Situated Software Development: Work Practice and Infrastructure are Mutually Constitutive

摘要

Software developers' work is much more interesting and multifarious in practice than formal definitions of software development processes imply. Rational models of work are often representations of processes defined as they should be performed, rather than portrayals of what people actually do in practice. These models offer a simplified picture of the phenomena involved, and are frequently confused with how the work is carried out in reality, or they are advocated as the ideal way to accomplish the work. A longitudinal ethnographic study (45 days of fieldwork over 20 months) of a group of professional software developers revealed the importance of including their observed practice, and the "infrastructure" that supports and shapes this practice, in an authentic account of their work. Moreover, this research revealed that software development work practice and the infrastructure used to produce software are inextricably entwined and mutually constitutive over time.