Roster-in-a-Box is a course management system. It includes modules for auto-graded course problems (currently for introductory statistics, introductory microeconomics, and intermediate microeconomics), as well as a mechanism for students to submit answers to text questions and have the instructor grade them online. Summaries of student performance (such as student problem areas or class averages by subject) are available to both the instructor and the student in real time, and the system can send reminders to students of late homework assignments.
Re: Already exist
XST had/has a lot of potential and I was really sad to see Ximian drop the project. It may have been overly ambitious in trying to support other unices (e.g., Solaris, HP-UX) in addition to Linux. But it had a clean separation of backend and frontend, as well as the ability to save old versions of the configuration, that gave its underlying structure a lot of potential. Although it has been developed for Gnome, there is no reason a KDE or web frontend couldn't be made, and in fact the developers have mentioned that they secretly hope someone might try such an undertaking. Although the proposed project here has the advantage of using a filesystem-like layout for configuration information, this does not seem like a big enough advantage to justify the duplicated effort.
Anyway, I agree with the authors that confusing and inconsistent configuration tools, not a lack of applications or features, are the biggest barrier to getting non-technical users to enjoy Linux. Mandrake shows some promise but RedHat is just awful -- even 8.0, while showing some improvement, has its setup tools lying all over the place (Some under System, some under Setup, some under "Start Here" -- which is not where I'd look for configuration tools unless I knew to) and lacks things like a decent firewall setup.
Whatever you guys end up doing, duplicated or not, will be a big help if you manage to bring it to completion.