What Software Builders Can Learn From Building Builders

How Buildings Learn

The basic thesis: we think of "building a building" as something that happens at a moment in time. But actually, buildings evolve and adapt over time.

The best architects plan for this adaptation up front.

Magazine Architecture

She calls her peers “magazine architects.” By which she meant image-driven and fad-driven architects, because architecture magazines probe no deeper than the look and style of the buildings they cover. They never interview clients or users. They never criticize buildings except, rarely, in terms of being bad art or off-trend. Articles consist primarily of stylized color photographs. Reports cover only new or newly renovated buildings, often in language that sounds like the “prismatic luminescence” school of wine writing. The subject is taste, not use; commercial success, not operational success.

A range of observers of architecture are now suggesting that the field may be bankrupt, the profession itself impotent, and the methods inapplicable to contemporary design tasks. It is further suggested that collectively they are incapable of producing pleasant, livable, and humane environments, except perhaps occasionally and then only by chance.

“The curse of architectural photography, which is all about the wonderfully composed shot, the absolutely lifeless picture that takes time out of architecture—the photograph taken the day before move-in. That’s what you get awards for, that’s what you make a career based on. All those lovely but empty stills of uninhabited and uninhabitable spaces have squeezed more life out of architecture than perhaps any other single factor.

Example: Fallingwater by Frank Lloyd Wright

Fallingwater in Pennsylvania, celebrated by the AIA poll as “the best all-time work of American architecture,” lives up to its name with a plague of leaks; they have marred the windows and stone walls and deteriorated the structural concrete. To its original owner, Fallingwater was known as “Rising Mildew,” a “seven-bucket building.”

Our present attitude is all reversed. What you have is extremely inexpensive structure and all this glitz on the surface. The structure rots after thirty years, and the glitz is so expensive that you daren’t even fuck with it

Feature Factories

I recall asking one architect what he learned from his earlier buildings. “Oh, you never go back!” he exclaimed, “It’s too discouraging.”


Buildings can't start learning until people move into them. Unfortunately, architects usually don't inhabit the buildings after they're created.

“Finishing is never finished,” but at some point you have to just stop, let the builders go away, and start living in the place. Inhabitation is a complex, thrilling, prolonged event…Inhabit early, build forever. Part of relishing a place is tinkering with it. You’re going to have to anyway, because no building works right at first. It will take a year to work out just the major bugs.

Good artisans also know or can figure out the details that will make a building work right for the inhabitants, so the toilet-paper dispenser can be easily reached and the shower doesn’t spray the bathroom floor or burn the hand that adjusts the hot water. Well-made buildings are fractal—equally intelligent at every level of detail.

Design → Engineering "Handoffs"

Large-lump development is based on the fallacy that it is possible to build perfect buildings. Piecemeal growth is based on the healthier and more realistic view that mistakes are inevitable…. Unless money is available for repairing these mistakes, every building, once built, is condemned to be, to some extent, unworkable…. Piecemeal growth is based on the assumption that adaptation between buildings and their users is necessarily a slow and continuous business which cannot, under any circumstances, be achieved in a single leap.”

Architects think of a building as a complete thing, while builders think of it and know it as a sequence—hole, then foundation, framing, roof, etc. The separation of design from making has resulted in a built environment that has no ‘flow’ to it—you simply cannot design an improvisation or an adaptation. It’s dead.

A man wants a house. He talks with a builder. Together they design the house out of their shared experience, their culture of what a house should be. There is no need for formal plans. Students of vernacular architecture search for plans, wish for plans, but should not be surprised that they find none. The existence of plans on paper is an indicator of cultural weakening. The amount of detail in a plan is an exact measure of the degree of cultural disharmony; the more minimal the plan, the more completely the architectural idea abides in the separate minds of architect and client.

Layers of Change

Different pieces of a building are designed to evolve at different speeds. Some are easy to change, while others are very difficult.

Product Layers

IMO a big problem in product development is that we treat everything like it's got the same lifetime. So we over-engineer surface things like the UI, and under-engineer data models that need to last.

The major difference in a “learning” building is its budget. There needs to be more money than usual spent on the basic Structure, less on finishing, and more on perpetual adjustment and maintenance

Planning for adaptation

Instead of planning for a single launch, we should expect that the products we build will evolve constantly over time. What would it look like to plan for adaptation from the start?

…a continuous series of adaptations—small, very small, and tiny, in ever larger quantities—so that by the time you get down to the smallest level, you’ve got hundreds of things that are getting tuned all the time. A bench here, a window here, a tree here, a couple of paving stones here. Under those conditions you could say, ‘This thing is actually going to get healthy.’ If that small stuff isn’t happening all the time, you’re not going to take care of it, and it isn’t going to come to order.”

Pave where the path is

An oft-told story (perhaps apocryphal) tells how a brilliantly lazy college planner built a new campus with no paths built in at all. She waited for the first winter and photographed where people made paths in the snow between the buildings. Next spring, that’s where the paving went. Some design is better if it’s postponed.

"Loose fit" and extensibility


New College, Oxford, is of rather late foundation, hence the name. It was founded around the late 14th century.

It has like other colleges, a great dining hall with big oaks beams across the top, yes? These might be two feet square, forty five feet long.

A century ago, so I am told, some busy entomologist went up into the roof of the dining hall with a penknife and poked at the beams and found that they were full of beetles. This was reported to the College Council, who met in some dismay, because where would they get beams of that caliber nowadays?

One of the Junior Fellows stuck his neck out and suggested that there might be on college lands some oak. These colleges are endowed with pieces of land scattered across the country. So they called in the College Forester, who of course had not been near the college itself for some years, and asked him about oaks. And he pulled his forelock and said, “Well sires, we was wonderin’ when you’d be askin’”.

Upon further inquiry it was discovered that when the college was founded, a grove of oaks ha been planted to replace the beams in the dining hall when they became beetly, because oak beams always become beetly in the end.

This plan had been passed down for one Forester to the next for five hundred years.

“You don’t cut them oaks. Them’s for the College Hall.”