Teaching rule‐based algorithmic composition: the PWGL library cluster rules

Abstract

This paper presents software suitable for undergraduate students to implement computer programs that compose music. The software offers a low floor (students easily get started) but also a high ceiling (complex compositional theories can be modelled). Our students are particularly interested in tonal music: such aesthetic preferences are supported, without stylistically restricting users of the software. We use a rule‐based approach (constraint programming) to allow for great flexibility. Our software Cluster Rules implements a collection of compositional rules on rhythm, harmony, melody, and counterpoint for the new music constraint system Cluster Engine by Örjan Sandred. The software offers a low floor by observing several guidelines. The programming environment uses visual programming (Cluster Rules and Cluster Engine extend the algorithmic composition system PWGL). Further, music theory definitions follow a template, so students can learn from examples how to create their own definitions. Finally, students are offered a collection of predefined rules, which they can freely combine in their own definitions. Music Technology students, including students without any prior computer programming experience, have successfully used the software. Students used the musical results of their computer programs to create original compositions. The software is also interesting for postgraduate students, composers and researchers. Complex polyphonic constraint problems are supported (high ceiling). Users can freely define their own rules and combine them with predefined rules. Also, Cluster Engine’s efficient search algorithm makes advanced problems solvable in practice.