... also on July 6 I'll give a longer talk on algorithmic complementarity (abstract below).
http://www.computing-conference.ugent.be/hapop12/programme
details:
http://www.computing-conference.ugent.be/file/11
If any of you could point me to video material that you'd associate with this, I'd love to integrate parts of it into the second part of my talk.
\abstract{
Today one can say that programming has not only osmotically infused scientific and artistic research alike, but also that those new contexts elucidate what it may mean to be an algorithm. This talk will focus on the `impatient practices’ of experimental programming, which can never wait till the end, and for which it is essential that the modification of the program in some way integrates with its unfolding in time. A contemporary example is \emph{live coding}, which performs programming (usually of sound and visuals) as a form of improvisation.
Early in the history of computer languages, there was already a need felt for reprogramming processes at runtime. Nevertheless, this idea was of limited influence, maybe because, with increasing computational power, the fascination with interactive \emph{programs} eclipsed the desire for interactive \emph{programming}. This may not be an accidental omission, its reasons may also lie in a rather fundamental difficulty, on which we will focus here.
In itself, the situation is almost obvious: not every part of the program-as-description has an equivalent in the program-as-process. Despite each computational process having a dynamic nature, an integration of programming into the program itself must in principle remain incomplete. As a result, a programmer is forced to oscillate between mutually exclusive perspectives. Arguably, this oscillation reveals a specific internal contradiction within algorithms, a necessary obstacle to semantic transparency. By calling this obstacle \emph{algorithmic complementarity}, I intend to open it up for a discussion in a broader conceptual context, linking it with corresponding ideas from philosophy and physics.
Here a few words about this terminology. Complementarity has been an influential idea in conceptualising the relation between the object of investigation, as opposed to the epistemic apparatus and the history of practice. Originating in the psychology of William James, where it referred to a subjective split of mutually exclusive observations, Niels Bohr used it to denote the existence of incommensurable observables of a quantum system (position vs. momentum, time vs. energy). Independent of the particular answer Bohr gave, complementarity raises the question of whether such a coexistence is induced by the experimental system or already present in the subject matter observed. Accordingly, in the case of programs, we may ask whether this obstacle is essential to their nature or whether it is a mere artefact of a specific formalisation. Algorithms, arguably situated between technical method and mathematical object, make an interesting candidate for a reconsideration of this discourse.
The existence of an obstacle to semantic transparency within algorithms and their respective programs need not mean a relative impoverishment of computation. Conversely, prediction is the wager and vital tension in every experimental system, as well as in interactive programming. After the conceptual discussion, I will try to exemplify this claim by introducing a few examples in the recent history of \emph{live coding}. Again and again surfacing in form of symptoms such as an impossibility of immediacy, I hope this practice will be conceivable in terms of having algorithmic complementarity as one of its driving forces.
}
Received on Wed Jun 20 2012 - 10:11:49 BST