On the 'correct prefix property' in precedence parsers (Q792099)

Precedence parsers are very fast and simple in operation. Unfortunately, the common opinion is that the precedence technique has many disadvantages. Among other ones, the absence of the correct prefix property (CPP) has been stressed. This paper attempts to refute this opinion. A practical transformation is presented that accepts a mixed strategy precedence (MSP) grammar as an input. The transformation results in two new grammars, both of them satisfying the MSP property. The precedence parser for the first grammar equipped with the extended error detection capability due to \textit{S. P. Rhodes} [Ph. D. Dissertation, Univ. of California at Berleley (1973)] is guaranteed to have the CPP. In the case of the latter grammar, the classical MSP parser obeys the CPP. The transformation algorithm can be extended to accept LR(0) or SLR(1) grammars as inputs. (From the author's introduction.)