Rolling cadence is the biggest development in LSST scheduling in a decade. It is also infinitely complex and the time available for studying it is dropping rapidly. I would like to see organized community input for rolling cadence studies. This would involve inviting cadence needs, and asking for several balanced proposals in key TVS domains, e.g. SN, variable stars, and fast transients at a minimum. Proposals should spell out cadence, timing, filters, duration, sky coverage, repetition pattern, etc. This is not terribly hard and could be done pretty well in breakouts, as long as the participants bring a clear idea of the science requirement. it would be timely to provoke some new thinking on this topic as the LSST simulations group is just developing the capability to implement proposed patterns.