Cyclical Iterative Design Process, Learning From Experience

Summary PhD Dissertation The Cyclical Iterative Design Process and Design Environment learning from experience

Martin Smit

The Cyclical Iterative Design Process and Design Environment Architects usually regard their designs as unique. Using elements from previous designs is generally considered to be a lack of creativity. A design is not usually looked upon as as being a product. Parts of buildings can be products, but whole buildings are not considered to be products. There are however design assignments with a repetitive character, in which large numbers are being designed and realised on the basis of a certain concept. In the retail business for instance, the same concept may be applied to a relatively large number of locations. The concept is standard and the design assignment for each location is also standard. The implementation of the concept, however, is seldom identical. Even with a strong concept and an equally strong implementation process, each location is unique with its own specific conditions, and each implementation has its own unique aspects.

Dissertation Ter verkrijging van de graad van doctor aan de Technische Universiteit Delft, op gezag van de Rector Magnificus prof. dr. ir. J.T. Fokkema voorzitter van het College voor Promoties, in het openbaar te verdedigen op,

From participation in practice in designing and implementing a retail concept, the idea evolved that there is no such thing as a standard design or a standard implementation process. From each design realised there is something to be learned. The ability to look at the finished product is an opportunity to learn. This way each product can be better than the last product.

maandag 11 februari 2008 om 12.30 uur door Martin Willem SMIT bouwkundig ingenieur geboren te Soest op 18-11-1957

Research Question The research question that evolved is whether it is possible to develop a design process in which an architect can acquire knowledge and experience efficiently from comparable design projects and use them in current and future projects without sacrificing creativity.

Dit proefschrift is goedgekeurd door de promotor: Prof. dr. ir. A.C.J.M. Eekhout

Existing design processes in architecture and other design disciplines Several design processes used by architects are analyzed in chapter 2. Except for the design process of Eekhout, they all contain a number of sequential phases and lack a degree of feedback. There is almost no possibility to improve the result or to learn from experience.

Samenstelling promotiecommissie: Rector Magnificus, voorzitter Prof. dr. ir. A.C.J.M. Eekhout, Prof. dr. ir. J.J.N. Lichtenberg, Prof. dr. ir. J.W.F. Wamelink Prof. dr. ir. M.J.L. van Tooren, Prof. ir. A. Beukers, Prof. ir. B. Molenaar, Prof. ir. D.E. van Gameren

Technische Universiteit Delft, promotor Technische Universiteit Eindhoven Technische Universiteit Delft Technische Universiteit Delft Technische Universiteit Delft Technische Universiteit Eindhoven Technische Universiteit Delft

Dr. ir. A.I.M. Voorbij heeft als begeleider in belangrijke mate aan de totstand­ koming van het proefschrift bijgedragen.

In chapter 3, several design processes from other design disciplines, such as industrial design and aeronautical design, are analyzed. There is usually a high degree of feedback due to the cyclical aspect of the design process. Improvements are an integral element of the design process. The conclusion is that the aerospace design methodology is the one most relevant to architectural design. As well as the already mentioned cyclical aspects, the customizing aspect plays a role. How a number of architects have used methods and technologies from the aerospace industry or were inspired by them is analyzed in chapter 4.

summary dissertation

1

Beside technological inspiration, a number of architects are inspired by the fact that in aerospace there is a strong and continuing focus on improvements.

A number of case studies in architecture, as well as in aerospace, were used to achieve a better understanding of these stages for architects:

The Cyclical Iterative Design Process Based on a design scheme from Thomke and inspired by aerospace, the socalled “Cyclical Iterative Design Process” is defined in chapter 5 (figure S-1). The design process consists of 4 stages: • Re)definition of the specifications and design development; • Production (Model, Mock-up, Prototype and Product); • Testing / Using; • Analysis and Evaluation.

(Re)definition of the specifications and design development The specifications for a retail project are analysed. The continuous improvement of the Boeing 737 and the development of housing design by the Dutch architectural practice Inbo are studied. From the study of the Boeing 737 it appears that by being sensitive to internal design aspects, such as technical developments and, more important, external factors such as socio-economic circumstances and legislation, a product can be updated time and time again. This way it has managed to stay a state-of-the-art product over more than forty years. In architecture, it seems to be especially important to define the right requirements and to incor­ porate a certain degree of flexibility in the specifications from the outset.

(Re)definition of the Specifications and Design Developement Producing

Analyzing and Evaluating

Iterations Model

Testing / Using

Figuur S-1 Basic Model Cyclical Iterative Design Process

Mockup

Prototype (Pilot)

Product (Location)

Production Different forms of production are studied. They can have several embodiments: Models, mock-ups, prototypes (or pilots), and the eventual product (at location). A set of definitions was developed that can also be used in architectural design. Testing / Using In the testing that took place during the development of the Boeing 777, a number of aspects became apparent, such as testing in a user environment and the emphasis on test criteria. A new test regime for architectural design was developed. Analysis and Evaluation Three cases are looked at. Each of the cases has its own specific evaluation method. For a retail project, the projects are evaluated in weekly sessions attended by most of the participants involved in the project. Each participant gives his opinion and suggestions for potential improvements. In addition to this process of feedback, an evaluation tour was organised. In this tour, which took place once or twice a year, a number of realised branches (locations) were visited by everybody involved. Each participant gave his or her opinion, and the results were collected in a booklet with comments and suggested improvements. A study of an upgrade of one of the locations in the retail project shows that the recommendations of the evaluation can be implemented for relatively little money and effort. The result is a considerably improved branch. The four stages are worked through successively. At the end of each cycle, the result is analyzed and evaluated. This serves as input for the next cycle. The iterative principle is an integral part of the Cyclical Iterative Design Process, in which each cycle should lead to an improvement of the quality. Maximum quality in the minimum number of cycles. If after the first cycle the result is analyzed and evaluated and the required quality has not yet been achieved, another cycle is started. Should the quality still not be at the required level at the end of the second cycle, another cycle is started. This process is repeated until the required quality has been achieved. Application of the Cyclical Iterative Design Process In chapter 6, the Cyclical Iterative Design Process is applied in a number of cases. The introduction of the Cyclical Iterative Design Process resulted in summary dissertation

3

continuous improvement in these cases. Development of an Internet reading room for the application in retail branches took place in a relatively short timeframe and for a relatively low budget. Mock-ups were produced for the facade elements and a laboratory of a large research facility. The mock-ups resulted in significant design improvements and refinements. The costs of producing the mock-ups were minor in comparison to the overall budget. This way, potentially costly failures could be avoided. A potential application was developed for a research project on industrial housing called the “Concept House”. For repetitve design assignments based on the Cyclical Iterative Design Process, an implementation process was developed that has proven to be applicable in real situations.

tion Management. The cyclical concept is applied to these four elements and the so-called Cyclical Iterative Design Environment is defined. Figure S-2 shows the relationship between the Design Process and the Design Environment. Tools are developed to optimize the four elements through a cyclical process to achieve the best possible environment for creating the best possible product. Contract Structure Contracts are usually written from a purely financial perspective and in the Dutch situation more or less dominate the entire design process. A designbased contact structure was developed. Office Organization By introducing the cyclical concept in the office organisation, the organisation takes a performance-orientated focus. There is a continuous process of improving by evaluating the results, as well as the factors that determine the result.

The case studies showed that the Cyclical Iterative Design Process can contribute to a significant improvement in the quality of the design and the final product. The Design Environment The Design Environment is defined in chapter 7. It consists of four main elements: Contract Structure, Office Organisation, Quality System and Informa-

Quality System With a relatively minor modification of the existing evaluation element of the quality system, the performance of the system can be improved dramatically.

Design Environment

Information Management

Design Process

(Re)definition of the Specifications and Design Developement

Contract Structure

Producing

Design Process

Analyzing and Evaluating

Iterations Model

Mockup

Prototype (Pilot)

Product (Location)

Testing / Using

Quality System

Office Organisation

Figuur S-2 Cyclical Iterative Design Process and Design Environment

Cyclical Iterative Design Process and Design Environment

summary dissertation

5

By emphasising the moment of evaluation, either internal or external, the organisation can acquire a more product-oriented attitude. The users will see the quality systems more as a means to an end and to achieve a certain performance level. The quality system is less a target in itself and becomes a tool. Information Management Chapter 10 deals with the cyclical aspects of information management. It describes an existing paper-based information system in a retail project, which is transformed into a digital information system. Where the paper system was sequential, the digital system is cyclical. The digital system contains strong feedback elements and all information in the system is continuously updated.

Conclusions and Recommendations Conclusions On the research method Extensive use was made of knowledge and experience in other design disciplines. It appears that this can generate new insights into the design discipline of the researcher. The research is qualitative rather than quantitative by nature. The aim was not so much to reduce the reality to a numbers-based set of rules, but to get an understanding of the processes that lead to the outcome. In the research there was a constant tension between on the one hand generating new insights on the basis of existing knowledge and on the other hand generating fundamentally new knowledge and insights that can be implemented. The cyclical interaction between the two approaches generated fundamentally new knowledge and insights. On the application of the Cyclical Iterative Concept The Cyclical Iterative Concept is not a system that, when used automatically, will lead to a better product. The reality of building is much too complicated for that. The process should be considered more as a way of thinking for everyone involved in the process of building that can be used to stimulate creativity, and in the end will lead to a better product. The Cyclical Iterative Concept is especially useful in recurrent design assignments. From their experience, clients seem to realise that the Cyclical Iterative Concept can lead to a significant improvement in their assignments.

Recommendations For a wider application of the Cyclical Iterative Concept In housing assignments there is a certain repetitiveness, in the projects themselves but also in the scope of different projects. The Cyclical Iterative Concept can be used to make the experience more explicit and useful not only for the architect but also for other participants in the project.

In utility assignments, the projects are one of a kind. Losses in reduced efficiency will not easily become apparent. On the level of building components, the Cyclical Iterative Concept is more applicable. Components can be built, tested and analysed. Specifications can be altered and concepts can be improved. Especially in high risk components, failure costs can be significantly reduced. For application of the Cyclical Iterative Concept by architects, clients, manufacturers and building managers It appears that once the participants have actively participated in a process based on the Cyclical Iterative Concept, they start to realise that they can achieve better results with less effort. Application of the Cyclical Iterative Concept, however, requires a change in attitude from the participants: Architect In the past, the architect was in a very strong position and was usually the determining factor in the design and building process. Nowadays, the role of the architect is much more limited. Especially in larger housing projects, the architect is hired to develop a pleasant concept. During the remainder of the process, they are merely regarded as aesthetics advisors. The cyclical way of thinking can give the architect greater involvement in the whole process and more influence on the final result. This overall aesthetics management will eventually improve the project, but it requires a more open approach from the architect to the input from the other parties. For most architects, organisation and especially management are negative and unfortunately necessary “sideeffects”. An architect working with a cyclical approach has to have at least basic organisation and management skills. Client (principal) The principal (client) usually prefers to divide the whole process into a number of rather distinct phases. For the principal, this setup looks more manageable. The downside is that each phase has its own participants and the interaction between participants in these phases is usually limited. This leads to a loss of information and the quality of the resulting product is less than could potentially be realised. The Cyclical Iterative Concept can reduce this loss of information and stimulate the interaction between the participants, making the whole process more transparent for the client as a result. Contractor (manufacturer) The manufacturer (contractor) is traditionally focused on producing that which others have defined. Using the Cyclical Iterative Concept requires a greater amount of involvement as well as responsibility, but also gives the manufacturer a greater chance to make a better product. Building Manager The management consultant is usually involved in dividing up the whole process as mentioned above. He will usually try to determine the structure of the process and through this control it as much as possible. This implies considering the reality as fixed. Nonetheless, it appears that the reality is very dynamic. By using the Cyclical Iterative Concept, the interaction between the different

summary dissertation

7

stages in the project and between the participants improves, and at the same time becomes more manageable. For an immaterial technology transfer Technology transfer is usually about the application of material and technologies from one discipline to another. The transfer is only material. Immaterial technology transfer is at least as valuable and potentially much more. It is not only about the product and processes themselves, but also about the thoughts, insights and vision behind those processes. For the architect as researcher Although architects are involved in research in day-to-day practice, that research is mostly project- and results-driven, whereby a result is sought in a limited timeframe. There seems to be no tendency towards a written or documented body of knowledge concerning learning processes. Project-unrelated research could lead to a body of knowledge, not only for the architect doing the research, but also for other architects, thus raising the professional level of the architectural profession as a whole. As we know from our own experience, project-unrelated research can be time consuming and therefore “budget consuming”. Only the very large architectural practices can therefore afford it. It is perhaps the duty of these large practices to engage in what can be called architectural research. My own PhD research can also be considered architectural research. There is a continuous interaction between (architectural) practice and (architectural) theory. Figure S-3 shows the research structure.

Colofon Practice

induction Case study

Case study Case study Case study Case study

Uitgave februari 2008 For more information please contact me: dr. ir. Martin Smit Martin W Smit Architects Valeriaanstraat 84 3765 ER Soest T +31 (0)35 58 80 381 F +31 (0)35 60 11 497 M +31 (0)6 57 55 92 52 [email protected] www.mwsarchitects.com

deduction

Figuur S-3 Research Structure

Theory

ISBN: 978-90-810140-4-5 Alle rechten voorbehouden. Overname uit deze opgave is alleen toegestaan met schriftelijke toestemming van de uitgever.