Someone needs to build a decent open source 3d brain model in shockwave, VRML or a stand alone OpenGL application (not pseudo 3d quicktime). As far as I can see, none exists. Believe me, when you’re trying to understand functional neuroanatomy, such a thing could not be more useful.
This is the sort of thing that education software does better than books or lectures, greatly accelerating the practical comprehension of complex 3d systems. Educational institutions would do well to develop such software for engineering, medicine, and physics modelling.

Start-ups, currently thinking about building yet another social bookmarking app or firefox extension, might find a market for an advertising sponsored (or paid institutional subscription) folksonomy application, which could allow lecturers to build three dimensional tours to accompany their lectures, or mail students links which would open specific interactive 3D representations of models discussed.

The whole copyright in education thing really came home to me in first year, when one of our lecturers (a brilliant speaker) delivered out lecture notes in PDF form, with all of the accompanying diagrams (most of complex brain regions) removed; making the notes literally useless for exam revision or essay writing. This is an area that needs to be urgently addressed, both legally (with intelligent copyright exemptions for educational use), and through the development of common open learning platforms.

Also, isn’t it obvious that universities should do their best to enable the creation of educational resources by their computer science / education / psychology departments – not in a service provision role, but a pragmatic brain trust style solution development role; by encouraging and financially facilitating interdepartmental projects where divergent expertise could be used to mutual benefit – sort of like an intra-faculty open source movement.