The case for experiment-oriented computing
Experimentation aspects (e.g., systematic observation, exploration of alternatives, formulation of hypotheses and empirical testing) can be found dispersed and intertwined in many software systems. Numerous examples are provided in this article. This suggests that experimental activity is a class of computation in its own right. By abstracting the relevant experimentation features, a general Experiment-Oriented Computing (EOC) approach, orthogonal to other Software Engineering issues, is formulated in this article. Through this separation of concerns, it is possible to clearly pursue both theoretical and applied research with respect to experimental aspects. Concrete directions for such research and development are also given to illustrate the value of the approach.