Why Procedural Generation?

In my exploring of thingiverse and ed printed terrain on Kickstarter, I have come across two major patterns of terrain models.

In one camp, there are fully completed models. Here, the artist has created a model of a set size and design. These look great, but if you want to fill a table, you will be printing multiples of the exact same building and your tabletop will look like it's come off an assembly line, all completely identical buildings. If you want variety, the artist needs to make several buildings with different variations to make it interesting. Another issue is that the size of each building is fixed. If you want different sized buildings, again the artist needs to make larger and smaller models. This is a very content heavy and tedious approach.

In the other camp, are modular designs. Here the artist makes a number of modular pieces that can then be combined, like Lego bricks, to construct any building you want. Dungeon tiles and other clip together systems all follow this pattern. The advantage here is that you can construct many different buildings and even change them up every game. However because all the modular parts need to be interchangeable, the results can be a bit blocky. They can also look like they're made from a construction set instead of a scale representation of a building or city. 

Procedural generation aims to give us the best of both systems. An intelligent algorithm can take various artists created models and textures and combine them to create some desired building specification. As we resize our desired specification, the algorithm can recreate the building automatically. And it can generate multiple variations of buildings too.

This is the Holy Grail I am after. Stay tuned to follow me on my quest!