The Map Generator

[F_Smith]

Outstanding:

One question, tho -- how will this map be rendered to the screen? Just so I can put a dummy version into the demo, and have 'hooks' ready?

Personally, I love the idea of generating the elevations first, then determining sea level. Beautiful!

[Gregor]

The Map Generator


I am beginning the Major portion of my work on the Map Generator this weekend (as I have four days off). One thing I may need is someone I can send graphics files to, so they can be linked to this thread.

Some of what I will be working on:

The generator will take in account continent formation in association with continental plates (plate tectonics). So major fault lines will be identified early and provide reasoning behind elevation differences (mountain ranges). I am playing with the idea of creating an entire land world, and then identifying sea level afterwards. This would affect the conduct of undersea warfare (as both underwater topography and salinity have a great effect on this type of warfare) and also if changes in sea level were to occur, the mountains could be identified in a reasonable manner (the balkan range extends well into and under the aegean, also Norway's fjords are nothing more the gaps between peaks).
Latitude will have a great effect on the land, as well as land-water distribution and
ocean currents (gyres). A climatological model will be constructed as every map is generated that will be able to handle changes in the land (farms, irrigation, and possibly river reorientation). Climate and land type will define the fertility of the land and the type of trade recources and food that it offers.
If I can convince the animator (artist), I would like to propose a method of rendering the maps. For the main playing map, for coastlines within a certain range of opposing coastlines (say, for example, all the coastline in the mediterranean) you would have between 15-30 random coastline graphics which diplayed inlets and bays and so forth. For, land on the edge of a main continent, far from other land (i.e. the edge of a main initial land mass) may have a different set to use). More on this later.
I will post the equations and a more detailed essay on what I am proposing on Monday night (GMT) 17may.
Cheers, Gregor

[JimC]

Please don't make any major assumptions about interface/class design without informing me, since I've pretty much finished the Map AI.

If you give me an e-mail / ICQ address I'll send you an outline of how things are organised for pathfinding etc...

Jim

[JimC]

btw A nearly-complete Applet showing the A* search is here:

A* Applet

After my exams have finished, I'll complete this Applet, and work on a 'path-intensity' Applet demo based on the 'understanding geography' stuff on Mark's site.

(estimated completion date: early June)

At the end of the day, somebody is going to have to integrate all of our map work into Clash, so it's best if we keep everything as uniform as possible. Please don't force me into a big redesign when I'm so close to completion.

Jim

[F_Smith]

oops . . .

[This message has been edited by F_Smith (edited May 14, 1999).]

[F_Smith]

Cat on the keyboard! Sorry . . .

[This message has been edited by F_Smith (edited May 14, 1999).]

[F_Smith]

So far, I've only worked on
the 'pre-game' interface.

Mark has a working game demo of early in-game code, but almost everything is likely to change at least a little bit.

My emails are jfalasch@csc.com (work) and barefoot@eden.com (home). I think I'm the fella that's going to do most of the merging of the code, and I totally agree about a uniform architecture. I'm a little concerned that there's some code written already without any discussion as to architecture to begin with, but we'll just have to make it work. As it stands, the pre-game Gui simply gathers user choices and then instantiates 'Game.class'. I'll be putting the source code up on the web sometime this weekend.

I saw the applet you put together. It seems solid, very good work. If I can be of any help, please let me know.

[JimC]

Well, I've tried to keep my map work as closely related to the Clash Map/MapSquare implementation as possible.

In the last discussion I had with Mark, he seemed to think that simply inheriting from my base Map/MapSquare would be best....

Jim

[F_Smith]

It'll work fine, I'm sure.

Is that 'BaseMapSquare' class the same as, or similar to, the one in the v1.05 code?

[Mark_Everson]

Gregor:

Thanks for taking this on. It looks really interesting so far .

All:

Gregor and I have talked and he knows the map we're going with is square-based and cylindrical.

We should be able to "dock" what Gregor's doing into the BaseMapSquare and descendent classes fairly easily.

-Mark