The Problem: How to perform a software development project in university settings, where there is high volatility of tasks and interrupts, necessity of handling several projects in parallel, and crew of part-time student developers instead of a full-time dedicated team.

My del.icio.us trail using "agile process university research project" tags

Crumbs of wisdom collected thus far

People, roles

• myself as Scrum Master or the like - act as lubricant, i.e. devote time to coordinating and use spare moments for programming [discussion with XP workshop at ESEC/FSE, infoq 2008/05/handling-interruptions]
• "you have 5 part-time people and the customer thinks you have a team while in reality you have 1 full-time person, if not less" [infoq 2008/05/handling-interruptions]

Process, activities

• iterations of 4-8 weeks (remember the team is working with 20-30% allocated time at the best, so real amount is 1-2 weeks of work)
• plan iterations with half-fixed deadline: set deadline, if really necessary move it during iteration but not too far - balance "cut features" and "extend time" reasonably [self]
• try hard to have weekly or bi-weekly meetings, roughly in SCRUM daily meeting style [self]

Tools

• use bugtracker as planning tool [self]
• whiteboard, digital camera, and wiki as meeting documentation tools [self, after agile modeling]

Resources

• Cornell Uni Project Management Methodology -- http://www2.cit.cornell.edu/computer/robohelp/cpmm/CPMM_Guidebook.htm
  • looks rather waterfall-like (had only a first glance, though)
  • nicely detailed -- this is how ASWI process in RMC could ultimately look like