Since the formulation of first development methods such as waterfall and later method innovations such as iterative and agile, specific forms of sequences of different tasks have been viewed instrumental for effective software development given the logical and temporal dependencies between tasks and their inputs or outputs. In open-source software (OSS) no normative guidelines are available to enforce a preferred temporal structure of activities. However, even in OSS contexts some patterns are likely to be more effective than others. Alas, not much is known of the effects of time on OSS development outcomes similar to studies in vertically integrated settings. We address this gap by drawing on the notion of temporal structuring of clock-time and event time based on coding and commenting activities referred to as “spells” on software development effectiveness defined as the rate of completing opened tasks. We validate the effects of specific spells on development activities by analysing temporal structuring of development activity in a well-known, successful OSS project Apache Hadoop over a ten-year period comprising of 9865 tickets, 157721 activities and 276 releases. We identify spells of coding and commenting for each ticket and calculate their average length and estimate the effect on task completion. Overall, we find that a unit increase in event-based workflow coding structure increases the rate of ticket success by 17% and that the effect is moderated by clock-based coding and commenting structure.
