I attempted to get my old director of studies to write a reference for me for my PhD application. It turns out that he has all but retired. He is the second of my DoSs to leave. Perhaps that tells you something. ANyway, the application process is slow to say the least. Just gathering all the little bits of paper together and posting them off. Hey, that gives me an idea...
I've also been having lots of ideas related to my PhD. Basically I would like to kick off the project by developing a series of networked mobile robots and static workstations and some API for cross-agent communication (and perhaps code migration). Then I hope to move towards an evaluation of various control methods across this network (hopefully with more than a passing look at fault tollerance) and develop what I think will turn out to be a good layered network for fault tollerant network operation. The only really tough bit that I can see is chosing the co-operative task that will be used as the benchmark. Hmm. Please send your ideas on a postcard.