Extruding Architecture

This journal is devoted to a quest that I began in 1969, the object of which was to create the built environment with construction robots rather than people. It has been a lifelong quest for me.

Wednesday, February 22, 2006

Rethinking Human Buildings

The massive structures erected by Crichton's termites offer a thought-provoking contrast to buildings erected by modern societies.

A typical building erected in Barnett's Integrating Core is made of hundreds to thousands of different materials brought together by dozens of trades using hundreds of different skills.

A termite hive, every bit as sophisticated in both design and execution, is produced by one kind of craftsman with no oversight employing one skill and using moistened pellets of soil as a building material.

Typically, innovators seeking to automate construction attack the problem piecemeal. The International Association for Automation and Robotics in Construction (IAARC)...


...typifies this approach.

When I began my intellectual quest for automated construction in 1969, however, I rejected this approach. I believed from the onset that for construction to be successfully automated the whole process of building had to be rethought and rebuilt. It was insufficient to build a robot that could lay bricks or build drywalls.

I was not alone in thinking these sorts of thoughts at that time. There were others who inspired me in this line of thought. To the best of my knowledge, however, I am the only one who has stayed with it to the present time.

Factories and Craftsmen...

Factories tend to be many magnitudes larger than the thing that they produce. Teams of craftsmen, on the other hand, can be very small when compared with the things they produce.

Factories tend to produce a single item in large numbers. Teams of craftsmen tend to product one-off items.

Factories can produce things cheaply while teams of craftsmen typically can not.

George Romney's idea was good. His notion of what a factory was, however, was a little too conservative.

There is another way.

Automate the craftsman.

Until the last two centuries the tools and materials that one used to build buildings had changed little since Roman times. Indeed, virtually the entire contents of a craftsman's tool box from as recently as sixty years ago could be readily identified on Roman frescoes and carvings two millenia old.

About a century ago, we began to power these tools. The productivity of individual craftsmen improved dramatically. That process has continued to this day. Indeed, a visitor employed in construction to a construction site in Europe or North America is invariably struck by the very small number of people on a building site. Powered tools have made that possible.

Craftsmen have always been highly trained, scarce and well-paid individuals. They can apply their skills effectively to a large number of very different situations. Contrast the craftsman with the situation of the classic factory worker described here...

As the individual work tasks became simple and repetitive this allowed the use of unskilled laborers who could be quickly trained for a single task (though it also removed most of the satisfaction that a worker performing multiple tasks may enjoy).

Factories enabled industrial designers to reduce complicated tasks hitherto attempted only by highly skilled craftsmen to a series of much simpler to learn skills that could quickly be taught to less capable, and thus cheaper to hire, workers.

Monday, February 20, 2006

Making Buildings

For most of human history the making of buildings has been an exercise in the coordination of teams of craftsmen. To a large extent, it still is. At various times down through history, though, the commonality of certain types of buildings and especially certain features of building has resulted in brighter craftsmen mass-producing those buildings or features.

Nearly a millennium ago Chinese civil servants were designing public buildings using standard plans and parts from manuals and drawing pre-cut wood columns and beams from government warehouses to build them.

The notion of palletized building parts that are transported to a building site and assembled there in a brief time by a team of artisans has been reinvented dozens of times before and since then. Probably the best example known to Americans were the kit houses marketed by the Sears and Roebuck company over a century ago.

Most people credit the factory production line to Henry Ford and his Model T production plant at Highland Park. Lewis Mumford dispelled that notion in his book, The City in History, in 1961. {the following description was largely taken from a now defunct website, copies of The City in History being very difficult to acquire}

In it Mumford credits the Venetians with the invention of "a new type of city, based on the differentiation and zoning of urban functions, separated by traffic ways and open spaces", and cites the island of Murano and the Arsenale as Europe's first examples of industrial planning. Of these two, the Arsenale most closely resembled a modern factory complex. Construction techniques in the Arsenale were the most sophisticated of their time: by the fifteenth century the Venetians had perfected a production-line process for equipping their warships, in which the vessels were towed past a succession of windows, to collect ropes, sails, armaments, oars and all their other supplies (ending with barrels of hard biscuits), so that by the time they reached the lagoon the vessels were fully prepared for battle. The productivity of the wharves was legendary: at the height of the conflict with the Turks in the sixteenth century, one ship a day was being added to the Venetian fleet. On the occasion of the visit of Henry III of France in 1574, the Arsenale workers put on a bravura performance - in the time it took the king and his hosts to work their way through a state banquet in the Palazzo Ducale, the Arsenalotti assembled and made sea-worthy a ship sturdy enough to bear a crew plus a cannon weighing 16,000 pounds.

In the USA, shortly after Richard Nixon's election as President in 1968, he appointed George Romney as the head of the Housing and Urban Development department. Romney, a highly successful former CEO of the the American Motors Corporation declared that housing would never be truly affordable until we made homes in factories just like we made cars.

The concept of factory housing dates from this period. There were two major obstacles to the realisation of Romney's dream. Both had to do, ironically, with transportation.

  • houses were not costly enough per unit volume to justify long distance transportation in finished form from a factory
  • existing modes of transportation simply could not economically accomodate an object as big as a house.
Housing factories, to an extent, worked in more densely populated countries. In the US, however, the economic production capacity of a housing factory was simply too large to be workable near any but the very largest metropolitan areas.


In any case, this was a deep human prejudice. Human beings expected to find a central command in any organization. States had governments. Corporations had CEOs. Schools had principals. Armies had generals. Human beings tended to believe that without central command, chaos would overwhelm the organization and nothing significant could be accomplished.

From this standpoint, it was difficult to believe that extremely stupid creatures with brains smaller than pinheads were capable of construction projects more complicated than any human project. But in fact, they were.

African termites were a classic example. These insects made earthen castlelike mounds a hundred feet in diameter and thrusting spires twenty feet into the air. To appreciate their accomplishment, you had to imagine that if termites were the size of people, these mounds would be sky scrapers one mile high and five miles in diameter. And like a skyscraper, the termite mound had an intricate internal architecture to provide fresh air, remove excess carbon dioxide and heat, and so on. Inside the structure were gardens to grow food, residences for royalty, and living space for as many as two million termites. No two mounds were exactly the same; each was individually constructed to suit the requirements and advantages of a particular site.

All this was accomplished with no architect, no foreman, no central authority. Nor was a blueprint for construction encoded in the termite genes. Instead these huge creations were the result of relatively simple rules that the individual termites followed in relation to one another. (Rules like, “If you smell that another termite has been here, put a dirt pellet on this spot.”) Yet the outcome was arguably more complex than any human creation.

Michael Crichton, PREY (2002)