Arno F. Granados (NRAO)
In order to analyze the current state of scientific software development in a qualifiable context, the author has written a ``Software Development Methodology Survey.'' The purpose of the survey is to obtain a formal baseline of understanding of the software development methodologies in use in scientific computing-what works and what doesn't. Survey topics include: To what degree do institutions separate science from coding? Are software groups creating formal requirements documents? How are developers obtaining requirements? Are software projects being tracked by project managers? Are developers using commercial tools (e.g., defect tracking, CASE tools, GUI tools), or building tools in-house? Do coding groups use techniques such as walk throughs and code-reviews to limit code defects and improve reliability and maintainability? Are formal development procedures helpful or just too much overhead? What percentage of software groups have formal training in programming or computer science?
This survey will be distributed to attendees of ADASS'98, and the results and analysis will be presented at ADASS'99.