This paper deals with actual construction of programs for experiments in simple precedence parsers based on the theoretical results of others that represents one of the most fruitful contributions in computer science. Specifically the report deals with (1) a practical general method for the generation of the precedence parser, which provides automatically generated precedence relation matrix (2) a generation of linearizing techniques (3) a syntactic error recovery method for the bottom-up analysis, based on that of Graham \& Rhodes (1975), which provides high-quality recovery with good diagonostic informations at relative low cost.