Research centre
Communication and Computing Research Centre

Date
2009 - 2014

Research Degree Project

I completed my PhD studies, investigating the communication practices of Software Engineers working in Agile teams.

There is a stereotype of the software engineers as a lone coder hacking away in a darkened room or cubicle surrounded by thick programming books and Star Wars or Red Dwarf memorabilia. In the popular imagination these engineers aren't interested in the users of their software and tend not to work well in teams. I'm not sure that this cliché has any validity today, and I'm not really sure that it ever did. The development of software is a communal activity which involve teams of designers, testers, coders and users.

Agile methods are predicated on the social aspects of software development. All of them base their practices on people working closely together rather than going off and doing their own thing. Think about the XP practice of pair programming. If two developers share a single keyboard and screen then they must talk about the code which they are writing. On-site customers are there to talk to the developers. And methods such as scrum are built on regular meetings within the team. You can't be an agile developer without also being a social developer.

Social scientists have studied the way in which people communicate in many aspects of life. By studying conversations they have gained new understanding of the working lives of doctors, drug counsellors, designers and many other professionals. Social scientists, typically ethnographers, have also sometimes been employed on software development projects to understand end-users.

There is a close relationship between work and sociology yet the actual working life of software engineers has had surprisingly little study. I am stepping in to that space. My starting point is the regular meetings which happen in scrum projects. How do developers communicate in these meetings as they negotiate and deliver their client's brief? Are they really helping each other, or are the meetings used just for management and control? If the latter then how do agile methods affect the way that management is done?

I have made ethnographic studies of several software houses observing the programmers' working practices. Analytically I am using Conversation Analysis to reveal how programmers talk to each other as they negotiate the coordination of their work and the meaning of their code. This will reveal patterns of interaction typical of working developers and show how they use talk to achieve their specific project-oriented goals.

Project Supervisors

• Dr Kathy Doherty  (Director of Study)
• Dr Karen Grainger (Second Supervisor)

Researchers involved

Dr Chris Bates - Senior Lecturer in Software Engineering