In this paper, we consider issues of visualization from a programming perspective, proposing fundamental concepts that we need for successful visualization systems and suggesting general principles. We shall argue that there is a close parallel to be exploited between observation of physical phenomena and the interpretation of programs - in the real-world rather than abstract computational sense. In particular, we suggest that developing programming methods well-suited to scientific visualisation is intimately connected with the problem of making programs intelligible.
The paper is in four main sections. Section 1 considers scientific visualization in the abstract. Section 2 considers visualization in its relation to computational concerns. Sections 3 and 4 introduce an experimental system we have developed for visualization, indicating the principles on which it is based and some of its applications. The experiments and observations in our title refer both to the nature of our software prototyping and to the fundamental principles upon which our approach to programming is based.