The study described in this paper is a timely topic. Gotel et al. describe the experience of having five globally dispersed software development teams--composed of students from various universities--complete a project.

The client is the Institute of Technology of Cambodia (ITC). The project is to build a library system for its computer science (CS) department. The five development teams that compete to fulfill this project are University of Delhi, India; Mahidol University, Thailand; Pace University, New York City campus; Pace University, Pleasantville, New York campus; and ITC itself. In addition, a Cambodian university--the Royal University of Phnom Penh (RUPP)--served as a socialization team that acclimatized students from other cultures to the Cambodian culture. There was also a team of 16 US graduates and information technology (IT) professionals that served as auditors, and five US graduate students who helped each development team inject quality into their processes and products--sadly, the paper does not reflect the role of these two teams in any detail. The software was developed in a modified waterfall model, with iterations thrown in when requirements needed to be further refined.

Of the 71 functional requirements, the ITC team implemented 24, the Pleasantville team implemented 33, the NYC team implemented 64, the Indian team implemented 66, and the Thai team implemented 68. The teams with the three highest implementation metrics were chosen as winners; no single winner was declared. It is extremely interesting that the ITC team implemented the least number of functional requirements despite being in proximity to the client. Gotel et al. do not posit why this may have been the case, beyond just attributing it to power cuts experienced in Cambodia and the subsequent lack of access to servers. If this is indeed true, were there no power cuts in Thailand or India? And if there were, why did the Cambodian team fare worse, despite other advantages? Did the auditors spot anything specific to the Cambodian team that made them perform worse? In the same vein, did the auditors note why the NYC, Indian, and Thai teams fared better? Unfortunately, these discussions are not included in the paper.

While reading the paper, I was also intrigued with other facts. There is almost no mention of Skype or Yahoo! Voice as communication mediums; this is understandable for the early phases of the study, from 2005 to 2006, but not for 2007 and 2008, by which time these tools were widely used in the IT community. The data shows that the local development teams used phone and face-to-face meetings to communicate, but the global team did not, on account of voice/video conversations being, supposedly, too difficult to set up. This struck me as particularly interesting, given that the diaspora of Asian students in the US was among the first to use voice/video communications on the Internet (such as Skype and Yahoo! Voice).

The paper concludes with a table, “Tooling Do’s and Don’ts for Educators,” and some recommendations for the industry on how to structure global software development projects. I think that this list will be beneficial to corporations interested in offshore projects. Indeed, this paper will give corporations a good idea of the challenges and possibilities faced when moving work offshore.