On the relationship between LL(1) and LR(1) grammars

Abstract

It is shown that every p-reduced LL(I) grammar is LALR(I) and, as a corollary, that every A-free LL(I) grammar is SLR(I) A partial converse to this result is also demonstrated: If there is at most one marked rule m the basis of every state set in the canonical collection of sets of LR(k) items for a grammar G in which S i~ S-/ts impossible, then G ~s LL(k).