Dithering brings a number of benefits, but it involves compromises that may not be recognized. Also, there are many models for dithering and not all are consistent. Dithering may best be context dependent: e.g. for main survey, deep drilling and rolling cadence. it would be timely to open up the topic for wider and more focussed discussion.
This is a good topic. I think Chris Stubbs was also considering a session on optimal tiling of the sky with the LSST footprint, which is pretty similar. The dithering/tiling scheme impacts a lot of science as well as calibration, so it would be great to review what work has been done and start discussing next steps on improving the dither schemes.