spherical maps

[coregames]

ORIGINAL: coregames
Has anyone seen a use of the rhombic triacontahedron in a projection?

I am going to superimpose a triangular grid and a hex grid on this unfolded shape, to see how it compares to the icosahedron. It is not as continuous in its unfolded form, but it should still project onto a sphere, perhaps with even less distortion.

[coregames]

ORIGINAL: coregames
I am going to superimpose a triangular grid and a hex grid on this unfolded shape, to see how it compares to the icosahedron. It is not as continuous in its unfolded form, but it should still project onto a sphere, perhaps with even less distortion.

rats... it seems that topologically using a hex or triangular grid on a rhombic triacontahedron is the same as the icos... 30 congruent faces, but where three come together on their sides it forms a hexagon, and where five come together on their ends they form a five pointed star. The shape is derived by combining the vertices of an icosahedron with the sides of a dodecahedron. Because the vertices of the "star" corners are still pentagonal, distortion is not lessened if using a trianglular or hex grid, so the icos projection still seems best.

[Mac_MatrixForum]

ORIGINAL: Cap Mandrake
Mac...have you seen the Homeworld 3D navigation interface?

I'm a fraid that I haven't. Similar games though. What I've been thinking is that maybe use the common FPS game controls (WSAD) combined with "mouse look" drag. I really should try this in action. The problem is that I also want to make it really easy to just type in whatever text you want.

Clearly you are talking about a much more ambitious scale.

Yeah I started small, got big and now I'm trying to scale it back to something that is possible. Although as I said, I'm developing technology to enable games and other applications to be developed eventually. I wont fix anything about the game mechanics yet.

In the case of a global Naval Sim for eg., surely you are not saying that there is insufficient data to say that point X,Y on the surface of the globe is Ocean vs Land at a 1 km resolution? I take it you mean handling the graphical burden of manipulating it in 3d on a PC. Superprecision is not key, only that the coast of Scotland look like Scotland and both sides have exactly the same map.

Oh I only meant that it makes nearly no difference to the appearance of the world whatever the height is. The Earth radius is 6300 km so you will not spot the Everest in the horizon. That's what I meant. But I think I can now see what you mean. I also want the world to be rather accurate representation and to look pleasing. But think about the 1 km resolution ... I can see it taking several gigabytes of memory. But no need to make the sampling uniform. I guess I really ought to do a new test with non-uniform sampling, editor and some new ideas for texturing. Now if I only could find the time [:(].

ORIGINAL: coregames
The issue of scale becomes important when you consider the idea of location. City locations should not be 300 miles on a side. By providing multiple map scale options -- not only a second smaller scale, but a third, and perhaps even a fourth -- a world can be mapped to the desired level of detail in each location. What I see is a continous zoom that can go all the way to the level of a street map of a city, or even floor plans of a building. This does not require mapping every square foot on the planet if you use multiple scales.

Ok, now I get it. You want a non-uniform distribution of locations on the sphere? Say, you want increased resolution around certain points-of-interest like cities. Just like Cap wants for shorelines? It is possible with a flexible subdivision scheme but I'm not quite sure what the gameplay would be. On one hand you command entire armies and on one hand you want to be able to zoom in to see the sewers? Within the same game? I can see some of that support for different settings.

One could track the exact location of units and say the terrain and other attributes come from the nearest vertex/vertices. That's really what should be done with non-uniform sampling anyway.

[coregames]

original: Mac
On one hand you command entire armies and on one hand you want to be able to zoom in to see the sewers?

Exactly... if you view the application as a system rather than merely as a stand-alone game, then its usefulness for many game ideas increases dramatically. The advantage of this would be that all games using the engine would be compatible with all the others. It could be at the center of a whole series of games and simulation tools. The initial project would be a difficult challenge, but the result (if the project could realize such a vision) would have huge implications throughout the gaming industry, not just in wargaming, but in economic/diplomatic games, roleplaying games, etc... Planettest seems to me like a step in that direction.

On a side note, I can envision single games that require a global perspective for the strategic aspect, but still zoom in to a third-person scale for certain key missions. For example, a WWII game could have the whole war fought on the macro, zooming in for certain key battles, and then zooming in even further for the mission to assassinate Hitler, or to blow up a key bridge, command post or lab. Also, in some instances, a larger strategic game could be used as a scenario generator for a roleplaying game.

[coregames]

original: Mac
Ok, now I get it. You want a non-uniform distribution of locations on the sphere? Say, you want increased resolution around certain points-of-interest like cities. Just like Cap wants for shorelines?

That is a good way to look at it. If you think of each location at the largest scale as having a terrain type, you can add an additional parameter that indicates which of a standard set of maps specific to that terrain is in use, along with which rotation to use. The mapping convention then places roads, rivers and rail based on some simple rules, whenever a "generic" space needs to have activity at the smaller scale, the computer generates the map on demand. Only cities, coast lines, possibly the vertices locations, would need to have unique maps at the smaller scale.

Some applications of the system won't need the smaller scale(s), but the possibility of using it creates interesting marketing opportunities. Imagine a grand strategic interstellar space game, complete with roleplaying tie-in. Each habitable world is mapped at the macro scale, allowing strategic-level operations on the surface and in near-orbit. Above near orbit, orbital dynamics around the star should hold sway. Outsystem, interactions could use a more euclidean grid.

As the interstellar game unfolds, an "official" version of the game could be run by Matrix, or whatever company releases the game. Key battles during the official game could be released as complete games in their own rights, scenarios at the smaller scale.

Finally, some people may wish to use the grand strategic space game as a backdrop for a roleplaying campaign. The system, if smoothly zooming upwards, could allow the characters to affect a skirmish that could turn the tide in a battle that affects results in the macro game.

I realize this is merely a vision.

[coregames]

I dusted off my copy of GDW's Invasion: Earth, which uses a large hexagonal grid on an icosahedral projection, to take a look at the map. Then, we played our weekly turn of World in Flames, and I found myself accutely aware of the map distortions from the cylindrical projection. I am curious why someone hasn't used the icosahedral projection already (after GDW and Bucky Fuller anyway), at least as far as I'm aware. Mac's test application seems so intuitive -- what has been the stumbling block for such a game? Is it the pentagonal vertices? The preference for non-discrete location (using great circles and pathfinding exclusively)? Possibly laziness and lack of creativity in the game industry? Any thoughts on reasons for the lack of a modern icos projection game would be much appreciated.

[Cap Mandrake]

But think about the 1 km resolution ... I can see it taking several gigabytes of memory

Not necessarily..a 1 km surface grid on the surface of the globe is about 5 x 10**8 data points..at it's most simplistic (ie 1=land, 0=water) that is 5 x 10**8 bits. If one allocated a full byte to each data point (to flesh out depth and altitude) that would be about half a gig. As you say that would be a burden on current machines.

As you suggested, the interior of vast oceans or land masses wouldnt need data points...as long as there was some systematic way to judge where on the globe the neighboring cell mapped to.

[rhondabrwn]

ORIGINAL: Cap Mandrake

But think about the 1 km resolution ... I can see it taking several gigabytes of memory

Not necessarily..a 1 km surface grid on the surface of the globe is about 5 x 10**8 data points..at it's most simplistic (ie 1=land, 0=water) that is 5 x 10**8 bits. If one allocated a full byte to each data point (to flesh out depth and altitude) that would be about half a gig. As you say that would be a burden on current machines.

As you suggested, the interior of vast oceans or land masses wouldnt need data points...as long as there was some systematic way to judge where on the globe the neighboring cell mapped to.

I think that doing the entire globe at 1 km is probably overkill. For a global war game I would think that the old Europa scale of 16 miles to the "hex" (or whatever is used) would be quite within the realm of possibility.

As far as using the global projection for localized tactical battles at a 1 K grid, we wouldn't be using the entire globe for the game, only a portion of it. So... I would see a completely mapped globe established as a standard resource using a variety of scales. Individual game designers would then pull from this multi-gigabyte "map" to suit their own purposes.

As a money making venture, the "Globe" would be a raw building block like various grahic engines currently in use for a variety of FPS games and so forth.

Just a few thoughts on this beautiful almost spring evening!

[coregames]

ORIGINAL: rhondabrwn
I think that doing the entire globe at 1 km is probably overkill. For a global war game I would think that the old Europa scale of 16 miles to the "hex" (or whatever is used) would be quite within the realm of possibility.

If the "resolution" of the engine is set to a higher number, more worlds could be managed at the same time without needing special hardware. Perhaps Mac's scale from Planettest is too wide however. I like 80 km per space, which gives 100 hexes or triangular vertices along each face of the icosahedron. The turns at that scale could be weekly. The circumference would be 500 spaces at that scale, and the total number of spaces would be about 100,000. If each space has a byte, as Cap Mandrake suggested, that's ten worlds for one MB rather than one world for a half a gig. As needed, smaller scale(s) could be available for certain important spaces.

As far as using the global projection for localized tactical battles at a 1 K grid, we wouldn't be using the entire globe for the game, only a portion of it. So... I would see a completely mapped globe established as a standard resource using a variety of scales. Individual game designers would then pull from this multi-gigabyte "map" to suit their own purposes.

This is where you and I disagree with Mac's preference for one unified scale Rhonda. By mapping special spaces -- such as cities, coastlines, possibly the 12 pentagonal vertices -- at a smaller scale, and including a mechanism for generating those spaces left (generic spaces) at the smaller scale, theoretically with the right network you can play globally at the smaller scale (daily turns, even 2 to 4 turns per day?).

One thing I like is your idea of more scale options. Rather than jumping right from 80 km per space all the way to 1 km per space, perhaps there is an intermediate scale break-down (10 km suggests itself). The zoom function in the game can recognise the scales in use and restrict zooming accordingly... if a smaller scale is in use, Mac is right that a continuous visual zoom from scale-to-scale would be the most appealing and convincing.

As a money making venture, the "Globe" would be a raw building block like various grahic engines currently in use for a variety of FPS games and so forth.

If the conception is forward-thinking enough, such a Globe project can indeed be useful as a development tool, but a a very user-friendly version should be available for the public as well, so gamers can edit the maps for use in roleplaying, as well as creating their own scenarios for wargaming, etc... If marketed properly, such an application could infiltrate many areas of gaming / simulations, probably in several forms. As a development tool, it could include integrated scale options in more increments than available in the public version, as well as programming features that let the application be tailored to individual uses. Of course such a "professional" version would be substantially more expensive.

[coregames]

I am not that familiar with programming requirements for an application like the one being discussed in this thread. What scale(s) do you all think would be the most useful for such an icosahedral projection, so as to cover as many game scenarios as possible?

[Mac_MatrixForum]

ORIGINAL: coregames
I am not that familiar with programming requirements for an application like the one being discussed in this thread. What scale(s) do you all think would be the most useful for such an icosahedral projection, so as to cover as many game scenarios as possible?

The effort is directly proportional to the exact features that are required. For a simple but flexible rendering scheme and the possibility to edit and save the terrain, maybe a week or two worth of coding if I were to do it. I'm sure it wouldn't have all the features thought useful or mentioned in this thread. However, polishing and adding features can be done for an eternity. For a game using this engine? Well games usually have a lot of rules, logic and user interface elements not to mention art assets, so much much longer.

I'm a bit wary of trying to make one engine or ruleset that tries to cover many types of games. It's the jack-of-all-trades argument. Generic engines have less "personality", sort of. But I'm fully committed to a generic open environment that can be extended or modified to cover any situation, game or not. I see the same environment as capable of handling my personal information management needs, movie and music collections, all database related work and strategy gaming needs. After all, what is a strategy game but an advanced database manipulation engine? It doesn't have to look like Excel. Maybe you get my point [;)].

[coregames]

ORIGINAL: Mac
But I'm fully committed to a generic open environment that can be extended or modified to cover any situation, game or not. I see the same environment as capable of handling my personal information management needs, movie and music collections, all database related work and strategy gaming needs.

So you want to use the world map as a visual data base management interface? I'm interested in how you use it for non-game stuff... do you listen to a lot of world music Mac? It certainly makes sense for keeping track of weather in an area (if the scale is small enough), possibly for travel plans.

[Mac_MatrixForum]

ORIGINAL: coregames
So you want to use the world map as a visual data base management interface? I'm interested in how you use it for non-game stuff... do you listen to a lot of world music Mac? It certainly makes sense for keeping track of weather in an area (if the scale is small enough), possibly for travel plans.

The grand idea is that there is this environment. The environment is good at managing data, showing it, general computing and as a base for building more. For example a planet is only a visualisation of a certain subset of a larger dataset which is a game that is only one possible application. There are rules that can be processed to modify the dataset (e.g. process one year of game time simulation). Visualisations can be built to show what the state of the world is like or in any way assist the player to make his decisions. The player or anybody else can extend the system to build abstractions as they like or change the rules etc.

So this is a generic environment to fulfill all my processing needs [:)]. This is mostly off-topic for this thread but may help to explain where I come from. And where I'm eventually going. If I had this environment now, I would be making applications with it. Since I don't, I have to first make it. I'll also do some smaller projects like Planettest. Nevertheless, my ultimate goal is to make the environment to make the real applications I want to be able to do.

Lets say the original motivation for it was the realization that a space strategy game I wanted to make many years ago would need like 100+ windows and making them with any modern technology absolutely sucks. And that's not the only thing wrong with modern software development. I do that for a living but I get paid for my troubles. Even then I would do things better, if I only knew how. I've looked at and tried quite a lot of things. I need something better and this is the best idea I've come up with.

[coregames]

It's not off-topic if the database application you are working on is closely associated with Planettest. I wish I was able to produce a viewing program of its sophistication, but it's also good to know you have real uses in mind for it.

[coregames]

I think I have a good implementation for this idea in mind... Subs and Satellites, a game of economics, idealogies, espionage and cold war tension ranging from 1950 through whenever one side wins (Soviet Bloc or NATO, possibly China as a third option). Using the projection, locations of individual subs and satellites can be tracked. Both subs and orbiting satellites would have accurate pathfinding with David Clark's idea of using a great circle and then back-mapping to the grid for discrete locations.

Because of nuclear tension (MAD from beginning to end), widespread open strategic warfare would never be an issue. Perhaps if one side made a lot of mistakes, a first strike by the other might be an option at some point, but that should be the exception. Wars would be fought individually at a smaller scale, each a scenario within the larger game. Some would be small conflicts that would be resolved within a single turn at the larger scale; longer wars would carry on, with reinforcements from spaces adjacent to the theatre of the conflict. When a war broke out, the theatre could be generated at the smaller scale, and at other times, all interaction could be at the global scale.

I'm interested to hear ideas about victory conditions, scope of the game, or any other notions that occur to any of you.

[coregames]

I guess this idea had no legs... ah well, I will still toy with it, just not in the Matrix forums.

[rhondabrwn]

ORIGINAL: coregames

I guess this idea had no legs... ah well, I will still toy with it, just not in the Matrix forums.

It sounds like it has potential. I'd keep posting on it as you come with more ideas. Just because no one jumped on it, doesn't mean people wouldn't be interested.

Just so much to do these days... and so many forums to follow... plus I'm a News/Blog Junkie (does that translate into "masochism"? At least I can come here and not leave depressed and angry!

[coregames]

Thanks rhonda... I'm glad someone still thinks the idea has merit. My most recent notion was to use a macro-scale of 250 miles across each space, which translates to 20 hex/triangle spaces across for each triangular face of the icos. This scale would be grand strategic of course, with the assumption that at least one smaller scale would be available -- perhaps 10 miles per hex as shown here:

[coregames]

In our current World in Flames game, I have been thinking about the contents of coastal hexes, and contemplating how it is that Malta can hold as many troops as a 100 km wide stretch of open terrain in France ( or even a 200+ km wide stretch of the Gobi desert). Perhaps the way to deal with this is to discriminate between coastal hexes, based on how much land they actually contain. The issue isn't graphic in nature, but rather, functionality and believability.

One way would be to categorize land hexes based on how many hex sides actually connect to another land mass (e.g., 0-6, or 0-3 perhaps) and use this component of the hex information to restrict stacking. I'd be interested to hear anyone else's ideas on this.

[Panzeh]

I don't even think you need hexes or spaces at all. Just use vectors like HTTR.