Software Complexity: Towards a Distributed Governance of a Production System

This PhD is interested in the complexity which arises in software production due to the divergencies of the organisational and market forces and aims to find out what are the existing or potential new ways of mitigating the complexity effects. Viewing the distributed kind of governance characteristic for most complex systems together with the value produced, I discover that the complexity is often caused by software capitalism – a valorisation regime which perceives every firm primarily as a technology firm and creates profits by keeping its software in a state of perpetual disrepair. The software in such a regime acts as an interface between the domain of market exchange and the sphere of organisational culture, with the market tending to increase complexity so that the production intensifies, and the organisation resisting this tendency to be able to optimise the audit of the production performance. I approach the study from the standpoint of development operations (DevOps) to reveal the mobilisation of the software system’s epistemology as a productive assemblage for planning and control, which becomes a key dynamic in the situation of uncertainty that complexity presents.

Coming from the experience of empirical work as a digital product lead, I try to view DevOps with diffractive and compositional optics, by explaining a software production system through the notions of the problem space and epistemic infrastructure. This makes it possible to clarify the performative role of a software system in the capitalist mode of production as simultaneously a service and a product. While the software system is at any moment too complex to be fully repaired, its dysfunctional condition is further aggravated by the phenomenon of the falling cost of computation. Acknowledging this trend, I argue that mitigation methods should be investigated in a diffractive way, effectively attempting a queer methodology for DevOps, to open a conversation about a more-than-human representation of the space of production that sees the negotiation of shared meanings topologically and immanently, rather than based on the dominant hierarchies, pre-existing assumptions or external evaluation.