Adding Live Coding Music Capabilities to the MicroBlocks Educational Programming Language

This paper describes the process of adding live music coding capabilities to a programming language for microcon-
trollers not originally designed for that purpose. MicroBlocks, both developed and used by the authors, was designed
from the ground up to be interactive, concurrent, and welcoming and intuitive to beginners. Those features, along with
a visual code representation that makes the code more readable by the audience in a live-coding performance, make
MicroBlocks attractive for live coding music. When designing a live coding music language, two of the most essential
components are the timing engine and the sound generation engine. Although MicroBlocks had neither of those com-
ponents, testing showed that its virtual machine is fast and efficient, with low-latency task switches. Those properties
allowed us to build rhythm abstractions in MicroBlocks with timing precision comparable to many production live
coding languages. Adding sound generation to MicroBlocks posed more of a problem. After exploring three different
methods, our preferred solution is to externalize sound generation by sending commands to a General MIDI sound
module. In this paper, we describe a set of libraries that make it possible to perform live-coded music with MicroBlocks.
We explain the reasoning behind each abstraction and show examples of its use. MicroBlocks has now been used in
numerous live coding performances in Barcelona and elsewhere. Overall, we’ve found MicroBlocks to be a viable option
for live music performances. Furthermore, since MicroBlocks is easy to learn, it may be useful in music education and
welcoming to experienced musicians with limited coding experience.