From software data to software theory: The path less traveled

From software data to software theory: The path less traveled

1. PATHWAYS OF SOFTWARE REPOSITORY RESEARCH

   • this software mass spectrometer changes our understanding of the world by making the previously invisible, visible, allowing us to ponder what it means.
   • observational research phase
     • frontline outputs of software repository analysis, the primary observations that researchers use as initial building blocks in their work
   • application of the observational knowledge toward the betterment of software engineers
     • information or visualizations that can support managerial decision making
     • pragmatic outcomes
   • theory formation.
     • use the frequency distributions and relationship networks to support fundamental understandings of what software is

2. FROM OBSERVATION, TO THEORY, TO PRACTICE

   • software power laws
     • preferential attachment
     • self-organized criticality
       • determines the size of software commits.
   • software entropy
     • “Preferential attachment” means that a trend toward large software files is inevitable in software systems, unless large files are continually refactored.
     • software entropy values reveal software as having a high degree of local regularity.

3. solving problems faced by practicing software engineers

   • pragmatic outcomes
   • fundamental theoretical questions that remain unanswered.
     • Consider the fact that most frequency distributions observed in software projects follow a power law or log normal distribution, to the extent that it is unusual to find one that is not. Why is that?