Translation of on object role model schema into the formal language Z

In the development of information systems for business, structured approaches are widely used in practice. Structured approaches provide a prescription and guidelines for how to go about the process of developing an information system, are relatively easy to learn and provide tools which are well suited to their task. However, the products of structured approaches are sometimes seen to be vague and imprecise since requirements are written using natural language or represented in the form of models which do not have a formal foundation. This vagueness or ambiguity can be the source of problems later in development of the information system. A possible solution to this is to represent requirements using formal methods since these are seen as precise and unambiguous. However, formal methods are typically only a mathematical language for representing requirements. They are often regarded as difficult to learn and use. Even though formal methods of one sort or another have been in existence for many years they are not popular and appear unlikely to become popular in the future. One possible approach to providing the advantages of structured approaches and formal methods is to provide translation procedures from the products of structured approaches to a formal description in a suitable formal language. The work in this thesis follows this theme and is aimed at the creation of a translation procedure from an Object Role Model (ORM) schema to a Z specification. An object role model schema is the end product of a process called the Natural Language Information Analysis Method (NIAM) which is used to produce an information model for an information system. NIAM is a method which has been used successfully in industry since the mid 1970s and continues to be used today. This thesis provides a translation procedure from ORM to Z which is less arbitrary and more comprehensive than previous conversion procedures in the literature. It establishes a systematic method for (i) choosing suitable types and variables for a Z specification and (ii) predicates that express all the standard constraints available in ORM modelling. The style of representation in Z preserves ORM's concepts in a way that aids traceability and validation. The natural language basis of ORM, namely the use of elementary facts, is preserved. Furthermore, an ORM schema differentiates between abstract concepts and the means by which these concepts are represented symbolically and this thesis provides a representation in Z that maintains the distinction between conceptual objects and their symbolic representation. Identification schemes of entity types are also translated into the Z specification but it is left as an option in the translation procedure. Guiding and evaluating the work conducted here are a published set of criteria for the evaluation of a conceptual schema. These have helped in making decisions regarding the translation procedure and for assessing my work and that of others