Improvisation, Computers, and Interaction : Rethinking Human-Computer Interaction Through Music

Interaction is an integral part of all music. Interaction is part of listening, of playing, of composing and even of thinking about music. In this thesis the multiplicity of modes in which one may engage interactively in, through and with music is the starting point for rethinking Human-Computer Interaction in general and Interactive Music in particular. I propose that in Human-Computer interaction the methodology of control, interaction-as-control, in certain cases should be given up in favor for a more dynamic and reciprocal mode of interaction, interaction-as-difference: Interaction as an activity concerned with inducing differences that make a difference. Interaction-as-difference suggests a kind of parallelity rather than click-and-response. In essence, the movement from control to difference was a result of rediscovering the power of improvisation as a method for organizing and constructing musical content and is not to be understood as an opposition: It is rather a broadening of the more common paradigm of direct manipulation in Human-Computer Interaction. Improvisation is at the heart of all the sub-projects included in this thesis, also, in fact, in those that are not immediately related to music but more geared towards computation. Trusting the self-organizing aspect of musical improvisation, and allowing it to diffuse into other areas of my practice, constitutes the pivotal change that has radically influenced my artistic practice. Furthermore, is the work-in-movement (re-)introduced as a work kind that encompasses radically open works. The work-in-movement, presented and exemplified by a piece for guitar and computer, requires different modes of representation as the traditional musical score is too restrictive and is not able to communicate that which is the most central aspect: the collaboration, negotiation and interaction. The Integra framework and the relational database model with its corresponding XML representation is proposed as a means to produce annotated scores that carry past ...