In the field of information processing, documentation and classification there is an almost universal bias towards text and terms, since publications have titles and normally contain text. This is a very persuasive argument in favour of word oriented computers and classification schemes. It is associated with the generation of a plethora of costly bibliographical tools, abstracts, directories and encyclopaedias (39).

But even if all items in the total body of literature were identifiable and available at low cost (which is the aim of those who favour this approach), there still remains the problem of how to improve the relevance of the questions asked to the problem complex faced by the policy-making process. Retrieval is not the problem, it merely aggravates this more fundamental problem. Retrieval systems focus queries in the light of the user's existing knowledge and biases. They do not orient the policy-oriented user to knowledge and issues with which he should also be concerned in relation to his current preoccupations (in the light of qualified or alternative opinions). They do not bring to his attention where his preoccupation may fit in relation to other preoccupations. He is given no sense of scale, proportion or orientation - he merely gets what he asked for however much difficulty he has in formulating his question in appropriate words.

Explanatory power of diagrams: It is ironical that within any book or article, whenever the point to be made is too complex to be expressed in words, the author resorts to a diagram of some kind. This ensures that various elements, are brought into appropriate relationship within a whole of which the reader has an overview, from that overview the reader can then select (a) how he wishes to explore the elements interrelated therein, and (b) those he considers significant as meriting further examination. Yet existing information systems are completely incapable of producing or manipulating diagrams as an aid to policy-making.

Computer - generated diagrams: The exceptions to this statement are interesting as indications of the kinds of technology not available to policy-making in relation to the development process :

• air-traffic control radar display screens
• computer-aided architectural and engineering design displays
• factory process control flow displays
• electronic circuit analysis and design displays.

In each such case there are complex problems of choice and decision analogous to those in a policy-making situation. The examples are given to show that a technology is in use to manipulate such information. Unfortunately, however, that technology cannot yet be used satisfactorily in relation to development processes because the information is in the wrong form. The information available to policy-makers is contained in a multitude of lengthy reports supported by tables and diagrams. These can of course be put straight onto sophisticated computer systems in toto. But the basic problem still remains how to ask the question relevant to the policy process - it is not a retrieval problem. Information systems give no assistance inthisrespect.

Media - oriented techniques: To go to another extreme, those concerned with facilitating understanding of complex issues by the public (and this may well include decision-makers) use media-oriented techniques. Great emphasis was placed on films at the UN Human Settlements Conference. Books attempting to describe social change make much use of mcluhanesque illustrations (40 ). But despite the gain in visual interest and emotional appeal, the value of such superficial displays for policy-making itself must be questioned. Aesthetic constraints too frequently conceal important issues.

Mathematical models : Another extreme is provided by the computer- based mathematical model interrelating hundreds or thousands of equations. These may be satisfactory where no policy problems have been avoided in constructing the model and there is consensus that it reflects the social reality it purports to model. This is rarely the case. Furthermore such models tend to be incomprehensible to all but their creators and critics. Again they do not help the policy maker to determine which questions to ask, but only answer those he chooses to ask (many "answers" having been built into the design of the model anyway).

Graphic models: An intermediate approach involves the use of graphic, two-dimensional, non-mathematical models. Such models are a symbolic representation of the various aspects of a complex event or situation, and their interrelationships (41). They are analogies which policy-makers may use to clarify their thinking about a relatively complex situation. They range from organization charts through to system flow charts, including the many kinds of schematic diagram that are prepared on flip-charts or slides for presentation purposes. They are widely and successfully used. Their main disadvantage is that only a limited number of elements and relationships can be incorporated in the model if their comprehensibility is not to be lost - the extreme case being the complex system or circuit diagram only comprehensible to the expert. None of these approaches is immediately relevant to improving the information problem in relation to the development processes. Each of them indicate: constraints and some offer clues to a new approach.