Levitated | Node Gardens, a computational sketch

An interesting computational system can be built up by allowing a collection of self-similar nodes to connect and communicate with each other. From the patterns of organization that emerge within, such a system is called a Node Garden.

figure a, the node object

The basic elemental building block of the node garden is The Node. In this particular case, the node is an extension of the MovieClip Object within Macromedia Flash MX. No doubt if you have arrived here, you are aware of what Flash is. - or - Flash is a general purpose programming environment capable of high quality graphic display, audio, and interaction across a number of different computational platforms both local and networked.

The general purpose node is described in detail in its own paper, The Node. For this project we will be extending our previous definition of the node to allow for greater functionality. Specifically, we will be adding advanced methods allowing the node to connect, communicate, and collaborate with other nodes.

figure b, a highly organized node garden generated using the high contrast areas of a painting as nourishment

The beautiful rose in figure b is a sixteen square foot painting created by Jay Long in Austin, TX titled Rose III. As an interesting experiment, the painting was used as the generating mechanism for a node garden of specialized edge detecting nodes. Over the period of about five minutes, areas of high contrast encouraged the development of new nodes. This process rendered some beautiful results, to the credit of Jay Long Studio.

Garden growth is a simple process. The node is programmed to automatically seek out connections to other nearby nodes. The resulting complex network of connections is simply a manifestation of the nodes' desire.

figure c, the node object is programmed to automatically seek out connections to nearby nodes

Once instantiated, each node moves to its destination under speed constraints imposed by the environment. Once a node has reached its destination, it searches for nearby nodes to connect with. The radius of the search is directly proportional to the size of the node. The search happens only once.

Nodes only connect to nodes that have also reached their destinations. No limit on the number of connections is imposed. With each connection, a line is drawn and a message is sent informing the other node that it has been connected to.

figure b, arbitrarily positioned nodes with a few connections

figure d, a completely arbitrary node placement scheme usually renders a fairly homogenous network

With some thoughtful node placement, this simple process can render intriguing networks, common and uncommon as seen in the following forms:

figure e, a cellular structure

figure f, a heavy concentration of nodes

figure g, local groupings

A node can send messages through its connections. In turn, it can receive messages. In this fashion, information can be carried throughout a network of nodes. Information becomes the energy of the network.

Nodes share and store information. To convey that information to the user, and to illustrate the larger concept of a dynamic network of information exchange, we can represent value using geometric shapes and color. The shapes are localized about the node by which it is contained.

figure h, nodes display interval values as colorized geometric shapes

figure. an information rich network super cluster

Some other node gardens illustrating internal data follow.

figure. a spiral node garden 'on-edge'

figure. corridor node growth

figure. another info illuminated node garden

Occasionally nodes will connect to themselves, or to other nodes, who in turn connect back. Instances of these connection schemes provides for a feedback loop with potentially external interference. Feedback loops with no external interference are considered to be self-modifing.

jtarbell, July 2002