, so what we've got is a map consisting of two spheres.

More spheres can be added in similar manner... the "edge" now works as a teleporter that always drops you down one level, except in the final level it puts you all the way to the top (or, if you are walking in the other direction, does exactly the opposite). It might be a little bit difficult to visualize, but I think that this way it may be possible to have a uniform map of regular pentagons.

Does this work?

[geeslaka]

This is how a rectangular map could have the topological properties of a Klein botle.

 ______________________________________
|        ||        ||        ||        |
|       a||       c||       e||       g|
|        ||        ||        ||        |
|        ||        ||        ||        |
|        ||        ||        ||        |
|        ||        ||        ||        |
|h       ||b       ||d       ||f       |
|        ||        ||        ||        |
 ______________________________________

The columns of lines represent smooth frictionless hemispheres that units slide over.without stopping.
If a unit goes left from point a it moves to point b.
c<->d e<->f g<->h
It might be a good idea after all.

[geeslaka]

Leland: How do you determine when someone crosses the edge?

Everyone: Keep in mind that there needs to be a way to display a reasonable amount of the map. If you can't visualize it, it can't be displayed. This is one of the reasons why I favor fooling around with square-based maps.

[Leland]

Actually, I realized that the scheme doesn't work out quite so easily. Back to the drawing board...

The "edge" was supposed to be just an arbitrary circumference of the sphere, like equator. It would have been totally invisible to the players. As for visualizing a map like that, well, I can think of multiple ways actually. One would be to allow gaps between tiles, and when scrolling the map the gaps would close so that they're least obtrusive at the exact center of the map. Another way would be to distort the map on the edges. Tiles would still be visible, but their shapes would be messed up. Next time, I think I'll try to come up with a visualization first and write the explanation around that. Maybe even make a little demo program or something.

To see weird topologies like this in actual games would probably require that the games themselves only have a very abstract representation of a map (e.g. as an arbitrary graph) and provides an API for plugging in new topologies and visualizations. I can't imagine that happening in commercial projects because it's beyond simple modding, but maybe an open-source project somewhere will take heed...

[Trifna]

Erm.... How would it be hard to make a map with vectors? I mean exactly like in many RTS: you move on a distance X in a certain direction, simply.

Would it work well? For the AI, I guess they should have some specific rules... like having a specific goal which forms a rally point, or otherwise simply have units that are wandering in a specific area.

[geeslaka]

Trifna: Which RTS has a vector-based map? AoE, Warcraft, and Starcraft are all based on squares.
It wouldn't be hard at all. A city would be at point A and draw resources from any object within a circle of radius 3. Terrain objects would affect the movement of any unit that passed within 1 unit of measurement of them.
Leland: If you are interested in working on an open source civ game go to http://freeciv.org

[Dauphin]

Originally posted by geeslaka
It might be a good idea after all.

It would be interesting, if nothing else it would help people learn about unfamiliar topography. I just don't think its a good idea for people trying to visualise the actual 'globe' you're playing on.

[geeslaka]

I was thinking about that and I realized that if you displayed those four sections side by side and flipped two of them to avoid confusing the user it would be indistinguishable from a cylinder.

[:) Smiley]

Screw spheres, I want a 3-D map!