Much improvised music that has developed since the advent of free jazz has been concerned with the imposition of structure, often through systems of directed improvisation, or through the use of rule-based approaches (e.g., game pieces). In this article, we explore the possibility of a networked live-coding system as a structural intervention mechanism par excellence, through the discussion of two pieces from the repertoire of the Birmingham Ensemble for Electroacoustic Research.The relationship between live coding and improvisation is generally understood in terms of the latter being an implicit aspect of the former, or as the former being a specific case of the latter (see, for example, the discussion in McLean and Wiggins 2010). Live coding provides one fertile solution to the problem of interface design for musical performance, with rich implications for improvisational practice. Unlike most computer music performance systems, which tend to consist of relatively fixed configurations of graphical user interface (GUI) widgets and/or gestural controllers, "code as interface" allows radical intervention and reconfiguration of musical systems while they are running. On the one hand, it becomes possible to do many previously unimaginable things; on the other, it is often not possible to do them very quickly.In terms of this flexibility at least, live coding might be seen as an improvisational interface par excellence, albeit one often lacking in agility and deftness. Like much post-free-jazz improvisation, however, the practice of live coding can present problems in terms of creating formal articulation. Moreover, as in free improvisation, a substantial