So I just looked at my calendar and went "gulp". PyCon.ca is just 3 (or 4, depending how you want to count) days away. I've got my presentation roughed out, but I don't have Coldshot to the point where it can be released as a 1.0 implementation.
I also still need to play with statprof, gprof2dot, Graphi, PyPy profiling, Py3K profiling, multiprocessing profiling, and then go back and review the talk from 2009 to make sure I haven't forgotten to address any of the questions that came up from there. The big problem is that I sit down to work on the presentation and find myself working on RunSnakeRun or Coldshot instead.
I've now got RunSnakeRun doing file-oriented views for Coldshot (as it always did for cProfile), but as of yet it doesn't do anything more interesting than it does with Coldshot, as every time I get into "interesting" stuff I get into gimbal lock with multiple different things to optimize for arbitrarily huge data-sets... premature optimization happens even when you're writing a profiler and should know better :) .