The practice of coding music live challenges computing conventions with regard to developmental agility. The computational representation of music likewise challenges musicians to articulate their practice in new ways. In this article we describe the development, teaching and use of the JythonMusic environment designed to meet these challenges head on. JythonMusic, written in Python, is an open source project for music making and creative programming activities intended for musicians and programmers, of all levels and backgrounds. JythonMusic supports algorithmic music composition, dynamic coding and musical performativity including live coding. This article examines the functionality of JythonMusic for composition and performance projects in the context of university courses that combine computer science and music. The advantages of this medium for music and programming instruction are demonstrated through several projects and a live coding case study, leading to a series of observations and proposals concerning the advantages of coding in music pedagogy.