Valid models are central to the existence of Computer science as in most other disciplines, but at what point can one say that a model is valid, correct and hence yield information? A model is often taken to be an abstraction and simplification of reality (real system being modelled) but reality (the nature of measured data, environmental and human factors) in itself, has a nature of abstract complexity; hence a 'correct' model could at best be judged as one which is 'closest' in representation to the real system, but the question is: just exactly how close should 'closest' be to be correct? In this paper, we examine the correctness criteria for models' validation and seek to relate them to various philosophical perspectives to see how much information and knowledge (content and truth) the basis of acceptance of such valid models could give. We conclude that models should not be used as substitutes or sole basis for critical thoughts, or major decisions but should be viewed just as tools for improving judgement and intuition.
展开▼
