Developing software across time zones: an exploratory empirical study

Abstract

The empirical study described in this paper was carried out as part of a research project investigating the potential for reducing the time needed to carry out software engineering tasks (and hence to deliver software products) through the use of around the clock working. The working style being addressed is one where a task is passed in a sequential manner from one software engineer to another ‘across time zones’. We call this sequential collaborative software engineering (SCSE). A general three-site scenario is illustrated in Figure 1, and shows: • the time differences between sites (which may involve some overlap); • the reporting time when, at the end of a working period or shift, a developer records progress made; • the catching up time, when a developer catches up with the work carried out during the previous shift. As well as undertaking an empirical study of SCSE, our research has involved the development of a set of equations which models the relationships between estimated development time using SCSE, estimated• distribution effort loss, which is the time lost when a developer at one site fails to complete a task on which a subsequent developer is depending. The aims of our empirical study were to illustrate the feasibility of employing this type of work pattern and to identify the critical success factors for SCSE. We also expected to obtain some values for the associated overheads (such as reporting time and catching up time) for the particular experimental context. The remainder of the paper is organised as follows. Section 2 and 3 describe the preparatory activities (such as the selection of subjects, the design of the study and document preparation) and the execution of the selected task. Section 4 and 5 summarise analysis and observations, critical success factors, and final remarks. 2 Preparatory activities