Renormalization-scale variations are among the most important ways theory uncertainties are evaluated in both phenomenological and experimental applications of MC generator calculations. I will give an overview of the various definitions of alphaS and associated scales / parameters that appear in MC generators, along with relations between them and global-fit extractions, arguments for and against specific schemes, the implications of matrix-element matching/merging at LO and NLO, and my own recommendations for variations. Since this is a topic on which no two persons fully agree, I will not presume to present a 'gold standard' for what is best, leaving instead the floor open for input and discussion.
