3pointD.com

Here’s some fascinating stuff that appeals to the twisted part of my mind that wants to systematize everything it comes into contact with (which is also probably why I’ve always enjoyed reading Donald Knuth, but that’s a subject for another post — or another blog altogether). Anyway, there’s an interesting article on the Directions magazine site about something called CityGML, an open-standard, XML-based “common information model for the representation of 3D urban objects” that’s being developed in conjunction with the Open Geospatial Consortium and ISO TC211, which is responsible for the ISO’s geographic information series of standards.

[CityGML] defines classes and relations for the most relevant topographic objects in cities and regional models with respect to their geometric, topological, semantic and appearance properties. “City” is broadly defined to include not just built structures, but also elevation, vegetation, water bodies, “sidewalk furniture” and more. Included are generalization hierarchies between thematic classes, aggregations, relationships between objects and spatial properties. These thematic information types go beyond graphic exchange formats and allow users to employ virtual 3D city models for sophisticated analysis tasks in different application domains such as simulation, urban data mining, facilities management, decision support and thematic inquiries.

In other words, if I’m understanding things correctly, CityGML attempts to describe a data set that can be used to answer questions like What’s the best place to set up a crisis management center during a disaster that’s centered on the downtown area? But on the way to that, the standard itself has to talk about things like What is a building? What is a park bench? What’s the relationship between the intersection of Sixth and Main and the firehouse two blocks away? Fascinating questions. But can they be answered by blocks of code?

CityGML defines five levels of detail, which appear to be cities, neighborhoods, blocks, buildings, and building elements. Here’s an image of 3 levels of detail represented at once:

Schema definitions, sample datasets and a free viewer app are available on the homepage. I’m not a big fan of systems that come with their own browser (I don’t need yet another client on my hard drive), but since CityGML is open-standard and the viewers being developed are open-source, it ought to be possible to usefully layer the data into another app, though you’d have to build a pretty formidable plug-in to interpret it. From the sound of the article, there’s a lot of work going on around the initiative.

As noted, I love this kind of thing. But is it really something that can be used to solve problems in urban planning? And how might something like CityGML work at the margins? Can I use for looking at a small town? Or does that take VillageGML? Why not just a semantic system for describing the entire world, or space in general? (I don’t know the answers to these questions; I’m just asking.)

Somehow, this all puts me in mind of the work of architect Paolo Soleri and sociologist Holly Whyte, as well as the min/max styles of many players of massively multiplayer online games (i.e., players who attempt to maximize their progress and minimize their investment of time).

A while back, I blogged about a project called Scalable City, which looked at computerized urban-planning algorithms (and seemed to find them wanting). This is min/maxing at work in the real world: how do you maximize population density while minimizing the impact on the environment (or other undesirable effects, like traffic)? This is the question that Soleri has spent his career trying to answer, with his response being “arcologies,” vast self-contained cities (like the one pictured at the top of this post) that fit the maximum number of people into the minimum amount of space. While I think something like CityGML could have very productive uses, I guess my alarmist fear is that it would lead to nightmare scenarios involving sardine-can cities that turn out to be totally unliveable.

Which is where Holly Whyte comes in. A sociologist, journalist and deep student of the life of the city, Whyte worked with the New York City Planning Commission in the 1960s and 70s, but his techniques were worlds apart from anything resembling a min/max algorithm. What Whyte did was simply watch what went on around him. His observations about how to plan urban spaces are simple and brilliant. (The ideal depth for a public bench, he asserted, was “two backsides wide.”) Whyte studied people and how they lived, proceeding inductively from the details of street life out to broader principles of how better spaces might be designed. And he did so with unrivalled charm. His narration of the documentary “The Social Life of Small Urban Spaces” is adorable. I think you can still order the book and the video here.)(If I ever go back to school, this is probably what I’d study: online worlds as public spaces.)

But to bring this back around to CityGML: I love the idea of a system that semantically describes what goes on in urban spaces. I imagine CityGML could be a very useful system in all kinds of contexts, and I’d be excited to see it reach full implementation. But I also imagine that the seductive power of technology could lead some people to rely on it too heavily. The antidote to that, I think, is simply to be part of your surroundings, to watch and listen, as Whyte did, and to draw conclusions from the machine that lives around you, without abdicating responsibility and wisdom to the algorithm. Min/maxing is fine in game worlds, but the real world will always be more complex, more nuanced, more fascinating still. It will be interesting to see how to the two can usefully converge (since that’s at the heart of 3pointD, after all), and whether CityGML will be part of getting there.