Towards GEEZMO: hiGh-frEquEncy Zest and Mood-pOlling for proactive software development problem-solving

Abstract

Development of software is happening on an increasingly distributed fashion. Individuals normally coordinate and interact over any combination of IRC rooms, mailing lists, private email, etc. Conversely, big players like Google Inc. employ yearly Googlegeist polls asking its employees how they feel about the company, its directions and its managers. As a consequence, the amount of time required for managers and management to feel the pulse of their subordinates is removed from actual work. We argue that software development status can be computed by eliciting just a few bits of information that can be anonymously extracted by a single poll - the poll can be completed very quickly and its administration can happen as a simple prompt inside windows that developers would be visiting anyway (e.g., the log-in screen, commit screens, etc.). This paper elaborates this idea and develops a prototype to articulate the idea in practice. Finally, the idea is discussed using a preliminary validation by means of statistical methods and simulations.