Greg Wilson has just posted his 3 rules for supervising student programming projects. Since he’s been running very successful student projects for years, it’s not surprising to see that his article is packed full of insights into what makes a project work, and the pitfalls ready to swallow up the unwary.
If you generalize Greg’s rules just a bit, they apply much more widely than to programming projects. For example, in my own undergraduate physics education I did several projects where the goal was to do a small piece of original research and write it up or present it in a talk, and I watched many colleagues attempt their own projects. Those whose supervisors obeyed Greg’s rules were generally the most successful and least angst-ridden for the students.