Conference PaperPDF Available

MAKING PAVILION AS PEDAGOGY

Authors:
Y. Ikeda, C. M. Herr, D. Holzer, S. Kaijima, M. J. Kim. M, A, Schnabel (eds.), Emerging Experience in
Past,Present and Future of Digital Architectur e, Proceedings of the 20th International Conference of the
Association for Computer-Aided Architectural Design Research in Asia CAADRIA 2015, 000000. ©
2015, The Association for Computer-Aided Architectural Design Research in Asia (CAADRIA), Hong
Kong
ARTICULATED TIMBER GROUND, MAKING PAVIL-
ION AS PEDAGOGY
PAUL LOH1
1University of Melbourne, Melbourne, Australia
paul.loh@unimelb.edu.au
Abstract. Designing and making a pavilion within a studio setting has
been undertaken by various educators and researchers as a valuable
pedagogy in the past 10 years. It aims to construct a collaborative en-
vironment that allows students to develop an integrated approach to
learning; through association, teamwork and creative collaboration.
Usually the tacit knowledge applied and acquired through making,
and the knowledge of design strategy and analysis are separated in the
way they are taught; it is often difficult to integrate these within the
same coursework which often leads to students using digital software
and fabrication tools as problem solving devices. This paper looks at
an integrated approach to learning computational design and digital
fabrication through the making of a pavilion by a Master level design
studio. The paper discusses the pedagogy of making through creative
collaboration and integrated workflow. It focuses on the use of digital
and physical prototypes as devices to stimulate an oscillating dialogue
between problem solving and puzzle making; a counterpoint for stu-
dents to develop and search for new knowledge in order to create per-
sonalised learning experience. The paper concludes with an examina-
tion on the limits of digital prototype when interfaced with physical
environment.
Keywords. Digital Fabrication; Collaborative Design; Design Work-
flow; Pedagogy, File to Production
1. Introduction: Making pavilion as pedagogy
Designing and making a pavilion within a studio setting has been undertaken
by various educators and researchers as a valuable application of pedagogy
in the last decade (Walker and Self 2011). It aims to construct a collabora-
tive environment that allows students to develop an integrated approach to
Y. Ikeda, C. M. Herr, D. Holzer, S. Kaijima, M. J. Kim. M, A, Schnabel (eds.), Emerging Experience in
Past, Present and Future of Digital Architecture, Proceedings of the 20th International Conference of the
Association for Computer-Aided Architectural Design Research in Asia CAADRIA 2015, 23–32. © 2015,
The Association for Computer-Aided Architectural Design Research in Asia (CAADRIA), Hong Kong
2 P. LOH
learning; through association, teamwork and creative collaboration (Kalay
and Jeong 2003). Usually the tacit knowledge applied and acquired through
making, and the knowledge of design strategy and analysis are separated in
the way they are taught (McCullough 1998, Salama 2008). It is often diffi-
cult to integrate these within the same coursework due to time constraints,
which leads to students using digital software and fabrication tools as prob-
lem solving devices.
This paper looks at an integrated approach to learning computational de-
sign and digital fabrication through the making of a pavilion by a Master
level design studio. It discusses the pedagogy of making with focus on the
use of digital and physical prototypes as devices to stimulate oscillating dia-
logue between problem solving and puzzle making in digital design and fab-
rication. Kalay (2004) suggested that design processes necessitate the need
to move between the two above mentioned paradigms in order to achieve a
balance between the goals and the solutions. In this paper, the author sug-
gests that this dialogue provides a counterpoint for students to develop and
search for new knowledge in order to create personalised learning experi-
ences.
2. Background
2.1 BRIEF
The studio set out to design and fabricate a pavilion in 12 weeks, with a con-
straint of 25 sheets of 18mm plywood supplied by our sponsor, OH&S re-
quirements and a site boundary of 4 x 4 x 3 metres high located within the
University grounds. The brief called for a pavilion that examined the rela-
tionship between ground and envelope while engaging with the context of
the campus. The objective was to generate a workflow that is manageable
and yet complex enough to stimulate an intense office environment for the
students. Formally, the project was developed from ruled surface geometry.
The vertical studio had 14 participating students and two studio leaders.
The design of the pavilion was selected through a two phase competition
structure; refer to section 3.0
2.2 FILE TO PRODUCTION WORKFLOW
The scale of the pavilion allowed students to be immersed in and experience
first-hand file-to-production workflow. There are two principal workflow
procedures that are explored in this project. Firstly, the continuous feedback
loops between digital design information and physical prototyping (Knapp et
al 2014). This took place from the early design stage and continued through
24
ARTICULATED TIMBER GROUND 3
to the final prototype stage. Secondly, the process of refining the digital in-
formation, free of representational annotation, allowed the students to inter-
rogate modelling information for fabrication sequence and assembly (Garber
and Jabi 2006). This included preparing files for production and checking
procedures. Both of these workflows are widely documented but rarely
taught in design studio concurrently.
Figure 1. Articulated Timber Ground pavilion within the University ground.
2.3 DIGITAL FABRICATION AS WICKED PROBLEM
While a typical design studio concludes at the representation of design pro-
posals, even among digitally focused studios, this studio challenged students
to take the design one step further to construction. It forced the students to
examine the constructability of their design in greater detail and remove their
perception of digital fabrication as model making techniques only. By inte-
grating buildability issues with the computational workflows demands a con-
tinuous feedback loop between digital analysis and prototyping with 1:1
physical testing and making. Students are confronted with the wicked prob-
lem of design (Kalay 2004).
In the context of learning digital design and fabrication, the wicked prob-
lem starts when digital data interfaces with the reality of construction, budg-
et and material constraints. Here, students had a preconception that the preci-
sion in digital fabrication and workflow naturally leads to a zero tolerance
construction process. The reality is a messy business of learning to balance
design integrity with constructability, budget, manoeuvring through universi-
ty OH&S requirements, variable site conditions, just to name a few. The
above pedagogy shifts the output of the design studio from the hypothetical
to the real; what Salama (2008) called the systematic pedagogy (as opposed
25
4 P. LOH
to the mechanistic pedagogy) where students grapple with the different as-
pects of the bodies of knowledge and restructure them to formulate new
knowledge and personalized learning experience.
3. Approaches to learning
A common tendency among students which the author observed from both
undergraduate and graduate teaching of digital design and fabrication is that
students generally have an understanding of the impact of digital fabrication
in contemporary practice through various reading. However, understanding
of the practical workflow itself is limited; this may be due to the fact that
knowledge gained is built on a foundation of limited duration (Boza 2006).
From this context, the studio focuses on the following aspect of learning:
x Collaborative design methodology. This is explored through a two phase
competition at the start of the studio. Four design teams are formed through
self-selection. By Week 3, two proposals are shortlisted; the eliminated teams
joined the shortlisted teams to continue to their bid for the project. By Week
6, a jury panel selects one winning scheme to be constructed by the entire
studio. After Week 6, the entire studio operates as a single office; refer to
Figure 2. The author role changes from tutor to consultants/client representa-
tive.
x Constructing a project specific design to fabrication workflow. This is ex-
plored during the two competition phases where each team formulates their
design based on the concept of ground and envelope learning through prece-
dent study. Each team grapples with their own design workflow with an em-
phasis on developing both digital and physical prototypes. This is the stu-
dents’ first attempt at constructing a feedback loop between design and
fabrication.
x Develop tacit knowledge of tooling and its relationship to the design process.
During the competition phases, all the design teams are introduced to a flat-
bed 3-axis CNC router; from tooling to toolpathing. This expands the fabrica-
tion repertoire of the students and forces them to make a leap from model
making to large scale prototyping. During this period, individual students
build up specialist knowledge of the machining process. This proves to be
exceptionally useful during construction phase.
x Examine the haptic relationship between fabrication, material, tooling and the
overall design process; the messy business of making. Here, students are
challenged to take on a variety of roles; as project coordinator, team leader,
quantity surveyor, designer and maker. The feedback loop between digital
and physical prototypes becomes multi-layered compounded by environmen-
tal and structural digital analysis as well as information gathered from physi-
26
ARTICULATED TIMBER GROUND 5
cal prototypes; buildability, construction sequence, fixing types and load test-
ing results of 1:1 prototypes at the Engineering department; refer to Figure 3.
Figure 2. Competition structure showing design selection
3.1 LEARNING WORKFLOW
Like similar digital fabrication research in recent years, this project utilised
McNeel Rhinoceros, with Grasshopper v0.9.0076. Structural analysis is car-
ried out using Karamba, a plug-in for Grasshopper. Solar analysis is under-
taken using Geko, a plug-in for Grasshopper. This suite of software has be-
come a standard tool kit for the design studio. By keeping within the same
modelling software, it facilitates a coherent collaboration (William et al 2014)
where feedback loops between design and analysis are direct. In contrast,
Autodesk Vasari is used for wind analysis. It remains 'external' to the Rhino
modelling environments where feedback into the design are primary based
on judgement through iterative analysis instead of direct response to geomet-
ric changes.
The studio recognised that it was not possible for students to learn and be
involved in every facet of the project. Students were encouraged to special-
ise towards the construction phase of the studio, developing a particular skill
set learnt during the Phase 1 competition stage. They were categorised into
the following team of expertise;
1) The scripting team was responsible for the Grasshopper definition
of the global form and the fabrication script - the digital prototype.
27
6 P. LOH
2) The prototyping team was responsible for the development of
physical prototypes; joints systems as well as learning the opera-
tion of the CNC machinery and generation of toolpath cut file.
3) The analysis team's task was to bridge between the digital and the
physical prototypes; providing feedback, constraints and opportu-
nities to the design team.
The three teams operated concurrently; refer to Figure 3. In order to
manage the design workflow, tutorial sessions became team meetings
chaired by the studio leaders; problems were aired, discussed and worked
through while new issues, analysis and problems were flagged, similar to a
coordination meeting in any building project. The meeting concluded with
the cost review and programme to ensure the project hit the required mile-
stones. An action list was issued so students felt accountable for their own
work. There became a sense of ownership over the project as opposed to au-
thorship. The organization was by no means perfect but nevertheless it
showed cross over and sharing of skillsets, resources and knowledge; what
Kalay (2003) called creative collaboration.
Figure 3. Diagram illustrates the crossover of roles between teams during Phase 2
4. Limits of prototypes
As Neil Denari (2012) explained, the digital workflow is not without its limi-
tation; the precision of the computational process engenders a myth or faith
in a perfect error free realization. He compares building workflow with those
of product design; where precision is achieved through serial iteration, build-
ing on the other hand has other factors such as tolerance, human error and
unpredictability in material to complicate the process. This means that the
accuracy in building the physical object is a matter of degree of closeness to
digital prototype quite unlike product design. Contemporary digital fabrica-
28
ARTICULATED TIMBER GROUND 7
tion workflow is akin to product design in the sense that building, as Denari
(2012) pointed out, exists as a digital prototype where the build-up of data
prior to construction represents a kind of laboratory version of the building
itself’.
4.1 OPEN VS LINEAR SYSTEMS
In this project, students were faced with the discrepancy between their digi-
tal and physical prototypes. The scripting team treated the digital infor-
mation akin to product design, inputting design data and designing its layers
of algorithmic structure to a near complete set of information. Up until the
stage of the final check of the digital model, the model consisted of the cut-
ting profiles and drill holes positions only. As the exact joint type varied
throughout the system, it was not assigned until the final stage of the ‘baked’
process. Here, the digital prototype remained an open system where data
could continue to be altered until the last minute. However, despite openness
of the system, it became obvious that once all the parameters were identified
and fixed, the digital prototypes risked becoming a problem-solving tool,
which required another set of dialogues to make it responsive again.
In contrast, the physical prototype team evolved the joints system as a tri-
al and error procedure. The aims of the physical prototypes were:
x To develop a rubber joint that allowed the seat of the structure to move local-
ly to absorb ergonomic movement when one sits on the structure; the rubber
insert bounces the joint back to its original location like a cushion.
x To develop a so called washer joint that dissipated the ergonomic movement
across the chain link surface; without this, the surface would remain static.
x To test the construction sequence. The structure was designed to be assem-
bled on site in chunks. The design team articulates the rhythm of the structure
so the ruled surfaces could have a varying level of porosity. This reduced the
amount of material required.
Although empirical in nature, the physical prototyping process is less
open and inherently linear as judgement and design decisions can only be
made once the prototype is completed and tested at each stage. A number of
reasons necessitate this linear procedure. Firstly, no computational software
as yet available to the team can deal with structure and movement analysis
on the localised level of the joint prototype. Secondly, the joint movement
requires the ruled surface to act as a network. It was more immediate to con-
struct 1:1 segments of the design to test the slight shift in movement; the
movement was in the order of 15-50mm. Lastly, the rubber and latex used in
29
8 P. LOH
the joint are non-standard and only through evolving the geometry through
1:1 testing could constraints be identified; the puzzle making paradigm.
From observation, all the above issues concerned the resolution of the pro-
ject where it is either too fine or too complex to reproduce in the digital pro-
totype.
The discrepancy between the physical and digital prototype became ob-
vious during the fabrication and installation process. Variations in material
thickness, consistency of plywood, existing ground levels and human errors
all contributed to the messy business of making.
4.2 CHECKING PROTOCOL
In this project, upfront investment in the Grasshopper definition as an open
system allowed the design team to maximise their design period. There is a 1
week overlap between design and fabrication; that is to say the design con-
tinued to be adjusted while other parts of the structure were fabricated at the
same time.
The concurrent workflow discussed in 3.1 enabled the above efficiency
but meant a robust checking protocol was necessary. While normally check-
ing protocols are on the validity of the data, because of the compressed time
frame for this project, the team (with the assistance of the studio leaders)
needed to undertake both design and information check simultaneously.
The aims of the checking protocol were:
x To identify geometric errors; both design and computational
x To check for completeness of information in fabrication cut files
It took 3 cycles of checks before the information was ready for fabrica-
tion. Each check was carried out manually through visual inspection of the
digital prototype, layer by layer. To focus on one particular issue that arose
during the design and fabrication process, namely the position of the drill
holes relative to the cut edge. This example illustrates the feedback process
between digital and physical prototypes and how the checking protocol al-
lowed the students to learn from a systematic pedagogy (Salama 2008).
Our structural consultant had given the team a rule of thumb for the fix-
ing position relative to the cut edge of each profile; typically 4 x diameter of
the hole. This is to ensure there is sufficient material strength in the joint.
When developed in the digital model, the radius of the input curve was too
big for the stepping joint resulting in large segments where the fixing posi-
tion is too close to the edge of the profile. The team compensated this
through articulating the joint connection profile resulting in the bone-like
30
ARTICULATED TIMBER GROUND 9
structure which was subsequently load-tested at 1:1 scale. This still resulted
with a small number of fixing positions shy of the parameter set by the engi-
neer. These small numbers of joints were identified for the team to resolve.
The team devised a simple and effective solution where the constraint di-
ameter is extruded to highlight areas of error using Boolean logic as mean of
visual inspection.
Figure 4. (Left) Joint design showing constrain diameter; (Middle) constrain diameter
implemented over design, (Right) Boolean intersection to highlight area of error
The above example questions the type of errors in a digital workflow.
Here, the word ‘error’ merely defines what is outside of a defined bandwidth
of judgement or criteria. For example, errors in geometry for this project can
only be judged as errors because they had not matched or satisfied the crite-
ria of structural integrity or fabrication constraints based on tooling. From
the above, it is useful to observe that errors in the digital prototype and phys-
ical prototypes are different in nature. In physical prototyping, the error is a
question of tolerance; numeric accuracy of digital fabrication against physi-
cal variant in material thickness and ground condition. Here, tolerance is a
range of data that can be resolved through anticipating the physical discrep-
ancy.
In digital prototyping, scripting errors are typically either geometrically
based, where a numeric discrepancy results in visually different geometry or
is consequence of input errors. While most input data can be constrained
within the digital prototype, it requires design judgement to question (and to
qualify) the information and its output as either acceptable or as error. As
this potential error is within the digital structure itself, it could only be iden-
tified through visual inspection of the digital prototype. To perform a digital
check procedure would require the geometry to be completely associative.
The authors’ opinion is that this will only work in product design where seri-
al testing and checks are embedded within the procedure. Returning to De-
nari’s cautions earlier, it is neither sustainable nor practical for buildings to
achieve a level of zero error.
31
10 P. LOH
5. Conclusion
This paper presented digital design and fabrication as a systematic peda-
gogy which, when operated concurrently and practiced in the form of mak-
ing a full scale structure, sets up an environment for integrated learning and
creative collaboration among students that is akin to the reality of an archi-
tecture practice. It fosters a personalised learning experience through struc-
turing dialogue and feedback loops between problem solving and puzzle
making as design paradigm. Prototypes, both digital and physical are devices
that enable this dialogue. This paper highlights the messy nature of making;
when digital information interfaces with the physical environment.
Acknowledgements
David Leggett, teaching partner. John Bahoric, structure consultant. Students: Norhasnani
Azman, Farheen Dossa, Jack Hinkson, Rachel Low, Siavash Malek, Amanda Ngieng, Ridho
Prawiro, On On Tam, William Varrenti, Bo Woon Wong, Bingquan Zhang, Qingyue Zhang,
Jia Jia Zheng, Shaobo Zhu. Dr Jane Burry for feedback and critic on this paper and project.
The pavilion is sponsored by Maxi Plywood, Melbourne School of Design and Power to
Make.
References
Boza, L.: 2006, (Un)Intended Discoveries Crafting the Design Process. Proceedings of the
25th Annual Conference of the Association for Computer-Aided Design in Architecture,
150-157.
Denari, N.: 2012, Precise form for an inprecise world. In: Marble, S. (ed.) Digital Workflows
in Architecture, Birkhauser, Basel.
Garber, R. A. J. W.: 2006, Control and Collaboration: Digital fabrication strategies in
academia and practice, International journal of architectural computing, 04, 121-143.
Kalay, Y. A. J. Y.: 2003, Collaborative design simulation game. CAADRIA 2003 Proceedings
of the 8th International Conference on Computer Aided Architectural Design Research in
Asia, 745-758.
Kalay, Y. E.: 2004, Architecture's New Media, The MIT Press, Cambridge, MA, US.
Knapp, C., Neslon J and Parsons M.: 2014, Constructing Atmospheres. Rethinking
Comprehensive Design: Speculative Counterculture, Proceedings of the 19th
International Conference on Computer-Aided Architectural Design Research in Asia
CAADRIA, Kyoto, 149158.
McCullough, M.: 1998, Abstracting Craft: The practiced digital hand, MIT Press
Massachusetts.
Salama, A., M.: 2008, A theory for integrating knowledge in architectural design education.
International Journal of Architectural Research, 2, 100-128.
Walker C. and Self M.: 2010, Making Pavilion, AA Publication, London, UK.
Williams, N., Bohnenberger, S., and Cherrey, J.: 2014, A system for collaborative design on
timber gridshell, Rethinking Comprehensive Design: Speculative Counterculture,
Proceedings of the 19th International Conference on Computer-Aided Architectural
Design Research in Asia CAADRIA, 441450.
32
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
The integration of digital tools currently being used in many schools and offices with Computer Numerically Controlled (CNC) hardware, has allowed architects to exert a far greater degree of control than they have previously been afforded. It is precisely this control that enables greater collaboration during design phases between architects and fabricators. However, the impact of this integration on academia and small practice is unknown. Several questions remain to be answered regarding teaching fabrication techniques and identifying strategies suitable for adoption in small firms. This paper investigates digital fabrication not as a software-specific set of capabilities, but as a design methodology that can allow schools to graduate young practitioners who can use these concepts to design and manage projects in more sophisticated ways. We outline six control and collaboration strategies and present several projects that explore those concepts through analog, digital, and hybrid methods.
Article
Full-text available
This paper argues for introducing a theory for knowledge integration in architectural design education. A contextual analysis of the reasons for developing a theory is introduced and reasons are categorized. The milieu of the theory is constituted in several contextual elements. The theory encompasses a number of underlying theories and concepts derived from other fields that differ dramatically from architecture. It consists of three major components: the disciplinary component; the cognitive-philosophical component; and the inquiry- epistemic component. Each of these components encompasses other smaller components integral to the building of the theory itself. Notably, the three components address ways in which knowledge can be integrated, how the desired integration would meet the capacity of the human mind, how such integration relates to the nature of knowledge and how knowledge about it is acquired, conveyed, and assimilated. Possible mechanisms for knowledge acquisition are an indispensable component of the theory, whose aim is to foster the development of responsive knowledge critical to the successful creation of built environments.
Architecture's New Media
  • Y E Kalay
Kalay, Y. E.: 2004, Architecture's New Media, The MIT Press, Cambridge, MA, US.
A system for collaborative design on timber gridshell, Rethinking Comprehensive Design: Speculative Counterculture
  • N Williams
  • S Bohnenberger
  • J Cherrey
Williams, N., Bohnenberger, S., and Cherrey, J.: 2014, A system for collaborative design on timber gridshell, Rethinking Comprehensive Design: Speculative Counterculture, Proceedings of the 19th International Conference on Computer-Aided Architectural Design Research in Asia CAADRIA, 441–450.
Constructing Atmospheres Rethinking Comprehensive Design: Speculative Counterculture
  • C Knapp
  • J Neslon
  • M Parsons
Knapp, C., Neslon J and Parsons M.: 2014, Constructing Atmospheres. Rethinking Comprehensive Design: Speculative Counterculture, Proceedings of the 19th International Conference on Computer-Aided Architectural Design Research in Asia CAADRIA, Kyoto, 149–158.
Un)Intended Discoveries Crafting the Design Process
  • L Boza
Boza, L.: 2006, (Un)Intended Discoveries Crafting the Design Process. Proceedings of the 25th Annual Conference of the Association for Computer-Aided Design in Architecture, 150-157.
Architecture's New Media
  • Y E Kalay
  • M A Cambridge
  • U S Knapp
  • C Neslon
  • J Parsons
Kalay, Y. E.: 2004, Architecture's New Media, The MIT Press, Cambridge, MA, US. Knapp, C., Neslon J and Parsons M.: 2014, Constructing Atmospheres. Rethinking Comprehensive Design: Speculative Counterculture, Proceedings of the 19th International Conference on Computer-Aided Architectural Design Research in Asia CAADRIA, Kyoto, 149-158.