DESIGN AND IMPLEMENTATION OF A LANGUAGE SCANNER GENERATOR (KnitAutoGen)

Abstract

The need for fast, efficient and simple scanner generator that has the primary responsibility to perform efficiently gave rise to this paper. This is due to the fact that, on daily basis new technologies arise which brings a great improvement on the design of computer architecture. However, attention was given to speed, run time and resource availability of the design machine to be used since lexical analysis has an impact on how the compiler works. This paper seeks to develop a lexical analyzer (scanner generator) automatically by specifying the lexemes patterns to a lexical analyzer generator and compiling those patterns into a code that functions as a lexical analyzer. The scanner accepts characters as input and breaks them down to produce tokens by grouping the characters and not deviating from specifications. The project employs one of the different methods of lexical analyzer generator to perform pattern-matching on text using regular expression over a global character set. The paper shows how input is matched and specifies what to do when a pattern is matched. This was achieved with the use of regular expressions (RE) which were converted to non-deterministic finite automata (NFA) or deterministic finite automata (DFA). The regular expression and the regular grammar were thus joined together mathematically. Various results are presented and further work on micro compilers was proposed. Keyword: Scanner generator, Regular expression, KnitAutoGen, DFA, NFA. DOI : 10.7176/CTI/8-02