Paper short abstract:

This paper asks how the "material agency" (Pickering) of computer programs manifests itself in the discussions of developers. It argues that developers explore the technological possibilities and limitations of computer programs in their work and that these evolve over the course of a project.

Paper long abstract:

This paper considers the collaborative work of software developers from an ethnographic perspective. It asks how the "material agency" (Andrew Pickering) of computer programs manifests itself in the discussions of developers. Modern digital computers provide possibilities, things that would not be possible without them, but also limitations, what is possible needs to be done in a certain way: e.g., digital computers require digitized inputs, and computer programs require unambiguous and explicit instructions formulated with the restricted vocabulary of a programming language. Software developers attend to these possibilities and limitations in their work. But whether a program does "what it is supposed to do" cannot be settled by the computer itself. This needs to be discovered over time by developers both by writing and testing the program and discussing what it does and does not do right (yet) in interaction with other developers and stakeholders. These discussions of developers in meetings are the focus of the paper. Empirically, the paper considers projects of computer scientists who specialize in semantic computing. While based on what is technologically possible, the work of development is far from being technologically determined, it is a practice where the interests of stakeholders and the conventions of computer science matter, with the computer program itself serving as a powerful mediator. This mediator evolves over the course of a project, in a way where its inherent limitations/possibilities evolve as well.