The controls would be something like:
- left-click and drag to rotate up/down/left/right
- right-click and drag to zoom in/out

Some cities are in orbit. They are outlined in yellow.

During zooming, elements would be representated by coloured dots (cities) or squares (units).

Big cities would have more dots clustered together.

After zooming, the map could regenerate cities and units to look more real.

I would take Real-time zooming and 256 colours over a flat map and 16-bit.

The benefit is that you don't need a second panel showing the world view... I found it wasn't too useful anyways, as the scale was funny.

It certainly looks good to me. Does it mean a farewell to the traditional square map?

In ToT, there was small panel showing the globe spinning slowly at upper right side and if I click the globe it became the flat map again. But this was not a true representation of the Globe anyway(You are right)

So how you going to define the basic land unit? Certainly not squares I reckon since their collection will be another hugh square.
Dots? how?(I'm not sure) or no more basic land unit at all?

Anyway good job slingshot. keep it up!

However, I'm one of these people who wants Civ3 to run on old systems so I don't have to upgrade (I've got a PII 233 so i should be right.) And even if I could run it, I don't want it to slow down come mid-to-late-game. I'd prefer to see it run smoothly and with little waiting, than get a 3D map.

Besides, if you start having a 3D map, the customisers are going to have a fit unless they get a top class editor.

Apolyton should be a medal for good feedback, and it should be called the "Spirit of Those Guys Downunder" Award.

Depending on the resolution, you should be able to fit polygons on the surface of the globe (pentagons, I think).

Maybe someone can help me here, but I believe the biggest processing obstacles to real-time zooming are:

- Just how "real-time" you want it (eg. smooth, with lots of regenerations versus choppy)

- The number of colours

- The resolution of surface-bound objects

One cool artistic trend is that Art-Deco look. Here, units would be mainly black and sepia (a brownish-yellow). A movie like GATACA, or a game like Fallout are examples of the style. Various highlights could be added that would use more primary colours. I think that this art syle could help optimize zooming, while keeping the art looking good.

Another option is simply to have objects represented by coloured blocks. All magnifications will look this way except for the Standard View. The magnification of the Standard view would be similar to what a Civer is used to.

The best part about a map on a sphere is how the relief (changes in elevation) would look.

Imagine your unit on the top of a mountain, with its field of view larger than if it were in a valley. And imagine that field of view being gradually darker with distance. This would really give Civ more of a 1st person, RPG feel.

I don't see why a lot of these effects could not be done on a Pentium II system!

Unexplored parts of the globe could be outlined with dotted polygons.

There's no doubt that 3D graphics would be cool. Again, after reading what you just wrote, it sounds even better. And if they can do it and keep the game fluid and fast without it being on a super-system, then I'm all for it. No doubt. I'm just not sure that they can.

Although the release will be far away yet, I'm thinking that perhaps they may have gone too far into their design to choose 3D now anyway. Unless they already have

Maybe we could call it something else, like Super Civ ideas.

You never know. Firaxis probably won't put it in Civ III, and Activision definitely won't put it in CTP II, but maybe another company will create a civ-like game that will.

I was going to present a long impromptu geometry lesson, but hopefully you'll just take my word for it that there's no _simple_ way to make anything like a sphere out of polygons. Yes it can be done, but at the expense of being very hard to program and not very fun to play. If you really want proof of this, just say so and I'll do my best to show why.

It seems to me - and if anyone has a better idea I'd love to hear it - that we can't have anything like a spherical world based on any reasonable arrangement of polygons. The only way to have a spherical world then - and in this case it could be perfectly spherical - would be to do away with the polygons altogether. (Based on something someone who's name I forget said, I think they may have done this in the game Populous. Sorry haven't seen it).

Spherical world without polygons:

pros:
-extremely realistic
-looks really cool

cons:
-hard to program, entire game concept re-worked
-probably requires really good computer to do it justice.

Can anyone else think of more pros and cons for the polygonless spherical world? Can we really give up on a polygon based spherical world, or is it just me? Hope this adds to the discussion...

Dienstag:

We can crunch the math if you want.

But think of a soccer ball. It has a surface made up entirely out of pentagons.

Take a sphere and look at it from the North pole. That sphere, of course, is now a circle. The circle can be aproximated by any type of polygon. We now have lines of longitude for a grid.

Now if we look at the sphere from the equator, it is possible to draw lines of latitude in the same way.

Of course, a perspective must be generated, but this could be simplified by allowing only one light source directly behind the line of sight. This way, the center of the sphere is highlighted and the edges are increasingly dark.

Does anyone have the system requirements for these games:

Populous - The Beginning
Homeworld

I don't know about you guys, but every "regular" soccer ball (I've seen a bunch of weird ones recently) is a combination of pentagons AND hexagons. Check me, but I'm pretty sure I'm right. This would make things harder, but even if they were all pentagons you'd only have what 20, 30 of them tops? Playing a game with only 20-30 spaces wouldn't work, and you can only cut up pentagons into triangles so that's not much help either. See what I mean?

As to Slingshot's idea with the lines of latitude and longitude: I think I see what you mean, but if I do it means having lots of different sized polygons. Am I right? Wouldn't the spaces get smaller toward the poles, and wouldn't this make things like movement unrealistic? ...just trying to figure this out.

Also, I don't think it will be much of a problem rendering the 3-D. The bigger problem will be having something to render.

Thanks for the info on Populous. That sounds like a neat way Civ III could go if Firaxis is willing to totally start fresh.

Slingshot-
I'm not entirely sure what math it was you were offering to crunch. Whatever it is, if you've got nothing better to do, go for it!

Populous:
P133 (recommended: 200 Mhz or higher)
100 MB HD space free (recommended: 110+)
16 megs RAM (recommended: 32 megs)
4x CD-ROM or better (recommended: 8x or faster)
Mouse, keyboard
Supports D3D on a variety of 3D cards, which they recommend using.

Maybe civ3 should allow players to build satlelite launch center in civ3... Such center would enable players to launch spy satelites, weather satlelites, etc. And if they introduce natural disasters in Civ3, satelites could monitor the earth and predict (some)disasters in advance, thus reducing the amount of damage the disasters. What do u think?

i loved your little globe graphic! if they ever make a civ4 globes are a must...i think that this would be a much more realistic system, all i wonder if how does it look when you zoom in? would it be flat? i think it needs to be flat when you zoom in if that is possible

Forget polygons! All we really need to do is mark the globe with points. At each point you would put a terrain icon. The land-based icons could be made with fuzzy edges so that they blend together. The ocean icons would be solid all the way to the edge. Each icon could have an octagon shape.

As you said, each point on your map has three coordinates (essentially altitude, lattitude, and longitude). Let's say a unit is at one of the points, and that the unit can move but only onto an adjacent point. This is the heart of the problem: If the points are arranged so that there is always a point to the north, south, east, and west, then the points get closer to each other at the poles, so that it takes the same amount of movement points to get around earth no matter what lattitude you're at. Then we have a cylindrical world that just "looks" like a sphere. Whats worse, this is what happens if the points are arranged in _any_ reasonable pattern that would make movement easy.

If, on the other hand, you arrange the points so that they are all about the same distance from each adjacent point, then you get a really messed up map (ever look at the little bumps on a basketball?), such that just moving a unit becomes obscenely complicated.

If we really want a sperical world with logical movement, I suggest the following:
1. forget polygons
2. forget points (as a basis for movement)
3. use the unit's actual coordinates

This may seem really weird, but I don't know any better. This would answer the movement question; just give the unit a direction and there it goes. Terrain could be distributed by points (assuming points evenly spaced, thereby making a neat map), but the affect of these "terrain points" on movement would be subtle. A unit is never really on a space, but for the purposes of movement we can consider it in the terrain of the nearest "terrain point." The distance the unit can travel in any direction is based on the MP's of the unit, and the coordinates and movement penalty factors for a few Terrain Points. The math would be somewhat difficult (just spherical trigonomtery, nothing the computer couldn't handle easily), so the player would never be certain how far they could get. This is just an extension of the current Civ II system of randomly being (un)able to enter the forest on your last MP, except it's less random so after a while the players will have a pretty good idea.

This idea needs a lot more developement before anything productive will come out it. I hope this idea isn't altogether dumb and that I've described it somewhat well. I'm not convinced of either at this point. I think its the best bet yet for a spherical world, but I'd really like your input. Also, does anyone know _exactly_ what Populous did, and how? That would be nice to know.

Thanks Gord Mcleod for that last post. It gives me the best idea yet for how Populous works, and it says pretty much what I tried to say, only much better.

Assuming the X-com thing is some version of the others, this leaves us with two basic methods of making a map:

1. Grid-based (cannot be spherical)
2. Coordinate-based (I'm guessing can be whatever shape you want - in this case: spherical)

So what would it take to make Civ III on a coordinate-based map? The movement question seems to have been answered. I suggested a means of creating landscape above, but that's open to debate. What happens to the city radius? What happens to special resource distribution? What about roads and railroads? What about all the other things that currently depend on a grid-based system but I can't remember them right now?

I realise that you are trying to make large changes for the better, but if you wanted a safer option, I think you'd be better to stick with a hex based system.

Sorry if I've dampened any spirits.

You are a top-quality Apolytoner!

Thanks for the constructive criticism.

I found an interesting site (made by hippies, I think) that talks about how to make geodesic domes (and two domes make a sphere!)

It can be reached at www.desertdomes.com

Note how six triangles make for a cool hexagon!

Check out their links for more information on the math.

I have tried to start a Civ 4 thread, and it didn't go anywhere.

It would be great if somebody could start one on this buletin board - or at least tell me why they won't do it.

Now would be a great time to say "Believe in Firaxis. They're creative and resourceful, and they probably end up surprising us!" But given the distinct silence, I don't feel so generous.

i think that the only reason people do not want to discuss civ4 is because civ3 is a year away so here is a civ4 time table

April 2001: Civ3 released
December 2001: Civ3 expansion pact released
September 2002: SMAC2 announced
March 2004: SMAC2 released
October 2004: SMACX2 released
December 2004: Civ4 announced
Febuary 2006: Civ4 cancled because of the demise of TBS games
July 2006: CivOnline announced
January 2007: CivOnline open beta test begins
September 2007: CivOnline released

ok of course my time table isn't correct...but i doubt that civ4 is going to be hitting store shelves any time soon...and when civ4 does finally go into a prototype phase who know what form it will take...

When you try to make a sphere out of regular polygons, it doesn't work. It seems to me that we were thinking backwards.

What we should be doing is to project a grid of regular polygons on a spherical surface. That should work.

Imagine a cube of clear jelly measuring 10 units on each edge. Inside the clear cube is a purple jelly sphere 10 units in diameter. Impose a 10 x 10 square grid on each face of the cube. Make cut marks 1 unit deep along the grid lines. Remove the clear jelly. Now the cut marks left on the purple jelly sphere is the projection of a square grid on a spherical surface. (Hm, I feel that something is missing here, what is it?)

It is just like etching out longitude and latitude lines on a sphere.

I don't think a hex grid could be projected in such a way to completely cover the surface of a sphere. Please correct me if I am wrong.

Geodetic domes are made of subtly varied triangles and cannot do this.

OK, here's how it could work.

You start with an icocehedron. No, its not what happens to your head after drinking too much. Its a 21 sided prismatic solid in the general shape of a sphere, formed entirely of equilateral triangles. (This is the shape of those big funny dice in D&D)

You map onto each triangle as many hex-tiles as you want. Because hex tiles form a perfect equilateral triangle, excepting that the three "lines" aren't perfectly straight.

These triangles would be mini-maps in themselves. To get a good size globe you would have about 240 hexes along one side of the triangle. This would give you about 10000 tiles, like a 80 x125 grid.

You would have 21 selectable views, which you could scroll between. Each would be straight "down" onto one "triangle, with a slightly oblique but very legible view of the six adjacent triangles. You could easily rotate your view in any of 3 directions, and quickly scroll around to the opposite side or the globe.

At the interface between two triangles, you would have only prependicular movement from your "own" triangle across the interface to the corresponding hex on the other triangle. Some graphical work here, but the geometry works.

Not as spherical as anyone would like, but lets you have correct movement, and send an ICBM over the ice cap if you want.

BTW, I don't believe there is such thing as a computer map without "tiles." Call them points or co-ordinates, they act as tiles. They are all mathematically defined relational space. Some programs just conceal them better than others. To try to emulate a gridless world at all you would need a ten by ten sub-coordinate system for each current civ "tile" - 1,000,000 points for your computer to track. For what?

Anyone with a picture to help me visualize the idea?

Movement points could be expressed in terms of miles or kilometers along the surface.

City radii would actually be just that, radii of perfect circles.

Resources would be continuous, rather than discrete. For example, a city could have 28239.9 hectares of forest, 1000293.6 hectares of grassland, and so on. Production of food and units and improvements would rely on a function translating hectares of these resources into grain storage and construction capacity. This function could actually be estimated by running regressions of population and economic growth on various climactic and geographical conditions of world regions. This would ensure realism. It would be very complex, to be sure, but given that turn-based strategy isn't very demanding on computer systems anyway, surely we can afford this kind of detail and complexity. We're not talking about a 3-D accelerated action game like Descent 3.

------------------
Curumbor Elendil
jason.sorens@yale.edu
ICQ 56126989

If we use a sphere, all terrain boundaries would be defined by coordinates. You would need at least 3 coords to define a region on the surface of your sphere, which the computer could then calculate the area of. Good thing it could, because I couldn't!

Units and cities locations would have to be defined by a point (a centre-point if you will) on the surface of the sphere.

You still must define an minimum increment for your units in order for the logic to work. (e.g. for the computer to know that two things are in the same location)

Movement would have to be iterative for a TBS game, that is, you click on a location and it tells you how much of your movement point allowance it would take before you commit to the move. I say this because having missiles fall short or planes run out of fuel takes the fun out of the game. For that matter, I want to know if a ground unit can get to a city in the current turn or not to plan my strategy.

How do you define the relationship between multiple points on the surface of a sphere? Latititude and longitude are comprehensible to us, but are awkward models for calculation the distance between two points, because the distance between two lines of longitude vary as your latitude changes. Calculations can be done on distance, but you still end up with as many points (possible locations) circling the north pole one degree south of it as you do along the equator. So you have 360 points about a mile apart located 70 miles south of the north pole, and 360 points 70 miles apart along the equator.