Herding Cats: Observing Live Coding in the Wild

After an eventful decade of live-coding activities, this article seeks to explore the practice with the aim of situating it in the history of contemporary arts and music. The article introduces several key points of investigation in live-coding research and discusses some examples of how live-coding practitioners engage with these points in their system design and performances. In light of the extremely diverse manifestations of live-coding activities, the problem of defining the practice is discussed, and the question is raised whether live coding is actually necessary as an independent category. This journal issue celebrates the tenth anniversary of organized live coding (see www.toplap.org). Numerous live-coding systems, festivals, conference tracks, journal issues, research projects, and club nights have appeared and introduced the practice to diverse fields of art, music, and science (in particular, computer science). As an arts practice, it has its roots in musical performance, but live coding has become common in visual arts, light systems, robotics, dance, poetry, and other art forms that operate with algorithmic instructions. In the context of this journal issue, it seems appropriate to stress the origin of live-coding practice in the arts, without failing to mention that the related term of “live programming” has been used for a considerably longer time in certain research tracks of computer science, particularly in the field of programming-language design. Researchers like Ungar and Smith explored live-programming practice as early as 1969 (Ungar and Smith 2013) and, in an article on the visual programming language VIVA, Tanimoto (1990) defined “liveness” as an attribute of programming. Live programming is thus perceived to be a useful method in activities ranging from re-programming factory production lines (Swift et al. 2013) to making music in nightclubs. This article will describe live coding as a unique practice in a strong relationship with live programming but stress its origins in live performance. Journal articles and conference papers on live coding, typically written by the protagonists themselves, have introduced the practice (Collins et al. 2003; Ward et al. 2004; McLean 2008), explored it in a computer science context (Blackwell and Collins 2005; Rohrhuber, de Campo, and Wieser 2005; Computer Music Journal, 38:1, pp. 8–16, Spring 2014 doi:10.1162/COMJ a 00216 c © 2014 Massachusetts Institute of Technology. McLean and Wiggins 2010; Sorensen and Gardner 2010), described particular systems and solutions (Sorensen 2005; Wakefield, Smith, and Roberts 2010; Freeman and Van Troyer 2011; Magnusson 2011a; McLean and Wiggins 2011; Roberts, Wakefield, and Wright 2013), explored live coding as musical scores (Blackwell and Collins 2005; Magnusson 2011b), and contextualized it as an embodied musical practice that requires practicing on a par with acoustic instruments (Sorensen and Brown 2007; Aaron et al. 2011; Collins 2011). After a decade of fruitful experiments, it is perhaps time to reflect on how live coding has operated within the performing arts. Discussing a selection of highly diverse live-coding systems, this article seeks to establish what they may have in common, resulting in a discussion of the problems of defining live coding, and, in the process, to introduce weak and strong criteria for the practice. On Naturalizing Live Coding Live coding does not have a particular unified aesthetic in terms of musical or visual style. Nevertheless, the practice has at times been perceived as a movement, akin to movements found in 20th century modernism. We find a strong emphasis on formal experiments, reductionism, and functionalism. There are manifestos, key texts, and custom coding platforms. A visit to a nightclub hosting a live-coding event might even conjure up images of exclusive avant-garde practices, where live coders perform for other coders already initiated into the wicked “sourcery” of programming computers. This art of writing algorithms for binary machines can be so alien and obscure to the audience that the situation almost recalls the difficulty people had understanding postwar European avant-garde music. 8 Computer Music Journal This image is as familiar as it is misleading. It is true that the birth of live coding draws from modernist practices (with its manifestos, rules, and imperatives) but this is inevitable, as formalism is—and formal experiments are—a necessary aspect in the exploration of a new medium; suffice it to mention video art experiments by Nam June Paik and the Vasulkas in the 1960s, or net.art work by artists such as Alexei Shulgin and Jodi in the late 1990s. Such a formalism involves a deep exploration of the properties of the medium at hand, and we find an analogy in the way live coders have designed and performed with their systems. In live coding, however, the investigation tends to be a formalism of thought and the language or system of encoding it, as opposed to artistic content, which may appear formalist or not. Perhaps the situation is better described by the modernist critic Clement Greenberg, who defines it as one in which the two aspects become inseparable: Content is to be dissolved so completely into form that the work of art or literature cannot be reduced in whole or in part to anything not itself [. . .] In turning his attention away from subject matter of common experience, the poet or artist turns it in upon the medium of his own craft (Greenberg 1961, p. 6). The performance-art elements of live coding differentiate it from the pure self-referentiality of formalist modernism, however—although there could, of course, exist a purely conceptual live coding without any other output than the code itself. I am not aware of such an approach in the area of live coding, although it exists in off-line coding, perhaps best exemplified by Pall Thayer’s Microcodes (Myers 2009). The live-coding activities that resemble the aforementioned modernist tendencies can be explained with a common trajectory that takes place when a new artistic format develops. The initial focus is on the formal part of the practice, on the medium or the tool, and it is only later, when the technology undergoes a process of “naturalization,” that the focus shifts elsewhere: The more naturalized the object becomes, the more unquestioning the relationship of the community to it; the more invisible the contingent and historical circumstances of its birth, the more it sinks into the community’s routinely forgotten memory (Bowker and Star 2000, p. 299). Live-coding practice has consciously put effort into expediting this naturalization process. It deliberately engages with the audience through various channels, such as sitting among them while performing (as with the band PowerBooks UnPlugged), allowing people to contribute to the coding of dance performers (e.g., Kate Sicchio’s work), or submitting code through Twitter (as in my own work with the live-coding environment “ixi lang”). One of the fundamental tenets of the TOPLAP manifesto (available online at toplap.org/wiki/ManifestoDraft) is “Show us your screens.” It is an explicit act of audience inclusion, responding to the common laptop-performance format where relatively simple interfaces are used but not shown, even though many are commonly known to the audience. This is taking the etymology of the word “program” seriously, as the Greek root, prográphein, signifies the activity of public writing (Hoad 1996). Live coders have also been prolific in explaining their practice and systems of writing code, with journal articles, online discussions, and similar activities (this article [and this paragraph] being an example of such recursion). Live Coding in the Wild Live coding is a heterogeneous practice and thus somewhat hard to define. It involves a multiplicity of approaches that have one thing in common: Algorithmic instructions are written in real time. Collins (2011, p. 209) states that “the more profound the live coding, the more a performer must confront the running algorithm, and the more significant the intervention in the works, the deeper the coding act.” For Collins, most performances fail to “live up to this promise.” The argument, also expressed by McLean (2008), is that a true “liveness” requires that the performer is not simply manipulating prewritten code in real time, but is actually writing and