Software development management: Empirical and analytical perspectives

Keumseok Kang, Purdue University

Abstract

Managing software development is a very complex activity because it must deal with people, organizations, technologies, and business processes. My dissertation consists of three studies that examine software development management from various perspectives. The first study empirically investigates the impacts of prior experience with similar projects (i.e., learning effects) and the impact of the length of time between similar projects (i.e., forgetting effects) on software development project productivity. Further, it focuses on the types of knowledge required for software development and how they impact the learning and forgetting effects. More specifically, it examines whether these learning and forgetting effects are componentized and differ by the type of software development knowledge—domain and development knowledge. This study tests the hypotheses using an extensive dataset on software development projects collected from an IT services firm in Korea and finds forgetting effects as well as differential learning effects. The second study examines the balance of specialized versus diverse experiences in software development. More specifically, it investigates the impacts of specialized experience (i.e., prior experience with the focal knowledge) and diverse experience (i.e., prior experience with other non-focal knowledge) on software development project outcomes. Using the same data set from the first study, this study tests the impact of these two types of experiences for different types of project members (i.e., project managers versus non-project managers) in different types of software development knowledge (i.e., domain and development knowledge) and their impacts on project productivity. The second study finds that the appropriate balance of these two types of experience varies by the project role and knowledge type. Lastly, since typical software organizations must juggle multiple projects at once and because there are numerous constraints to contend with, selecting the right software projects and assigning appropriate resources to selected projects is an important decision for software organizations. The final study proposes an analytical model of this multi-project selection and assignment problem and provides a solution for it. The results of the three studies provide managerial implications to the organizational issues of software organizations such as project staffing, knowledge development, career development, and project selection.

Degree

Ph.D.

Advisors

Hahn, Purdue University.

Subject Area

Management|Information Technology

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS