[分享] THE RATIONALIZING OF DESIGN
講義的文字檔
不過由於是掃描進電腦再加以文字辨識 可能會有部分出錯
我已經校對過一次 但是還是有些小錯字 要注意一下囉
有了文字檔 就可以......丟進翻譯軟體偷懶了 XD
--
THE RATIONALIZING OF
DESIGN
Anton C. Harfmann
School of Architecture and Environment Design
State University of New York at Buffalo
Buffalo, New York
The study of design attempts to externalize intuitive knowledge held by
designers. As more knowledge is externalized, the process becomes more
complex and fragmented. Computers offer the ability to reconstruct design
as a wholistic process, but what is still unclear is the degree to which com-
puters can or should be involved. The issues raised in the paper address the
core of the creative process itself and questions its computability.
□INTRODUCTION
Design has long been considered a uniquely human process that relies on
intuition, experience, and judgment of the designer. The study of design as
a process can be viewed as an effort to rationalize and externalize (or expli-
cate) the knowledge necessary for its conduct, Externalization of design
knowledge has increased the predictability and consistency of design proc-
esses, but has also tended to fragment the wholistic process or design into
discrete sub-disciplines. Efforts to define how design is actually carried out
can be understood as attempts to shed light on the "black box" process that
transforms a problem statement into a design solution [Jones 1969] (Figure
1).
A review of the major developments in design methodology illustrates
how the externalization of designers’ knowledge has brought about rational-
ization of the process.
□UNITY OF PROCESS AND KNOWLEDGE
Before the 15th century, design was a craft process of solving a problem
through the actual creation of an artifact. The knowledge required to cre-
ate the design was held solely by the craftsman or the artist. The process
was controlled intuitively and followed the simple sequence of problem
definition - design solution.
In complex situations, such as the creation of large buildings, this
method of solving problems "in the field" led to the development of a series
of sub-problems (Figure 2). These sub-problems were solved by experienced
craftsmen as they emerged based on knowledge accumulated through years
of practice and techniques passed down through the history of their crafts
(masonry, carpentry, goldsmithing, etc.). Each of these craftsmen individu-
ally held very specific knowledge about products and the processes relating
to their area of expertise.
□ ABSTRACTION AND FRAGMENTATION
Externalization and transmission of design knowledge has occurred
throughout history by apprenticeship or the documentation of rules, such as
the Ten Books of Architecture by Vitruvius. However, the first major step
in externalizing the design process itself occurred during the Renaissance,
when theoreticians sought to synthesize humanistic, philosophic ideals with
practical knowledge about the building crafts. With the emergence of per-
spective, this era saw architects like Brunelleschi using accurate drawings
and scale models to examine solutions prior to the actual construction of
buildings. Renaissance architects solicited professional technical consulta-
tion from various trades to improve the quality and utility of their proposed
designs [Ettlinger 1977]. The separation of the architect from the builder
marked the beginning of design problem-solving as an abstract process
removed from the manufacturing of the object (Figure 3). The use of scale
drawings and models to represent design intentions resulted in a major
change in the sequence of the design process as previously practiced; a solu-
tion could now be generated, at least conceptually, in an. abstracted, accu-
rately proportioned form, and reviewed by the related craftsmen prior to its
actual construction.
With problems or increased complexity, the craft methods or solving
problems proved to be too expensive and too slow [Alexander 1964].
Abstraction of the designed object, therefore, allowed a series of iterations
to be performed for the purpose of refining the product. This iterative
process suggests a further externalization of design, which includes the abil-
ity to change tile solution based upon the analysis of experienced craftsmen.
Iterative design, which evolved from the trial and error methods of the
past, has been modeled as a generate/test cycle within the abstraction space
of the proposed solution [Simon 1969] (Figure 4). Iterative methods were
used, although perhaps not formally, by Sir Christopher Wren in the design
of St. Paul's Cathedral, where several physical scale models were built and
studied to improve and develop design ideas (Wilton-Ely 1977].
The development of means to mathematically describe the laws of nature
increased the architect's ability to externalize and model design knowledge.
The formulation of very specialized, sophisticated knowledge, and the
development of engineering disciplines, brought further levels of separation
to the design process. The expansion of design knowledge has resulted in
the sub division of the process into discrete disciplines controlled by
"experts." The effect this compartmentalization has had on the design proc-.
ess is made obvious in the numerous disciplines currently required to
cooperate in the design and construction of a complex building (Figure 5).
□ DESIGN METHODS
The next phase in the externalization process was brought about by the
development of design methods that formalized and structured design
knowledge itself. The design process has been categorized into basic meth-
ods for the evaluation, generation, and selection of alternative solutions.
The simplest (and hence the first) areas to be formalized are the evalua-
tion procedures that are based on mathematical and physical principles.
These include cost, structure, lighting, and energy analysis.
Methods that have been developed to generate alternative solutions range
from very structured and systematic procedures, such as morphological
analysis, to intuitive and less structured methods such as brainstorming
[Jones 1969]. Other methods attempt to employ rules or patterns that define
solutions to certain specific design problems [Alexander 1977, Koning
1981].
The formalization of the process in which an alternative is selected
requires a search for the solution that best satisfies the overall goals and
objectives of the problem. Techniques that have been developed for this
purpose involve the search through a network of possible alternative solu-
tions for an acceptable solution [Winston 1984].
The strengths of formal design methods lie in their ability to structure
and document various specific aspects of design. However, the vast amount
of information that must be made explicit has hindered the designer's abil-
ity to maintain control over the process as a whole [Jones 1970].
□ COMPUTER AIDED DESIGN
Computers have been introduced to aid in the management of profuse and
fragmented information externalized through the study of design. Some of
the first intentions were to use the computer for the total automation of
design [Cross 1977]. The first "successful" use of computers was in proce-
dures where large amounts of numbers and equations were easily managed,
such as structural and energy analysis, relieving designers from tedious
tasks.
The next major advancement in CAD tools came about with the develop-
ment of computer graphics [Sutherland 1963], which resulted in tools that
are currently used for drafting by many designers. This use of computers,
however, is very limited, and does not affect the actual process of design,
except by reducing the time required to communicate the design ideas [Orr
1985]. Because this use was recognized as an underutilization of the poten-
tial of computers, three dimensional modeling systems were developed.
These systems increase the efficiency of the design process by allowing the
designer to review design solutions in three dimensions without the need to
construct physical scale models. The computer is still viewed in this use as
merely another medium for the representation and review of design solu-
tions.
The development of CAD tools to date has utilized the existing body of
externalized design knowledge for computing various aspects of the process.
It is obvious that computers are not a new method for design, but rather a
means to facilitate the existing established methods. It is also clear that
design knowledge must be externalized before it is computed.
In reviewing the current status of the design process, we notice a much
more comprehensive and complicated process than the first craft model pre-
sented. Instead of one person controlling the process and tile creation of the
product, the current model involves control and integration of several areas
of knowledge held by different professionals. Additional complications arise
in that the product itself is constructed by a different set of professionals
who have had very little input during tile abstract design phase. In order to
cope with this separation of knowledge, several researchers have suggested
an extended use of computers that would facilitate the integration of large
amounts of information generated during tile design process [Mitchell 1977;
Eastman 1985; Kalay 1985; Coyne & Gero 1986].
To be effective, these approaches must work at several levels of abstrac-
tion simultaneously and must be as flexible as human design processes.
These capacitics would transform computers from their current use as
design aids to "intelligent" machines capable of assisting in the creative
aspects of design and would aid in the reconstruction of design as a
wholistic process.
□FUTURE DIRECTIONS
The progression of the study of design has been presented as a continuum
from an intuitive, unstructured activity, to present attempts to explicitly
define and compute certain aspects of design knowledge.
It seems inevitable that further externalization of the design process will
occur, and that the use of computers in design will expand. Many issues
concerning the incorporation of computers into design remain obscured by
controversy over the "computability" of inherently human processes.
If we accept the potential computability of the design process, will the
final step in the externalization be the total automation of the process?.
Some argue that as long as design is viewed as problem solving, all aspects
could be computed. Others maintain that design (when viewed as an art)
defies computability, since art by. definition cannot be computed. The issues
raised here address nothing less than the core of creativity. Can or should
creativity be externalized? Can or should it be computed?
--
╭══════════════════════════╮
║ Thinking εξ ║
║ Working ●-ζ ●- -●/ ║
║ Living （ <■> ■_ ║
║ Keep Smoking......∕） /\ 〒) PTTTobacco ║
╰══════════════════════════╯ CoffeeOldMan
--
※ 發信站: 批踢踢實業坊(ptt.cc)
◆ From: 220.133.153.166
推
02/21 21:14, , 1F
02/21 21:14, 1F
→
02/21 22:10, , 2F
02/21 22:10, 2F
推
02/22 08:46, , 3F
02/22 08:46, 3F