In my lecturer's notes, the following definition for a 2-CNF wff is given:
A 2-CNF formula, or Krom formula is a CNF formula F such that every clause has at most two literals.
However, there is no mention in the notes of how one might represent clauses containing a single literal L in the wff's corresponding implication graph (although I suppose one could add an edge from every other literal to L). Looking at Krom's original paper, the definition he seems to use is:
A 2-CNF formula, or Krom formula is a CNF formula F such that every clause has exactly two literals.
This definition would seem to make a lot more sense. Which definition is correct? Am I missing something?