This chapter describes the process of creating an interactive web-based tool, named Audit, which allows students of the Royal College of Music, London, to practice a particular task from their aural-skills syllabus: the singing-back, upon request, of specified notes from four-part chords played on the piano. Audit generates practice questions dynamically according to a set of programmatically defined rules and procedures, creating an effectively unlimited resource.

Although it has been difficult to draw meaningful conclusions about Audit’s value as a means of facilitating the acquisition and development of relevant skills from initial quantitative assessments, reassurance may be found in the fact that the software has already been shaped by a more qualitative evaluation process: as its generative rules and procedures were gradually refined, the suitability and difficulty level of the questions it produced was compared to that of questions used in previous years’ assessments. This process, as well as decisions taken regarding Audit’s user interface and technical implementation, involved probing and articulating largely unwritten knowledge and reasoning, informally accumulated in the classroom, revealing a number of hypotheses as to what makes chords more or less difficult to analyze in this way, what perceptual processes the task entails, what makes a chord ‘well-spaced,’ and what constitutes helpful practice.

While some of these hypotheses and intuitions are supported by existing research (concerning, for example, the relative ease of processing consonance versus dissonance), others (that a greater number of similar intervals within a chord make its constituent notes harder to distinguish, say) have not yet been subject to such empirical investigation, and therefore suggest possible lines of future enquiry. The ‘deeper understanding of the subject matter’ (Jonassen & Wang) that has been gained from developing Audit can also now be fed back into ongoing teaching practice.