Advice to the course staff

We have several suggestions:

• More dedicated machines are desperately needed. It is impossible to apply meaningful tests to an application that requires sound hardware without access to that hardware. Likewise, it is foolish to try to test a network application on only one machine (in the absence of a simulator like Entrapid). Why, then, was each group allocated only one audio-capable workstation, over which ten or eleven teams were forced to fight? Not everyone required the audio hardware, but many groups needed the administrator access we were granted on those machines. It is one thing to contend with an entire class for CPU cycles on one easily overloaded Unix box. It is quite another to compete with ten other teams of students for use of a single NT box to which access is possible only from the console. It was not an uncommon sight in the last few days before deadline to see three people's hands hovering over one keyboard, engaged in a poignant demonstration of the multiple access problem.

• If the project is expected to simulate a real-world development experience, then there needs to be some incentive for writing scalable and maintainable code. It was far too tempting to make huge stylistic blunders in the interest of time, since we knew we would never have to look at our code again (much less document it). Such circumstances are rare in the professional world.

• There was far too much interdependence among teams--we felt practically deadlocked at times. Because almost no team could finish its testing without final working code from several other teams, practically our entire group was in the lab until 6 a.m. or later the night before our demo. It would have been nice if crucial parts of the system could have been completed well in advance of the deadline, but the inevitable end-of-semester workloads made that impossible. We appreciate the philosophy of learning by doing, but in this case it seemed that the we did not learn as much about computer networks as we did about working in large groups with insufficient resources.

• Although the idea of teamwork is a novel notion. We personally believe that this project might have taken that to an extreme especially considering the college environment that we are in, where everyone has different paper, exam, and project schedules. It also makes certain teams such as ours too highly dependent on others.

• Given the vast amount of work required for this course, it might be beneficial to separate the course into (perhaps corequisite) lecture and practicum components. Regardless, four credit hours certainly do not suffice.