SP18 Generative Design - Week 2 - Introduction to computational design

ARCH A4845
Generative design
Columbia University GSAPP
ARCH A4845: Generative design

Introduction to
computational design

1. Computers in design

Origins of CAD - Steven Coons and the Computer-Aided Design Project at MIT (1959-1967)
1963

Origins of CAD - Ivan Sutherland and the Sketchpad system (1963)
Ivan Sutherland’s Sketchpad system is demonstrated on the console of the TX-2 at MIT (1963).

Steps for drawing straight lines and circle arcs Manipulation of complex geometry

Geometric relationships and constraints

Origins of CAD - Timothy E Johnson and Sketchpad III (1963)
Sketchpad III, a computer program for drawing in three dimensions.
Generalized space allocation program, showing plan view, two perspectives, and two constraints.

Origins of CAD - Lawrence Roberts
Compound object construction (1963) Camera transformation (1963)

Origins of CAD - Nicholas Negroponte and the Architecture Machine Group (1969)
SEEK (1969-70)URBAN 5 (1969)

Origins of CAD - Nicholas Negroponte and the Architecture Machine Group (1969)

“There are three possible ways of having machines assist the design
process:
1. Current procedures can be automated
2. Existing methods can be altered to fit within the specifications and
constitution of a machine
3. The process, considered as being evolutionary, can be introduced
to a mechanism (also considered as evolutionary), and a mutual
training, resilience and growth can be developed
NICHOLAS NEGROPONTE, TOWARDS A HUMANISM THROUGH MACHINES (1969)

2. Design by algorithm

Victor Vasarely at work in 1948 Caopeo, 1964
Algorithms in art

Yoko Ono, Instruction Paintings, 1961
Algorithms in art

Sol Lewitt, Wall Drawing (1974)
Algorithms in art

Sol LeWitt, Wall Drawing #960 being executed at Site Gallery, Sheffield, 8 May 2010
Algorithms in art

Christopher Alexander, Notes on the Synthesis of Form (1964)
Algorithmic design

D’Arcy Wentworth Thompson, On Growth and Form (1917)
Morphogenesis in nature

Ernst Haeckel, Embryos (1870)
Morphogenesis in nature

Michael Hensel and Achim Menges (eds.) - AD Morphogenetic design series (2004, 2006, 2008, 2012)
Morphogenesis in design

Mark Burry, Sagrada Familia (1994)
Mark Burry, Architecture and Practical Design Computation. In
“Computational Design Thinking” (Achim Menges, Sean Alquist, ed.) 2011

3. Algorithm design

1. Variable
2. Conditional
3. Loop
4. Function
5. Object (Class)
5 elements of computation

1. Variable

1. Variable x = 1
y = ‘hello world’
myList = [1, ‘b’, 3]
myDict = {’a’: 1, ‘b’: 2}
2. Loop

1. Variable x = 1
myList = [1, ‘b’, 3]
myDict = {’a’: 1, ‘b’: 2}
for i in range(10):
# do something 10 times
for element in myList:
# do something for all elements
2. Loop
3. Conditional

1. Variable x = 1
myList = [1, ‘b’, 3]
myDict = {’a’: 1, ‘b’: 2}
for i in range(10):
if x > 10 and y > 10:
# do something
elif x > 5 or y < 5:
# do something else
else:
# do something else
2. Loop
3. Conditional
4. Function

1. Variable x = 1
myList = [1, ‘b’, 3]
myDict = {’a’: 1, ‘b’: 2}
for i in range(10):
if x > 10 and y > 10:
# do something
# do something else
else:
# do something else
def myFunction(input1, intput2):
# do something
return output
c = myFunction(a, b)
2. Loop
3. Conditional
4. Function
5. Object
“encapsulation”

1. Data-mapping (Grasshopper)
2. Procedural (scripting)
3. Object-oriented (OOP)
Types of programming

GH - single variable
GH Node
(“function”)
inputs outputs

GH system
Variable
geometry from Rhino
Variable
data in Grasshopper
Function
transform
geometry
Function
create
data type
“data flow”

GH system
Function
transform
geometry
Function
create
geometry
Function
create
data type

GH system
Function
transform
geometry
Function
create
geometry
Function
calculate
data
Function
create
data type
display data

GH system

GH - multi-data streams
Function
create list
of numbers
single data
multi-data

Inherent loop
(executes once on
each piece of data)

Conditional
(creates “pattern” of
True/False booleans)
Conditional
(separates data based
on pattern)

GH - data trees
Mode Lab - Grasshopper Primer V3.3 [http://grasshopperprimer.com/]

GH - data mapping - one to one

GH - data mapping - one to many

GH - data mapping - many to many (flat list)

GH - data mapping - many to many (data tree)

Limitations of Grasshopper
1. Hard to deal with variables having data structures beyond a 1-d flat list
2. Can’t make arbitrary loops (only data mapping)
3. Complex conditionals get messy because each one requires separate conditional and
dispatch nodes
4. Can’t create custom functions
5. No support for classes or object-oriented programming

Python in GH

What is a programming language?
HARDWARE
PROGRAMMING
LANGUAGE
USER
INTERFACE
SOFTWARE

Elements of computer programming
1. Variable x = 1
myList = [1, ‘b’, 3]
myDict = {’a’: 1, ‘b’: 2}

1. Variable x = 1
myList = [1, ‘b’, 3]
myDict = {’a’: 1, ‘b’: 2}
for i in range(10):
2. Loop

1. Variable x = 1
myList = [1, ‘b’, 3]
myDict = {’a’: 1, ‘b’: 2}
for i in range(10):
2. Loop
3. Conditional if x > 10 and y > 10:
# do something
# do something else
else:
# do something else

1. Variable x = 1
myList = [1, ‘b’, 3]
myDict = {’a’: 1, ‘b’: 2}
for i in range(10):
if x > 10 and y > 10:
# do something
# do something else
else:
# do something else
def myFunction(input1, input2):
# do something
return output
2. Loop
3. Conditional
4. Function

1. Variable x = 1
myList = [1, ‘b’, 3]
myDict = {’a’: 1, ‘b’: 2}
for i in range(10):
if x > 10 and y > 10:
# do something
# do something else
else:
# do something else
class MyClass:
def __init__(self, input1):
self.localVar = input1
def myMethod(self, input2):
# do something
return output
classInstance = MyClass(a)
c = classInstance.myMethod(b)
2. Loop
3. Conditional
4. Function
5. Object
# do something
return output
“encapsulation”

1. Variable x = 1
myList = [1, ‘b’, 3]
myDict = {’a’: 1, ‘b’: 2}
for i in range(10):
if x > 10 and y > 10:
# do something
# do something else
else:
# do something else
2. Loop
3. Conditional
4. Function
5. Object class MyClass:
def __init__(self, input1):
self.localVar = input1
def myMethod(self, input2):
# do something
return output
classInstance = MyClass(a)
c = classInstance.myMethod(b)
# do something
return output

4. Computational design strategies

Control types
1) Morphological
AdvantageDisavantage
2) State-change 3) Rule-based 4) Behavioral
• good top-down control over
design
• can create discontinous
design spaces
• control over individual
elemenst
• L-system, shape grammers,
1d CA (single-state)
• object-oriented, agent-based
behavior models (dynamic)
• continuous measures • choices, categories
• reduced number of inputs
(abstraction of inputs into
rule sets)
• can create complexity
• reduced number of inputs
(abstraction of inputs into
agent behaviors)
• can lead to emergence
• only top-down control
• can’t control individual
behavior
• can’t create emergence
• potentially redundant or
incomplete design space
• little intuitive control over
macro design
• potentially redundant or
incomplete design space
• can usually only generate
simple and design spaces
• many inputs (each element
needs to be controlled
seperately)

ARCH A4845/52
Generative design

SP18 Generative Design - Week 2 - Introduction to computational design

More Related Content

What's hot (11)

Similar to SP18 Generative Design - Week 2 - Introduction to computational design (20)

More from Danil Nagy (8)

Recently uploaded (20)

SP18 Generative Design - Week 2 - Introduction to computational design