7 Laying out a very large building complex: the Eishin Campus

The creation of the site and volume plan for the Eishin Campus in Japan followed a similar but more complex sequence of structure-preserving transformations. The process had two components.
First system of centers: defined by the pattern language.
The first component was a pattern language worked out by our team after extensive interviews with teachers and students, and then approved by the school as a whole in a general assembly meeting. The pattern language defined, in generic terms, which new centers ought to exist in the new campus.
[…]
Second system of centers: defined by the land.
The second component of the process was a system of centers that existed in the land as it was before we started. This system of centers was defined by the site, by the land itself.

Thus there are two quite different systems of key centers.
First, there is the system of centers which is defined by the pattern language. These major centers are the building blocks of the new project. […]
Second, there is the system of centers which exists in the land. This system is created by the land forms, by the roads, by directions of access, by natural low spots, natural high spots and by existing trees.
It must be emphasized that both systems of centers always exist at the time one starts a site plan.
The first system is generic; it exists in our minds and in the day-to-day experience of the people who are going to have the new school. The second system exists in the land, on the particular site where the project is to be built. Each of the two systems of centers is real.
Together, when fused, they will govern the plan which has to be made. The process of site-planning is the process of, somehow, finding a way to make these two systems of centers become one — a way in which the systems of centers defined by the pattern language can be placed, so that it enhances, preserves, and extends, the system of centers which is already in the land. In this specific case, and in general, the crux of the problem of making the site plan lies in the task of reconciling the two systems of centers — that means finding a new structure which unfolds from the existing wholeness, and which then embodies the centers of the pattern language within the system of centers that exist on the site.
Another way to put this is to say that in order to preserve the structure of the land, the new centers coming from the pattern language must be established in such a way that they fall naturally in places that coincide with, or enhance, the naturally occurring centers of the site.

How can this combination of centers from a pattern language and from the environment be translated into software? The pattern language approach is powerful and could be translated into centers that form from the most important use cases. The environment seems to map to what is in place today, which in software can only be the (inferior?) solutions in place today, and the technical platform to build on.

This roughly represents how I have been approaching the design process for my thinking tool: I wanted to deeply understand the problem first, so I spent an unbelievable amount of time learning from others what they do today, how they do it today, and what they want to achieve. In parallel, I looked at research explaining our creative thought processes to learn more about the “platform” this all gets built on — our cognitive capabilities and processes as much (or perhaps even more) than the technical platform where the actual building will take place.

Consequentially, a great next step would be to describe my thinking tool with a pattern language that generically describes how it could work in an abstract sense, just precise enough to enable people to imagine what certain parts of this tool look and feel like.

One aspect of what is the equivalent of site in software is the software environment that already exists. Thinking about an iOS application, for instance, a new app should respect and enhance centers that are already there — operating system features that get better the more apps are using them, and which at the same time enhance the app, because they offer functionality beyond what would be achievable otherwise. This speaks for deep integration into the system, and it opposes approaches where the environment is ignored in favor of more convenient (and more economically efficient) ways of implementing features. Platform-independent frameworks violate this principle, for instance.

We had pieces of balsa wood cut roughly to the size and shape of typical buildings or building wings. We played with them on the model, constantly trying the same variations I have shown in the diagrams, trying to reconcile the centers of the pattern language with the five key centers on the site.
Gradually one element emerged: the fact that the home base street might be more powerful as an approach to the university center, than as an entity approached from it. This was hard to see, because it implied reversing the main sequence of centers given by the language. But when we tried it, it was clear that the sequence of centers got much better from doing it. Playing confirmed it strongly.

All the time we had been imagining the university center as square in shape. It had appeared this way from the time of our earliest diagrams and we had continued to imagine it like this. But the problem was that the main center on the site “the most beautiful spot” was not square. It was the ridge along the south edge of the property. Suddenly we realized that the university center could actually be this ridge. The moment this second break in perception was made, the whole thing fell into place immediately.

This is how the design process is supposed to feel.

#book/The Nature of Order/3 A Vision of a Living World/5 The positive pattern of space and volume in three dimensions on the land#