I'm beginning to define my white paper idea as "general principles that govern the interaction between software design and social practices".
Software design influences social practices of the user community. Can we define some general principles or guidelines of what makes a software package foster desirable social practices (e.g. exploration, collaboration, and ownership)?
I'm working with Pat Marsteller -- we've set up a page on the coursewiki at UMass for our paper: 
http://bcrc.bio.umass.edu/coursewiki/index.php/CrossingBoundariesFinal
You can see all of the projects people have submitted. It looks like an interesting cross-section of what we've been talking about.
We just saw a demonstration of Access Grid running over a machine with 3 projectors attached. It quit working, so they rebooted it, but then the multicast traffic would drop off every 3 or 4 minutes anyway. I'm fairly skeptical of these technologies: I think you can get a lot farther with simpler technologies. Still, it would be fun to have a room with three projectors in it,
even if we never used the Access Grid aspect of it.
One project I've been meaning to work on is the next phase of the Learning Goals project. I've been intending to advocate that the department begin trying to find ways we can assess how our students reach our learning goals. I just found out about the field tested learning assessment guide, which looks like it might be a good start.
The UMass Chancellor has released his first set of proposed cuts for next year. They include a variety of services on campus, including the Campus Chronicle and the Academic Instructional Media Services. The one I'll miss the most is the University Club, which has been zeroed out.