In 1977, Christopher Alexander co-authored the book A Pattern Language: Towns, Buildings and Construction. Alexander and his co-authors lay out a series of generative patterns that form a language of solutions and problems in the built environment. While one can argue about the patterns themselves, the thinking behind this book is what interests me the most. This systems thinking is at the basis of many computational operations that rely on understanding the parts and the whole that they create (which is much greater than just summing up the parts). Generative grammar applies similar thinking to language and shape grammars are at the basis of computational approaches to design (and may be most closely associated to Alexander’s pattern thinking).
Why this nerdy, esoteric introduction? Well, at the Esri User Conference plenary, Jack Dangermond announced the acquisition of Procedural, the Swiss makers of CityEngine. CityEngine uses shape grammars and procedural modeling to rapidly generate 3D cities from 2D data. Users can apply rules and iteratively generate patterns of development. This may sound like something for someone’s PhD thesis, but really it is readily applicable to the real world. You see, since the beginning of zoning and land use regulations, planners and cities have been engaging in a sort of defunct systems urbanism. Buried in the legal language of land use codes and regulations are rules that give birth to the functional forms we see in our cities and towns across the country. Planners know these rules as things like setback requirements, floor area ratio, permitted uses, and so on.
While there are many ways to go with the use of this tool, I’m most interested in its use in a real geographic context to help non-planners understand the relationship between the legalese of codes and regulations and the potential outcomes on design. What CityEngine may allow us to do in the public context is to work interactively with the public to both collect ideas or feedback and to educate. This is not a panacea per se of working with the public in this context, but it provides a tool that can give us the kind of feedback loops we need to work meaningfully with the public in the limited time we normally have with them.
In an immediately practical sense, this announcement signals to the industry that Esri is serious about 3D. Autodesk has project Gallileo among there many 3D tools, Google is beefing up it’s Google Earth platform and continues to improve Sketchup, and now Esri has CityEngine. This is good for consumers of 3D geography as it should drive costs down, improve features and help with our clunky workflows. CityEngine is the perfect complement to the attribute-based 2D ArcGIS, and I’m excited to see how Esri folds this into their flagship product.