ANALYSIS AND ORGANIZATION OF KNOWLEDGE FOR RETRIEVAL

In a university, the mode of research is usually what is called ‘pure’ or ‘basic’ research; since I am keeping in mind primarily the applications of information science, I will prefer the word ‘basic’, although there is not so much difference. In such research, and really in any good research, one should not be collecting data haphazardly. One must isolate and define a problem and, as far as possible, control other conditions so that interfering factors are eliminated. Preferably one will narrow down the problem to manageable proportions. It is then essential to approach the problem with some sort of hypothesis or theory of the situation, and to concentrate on obtaining evidence for or against that hypothesis. The important task is to devise just that crucial experiment which will give the answer as efficiently as possible. If the answer disproves the hypothesis, one has at least further evidence upon which to construct a different hypothesis; if it confirms the hypothesis, one is ready for a further step forward, and so on. Research is easier in a fully controlled and reproducible situation; in a biological or human situation one must often have recourse to statistical methods, but this does not alter the general methodology. On the whole, I find a clear methodology lacking in much that is being done in the field of information science today.