In this paper, the author reflects on his experiences of compositional activity in the field of computer music in order to penetrate the twists and turns of composition in the Western world (including what was then the USSR) and to propose a compositional model. His discussion focuses on the general environment in which music exists and the status of the musical sign on the then current (1985) musical culture. This focus on the semantic aspects introduces two models that the author feels offer the best models for a semantic representation of works, namely, the MSR (Semantically Reformulated Model) and, Q (after Ross Quillian) models.