The twofold variety of nodes in KnitbITs®

Information Dynamics

Underlying the construction of KnitbITs is a relational shift. What are traditionally considered relationship instances have become object instances in their own right.1 It resembles the synthesis of for example the two approaches to so-called network planning: 1. activity on node and 2. activity on line.

For the sake of presentation, let's start with a more or less traditional object node, i.e., a node representing an object instance. Its immediate properties can be expressed by two different kinds of objects. One kind is another more or less traditional object node. The other kind of property is a node reflecting a relationship-as-object.

A sketch might help. A more-or-less-traditional-object instance is symbolized by a circle. An instance of the relationship-as-object by a rectangle. Again starting from an object as traditionally viewed, at its most general properties can be allocated as illustrated below (please 'read' figure 1 from top to bottom).

Figure 1.

Both kinds of 'objects' appear related. The concept of relationship therefore hasn't disappeared. But now there is a subdivision of relationships. One is the objectified variety, the other remains the more-or-less-traditional-relationship. The latter is — still — shown as a line connecting nodes annex objects.

The progression of properties is practically limited — there certainly is no theoretical ground for it — as far as 'properties' of a relationship-as-object are concerned. It is only admitted to suspend a more-or-less-traditional-object. Figure 2 sketches the differentiation in progression.

Figure 2.

Seen from bottom to top, a choice presents itself for 'constructing' one more-or-less-traditional-object as the property of another more-or-less-traditional-object. They may entertain a direct, i.e., more-or-less-traditional-relationship, as appears on the left of figure 1. Or a relationship-as-object may be interposed; see the right of figure 1. Relating more-or-less-traditional-objects through an explicit relationship-as-object adds a degree of freedom to the construction of structure.

Including the bottom-to-top referential variety inside the two kinds of nodes extends the individual symbols as shown in figure 3. (Inside the bigger circle, 'U' stands for 'or.')

Figure 3.

Figure 2, for example, can now be redrawn as figure 4. This brings out that the connecting lines only appear for convenience's sake in such visualizations. There is no equivalent for them in KnitbITs where the more-or-less-traditional-relationships are established through — unique — functional references within the nodes.

Figure 4.

This leads to the question of whether 'node' is an apt metaphor for the structure of information sets. It seems a multidimensional brick wall or jig-saw puzzle, whatever those 'really' are, is a more suitable image; the stones, building blocks or pieces are tightly stacked, their relative positions guided by matching elements strictly contained within them.

 

 

1. This paper deals with instances of objects and relationships. Typing such instances is not treated here.

 

 

© web edition 2002.