We seem to have veered off into a discussion about arithmetic errors. The original question was: when considering a DC-rated resistor for use with AC, should we look at VAC (rms) or VAC (peak). surfsup holds that rms matters because that is the conventional way of equating the amount of "work" that can be done by AC versus DC (heating, for example). The other point of view holds that we should worry about AC peak because max voltage ratings are concerned with insulation resistance/flashover/arcing which happen instantaneously, without respect to time.
This seems to be an academic discussion, since I can't think of a practical case when we'd want to shunt AC voltage to ground, although as diagrammatiks points out, when we have AC riding on DC the peak voltage reached is DC plus AC. This is why triode/pentode switches on high-power amps have to be insulated so well.