Software Development by LL(1) Syntax Description

A method for syntax-directed program development is described in several steps resulting in a series of so-called generation schemes with increasing complexity. Each such generation scheme defines the mechanical transformation from syntax-directed descriptions (called translation syntaxes) to transducers. Many kinds of translation syntaxes and transducers are treated in detail. The final generation scheme of the series forms the heart of an industrial product called MIRA, which is a tool supporting syntax-directed program development as described in the book. What distinguishes this book from most books on compiler construction is that here the emphasis is on general syntax-directed methods rather than on specific compiler implementation techniques. In this sense, topics such as run-time data representation, run-time memory management, organization of a symbol table, code generation and optimization techiques are not treated in detail. Another important point that characterizes the book is the presence of a large number of examples and case studies illustrating how syntax-directed program development can result in programs of high quality.
In these case studies traditional techniques of software engineering are used, such as separation between specification and implementation, modular design, information hiding and data encapsulation. One of the main goals of the text is to be conceptual and practical at the same time. This is achieved by starting from mathematical objects (such as syntaxes and automata) and gradually transforming them into practical translation syntaxes and transducers. This process is an important aspect of engineering. In this respect, the book is filling the gap between theory and practice on the area of syntax-directed program development. Therefore, the book is of interest for academics as well as for practitioners. It also serves as a textbook in graduate courses on language definitions and language processors. The syntax-directed development method is based on top-down recursive descent parsing, in particular an extended form of LL(1) parsing.