The design process when using b(), from grammar specification to a working compiler or interpreter, consists of the following steps: itemization( it() First a grammar is apecified in a form recognized by b() (see chapter ref(GRAMMARFILES)). For each grammatical rule in the language, an action can be defined that is performed when a production rule has completely been recognized. The action is described using bf(C++) statements. it() Run b() on the grammar, producing the parser class and several of its member functions (among which the parsing member function `tt(parse)'). it() Design a lexical scanner to process input, passing tokens to the parser. The lexical scanner may be written by hand (see section ref(LEX)), but it can also be produced by, e.g., bf(flex)(1) (using bf(flex)(1) is not covered in this manual). it() All the parser's members (except for the member tt(parse()) and its support functions) 1must be implemented by the programmer. Commonly additional member functions are also be declared in the parser class' header. At the very least the member tt(int lex()), calling the lexecal scanner to produce the next available token, em(must) be implemented (although a standardized implementation can optionallu be generated by b()). The member tt(lex) is called by tt(parse's) support functions to obtain the next available token. The member function tt(void error(char const *msg)) may also be re-implemented by the programmer (a simple in-line implementation is provided by default). The member function tt(error) is called when tt(parse) encounters (syntactic) errors. it() The parser can now be used in a program. A simple example would be: verb( int main() { Parser parser; return parser.parse(); } ) ) Once the program's source files are available, they must be compiled and linked togrether, producing the final program.