|
|
| Introduction: |
| 'Modular
Algorithm' is an Autodesk Maya modeling script for exploring pattern
making developed by Zachariah Muñoz
in Spring 2006. It was originally conceived as an academic project
for a seminar on iterative tools, taught by Billie Faircloth at
the University of Texas at Austin, School of Architecture. Several
patterns generated with the script were used in the design of the
building shown to the right. |
|
|
| Methodology: |
| A
range of architectural criteria determine which values in the script
are selected to become variables. Once these values have been determined
a block is created and the process repeats itself. The method by
which the blocks are assembled is controlled by translation and
rotation values that may also become scripted, depending on the
desired output. A randomness operator, furthermore, is included
in the script in order to generate a variety of solutions to a single
design goal. |
|
| The
following list offers a brief description of selected variables
from the script: |
| $Opacity:
$BlockDepth:
$BlockWidth:
$BlockRotate:
$Random:
$YTrans:
$ZTrans:
$RunSum: |
Probability,
ranging from 0% to 100%, that determines whether to remove a
block.
Depth of block.
Width of block.
Rotation of block in x-axis.
A random number generated between 0 and 10.
Translation of block in the y-axis.
Translation of block in the z-axis.
A sum of translation values in the X-axis that controls pattern
iterations.
|
|
| Each
of these variables, except $RunSum and $Random, may be reconfigured
as a function or constant number. The following scripts offer examples
of how different pattern types may emerge by adjusting the function
of any given variable. The $Opacity variable, furthermore, is operable
in each illustration in order to show its influence over the pattern
generated. Each script may be copied and pasted directly into Maya’s
MEL script editor for execution. |
|
|
| Pattern
01 : |
|
|
int
$i;
for
($i = 0; $i < 500; $i++)
{
int
$Opacity = 20;
float
$BlockDepth = 1;
float
$BlockWidth = ceil(rand(0, 8));
float
$Random = ceil(rand(0, 10));
float
$YTrans;
float
$RunSum;
if
($i == 0)
{
$YTrans
= 0;
$RunSum
= 0;
};
if
($RunSum >= 50)
{
$YTrans++;
$RunSum
= 0;
};
$RunSum
= ($BlockWidth / 2) + $RunSum;
polyCube
-w $BlockWidth -d $BlockDepth;
move
-r -ls -wd $RunSum $YTrans 0;
if
($Random > ((100 - $Opacity) / 10))
delete;
$RunSum
= ($BlockWidth / 2) + $RunSum;
};
select -cl; |
|
|
| Pattern
02: |
|
|
int
$i;
for ($i = 0; $i < 500; $i++)
{
int
$Opacity = 40;
float
$BlockDepth;
float
$BlockWidth = ceil(rand(0,8));
float
$Random = ceil (rand(0,10));
float
$YTrans;
float
$RunSum;
if
($i == 0)
{
$YTrans
= 0;
$RunSum
= 0;
};
if
($YTrans == 0)
$BlockDepth
= 25;
if
($RunSum >= 50)
{
$YTrans++;
$RunSum
= 0;
};
$RunSum
= ($BlockWidth / 2) + $RunSum;
polyCube
-w $BlockWidth -d $BlockDepth;
move
-r -ls -wd $RunSum $YTrans 0;
if
($Random> ((100 - $Opacity) / 10))
delete;
float
$BlockDepth = 25 - ($YTrans / 2);
$RunSum
= ($BlockWidth / 2) + $RunSum;
};
select -cl; |
|
|
| Pattern
03 : |
|
|
int
$i;
for ($i = 0; $i < 500; $i++)
{
int
$Opacity = 60;
float
$BlockDepth = ceil(rand(0,10));
float
$BlockWidth = ceil (rand(0,8));
float
$Random = ceil(rand(0,10));
float
$YTrans;
float
$RunSum;
if
($i == 0)
{
$YTrans
= 0;
$RunSum
= 0;
};
if
($RunSum >= 50)
{
$YTrans++;
$RunSum
= 0;
};
$RunSum
= ($BlockWidth / 2) + $RunSum;
polyCube
-w $BlockWidth -d $BlockDepth;
move
-r -ls -wd $RunSum $YTrans 0;
if
($Random> ((100 - $Opacity) / 10))
delete;
$RunSum
= ($BlockWidth / 2) + $RunSum;
};
select -cl; |
|
|
| Pattern
04: |
|
|
int
$i;
for ($i = 0; $i < 400; $i++)
{
int
$Opacity = 40;
float
$BlockDepth;
float
$BlockRotate;
float
$BlockWidth = ceil(rand(0,8));
float
$Random = ceil(rand(0,10));
float
$YTrans;
float
$ZTrans;
float
$RunSum;
if
($i == 0)
{
$YTrans
= 0;
$RunSum
= 0;
};
if
($YTrans == 0)
$BlockDepth
= 15;
if
($RunSum >= 50)
{
$YTrans++;
$RunSum
= 0;
};
$RunSum
= ($BlockWidth / 2) + $RunSum;
float
$CvStep = $YTrans / 10;
$BlockDepth
= ceil(rand(0,10));
$BlockRotate
= (3 * pow($CvStep, 2) * sin($CvStep))
+
(cos($CvStep) * pow($CvStep, 3));
$ZTrans
= pow($CvStep, 3) * sin($CvStep);
polyCube
-w $BlockWidth -d $BlockDepth;
move
-r -ls -wd $RunSum $YTrans $ZTrans;
rotate
-r -os $BlockRotate 0 0;
if
($Random > ((100 - $Opacity) / 10))
delete;
$RunSum
= ($BlockWidth / 2) + $RunSum;
};
select -cl; |
|
|
| Pattern
05: |
|
|
int
$i;
for ($i = 0; $i < 500; $i++)
{
int
$Opacity = 50;
float
$BlockWidth = ceil(rand(0,8));
float
$Random = ceil(rand(0,10));
float
$YTrans;
float
$RunSum;
if
($i == 0)
{
$YTrans
= 0;
$RunSum
= 0;
};
if
($RunSum >= 50)
{
$YTrans++;
$RunSum
= 0;
};
$RunSum
= ($BlockWidth / 2) + $RunSum;
polySphere
-r 1 -sx 4 -sy 3 -ax 0 1 0 -tx 1;
move
-r -ls -wd $RunSum $YTrans 0;
scale
-r $BlockWidth 1 1;
if
($Random > ((100 - $Opacity) / 10))
delete;
$RunSum
= ($BlockWidth / 2) + $RunSum;
};
select -cl; |
|
|
| Pattern
06: |
|
|
int
$i;
for ($i = 0; $i < 2000; $i++)
{
int
$Opacity = 50;
float
$BlockDepth;
float
$BlockRotate;
float
$BlockWidth = ceil(rand(0,1));
float
$Random = ceil(rand(0,10));
float
$YTrans;
float
$ZTrans;
float
$RunSum;
if
($i == 0)
{
$YTrans
= 0;
$RunSum
= 0;
};
if
($RunSum >= 50)
{
$YTrans++;
$RunSum
= 0;
};
$RunSum
= ($BlockWidth / 2) + $RunSum;
float
$CvStep = $YTrans / 10;
$BlockDepth
= 1;
$BlockRotate
= (3 * pow($CvStep, 2) * sin($CvStep))
+ (cos($CvStep)
* pow($CvStep, 3));
$ZTrans
= pow($CvStep, 3) * sin($CvStep)
+
pow(($RunSum / 10), 2);
polySphere
-r 1 -sx 4 -sy 3 -ax 0 1 0 -tx 1;
move
-r -ls -wd $RunSum $YTrans $ZTrans;
rotate
-r -os $BlockRotate 0 0;
if
($Random > ((100 - $Opacity) / 10))
delete;
$RunSum
= ($BlockWidth / 2) + $RunSum;
};
select -cl; |
|
|
| Conclusion: |
| The
patterns that have been generated thus far are not critically informed
by a system of construction. Many of the blocks in the preceding patterns,
for example, remain suspended in space, unsupported by a structural
frame. Further work on these scripts would involve the inclusion of
a function that generates a type of scaffolding system that supports
the blocks. This system would likely be influenced by new variables
relating to physical properties such as weight and materials.
Regarding
the format of the algorithm itself, the next improvement would involve
the development of a graphical user interface that enables users
to generate new patterns through a dialog box, without the burden
of modifying a text file. This type of interface would be useful
not only as a pattern-making instrument, but also as an educational
tool for introducing users to a basic understanding of the MEL scripting
language.
|
|
|
|
