How can we build a parsing table for a predictive parser?
Nullable, FIRST, and FOLLOW sets. Calculated as fixed points.
