A similar reduction could be done for France as well... it has 36,000 communes but that is equivalent to a singular city/town and is by no means comparable to an American county. Instead, the French canton would be a better comparison, of which there are 3883.

http://en.wikipedia.org/wiki/Administrative_divisions_of_France

Well, I took the lowest level administrative divisions for the comparison/calculation.

question on you guys...
Are you fueled by pure desire of extremly big map or you only want realistic maps to the tinies detail ?

Because if it is the first one, it could be quite "easily" achived by computer generation.

If there were some kind of software you could just put an image in and it gave you all the territories, Warlight would have a billion maps.

that would be also posible, and it even would not be that hard. but you would need map in wery big quality for that.

Anyway, generator is something else. It would just make up maps....

fizzer approves maps and categorises them before they are selectable on the game/tournament creation thing.

It would be cool to have a big huge tournament where people pick a continent and play a mini-tournament on that map, and then the winners of each continent's tournament play a big championship game where they start out on the big earth map with their continent already under their control (combine australia and antarctica,) the bonuses are tweaked so that each player starts with identical amounts of bonus (perhaps Asia would have to be split in 2 to make this feasible,) each player gets 100 bonus armies on the first turn to distribute how they please, and then they can play for the world championship.... Or, something along those lines..

Lurtz:
> that would be also posible, and it even would not be that hard. but you would need map in wery big quality for that.

I am reasonably certain that it would not be possible, and if it was it'd be hard as hell. It *is* possible to ease the development process a little using special software (which I proved by doing it and posting how-tos here), but if you create something that converts, say this map: [[1]](http://upload.wikimedia.org/wikipedia/commons/0/03/BlankMap-World6.svg) (SVG, infinite quality; rendered as ultra-high quality PNG: [[2]](http://upload.wikimedia.org/wikipedia/commons/thumb/0/03/BlankMap-World6.svg/5000px-BlankMap-World6.svg.png)) into a complete map, or a map with connections defined at least, or damn, even just the centerpoints, and then uploads this to Warlight, you'll be my hero.

I'm no scripting expert, but a script to define connections can't really be that hard...
I think the script will just need to search for territories within lets say 5 pixels of the given territory, and connect those two. Over-sea connections will have to be done manually, but that's a piece of cake.
And for centerpoints, just a script that can define the point in a territories that's located the farthest from any border.

> And for centerpoints, just a script that can define the point in a territories that's located the farthest from any border.

Yeah, seems simple enough. Enter non-convex territories. Enter multipart territories. Enter territories with holes. And don't forget we're not dealing with polygons, but with paths composed of bezier curves (and possibly a bunch of other types of curves, too), which can by the way intersect with themselves, or with other ones. Suddenly it's not as pretty; in best case, there will be (a lot of) cases where you still have to fix stuff by hand. Also, if you do it, you can't "input raw data" or anything, you have to make your script click all over the design screen.

Oh, so maybe you want to do it by screen-scraping, in the editor? Well, most of the problems still stand, and we get all new ones when territories are too small to be clicked without zoom (are you going to be zooming in and out and moving the map for the script?). Also, if you start comparing a couple of millions of points to another couple millions, things are going to get slow (and you will have to do it, you can't just magically find a territory when all you have is the pixels).

Maybe it's not impossible, but coding it would suck. Certainly not something one could do over the weekend.

It would of course suck much less if we had a direct API for writing centerpoint/connection data to the server (Fizzer, I know you can do it! Can we have an input box labelled "Magic goes here" in the map designer, where you could copy/paste the data in any format?), or possibly if Fizzer decided to do it by himself (a man can dream... ;) ).

@Matma Rex
I'm not sure if finding the point farthest from the edge (essentially, the centre of the inscribed circle) is computationally hard or not. However, finding the bounding box is easy enough, finding its centre is trivial, determining whether that centre is at least some specified distance away from any point of the edge is relatively easy as well. Sure, that won't give you all centre points, but it will give you many of them. Sure, they won't be "perfect", but they'll be good enough.
The way I see it, you shouldn't try to solve it fully automatically; that's way too hard (it would only make sense if you're gonna make many dozens or hundreds of maps, automatically). Instead, you should aim at doing most of the grunt work automatically (which keeps the effort to create the automation in the first place manageable), than solve the corner cases manually.

For a possible solution to the "different scales" (having a map with both the, generally small, European countries and the Americas, Australia and Russia) you could take a look at the "Canada Big" map. Of course that's not something which can be done automatically, but if you could create a manually-preprocessed SVG and feed *that* to the script.