In this article, the author, having surveyed the state of the art by reviewing some of the systems for computer assisted composition, algorithmic composition, and music signal processing of the past twenty years up to 1996, notes that often the hardware technology of the 1990s uses software technology of the 1970s and 1980s to realise the musical concepts of the 1960s. He argues that new compositional concepts are needed that keep pace with the developments in computer music systems.