I’m really, really excited about the
Solaris 10 University Challenge
Contest. In short– you can win yourself $5000 and an Ultra 20 workstation (and $100,000 in credit
towards Sun gear for your university, for which you will presumably be fêted and immortalized).
If you work in a team, each of you (up to 4) can win $5000!

Anyway, I wasn’t terribly impressed with the list of suggestions
we posted, and then Adam beat me to
posting a new list of ideas
by just a couple of hours. Here are some ideas I’ve had. Perhaps we should start a master list of
ideas somewhere?

Hardware Support

• Write a kernel framework for USB network adapters (and USB wireless adapters) and some drivers.
• Add Intel Speedstep support (esp. tricky on MP systems)
• Add a Bluetooth stack, and some useful apps
• Add an IRDA stack, and some useful apps
• Port GnomeMeeting to work properly with the Solaris 1394 camera driver (available for x86 in
  the latest express builds).

Other Interesting Technologies

• Demonstrate placing a real VoIP call using Solaris (including porting a
  decent client. Note that GnomeMeeting is working on SIP support– see previous
  bullet)
• A port of User Mode Linux such that it can run atop Solaris
• Port ValGrind to Solaris x86
• Port Yast2 to Solaris
• Bring SMBfs, ReiserFS, NTFS or EXT3 support to Solaris.
• Allow file event monitoring from userland via event ports (more info here).
• Write a driver for the Microsoft Fingerprint Reader
  (or similar) and plumb it into PAM, gdm2, etc. sufficiently to enable fingerprint authenticated login.

• Port Audacity to Solaris.
• Port gdesklets to Solaris.
• Create a cross-platform (Linux, Solaris, BSD, etc.) library/framework for accessing ACPI
  information from user programs, and port some interesting programs to utilize this information.

• Create a system which allows “fast user switching” a la MacOS and Windows– this would likely
  benefit all of the OS’s which use X11.

Grand Challenges

• Demonstrate getting to userland in a port to a new platform. Suggestion: ARM.
• Build a working– and in some way novel– set-top box, using Solaris as the primary OS.
• Port the Planetlab framework (and modify Solaris, probably)
  so that it works using Solaris Zones as a primitive.