Hi everyone :) And welcome to my first dev log!

The topic of this post is to discuss my thought process for map development and the research and development of developing the initial process for generating the world map.

The most important thing in my mind is to produce unique maps. What I mean by unique is that I want the maps to have life and character, but they still need to make logical sense. When the player creates a new world, I want them to see it and say, "This world makes sense."

Over the course of the development of the world generator, I have tried several approaches, each of which I will outline why I tried and why it didn't work in my case. In a following post, I will give further detail on how I generate the world in the current version of the game.

The first step that I took in this process was research. And when I say research, I mean research everything and use all resources. I researched standard applicable techniques for procedural map generation, what other games have used, and what makes a map interesting.

The most common response I got for techniques for procedural map generation was using noise effects, such as Perlin or Simplex. Both of these methods are good for creating randomly generated maps, but they both, in my opinion, require a lot of tuning and layering of effects to get a result that resembles a real map. Because of this fine-tuning and layering, I did not use Perlin or Simplex.

Through my research, I found a blog, Polygonal Map Generation for Games (stanford.edu). The blog I linked is a treasure trove of map generation ideas and structures from generating landmass, elevation, moisture, rivers, biomes, and lakes. Halfway through the development process of using the Voronoi polygons/Delauney approach outlined in the blog post, I questioned whether this approach fits into my game. See, I had the idea that my game is tile-based and while the approach I was taking generates beautiful maps, it doesn't lend itself well to a tile system as accessing the underlying information of each tile would be difficult. But there were some variations of ideas that I kept from this blog post that were useful, such as:

1. Keeping the outer edges as water
   1. simplifies development as I do not have to worry about wrapping the world.
2. classifying water types
   1. freshwater, enclosed by land
   2. saltwater, toches the map's border
3. Elevation is a simple calculation of distance from the coast. 
   1. This simplified calculation will change when adding complexity to the game.
   2. This calculation allows for a constant gradient, which eliminates local minimums rivers might experience
4. Rivers, calculated from random points at a high elevation and terminate at the ocean
5. moisture, distance from water

I still needed to find a way to properly represent my map naturally to tile-based maps but still generate interesting and unique maps; back to the drawing board! As I did more research to find appropriate map generation techniques, I came across an old tried and tested approach, cellular automata. Through some slight modification to the process, cellular automata allow for various natural-looking map designs and the application of techniques such as water classification, elevation, rivers, moisture, biomes, temperature, etc.

In the next blog, I will detail the development process of cellular automata and the first steps of generating a map!

S.