Towards a userful grammar implementation: beginning to learn what designers want. In Proceedings of CAADRIA 2020. Hong Kong: CAADRIA.
Abstract. We present a general shape grammar implementation that supports subshape detection and handles lines and labeled points in three-dimensional space. Its front and back ends are both set in the CAD application Rhinoceros3d. Information observations of designers using the implementation suggest that they are more interested in producing designs to work with than in using the more specialized features of shape grammars. This in turn suggests that researchers who create such implementations have more to learn about the designers who use them.
Final manuscript version (PDF, 137 KB)
Learning from users and their interaction with a dual-interface shape-grammar implementation. In Cultural DNA 2019, ed. Ji-Hyun Lee. Singapore: Springer.
Abstract. We present a shape grammar implementation with two new characteristics. One is that it visual and directly manipulable: users draw the shapes and rules in a modeling application. The other characteristic is advanced technical capabilities, such as non-visual attributes and higher-order elements like surfaces. It consists of three components running in Rhinoceros3d. We also report on workshops that introduced the implementation.
Final manuscript version (PDF, 199 KB)
A whole-grammar implementation of shape grammars for designers. AIEDAM, Artificial intelligence for engineering design, analysis and manufacturing 32(2): 200–207.
Abstract. I present an implementation of shape grammars that is aimed at supporting designers. It has two parts: a grammar editor and a stand-alone interpreter. The editor is the modeling application Rhinoceros3d using Python scripts. The interpreter is general, is three-dimensional, and supports subshape detection. A grammar is a Rhinoceros3d model; thus users can manipulate all its parts directly and immediately. That is, they can modify any shape without selecting or invoking an editor, and they can lay out the parts of the grammar in any way they find meaningful. Using this approach, which I call a whole-grammar approach, users are shielded from most subdomain tasks, like typing text files or specifying transformations. Informal observations suggest that users of this implementation can work effectively.
Paper (corrected proof; 693 KB)
Implementing shape grammars for designers. In Morphological analysis of cultural DNA: tools for decoding culture-embedded forms, ed. Ji-Hyun Lee, 165–176. Singapore: Springer.
Abstract. For the computational understanding of visual artifacts, shape grammar provides an important theoretical framework. In addition to the theory, there have been numerous computer implementations; these have tended to be proofs of concept. As such, they are essential steps in development, but do not directly help researchers do the kind of analyses seen in the literature, which were done by hand. That is to say, we have a theory but not yet a sturdy tool.
We present a prototype implementation to help designers and design researchers work with shape grammars. This implementation allows users to focus on domain tasks – editing and testing grammars – by shielding them from sub-domain tasks – mechanical tasks like matching shapes and applying rules. A grammar is displayed as a collection of shapes in 3D space that users can manipulate directly; a commercial 3D modeling application is used for this purpose. The components of the implementation are designed to make it easy for users to switch between editing and testing their grammars. The implementation handles emergence and is general. We report on users’ experiences with the implementation in workshops on grammatical design and analysis.
Paper (uncorrected proof; PDF, 420 KB)
Computing style. Nexus network journal 13: 183–193.
Abstract. Shape grammars are frequently used in analyzing style in architecture and other areas of design. But this is a more subtle task than is usually realized, and some grammatical approaches to design analysis are logically suspect. We examine the framework articulated by Stiny and Mitchell in 1978, fill in the operational gaps, and propose a more comprehensively considered approach to using grammars to compute style.
Paper (PDF, 480 KB)
Editing shapes in a prototype two- and three-dimensional shape grammar environment. In Computation: the new realm of architectural design: proceedings of the 27th conference on education and research in computer aided architectural design in Europe, eds. Gülen Çağdaş and Birgül Çolakoğlu, 243–249. Istanbul: eCAADe.
, and . 2009.
Abstract. Recently we developed a prototype general shape grammar system, called Grammar Environment (Li et al. 2009). It differs from other systems in that it aims to support designers who design with shape grammars. One task of such a system is to support users in editing shapes. The guidelines that we followed in developing Grammar Environment suggested that the shape editing system should both be integrated into the system and be powerful as a drawing tool. This seemed to be contradictory. We decided to make two shape editors: one stronger on integration, the other on drawing power.
Paper (PDF, 2 MB)
A prototype system for developing two- and three-dimensional shape grammars. In CAADRIA 2009: proceedings of the 14th international conference on computer-aided architecture design research in Asia, eds. Teng-wen Chang, Erik Champion, Sheng-Fen Chien, and Shang-Chia Chiou, 717–726. Touliu, Taiwan: Department of Digital Media Design, National Yunlin University of Science and Technology.
, and . 2009.
Abstract. A number of researchers have developed shape grammar systems, with a variety of aims. These systems all help users (to varying degrees) to run grammars, but not to develop grammars. However, we believe that developing grammars is also work and needs to be supported. A system to do this would make it easier and more convenient for people using grammars to do design work. Following the generate-test model, we design and implement a prototype system that supports the user in editing grammars, testing grammars, and switching easily between the two types of activity. We emphasize the graphic nature of the task: the user is all the time working with graphic objects, namely shapes.
Paper (PDF, 227 KB)
Computing the Yingzao fashi 營造法式的運算解析. In Shuzi yingzao: jianzhu sheji, yunsuan luoji, renshi lilun 數字營造：建築設計，運算邏輯，認識理論, Computational constructs: architectural design, logic and theory, eds. Chen Shouheng 陳壽恆, Li Shuyi 李書誼, and Josh Lobel, 113–125. Beijing: Zhongguo jianzhu gongye chubanshe.
Paper (PDF, 864 KB)
Thoughts on a designer-friendly shape grammar interpreter. In eCAADe 23: digital design: the quest for new paradigms, eds. José Pinto Duarte, Gonçalo Ducla-Soares, and A. Zita Sampaio, 523–528. Lisbon: Education in Computer Aided Architectural Design in Europe.
Abstract. Discussions of shape grammar interpreters overlook a fundamental issue: the model of the designer’s work. Such a model would provide guidance for developing an interpreter with an appropriate interface. In this paper, I first propose a model in which the designer’s work is to create and test generative specifications of languages of designs. I call this model designer-centered generative design. Then, I examine the characteristics of shape grammar and how they support or impede this model of work. Finally, I discuss the implications for the design of an appropriate shape grammar interpreter. These provide guidelines for implementing such an interpreter for testing.
Paper (PDF, 96 KB)
Styles, grammars, authors, and users. In Design computing and cognition ’04, edited by John S. Gero, 197–215. Dordrecht, Netherlands: Kluwer Academic
Abstract. Using a grammar to understand
style can be seen, for students, as involving two main tasks: cultivating
a standard of stylistic correctness and converging the language defined
by the grammar and the language of stylistically correct designs. I
discuss a framework for organizing such an experience, consider how
it informs the way we write grammars, present an example (including
a grammar), and report on a classroom experience.
Paper (PDF, 828 KB)
A set-based shape grammar
interpreter, with thoughts on emergence. Paper read at Workshop
3, Implementation issues in generative design systems, “First international
conference on design computing and cognition,” Massachusetts Institute
of Technology, 17–19 July 2004.
, and . 2004.
Abstract. We present a set-based interpreter,
implemented in Multimedia Flash, of a shape grammar for teaching the
language of twelfth-century Chinese wood-frame building sections. We
discuss the implementation of various aspects of the shape grammar formalism
– shape representation, the part relation, and so on – the usefulness
of the interpreter, and thoughts on the role of emergence in grammar
Paper (PDF, 256 KB)
The Yingzao fashi in the information age. Paper presented
at “The Beaux-Arts, Paul-Philippe Cret, and twentieth-century architecture
in China,” The University of Pennsylvania, 3–5 October 2003.
prototype simulated interactive shape grammar. In Design
e-ducation: connecting the real and the virtual, Proceedings of the 20th
conference on education in computer aided architectural design in Europe, edited by Krzysztof Koszewski and Stefan Wrona, 314–317. Warsaw: eCAADe.
Abstract. Existing shape grammar interpreters
have implemented some of the distinguishing features of the formalism,
notably emergence or rule application under multiple transformations.
By contrast, I implement a grammar in which these capabilities either
are not required or can be simulated. This frees a user from mechanical
issues, allowing him to pay more attention to the process embodied in
Paper (PDF, 75 KB)
architecture in twelfth-century China: the Yingzao fashi. In Nexus IV: architecture and mathematics, edited by José Francisco Rodrigues and Kim Williams, 141–150. Fuccechio,
Italy: Kim Williams Books.
Paper (PDF, 739 KB)
style grammatically, with an example from traditional Chinese architecture. In The proceedings of Mathematics & design
2001: the third international conference (3–5 July 2001, Geelong,
Abstract. Most analytical shape grammars
have aimed at an authoritative definition of style. This approach deprives
the student of an important way of understanding the style, by formulating
its definition. But this experience can be understood formally and incorporated
into the grammar. An example is given from the 12th-century Chinese
building manual Yingzao fashi.
Paper (PDF, 1.2 MB)
shape grammar for teaching the architectural style of the Yingzao
fashi. PhD dissertation, Department of Architecture, Massachusetts
Institute of Technology, Cambridge, Mass.
Abstract. The Yingzao
fashi [Building standards] is a Chinese building manual written
by Li Jie (d. 1110) and published in 1103. I present a shape grammar
for teaching the architectural style – the language of designs – described
in this manual. This grammar is distinguished by two objectives, and
the technical means used to accomplish them.
First, the grammar is for teaching. Usually, the author of a grammar
of a style aims to generate all and only the designs in the language.
To do this, he not only writes the grammar, but also judges whether
the designs it generates are members of the language. In the Yingzao
fashi grammar, on the other hand, I want to generate all and more than
the designs in the language. It is then the student who evaluates the
designs – does this design belong to the language? – and adjusts the
grammar accordingly. Thus the student participates actively in defining
the language of designs, and learns that style is a human construct.
Second, the grammar is designerly. As already observed, most authors
of style grammars focus on the language of designs; they do not consider
how to structure the user’s interaction with the grammar. By contrast,
I consider explicitly what the user decides and when he decides it,
and organize the grammar accordingly. In other words, I consider process
as well as products.
The grammar exploits several technical devices for the first time:
the design as an n-tuple of drawings, descriptions, and other
elements; the generation of descriptions in the n-tuple; and
techniques that are made possible by these devices.
Volume 1, text (PDF, 199 KB)
Volume 2, images (PDF, 1.38 MB)
symbolic and spatial information in shape grammars, with an example from
traditional Chinese architecture. In CAADRIA
2000: proceedings of the fifth conference of the Association of computer
aided architectural design research in Asia (CAADRIA) (18–19 May
2000, Singapore), 245–253.
Abstract. Stiny’s (1981) formulation
of descriptions is applied to building sections and their descriptions
found in the twelfth-century Chinese building manual Yingzao
grammars, and designerliness. In Proceedings
of Greenwich 2000: digital creativity symposium (13–15 January
2000, London), 81–85.
Abstract. From the user’s point of
view, what characterizes an appropriate interaction with a grammar?
What criteria can be used to determine this appropriateness? Are these
criteria fixed or changeable? These questions will be considered and,
to the extent possible, formalized.
parametric dependence in shape grammars, with an example from traditional
Chinese architecture. In Proceedings of
the fourth conference of the Association of computer aided architectural
design research in Asia (CAADRIA) (5–7 May 1999, Shanghai), 265–274.
Abstract. Shape grammars traditionally
generate one product at a time. This leads to difficulties when dependent
parameters are involved. Parallel grammars are proposed as a solution.
As an example, a grammar is shown which generates plans according to
the 12th-century Chinese building manual Yingzao
Sicheng’s A pictorial history of Chinese architecture. In On what ground(s)? Refereed proceedings of
the 1997 annual conference of the Society of architectural historians,
Australia and New Zealand (SAHANZ), ed. S. Pickersgill and P. Scriver,
110–119. Adelaide: SAHANZ.
Structural carpentry in the Yingzao fashi. Architronic 5 (2). Dead link. Sorry.
Using virtual models to teach
traditional Chinese wood construction. In The
introduction of technology: CAAD teaching and research directions in Asia:
proceedings of the first conference of the Association of computer aided
architectural design research in Asia (CAADRIA) (25–27 April 1996,
Hong Kong), ed. Thomas Kvan, 119–130.
Abstract. In this paper we discuss
our experience in using virtual models to teach traditional Chinese
wood construction. Although our approach is technically simple – we
use a kit of modedl parts made with the Solid Modeler of AutoCAD, Release
12 (now Release 13), and customized commands in AutoLISP – we have had
excellent results. This is because of the remarkable match between the
modelling medium and the highly systematized nature of traditional Chinese
wood construction. It is this crucial – and interesting – characteristic
that we want students to understand and appreciate.
In our first teaching experience, in the fall term, 1994–95, despite
unexpected drawbacks, our approach succeeded. In fact, our students,
all Hong Kong Chinese, were surprisingly enthusiastic and even took
pride in the sophistication of this uniquely Chinese construction system.
In 1995–96, we have used the same kit of parts in two courses: an introduction
to Chinese architecture (spring term) and an advanced course in Song
dynasty wood construction (fall term).
We first discuss briefly the theoretical basis for our approach. We
then describe the assignments, the kit of parts, and supporting materials
used in our teaching experiences. Finally, we discuss our findings and
consider directions for the future development and improvement of our
The rule-based nature of wood
frame construction of the Yingzao fashi and the role of virtual
modelling in understanding it. In Computing
in architectural research: proceedings of the International conference
on Chinese architectural history (7–10 August 1995, Hong Kong),
Abstract. The wood frame construction
system of the Yingzao fashi is rule-based.
In this sytstem, the text can be understood as rules, construction as
the execution of those rules, and the building as the output of the
rules. To illustrate, we discuss a three-bay and some of the rules which govern its overall form and
size. This three-bay is the smallest
hall defined in the Yingzao fashi.
Virtual models offer a critical advantage over real models and drawings.
This type of representation allows us to ask questions which are particularly
relevant to the rule-based system. We can execute the rules quickly
and thus concentrate on the rules, their output, and the relation between
them. We propose some areas of research made possible with virtual modelling.
One is a study of roof curvature as a product of several rules, including
Chinese wood frame construction
of the twelfth century: The construction system of the Yingzao fashi. CD-ROM.
A computer-based teaching tool
for traditional Chinese wood construction. Preconference
proceedings: advances in computer-based building design systems: 7th international
conference on systems research, informatics and cybernetics (15–21
August 1994, Baden-Baden, Germany), 149–59.
Abstract. In this paper we discuss
a method of using virtual modelling to teach and study aspects of traditional
Chinese wood construction as prescribed in the building manual Yingzao
fashi, published in 1103. There are two types of argument for
using virtual modelling. First, there are important conceptual commonalities,
the most compelling of which is the distinction between primitive and
instance. The Chinese construction system, in its treatment of building
components and their relations, makes precisely the same distinction.
Second, the precision of the virtual model and its ease of duplication
and modification – its very virtuality – seem to open up new areas of
investigation through comparative models. The amount of labor required
to make a real model is a serious impediment to comparison studies,
which involve multiple models. The meaning of the Chinese construction
system lies not in the manufacture of repeated components, but in their
The first phase of this project culminates in the demonstration of
a prototype system. We found that users had many operational difficulties,
which tended to distract them for more important issues. Based on this
demonstration, we propose directions for expansion and refinement in
the second phase, which will include use of the system in an advanced
undergraduate elective course in architectural history in the fall of