23.02.2014 Views

Shape

Shape

Shape

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

291 Erasing and Identity<br />

to produce the third decomposition. (Like results follow in the algebra U 12 þ W 22 with<br />

the schema bðxÞ fi x for lines and the shapes they bound. I can do the third decomposition<br />

exactly as before in six steps or, alternatively, in two steps. The fewest steps for<br />

each decomposition is the number of gray tones it contains. Boundaries have some<br />

nice uses.) As I’m going to show next—this is the second point of interest—the number<br />

of decompositions for each of the shapes in the series<br />

is fixed in a definite way, but how these decompositions are represented as shapes can<br />

vary freely. Decompositions are abstract in the same way words and sentences are in<br />

generative grammars, and they’re abstract in the same way as Turing machines, cellular<br />

automata, etc. How I show them is independent of what they do—it’s a question of<br />

design. Playing around with algebras and rules is an effective means to explore the<br />

possibilities.<br />

The number f ðnÞ of distinct decompositions for any shape in the series<br />

when the shape is divided into squares and triangles can be given in terms of its<br />

numerical position n. The first shape in the series is described in one way—it’s a<br />

square—and the second shape can be described in two ways—it’s either two squares<br />

or four triangles. Moreover, the number of possible descriptions for each of the succeeding<br />

shapes—so long as I keep the remainder empty—is the corresponding term in<br />

the series of Fibonacci numbers defined by f ðnÞ ¼ f ðn 1Þþ f ðn 2Þ. This works in<br />

the following way

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!