K
Klaus Philipsen
Guest
- Thread Author
- #1
During my time studying architecture the architect-engineer and Pritzker Prize winner Frei Otto was a full time professor at my alma mater, the University of Stuttgart and his Institute of Institute for Lightweight Structures was a flagship for innovation. He designed structures guided by gravity not ego, and like Buckminster Fuller before him he was guided by the credo of doing more with less. In spite of the similarity with the modernist credo of "less is more", he also broke with the prevailing modern dogma of the Bauhaus box.
This was well before desktop computers were at hand to calculate loads of complex shapes. Even though Otto employed main frame computing for architecture in a first of his time, he was mainly an experimenter who tested efficient shapes via models and natural precedents such as spider webs (the ultimate tensile structures) or soap bubbles (which enclose a maximum volume with a minimum of material). He filtered into the minds of us architecture students almost via osmosis propelled by his charismatic persona. He represented categories that were intriguing but also foreign to most of us at the time and only would become popular decades later such as sustainability, biomimicry or biometric design, and parametric design. One could argue that Patrick Schumacher's Parametricist Manifesto from 2008 which calls parametric design a style is really a misunderstanding of the term.
Regardless, with the help of the computer and the softwares that architects now use, especially building information modeling (BIM) and now AI parametric design is now commonplace.
Frei Otto who died age in 2015 aged 90 remained close to me by proxy when my first job took me to the first floor of a building in which the Stuttgart architect Guenther Behnisch 's office occupied the second floor.
Behnisch, like Frei Otto a promoter of lightness and airiness, amplified Frei Otto's international fame when they both designed the venues for the 1972 Munich Olymics, the largest tent structures ever tried making the Bavarian metropolis a mekka of architectural innovation, a clear break with the countries Olympic history in nearby Nuremberg. Once again the ideas of lightness and tensile design floated through the building.
That architecture is in large parts the result of technology and the result of interdisciplinary collaboration with creative minds taking loans from "adjacent" disciplines thus became part of my credo. This wasn't a novel insight by any means, yet one that kept me from being easily washed up into the architectural fashion of the time (post-mdernism) and kept me investigating how the means of architectural production shape architecture. But the history of architectural form is not a straight and upward trajectory from primitive to complex as Schumacher wants to believe.
In some way parametric architecture comes full circle, i.e architects re-learning what builders before there were architects had instinctively known, namely: learning from nature. Architecture without architects was the prevailing mode of construction for millennia. Complex structures include tents, the round "trullis" of La Puglia, Italy with their pointed roofs, the Egyptian and Mayan pyramids or the ancient stupas of Sri Lanka.
Western architecture, as every architecture student learns, developed from the simple post and beam structures of ancient Greece to round and pointed arches all the way to the soaring vertical slenderness of marvels such as Notre Dame. Then came the elevator and steel construction, the seperation of structure from envelope, the shedding of the ornament, the Bauhaus, post modernism, deconstructivism and now, if one believes Schumacher, parametricism.
I don't want define parametric design as a style but rather as the method of getting to design, hence my reference to Frei Otto and to the time before architects when construction didn't spring from drawings. Even when architects began to document their visions by drawing structures before they were built using the T-square, a ruler, a compass, and a triangle, some didn't stick with designing the obvious forms consisting only of right angles or circles. Innovators such Brunelleschi (actually a goldsmith and clockmaker) created technological break-throughs such as the giant dome of the Florence cathedral, an architectural, engineering and a production miracle by combining ingenuity, experiment, geometry and physics. In a broader sense this could be called parametric design. Neither Buckminster Fuller nor Frei Otto, and certainly not Zaha Hadid invented parametric design. Maybe the inventor of parametric modeling was Antonio Gaudi who designed his structures including load testing straight from models, using a biometric design approach, including measuring tension in key structural members by inserting measuring scales into the models very similar to how Frei Otto and Behnisch still did it for the Olympic tents where the bulki and slow main frame computers were to sluggish or the right coding did not yet exist.
But the personal computer and the ability of structural and visual modeling software finally lifted the boundaries of drawings and physical models and brought about a revolution in design. The computer that can build a digital twin of the envisioned building as a virtual sibling. This brings ultimate freedom from the limitations of the basic two-dimensional geometry of lines and 90 degree angles or the abilities of paper, wood or clay models.
This leap didn't happen quickly. Initially the computers only automated the same 2D drawings that had been drawn for centuries. The architect Frank Gehry was the first who fully explored the new opportunities by introducing the 3rd dimension right from the start by borrowing from the aerospace industry by adopting their software and their approach of thinking form, production and constructability as integral to each other. In that Gehry's approach differes from Hadid's office where the form isn't always the result of optimization but rather a style.
The new architectural design freedom did not end with creating virtual versions of the real buildings. It practically exploded with the use of AI based software such as Midjourney. Schumacher's idea of parametric design became common place when it was no longer necessary to imagine the design in one's mind or derive it from the algorithms of physics and geometry as Fuller, Otto and to some extent, Gehry had done it. Now architectural from can be created with words. It is enough to say "create a design that looks like Zaha Hadid's designs". Or ask for a building that looks like a spider or whatever comes to mind. Within seconds the rich visual database of AI will mine the precedents and compose a concept in an all new, never before done manner. With some tweaking it can even be specific to a particular site.
However, this freedom from restrictions should not be confused with doing architecture. As with any freedom, this newfound design freedom begs the question beyond what we gained "freedom from" (in this case the box). It urgently needs also the second question which is "freedom to what end"? Unlimited freedom requires more discipline and restraint. When the T-square and the ruler no longer restrain form-giving, physics still exists and many other considerations have to prevent an arbitrary and capricious folly.
Even in a time of seemingly unrestrained options for form-giving options it has to be understood that form giving isn't yet design, and design isn't yet architecture. Successful design and architecture require many considerations. For example, in any conglomeration of buildings (village, town, city) not every building can be a landmark that calls out "look at me"! In fact, most city buildings must blend in and be part of a fabric for a city to become more than the sum of parts and more than a circus of architectural vanity. Drawing a wild looking structure doesn't make it functional, sustainable, cost effective or buildable. To check all those boxes still requires experience in site observation, user needs and habits, production methods, knowledge of building materials and how they are fabricated, put together or get to the site. Sustainability demands require this knowledge more urgently than ever before. So far AI can generate renderings and even structures but mostly cannot combine all those skills or replace years of actual physical experience in applying those skills to physical world.
In training future architects, it would be wise to make them aware of the design process that Frei Otto used, for his approach was not arbitrarily imposing form on a building but deriving the optimal form from the laws of physics and geometry. His goal was not to create beauty but to use as little material as necessary to build a structure that could fulfill its function. He didn't do form follows function but form follows physics. The beauty came from the optimization. His tensile structures won't work for every purpose, in fact, they are probably not applicable for most building structures. His work influenced other architect giants such as Norman Foster, Richad Rogers, Zaha Hadid, and more recently Bjarke Ingels (Google HQ). Their applications, too are limited.
But for those who aspire to design beyond the box, Otto's restraint would be a good lesson anyway, especially now when everyone has access to tools capable of generating virtually unlimited formal variations. It is key to remember that form is not yet architecture. In this new environment the scarce resource is no longer imagination but judgment.
Klaus Philipsen, FAIA
Postscript: This essay suggests an antagonism between form and function, or pure form and structural optimization, which isn’t necessarily a result of AI or inherent to it. On the on the Digital Futures series Patrick Schumacher and his collegue Shajay Bhooshan demonstrate that AI can achieve structural logic and parametric design processes similar to those used by Frei Otto and Buckminster Fuller by applying tectonics and harnessing LLM-based AI for architectural and structural applications. (9-20-26)
Related on this Blog:
Technology in the Architecture Office
[td]
[/td]
[td]
Munich tent structures for athletic facilities for the 1972 Olympics
(Guardian)
[/td](Guardian)
This was well before desktop computers were at hand to calculate loads of complex shapes. Even though Otto employed main frame computing for architecture in a first of his time, he was mainly an experimenter who tested efficient shapes via models and natural precedents such as spider webs (the ultimate tensile structures) or soap bubbles (which enclose a maximum volume with a minimum of material). He filtered into the minds of us architecture students almost via osmosis propelled by his charismatic persona. He represented categories that were intriguing but also foreign to most of us at the time and only would become popular decades later such as sustainability, biomimicry or biometric design, and parametric design. One could argue that Patrick Schumacher's Parametricist Manifesto from 2008 which calls parametric design a style is really a misunderstanding of the term.
[td]
[/td]
[td]
Frei Otto in Stuttgart
[/td]Regardless, with the help of the computer and the softwares that architects now use, especially building information modeling (BIM) and now AI parametric design is now commonplace.
Frei Otto who died age in 2015 aged 90 remained close to me by proxy when my first job took me to the first floor of a building in which the Stuttgart architect Guenther Behnisch 's office occupied the second floor.
[td]
[/td]
[td]
Zaha Hadid in Baku, Azerbajan (Guardian)
[/td]Behnisch, like Frei Otto a promoter of lightness and airiness, amplified Frei Otto's international fame when they both designed the venues for the 1972 Munich Olymics, the largest tent structures ever tried making the Bavarian metropolis a mekka of architectural innovation, a clear break with the countries Olympic history in nearby Nuremberg. Once again the ideas of lightness and tensile design floated through the building.
That architecture is in large parts the result of technology and the result of interdisciplinary collaboration with creative minds taking loans from "adjacent" disciplines thus became part of my credo. This wasn't a novel insight by any means, yet one that kept me from being easily washed up into the architectural fashion of the time (post-mdernism) and kept me investigating how the means of architectural production shape architecture. But the history of architectural form is not a straight and upward trajectory from primitive to complex as Schumacher wants to believe.
In some way parametric architecture comes full circle, i.e architects re-learning what builders before there were architects had instinctively known, namely: learning from nature. Architecture without architects was the prevailing mode of construction for millennia. Complex structures include tents, the round "trullis" of La Puglia, Italy with their pointed roofs, the Egyptian and Mayan pyramids or the ancient stupas of Sri Lanka.
[td]
[/td]
[td]
Stupa in Thuparamaya, Sri Lanka
[/td]Western architecture, as every architecture student learns, developed from the simple post and beam structures of ancient Greece to round and pointed arches all the way to the soaring vertical slenderness of marvels such as Notre Dame. Then came the elevator and steel construction, the seperation of structure from envelope, the shedding of the ornament, the Bauhaus, post modernism, deconstructivism and now, if one believes Schumacher, parametricism.
I don't want define parametric design as a style but rather as the method of getting to design, hence my reference to Frei Otto and to the time before architects when construction didn't spring from drawings. Even when architects began to document their visions by drawing structures before they were built using the T-square, a ruler, a compass, and a triangle, some didn't stick with designing the obvious forms consisting only of right angles or circles. Innovators such Brunelleschi (actually a goldsmith and clockmaker) created technological break-throughs such as the giant dome of the Florence cathedral, an architectural, engineering and a production miracle by combining ingenuity, experiment, geometry and physics. In a broader sense this could be called parametric design. Neither Buckminster Fuller nor Frei Otto, and certainly not Zaha Hadid invented parametric design. Maybe the inventor of parametric modeling was Antonio Gaudi who designed his structures including load testing straight from models, using a biometric design approach, including measuring tension in key structural members by inserting measuring scales into the models very similar to how Frei Otto and Behnisch still did it for the Olympic tents where the bulki and slow main frame computers were to sluggish or the right coding did not yet exist.
But the personal computer and the ability of structural and visual modeling software finally lifted the boundaries of drawings and physical models and brought about a revolution in design. The computer that can build a digital twin of the envisioned building as a virtual sibling. This brings ultimate freedom from the limitations of the basic two-dimensional geometry of lines and 90 degree angles or the abilities of paper, wood or clay models.
[td]
[/td]
[td]
BIM can create a complete virtual model of a building (KennMar)
[/td]This leap didn't happen quickly. Initially the computers only automated the same 2D drawings that had been drawn for centuries. The architect Frank Gehry was the first who fully explored the new opportunities by introducing the 3rd dimension right from the start by borrowing from the aerospace industry by adopting their software and their approach of thinking form, production and constructability as integral to each other. In that Gehry's approach differes from Hadid's office where the form isn't always the result of optimization but rather a style.
The new architectural design freedom did not end with creating virtual versions of the real buildings. It practically exploded with the use of AI based software such as Midjourney. Schumacher's idea of parametric design became common place when it was no longer necessary to imagine the design in one's mind or derive it from the algorithms of physics and geometry as Fuller, Otto and to some extent, Gehry had done it. Now architectural from can be created with words. It is enough to say "create a design that looks like Zaha Hadid's designs". Or ask for a building that looks like a spider or whatever comes to mind. Within seconds the rich visual database of AI will mine the precedents and compose a concept in an all new, never before done manner. With some tweaking it can even be specific to a particular site.
[td]
[/td]
[td]
AI can create building renderings in seconds based on
word commands (Archistar)
[/td]word commands (Archistar)
However, this freedom from restrictions should not be confused with doing architecture. As with any freedom, this newfound design freedom begs the question beyond what we gained "freedom from" (in this case the box). It urgently needs also the second question which is "freedom to what end"? Unlimited freedom requires more discipline and restraint. When the T-square and the ruler no longer restrain form-giving, physics still exists and many other considerations have to prevent an arbitrary and capricious folly.
Even in a time of seemingly unrestrained options for form-giving options it has to be understood that form giving isn't yet design, and design isn't yet architecture. Successful design and architecture require many considerations. For example, in any conglomeration of buildings (village, town, city) not every building can be a landmark that calls out "look at me"! In fact, most city buildings must blend in and be part of a fabric for a city to become more than the sum of parts and more than a circus of architectural vanity. Drawing a wild looking structure doesn't make it functional, sustainable, cost effective or buildable. To check all those boxes still requires experience in site observation, user needs and habits, production methods, knowledge of building materials and how they are fabricated, put together or get to the site. Sustainability demands require this knowledge more urgently than ever before. So far AI can generate renderings and even structures but mostly cannot combine all those skills or replace years of actual physical experience in applying those skills to physical world.
[td]
[/td]
[td]
Todays' construction is still a fairly mundane and slow affair with
much manual labor. It has to be understood in order to design.
[/td]much manual labor. It has to be understood in order to design.
In training future architects, it would be wise to make them aware of the design process that Frei Otto used, for his approach was not arbitrarily imposing form on a building but deriving the optimal form from the laws of physics and geometry. His goal was not to create beauty but to use as little material as necessary to build a structure that could fulfill its function. He didn't do form follows function but form follows physics. The beauty came from the optimization. His tensile structures won't work for every purpose, in fact, they are probably not applicable for most building structures. His work influenced other architect giants such as Norman Foster, Richad Rogers, Zaha Hadid, and more recently Bjarke Ingels (Google HQ). Their applications, too are limited.
But for those who aspire to design beyond the box, Otto's restraint would be a good lesson anyway, especially now when everyone has access to tools capable of generating virtually unlimited formal variations. It is key to remember that form is not yet architecture. In this new environment the scarce resource is no longer imagination but judgment.
Klaus Philipsen, FAIA
Postscript: This essay suggests an antagonism between form and function, or pure form and structural optimization, which isn’t necessarily a result of AI or inherent to it. On the on the Digital Futures series Patrick Schumacher and his collegue Shajay Bhooshan demonstrate that AI can achieve structural logic and parametric design processes similar to those used by Frei Otto and Buckminster Fuller by applying tectonics and harnessing LLM-based AI for architectural and structural applications. (9-20-26)
Related on this Blog:
Technology in the Architecture Office
Can Information Modeling Make Cities Smarter?
Governance in the Digital City
Technology in the Construction Industry