Amorphous Constructions 7/97

presented at morphe:nineteen97

In the video I mention that the 3D data files represent the original artworks just like score sheets in music. And just like an original music score can be interpreted in a variety of ways ranging from classical instruments, jazz versions to fully synthesised versions, the 3D data file can be visualised through line drawings, computer generated still images, fly-throughs, print- outs on paper or projected images. For art lovers with a traditional mind-set this raises the interesting question of what constitutes the original artwork - for architects the question is irrelevant. The more I think about the data file, however, the more I see it as something more than just a file. The data sits there in a complex and highly encoded form and it should be possible to use the data not only to view it with the application program that created it, but to use it for other purposes such as extracting hidden structures which, in turn can generate visual effects, shapes or sounds.

In his presentation Ben van Berkel utilises a similar methodology in a traditional urban planning way by analysing the traffic flow around a train station and then utilising this data for the basic organisation of spaces. Greg Lynn in his Vienna project goes one step further in that he utilises the motorway traffic that by-passes the building to generate its shape. Every car that passes represents a force acting on the relatively arbitrary kinematic structure set up in his animation program. Like a sail in the wind, the building reacts to the density of traffic on the motorway - the final, built form is the result of a selection process of a series of configurations, similar yet different, arranged one next to the other, suggesting animated form through rhythm.

Researchers at the CSIRO had the idea that by visualising abstract, statistical data it would be possible to detect certain structures within the data itself. They market their activities with the catch phrase:

"Information is 21st Century Gold"
which they describe in more detail as:

CSIRO Australia - Division of Information Technology "Data Mining" PR-sheet.

"Many organisations collect data from daily business transactions and from servicing customers. These organisations can now 'mine' their data to gain valuable insights into their business, such as discovering customer buying patterns, detecting fraud and improving risk management. CSIRO is developing data mining techniques to uncover patterns or trends in company data, transforming the data stockpile into an information goldmine."

Taken literally, especially with 3 dimensional datasets or 3D data-subsets, interesting structures - quite possibly with a sculptural quality - should emerge. With my general interest into alternative shape generation this is one area I hope to experiment with - either revealing certain hidden structures or using them to generate shapes. Preferably the data would be compiled from society.

Another way of alternative shape generation which bypasses traditional drawing board techniques (and a few words about that later) is what I have called a growth process penetrating House II in the 'Eisenman Re-visited Project' which we just saw in the video.

The growth process was a relatively simple algorithm that ran on a 1MB Atari with no harddisk. It was an interactive process in that I had to type in the vertex point for each new growth step. The program would then calculate the new geometry, unfold the newly grown area and print out a cutting pattern in true size. The pattern was then transferred to cardboard, folded together and glued into place. The process was quite enjoyable - especially whilst watching MTV and drinking glasses of wine. Theoretically the whole thing could have worked in the computer, the visual representation of the model was too poor, however, to make any sensible judgement.

Trying to grow shapes is quite common since the successful use of L-grammar systems for the growing of plants in 2-dimensional computer graphics. In 3D this process is not that straightforward due to the self-intersection of surfaces.

Other people, like John Frazer and Marcos Novak have taken the 'growth' methodology a lot further and subject their developing shapes to evolutionary criteria - having seen Marcos Novak 's talk this morning I better shut up.

Just one comment though:

My simple growth program has since been re-written by Van Dung Ly a student at the University of New South Wales. It runs on SGI now and grows tetrahedra rather than triangles. The virtual model looks a lot better but judging size, proportion and overall geometry is still very hard. The real model wins hands down!

[show real model]

But is this shape amorphous?

Well, what is amorphous anyway? In its most basic sense it can be translated as shapeless and I've defined it in so many different ways that the term has become truly blurred if not shapeless for me. What I have learned though is to accept that things change and we do not need to try and define everything in fixed and authoritative ways for eternity. One explanatory image that has survived all these metamorphoses is the image of clouds.

Clouds appear, they disappear, they constantly change shape, they block our view, yet we can fly through them, they have clear outlines or they dissolve into the surrounding sky.

According to a quote from one of Bernard Cache 's (www.objectile.com unfortunately no longer on-line) essays clouds caused some trouble in the Renaissance already:

Bernard Cache "THE TABLE AND THE FIELD - DRAWING BOARD, CONSTRUCTION SITE"	"For example, take the difficulties which were immediately experienced by the theoreticians of Alberti's day in trying to find a way of fitting clouds into the system of perspective which was emerging at the time".

Are we surprised that they had difficulties?

Clouds consist of large numbers of water molecules subjected to basic chemical and physical laws. The fact, that we can easily recognise a cloud as a cloud, however no two clouds are ever identical is due to the astronomically large number of construction elements (water molecules) opening up a vast playground for permutational and morphogenetic processes.

Humans in the Renaissance were not equipped to think in such large numbers - and neither am I - but computers these days should be able to handle elements in such quantities. And they are, if you use so-called particle systems in animation packages which in turn can be used to generate surfaces (we've seen a concrete example of this in Greg Lynn 's presentation if you remember the white stuff that came out of the slab used to generate the shapes for the Triple Bridge Gateway to 9th Avenue). But architects use CAD programs and why aren't these tools available in architectural CAD packages? Thus we arrived at the few words about drawing board methodology that I had promised earlier:

Most of the invited key-note speakers at this conference have done some major CAD bashing and I am more than happy to join in. [Darren Knight Gallery Project, Documentation 1996/97, 8 minutes in the background] - if you heard it all before just focus on the video.

In an authoritative volume on "Computer Aided Design: An Integrated Approach" - admittedly quite old from 1992 - the authors Hsu and Sinha praise CAD technology as a design optimization tool:

Hsu, Tai-Ran, Sinha, Dipendra K. "Computer Aided Design: An Integrated Approach" St. Paul. West Pub. Co. 1992, p.29	"Although CAD appears to be a new technology to many people, its real essence, in the authors' view, does not deviate significantly from traditional design philosophy and methodology. CAD gives designers much better and more effective tools with which to produce better-quality design work."

'Better quality design work' is seen as an optimised, yet traditional design consisting of geometric primitives and produced in an environment that mimics the working methods of the traditional architect/designer, i.e. 3 orthogonal and perspective views. Hsu and Singha are missing the point and have opted for a totally unimaginative and inadequate way in trying to assist architects by simply transferring their work-methods into the computer environment. It is no surprise then that architecture so far has not been able to free itself form the straitjacket of rectilinearity and reductivist geometry (Whilst this was certainly true 5 years ago with Frank Gehry being a notable exception - but then he is using aerospace software rather than architectural software - we have this amazing conference in which a great number of innovative practicioners have come together, i.e. Bernard Cache, DECOI architects, foreign office architects, Itsuko Hasegawa, Greg Lynn , Marcos Novak , Kas Oosterhuis/Ilona Lenard and Antoni Gaudi as 're-constructed' by Mark Burry).

John Frazer, in the following quote, is not so much criticising the interface design, but the fact that the construction elements offered to the architect through CAD programs are mostly reduced to primitive shapes:

John Frazer "An Evolutionary Architecture", Architectural Association, London, 1995, p.66	"It is ironic that the fixed ways of representing and abstracting building form which developed within the limitations of the drawing board and the techniques of projective geometry should have been carried over so directly into the computer. Geometrical forms could have remained plastic and fluid in the computer; instead they have become rigidified."

The same argument in a slightly broader context is presented in the editorial of 'assemblage 26':

Editorial "Computer Animisms (Two Designs for the Cardiff Bay Opera House)" in "assemblage", No.26, MIT Press, 1995, p.8

"The computer seems to be at its most powerful in the fields of biology and physics and, further, aspires itself to the condition of biomorphic intelligence. Computer discourse is replete with biospatial metaphors: pathways, networks, environments, structures, mutational systems, morphing, cyborg personas, prosthetic apparatuses, and so on. However, very little of this computer intelligence (technical or otherwise) has found its way into architecture. Form Z and Photoshop are now fairly widespread in architecture, as is AutoCad, but numerous technologies such as computer animation, virtual reality, stereo lithography, and other software/hardware applications are seldom used and, when they are, are frequently reductive and clumsy. So-called interactive computer technologies in architecture are mostly bad walk- throughs of the Parthenon."

And animators do not seem to have it much easier as Mark Dippe (Assistant Effects Director for "Terminator 2") explained in an interview:

Stephen Perrella "Interview with Mark Dippe - Terminator 2" in AD "Folding in Architecture", Vol. 63, 3-4, 1993, p.91

"One quality of human form, of any living form, is that they are difficult to represent in traditional computer animation because they are soft and have tissue that reacts and changes in very subtle ways. For instance, when we run, our muscles shake each time our feet hit the ground and impact our thighs. These soft-tissue, muscle and bone dynamics of living creatures in motion are very difficult to model because conventional computer graphics are essentially Euclidean; everything is rigid, polygonal and flat."

Since Mark Dippe's interview a number of tools have been developed to assist the animator in overcoming the 'Euclidean' problems. Common to all these new tools is a physical component affecting the basic geometry. Some of the first tools of the new generation that made it into commercial animation packages were particle systems which have the physical dimension of velocity as an attribute. Then came the metaballs with some form of gravitational forces ( Greg Lynn 's "blobby things") and eventually simple finite element systems - now being used to animate the bounce of hair when walking or skin deformation in the program ALIAS.

Ok., I think we have laboured the point enough and there is hope that things will change. We have established that CAD programs are not ideal for creating amorphous shapes: - structures like frozen clouds or the blue veins in blue vein cheese; buildings with the sensuously curved characteristics of rolling hills or the washed out shapes of driftwood - but why would we want to create them anyway?

First of all I think that they look better - just compare a sportscar with a delivery van.
Why is 99.9% of architecture designed like a delivery van?

Peter Cook in "Primer" has the answer (just one sentence but very much to the point):

Peter Cook "Primer" Academy Editions, London, 1996, p.30	"Much architecture is based upon geometrical habit (almost all rectilinear - it's so sensible, isn't it?)."

Greg Lynn , in the essay "Multiplicitous and in-organic bodies" attributes it to a prejudice of geometric idealism.

Greg Lynn "Multiplicitous and In-Organic Bodies" Architectural Design, Vol.63, No 11/12, 1993, p.31	"It is in fact most likely not utility, economy or structure that fixes architecture statically, but it's prejudice in favour of geometric idealism. Geometry provides the apparently universal language with which architecture assumes to speak through history, across culture and over time."

In a similar way Mathilde Lochert has identified an interesting relationship between geometry and cultural prejudices as quoted by Kim Dovey in a recent issue of Architecture Australia:

Kim Dovey "Architecture about Aborigines", Architecture Australia, July/August, 1996, pp.98-103

"Some of these issues emerge again in relation to the Brambuk Living Cultural Centre, designed by Greg Burgess with others. The centre was completed in 1991 to house and represent Aboriginal culture of the Gariewerd region in western Victoria (the Grampians). The written brief called for the use of "curvilinear forms and natural materials" and a participatory design process. Burgess' work has long been characterised by non-orthogonal geometries, curvilinear forms and the use of timber."

That's the intro about the building concerned and now the quote by Mathilde Lochert:

Kim Dovey "Architecture about Aborigines", Architecture Australia, July/August, 1996, pp.98-103

"In its simplest form, this criticism is that the building is problematic because its use of natural materials and curvilinear forms meets white preconceptions of a 'primitive' culture. To white eyes, a set of conceptual oppositions are set up around the building; white/black, sophisticated/primitive, culture/nature, straight/irregular. The building is said to reinforce a construction of Aboriginal people as primitive, natural and irregular. This critique has sources in postcolonial theory which suggest that in such a power structure, the native 'other' finds a voice only within the framework of a dominant discourse. And the State has an interest in seeing Aboriginal identity 'fixed' in built form; its dangerous, amorphous, power 'arrested'. From this view, Brambuk becomes a gesture of reconciliation which is at odds with the unresolved conflicts it represses."

This is a very interesting point. From my personal experience of the difficulties encountered in trying to create curvilinear shapes and from musings about people's fascination with the Sydney Opera House - yet reading about the problems of constructing it, I am almost tempted to turn some of those dualities around and claim that modernist buildings with reductivist rectilinearity represent the primitive way of building compared to the complexity and sophistication of curvilinear shapes, furthermore I am tempted to postulate that culture has been pushed in the wrong direction by incompetent people who were confused and threatened by nature's complexities and praised square-mindedness over irregularity.

Leaving this political argument aside, we should go back to the question whether there is a more profound justification for amorphous shape options than merely the preference for sensuously curved sportscars?

Throughout history there have been attempts at organic architecture. It hasn't been fashionable lately and therefore I'd like to read out the following quote by Imre Makovecz:

Imre Makovecz "ANTHROPOMORPHIC ARCHITECTURE - The Borderline between Heathen and Earth", Architectural Design, Vol.63, No 11/12, 1993, p.15

"Those who think they understand architecture usually approach organic architecture in terms of 'form'. But in viewing the architecture of Frank Lloyd Wright, Bruce Goff, Herb Greene, Rudolf Steiner, Lechner Odon, Antoni Gaudi, Kos Karoly or William Morris we are looking at completely different worlds from the point of view of form. I think it is the mode of thinking of their view of life which makes these people representatives of organic architecture. One of the significant things about this view is that they are searching for metaphors between the meta- nature of society and the created world. The abyss which opens again and again between these two is unbearable for them. To this end, Gaudi and Morris used plant metaphors, Greene used animal metaphors and Rudolf Steiner used human metaphors. It was absolutely essential for these architects to view the world as a continuity."

Taking this idea of the meta-nature of society beyond a metaphorical source for organic architecture, it is possible to see society itself as an amorphous construct; especially today's global, highly dynamic societies, capable of forming fast changing networks of constantly morphing configurations.

In seeing society itself as an amorphous construct, the individual being a basic construction element like the water molecules in a cloud, we might be able to derive more differentiated and complex thinking patterns which will describe our multicultural societies more appropriately than the current thinking patterns based on rigid classification systems and simplistic geometries.

But how can such abstract considerations take architectural form?

Apart from the design approach employed by me, i.e. highly irregular, non-symmetrical and non-rectilinear, the approach and verbal justification for Foreign Office Architects' Yokohama project represents a step in this direction: a dynamic, amorphous zone of "differential mediation" instead of a rigid border between

"Foreign Office Architects - Yokohama International Port Terminal" in Architectural Design, Vol. 65, No. 5/6, 1995, p.XIX

"... the 'local' city of Yokohama and the 'global' pacific cruise-liner network. It will operate as a mediating device involving the two large social machines that the new institution includes: Yokohama City Waterfront and the organisation of the cruise passenger flows. Both components of the programmes are used as devices for reciprocal 'deterritorialisations': a public space that wraps around a terminal, neglecting its symbolic presence as a gate, decodifying the rite of passage, and a functional structure that becomes the cast of an a- typological space, a 'landscape' without instructions for occupation, rather than a 'place'. The aim is to produce a mediation of a differential nature; a machine of integration that allows us to move imperceptibly through different states into degrees of intensity, diminishing the importance of the rigid segmentation that social machines - especially those dedicated to the maintenance of borders - usually produce. The proposed artifact will minimise the energy required to change state, articulating in a differential mode the various segments of the programme through a continuous variation of form; from local citizens to the foreign visitor, from 'flaneur' to business traveller, from voyeur to exhibitionist, from performer to spectator."

Thus a design rhetoric based on society as an amorphous construct might lead to built environments visualising the complexities in society and hopefully stimulating more differentiated thinking patterns.

One last question related to this amorphous business is the question

whether and how we can actually build these complex designs?

Well, the answer lies in automated fabrication technologies. Marshall Burns has written an authoritative book on the subject which gives a good overview of what is possible now and what will be possible soon. In a more trendy form William J. Mitchell mentiones automated fabrication with respect to a virtual clothing boutique.

William J. Mitchell "CITY OF BITS: Space, Place and the Infobahn", MIT Press, 1995, p91.	"Your detailed measurements are stored in a database somewhere. There is no inventory: when you place an order, computer-controlled machinery accesses these measurements and fabricates the item perfectly to your size. (Fitting rooms become unnecessary, and your size is never out of stock.) It is then delivered to your door."

Whilst I think that this will be a possibility technically - I think it is questionable whether society would accept to have their personal measurements stored in a database. But that is not the point: The point is that we have technologies available that mass-manufacture individualised products.

Leaving the clothes scenario behind, the book 'Earth Moves' by Bernard Cache has a few pages on how manufacturing technology will affect society in what I have called a major shift from a rigid, mathematical to a dynamic, physical paradigm which means that the shift towards alternative shape generation - 'calculating shapes' in Bernard Cache 's words - will overcome the architect's difficulties in creating amorphous architecture - automated fabrication techniques will then overcome the architects difficulties in getting them built.

A concrete example: We all know about stereo-lithography which could be called an additive fabrication method in contrast to computer controlled milling which would be a subtractive fabrication method. Other additive fabrication methods are for example 3D welding and droplet deposition. One material used for droplet deposition processes is ceramics and it is quite easily imaginable that this method could produce thousands of individually shaped bricks as part of a highly curved and twisted wall. All it needs then is a brick laying robot which is currently being developed - and to some degree functional already - at the Technical University in Delft, Netherlands. This is one possible scenario of how your design could be realised in the future. Another scenario would be an improved version of folded sheet material as used in my Darren Knight Gallery Project.

Seeing that I have just started to work at the CSIRO I would like to say a few words about what I will be doing there. Well, I have strongly criticised the drawing board approach as a CAD user interface so what would be my preferred design methodology?

Ideally I would like to walk around an empty room, waving my arms, clicking my fingers, pushing a control vertex out here, go across to the wall and pull it in by half a meter, i.e. design through body movements on a life-size scale. Interestingly enough looking at certain developments in music, I realised that the sensor technology is available and that it will not be long before this approach could become a reality.

A collaboration of musicians performing under the name "Sensorband" have develped a set of virtual instruments described in the following way:

http://www.sensorband.com

"Each musician uses interactive electronics in a different way - making music from ultrasound, infrared, and bioelectric signals. [...] Edwin van der Heide is the "vocals", working with ultrasound based 3D handheld controllers to perform infinite time stretching on digitally sampled vocal sources. Atau Tanaka is "guitar-bass" using muscle (EMG) and gesture to play air guitar, and low bass parts. Zbigniew Karkowski is "drums", playing virtual percussion in a cage armed with infrared beams."

Within this group it is especially the way in which Zbigniew Karkowski performs which, I think, could be successfully transferred into an architectural full scale laboratory equipped with a virtual immersion system.

The main difference between music and 3D modelling, and probably the main stumbling block so far for not transferring the technology to architectural modelling and design, is the fact that once the movements of the performers have been detected, sound generation or sound manipulation of pre-recorded material is relatively straightforward whereas the 3D designs need to be processed into rendered imagery and communicated back to the performer via virtual reality goggles or similar devices in real time. Render algorithms and computing power have been prohibitive so far.

Once it is possible to create shapes and designs in the above way through natural movement, I think that the result will show its human involvement in a similar way to the sound creation as described by Sensorband when performing on the instrument called 'Soundnet':

http://www.sensorband.com/soundnet/index.html

"It is a giant web 11 meters x 11 meters, created with 16mm thick shipping rope. At the end of the ropes are eleven sensors that detect stretching and movement. We perform the instrument by climbing on it, all three of us at once. [...] The rope, the metal, and the humans climbing it take on an incredible physicality, and focus more on the organic nature and the human element of interaction rather than on mouseclicks and screen redraws. This puts the emphasis in man-machine interaction back towards the human side. [...] The physical nature of our movement meeting the virtual nature of the signal processing creates a dynamic situation where we deal directly with sound as our material. Through gesture and pure exertion, we sculpt the sound to create sonorities emanating from the huge net."

Creating, sculpting in such a way, I think, is well worth working towards.

Thank you

Back to Entry Page

For general feedback or specific questions please contact me on
horst@vislab.usyd.edu.au,