Software rarely works as intended while it is being written. Things go wrong in the midst of everyday practice, and developers are commonly understood to form theories and strategies for dealing with them. Errors in this sense are not bugs left behind in software, they are actively encountered and experienced. This paper reports findings of an ethnographically-informed study undertaken to examine error encountered at the desk. Films depicting paired open-source development practice over the course of a month were analyzed to identify and delineate instances of active error. Instances were interpreted within a framework of error handling drawn from psychology research. Analyses of representative instances are given and discussed in relation to software engineering research that examines practice at the desk. Findings demonstrate that the significance of active error in software development is personal, shaped by passing time, the emergence of preferred practices and environmental changes.