1. Introduction

MSTA is syntax description translator analogous to YACC. Although MSTA can fully emulate YACC, MSTA is aimed to solve some drawbacks of YACC. Therefore MSTA has the following additional features:

• Fast LR(k) and LALR(k) grammars (with possibility resolution of conflicts). Look ahead of only necessary depth (not necessary given k). Originally LALR(k) parsers are generated by modified fast DeRemer's algorithm. Parsers generated by MSTA are up to 50% faster than ones generated by BISON and BYACC but usually have bigger size.
• Extended Backus-Naur Form (EBNF), and constructions for more convenient description of the scanners. More convenient naming attributes.
• Optimizations (extracting LALR- and regular parts of grammars and implementing parsing them by adequate methods) which permit to use MSTA for generation of effective lexical analyzers. As consequence MSTA permits to describe easily (by CFG) scanners which can not be described by regular expressions (i.e. nested comments).
• More safe error recovery and reporting (the 1st additional error recovery method besides error recovery method of YACC).
• A minimal error recovery and reporting (the 2nd additional error recovery method besides error recovery method of YACC).
• Fast generation of fast parsers.