Return to Gossman Consulting, Inc Library

Return to David Gossman's home page

Return to Table of Contents


A Thesis Presented to the Faculty of the College of Natural Science Michigan State University

In Partial Fulfillment of the Requirements for the Degree Master of Science

by David Gordon Gossman

June 1979



The primary determinant of information and specialty needs is the problem which the R&D organizations will be trying to solve. This problem must be carefully analyzed in order to determine what specialists are needed and which specialist is going to need information from which other specialist. Depending on the breadth and magnitude of the problem, this initial determination of information and specialty needs may be done by one individual or by a group of not more than five individuals from a variety of areas. If it is done by one person, that person should be the one who will be managing the R&D group throughout the group's lifetime.

The determination of information and specialty needs has been divided into the following four basic steps by the author:

1. Determination of specialists needed.

2. Determination of information needs for each specialist.

3. Production of a topological graph of the specialists and their information needs.

4. Reduction of the topological graph to a quantitative matrix.

Each of these steps is closely interlinked, and depending on an evaluation of the results at the end of each step, it may be necessary at times to go back to previous steps for reanalysis.

Determination of specialists needed. The first step involved in the determination of information and specialty needs is to produce a list of specialists who will be needed to solve the particular problem. This list must be as inclusive and specific as possible. When defining the specialist, it is necessary to identify the specialty areas as narrowly as possible. In addition, if two or more individuals will be needed in one specialty area, they should be listed separately.

Determination of information needs for each specialist. The next step in the analysis involves determining for each specialist what his information needs will be. In other words, for a given specialist, which other specialty areas will he need information from. After this is done, the lists should be cross-checked to determine if there are any information needs in a specialty area for which there are not specialists listed. If this is true, it is necessary to return to the first step, add these specialists to the initial list and then determine their information needs.

It should be noted that in many cases, a particular specialist may have no need for information from any other, or some specialists may have need for information from half a dozen or more other specialists. At this time, it is not necessary to worry about whether one specialist will be able to communicate the necessary information to another. That will be dealt with later. At the present, the only concern is information needs.

Production of a topological graph of the specialists and their information needs. In order to show how to convert the list of specialists and their information needs into a topological graph, sometimes referred to as a sociogram, it is first necessary to introduce a few new terms from the area of topology. Figure VI-1 (a) shows two nodes connected by a non-directional edge. Nodes will be used to represent individual specialists and edges will, at least at first, be used to indicate information needs. They will be used later to indicate actual information flow. The use of directional edges is needed since specialist "X" may need information from specialist "Y", but not vice-versa. Figure VI-1 (b) shows two nodes connected with two directional edges.

If for any two nodes in a graph, there is a path along the edges between them, then the graph is said to be connected. Figures VI-1 (c) and VI-1 (d) show a connected and non-connected graph, respectively. A "completely" connected graph is defined as a graph where there is at least on edge between any pair of nodes. This is analogous to the interdisciplinary "committee approach to R&D.

In order to produce a graph from the lists of specialists and information needs it is first necessary to represent each specialist as a node on the graph. Each specialist must then be connected to other appropriate specialists by an edge which represents the information need. The edge should point towards the specialist who has the particular information need and, therefore, away from the specialist who will provide that information.

(a) Two nodes connected by a non-directed edge.    (b) Two nodes connected with two directional edges.

(c) Connected graph                                       (d) Non-connected graph.


After this is complete, it may be found that the graph is non-connected. If this is true, a brief reassessment of information needs should be done to determine if there is a need for additional edges. If the graph is still non-connected, it is now necessary to add R&D management personnel to the graph as additional nodes. These nodes should be linked to the other nodes (specialists) by new pairs of edges in such a manner as to produce a connected graph. The number of R&D management personnel added to the graph and the nature of these additions is primarily determined by the nature of the problem and the individual backgrounds of the management personnel involved.

Reduction of the topological graph to a quantitative matrix. It is now necessary to reduce the topological graph to the form of a square matrix. This is done by listing the nodes of the graph along both the horizontal and vertical axis of the matrix. The vertical axis is considered to be the sender of the information and the horizontal axis represents the receiver of information. A one or a zero is used to represent the presence of lack of an edge in the graph. Along the diagonal of the matrix, a one is used for a specialist since he obviously will need information from him own specialty. Graphically, this is called self-looping. The management personnel that may have been added to the graph by this time are not self-looped and have zeros along the diagonal of the matrix. Figure VI-2 shows a small example graph and its matrix.


  1 2 3 4 5 6 7 8 9
1 1 1 0 0 0 0 0 0 0
2 0 1 1 0 0 0 0 0 0
3 1 0 1 0 0 0 0 1 0
4 0 0 0 1 0 1 0 0 0
5 0 0 0 1 1 0 0 0 0
6 0 0 0 1 1 1 0 1 0
7 0 0 0 0 0 0 1 0 1
8 0 0 0 0 0 0 0 1 1
9 0 0 0 0 0 0 1 1 0

  Figure VI-2


Continue to Chapter VII