Grr, I cannot say I really understand why three wiki engines  got accepted, or a UI frontend for nmap - but not important infrastructure projects like GStreamer or Avahi. Mhmm, maybe I am just envious, and considering these two projects important is just hybris...
Anyway, we had already prepared a list of exciting  GSoC project ideas for Avahi. If anyone is interested to work on one of these there might be a small chance to get this done under the GNOME umbrella. Feel free to contact either me or Trent if you are interested!
 If there is something we already have enough of in Free Software - then it is Wiki engines. just check the output of apt-cache search wiki | wc -l on a recent Debian system.
 In our definition of exciting, of course - which doesn't seem to be the same as Google's. Grrrh!