Re: world as a globe

On Wed, 2003-01-08 at 03:29, John Wheeler wrote:
> 
> 
> --- Christoph Petschnig <cpet0101-/[email protected]> wrote:
> > [snip]
> > Wouldn't it be cool to have bombers fly over the
> > poles to attack
> > targets unreachable in a cylindric world? I also
> > think the freeciv
> > poles are just to big for a realistic world.
> 
> As I see it, these are two completely seperate issues,
> which have different solutions.
> 
> I see a simple solution for "flying over the poles":
> say you are at 40,0 on a world that is 120 wide.  If
> you go north (assuming Y=0 is the north pole) from
> there, you end up at 100,0 -- in other words, exactly
> half way around the world.

This has been suggested, but it is both problematic and unrealistic.

It's problematic because you're basically implementing a wrap onto the
same line, but flipped.  The reversal of the orientation is difficult
both to display and for internal calculations.  This is possible, but
takes a lot more work than other solutions that are IMO better.

It is unrealistic because although it intends to make the pole small, it
doesn't.  Although you can go halfway around the world in one step, it
still takes you many steps to get a quarter of the way around the world.

> To make the poles more realistic in size, you might
> want to turn the map on a diagonal, and only two
> corners the poles.  (The line connecting the other two
> corners would be the equator.)  This would make moving
> off the edge somewhat more complicated, and would
> require rewriting the world generation routines, but I
> think the real nightmare would be in trying to display
> the results seamlessly.

This is unworkable; it is impossible (or rather, nobody has demonstrated
it to be possible) to fit the edges together.  You might be tempted to
set things up like:

    <- X ->
  <- X X X ->
<- X X X X X ->
<- X X X X X ->
<- X X X X X ->
  <- X 3 1 ->
    <- 2 ->

with wrapping shown by the arrows.  But what happens then if you move
south from position 1?  Do you end up at position 2?  If you move north
from position 2, do you end up at position 1 or 3?  The correct answer
should be both, but this is obviously impossible.

> > It might be necessary to replace the squares by
> > circles, but maybe
> > there are also other solutions.
> 
> Games which put an emphasis on realism generally use a
> hexagonal grid.  

The only advantage of the hexagonal grid that I can see is that the
distance to the next tile is equal in all directions.

> > My questions: does anybody else like this idea?
> > Would this mean to
> > change 90% or only 10% of the code? (I do know that
> > 10% is still
> > rather huge than small).
> > 
> > Regards Christoph
> 
> Do I like this idea?  Flying over the poles sounds
> like fun to play and not too hard to implement. 
> Making distances more realistic not only sounds like a
> real bear but also might even have the cardinal sin of
> being less fun to play.  (Bicylindrical worlds would
> be a little weird, too.)

A torus (which I guess is what you mean by bicylinder) may seem weird
when you think about the topology, but for the typical player I suspect
it is indistinguishable from a sphere.

jason