Hey Danny! Thanks for putting this link up. There hasn't been much reaction here, but the linked article shows a major paradigm shift in terrain creation, and one that should dramatically change how we work. If I understand the concepts correctly instead of a "Shader" based process (create-filter-erode in a node connected progression of perlin/voronoi noises), to a "Goal" based process of having a defined terrain type (some sort of library of valleys, rivers, ridges, etc) and then placing and adapting the elements into the desired terrain. As the author points out; the skills to create believable terrains from networks of modified 'noises' are difficult to master, and don't lend themselves to an "artistic" workflow. To be able to just place a 'valley' in a given spot and then put a 'river' off to one side of it beneath a 'cliff' and then tweak with types of erosion, displacement and debris forming a desired shape would be truly awesome. While TG contains a nodal language to create virtually any terrain feature, it ain't easy (understatement)
.
So Danny, thanks again, and this bump is because I think this is a very important development!