The Generator Project
Cedric Price’s proposal for the Gilman Corporation was a series of relocatable structures on a permanent grid of foundation pads on a site in Florida. Cedric Price asked John and Julia Frazer to work as computer consultants for this project. They produced a computer program to organize the layout of the site in response to changing requirements, and in addition suggested that a single-chip microprocessor should be embedded in every component of the building, to make it the controlling processor. This would result in an “intelligent” building which controlled its own organisation in response to use. If not changed, the building wold have become “bored” and proposed alternative arrangements for evaluation, learning how to improve its own evaluation, learning how to improve its own organisation on the basis of this experience.
Generator (1976-79) sought to create conditions for shifting, changing personal interactions in a reconfigurable and responsive architectural project.
It followed this open-ended brief:
“A building which will not contradict, but enhance, the feeling of being in the middle of nowhere; has to be accessible to the public as well as to private guests; has to create a feeling of seclusion conducive to creative impulses, yet…accommodate audiences; has to respect the wildness of the environment while accommodating a grand piano; has to respect the continuity of the history of the place while being innovative.”
The proposal consisted of an orthogonal grid of foundation bases, tracks and linear drains, in which a mobile crane could place a kit of parts comprised of cubical module enclosures and infill components (i.e. timber frames to be filled with modular components raging from movable cladding wall panels to furniture, services and fittings), screening posts, decks and circulation components (i.e. walkways on the ground level and suspended at roof level) in multiple arrangements.
When Cedric Price approached John and Julia Frazer he wrote:
“The whole intention of the project is to create an architecture sufficiently responsive to the making of a change of mind constructively pleasurable”
They proposed four programs that would use input from sensors attached to Generator’s components: the first three provided a “perpetual architect” drawing program that held the data and rules for Generator’s design; an inventory program that offered feedback on utilisation; an interface for “interactive interrogation” that let users model and prototype Generator’s layout before committing the design.
The powerful and curious boredom program served to provoke Generator’s users. “In the event of the site not being re-organized or changed for some time the computer starts generating unsolicited plans and improvements,” the Frazers wrote. These plans would then be handed off to Factor, the mobile crane operator, who would move the cubes and other elements of Generator. “In a sense the building can be described as being literally ‘intelligent’,” wrote John Frazer—Generator “should have a mind of its own.” It would not only challenge its users, facilitators, architect and programmer—it would challenge itself.
The Frazers’ research and techniques
The first proposal, associated with a level of ‘interactive’ relationship between ‘architect/machine’, would assist in drawing and with the production of additional information, somewhat implicit in the other parallel developments/ proposals. The second proposal, related to the level of ‘interactive/semiautomatic’ relationship of ‘client–user/machine’, was ‘a perpetual architect for carrying out instructions from the Polorizer’ and for providing, for instance, operative drawings to the crane operator/driver; and the third proposal consisted of a ‘[. . .] scheduling and inventory package for the Factor [. . .] it could act as a perpetual functional critic or commentator.’ The fourth proposal, relating to the third level of relationship, enabled the permanent actions of the users, while the fifth proposal consisted of a ‘morphogenetic program which takes suggested activities and arranges the elements on the site to meet the requirements in accordance with a set of rules.’ Finally, the last proposal was [. . .] an extension [. . .] to generate unsolicited plans, improvements and modifications in response to users’ comments, records of activities, or even by building in a boredom concept so that the site starts to make proposals about rearrangements of itself if no changes are made. The program could be heuristic and improve its own strategies for site organisation on the basis of experience and feedback of user response.
In a certain way, the idea of a computational aid in the Generator project also acknowledged and intended to promote some degree of unpredictability. Generator, even if unbuilt, had acquired a notable position as the first intelligent building project. Cedric Price and the Frazers´ collaboration constituted an outstanding exchange between architecture and computational systems. The Generator experience explored the impact of the new techno-cultural order of the Information Society in terms of participatory design and responsive building. At an early date, it took responsiveness further; and postulates like those behind the Generator, where the influence of new computational technologies reaches the level of experience and an aesthetics of interactivity, seems interesting and productive.
- John Frazer, An Evolutionary Architecture, Architectural Association Publications, London 1995. http://www.aaschool.ac.uk/publications/ea/exhibition.html
- Frazer to C. Price, (Letter mentioning ‘Second thoughts but using the same classification system as before’), 11 January 1979. Generator document folio DR1995:0280:65 5/5, Cedric Price Archives (Montreal: Canadian Centre for Architecture).