A cognitive dimensions analysis of interaction design for algorithmic composition software

This paper presents an analysis of the user interfaces of a range of algorithmic music composition software using the Cognitive Dimensions of Notations as the main analysis tool. Findings include the following: much of the reviewed software exhibits a low viscosity and requires significant user knowledge. The use of metaphor (staff notation, music production hardware) introduces multiple levels of abstraction which the user has to understand in order to use effectively: some instances of close mapping reduce abstraction but require the user to do more work. Significant premature commitment is not conducive to music composition, and there are clear opportunities for the greater provisionality that a piece of structurally-aware music software could provide. Visibility and juxtaposability are frequently compromised by complex design. Patching software reduces the hard mental operations required of the user by making the signal flow clear, although graphical complexity can have a negative impact on role-expressiveness. Complexity leads to error-proneness in several instances, although there are some tools (such as error-checking and auto-completion) which seek to ameliorate the main problem