Working session: empirical studies of programming-in-the-large: how?

Abstract

Studying software design is a juggling act of tradeoffs and constraints. There are good, pragmatic reasons why most empirical studies of programming and software engineering have focused on programming-in-the-smalleven programming-in-the-miniature. Such research has certainly revealed some useful things. However, have we made our focus so small that we have not even noticed the real design issues? The difference between programming-in-the-small and programming-in-the large is measured, not just in lines of code, but also in data, roll calls, engineering processes, tools and environments, and - crucially - time.How can we scale up the focus of our empirical studies without exceeding feasibility? Is it enough just to change the granularity of our examination? How can we take a longitudinal view, examining software development over the whole project lifetime, rather than within a given hour in the life of a program? How at least can we look enough at programming-in-the-large to identify issues that are small, tractable, and critical? This session will begin to address the application of empirical methods to the study of programming-in-the-large, and will consider what lessons might be drawn from past research experience.