We rated the projects by secret ballot on each of a number of attributes. Note that most of the attributes don't have simple good/bad semantics - there are reasons for and against choosing an easy project. The fact that one will be used by lots of people and another might provide proof of concept for just one does not make one better than the other. Some of course do equate to good/bad - for example "is this fun?".

Note too that these are the opinions of prospective software engineers, not experienced managers.

After we collected this information we talked about it, hoping that some consensus about the least desirable project would emerge, since a formal vote seemed inappropriate. But none did.

So we proceeded to provide individual first and second choices to see what pattern might emerge. The results showed that the only way in which everyone could get one of his or her choices was to drop Forensic Botany. So we did.

Then after a sequence of simulated job interviews followed by lots of juggling of index cards to meet student preferences, balance teams and provide each team with the skills needed to get the job done, we arrived at the following arrangement. The teams are listed alphabetically, with team leader starred*