A CASE methodology companion for batch manufacturing control software for programmable logic controllers

Abstract

An investigation of the effectiveness of a domain-specific computer aided software engineering (CASE) tool to support the software development life cycle of control software for programmable logic controllers (PLCs) for manufacturing is described. An approach to guide the development of domain-specific CASE tools is articulated. These ideas are combined with past research in automatic programming for PLC software development to produce a prototype CASE tool, called a methodology companion, that is used as the vehicle to assess the potential of the CASE technology to improve industrial practice and the effectiveness of the individual features incorporated into it. The assessment is carried out using a series of structured interviews with practitioners to obtain their informed judgment concerning the potential and limitations of the CASE technology. The results indicate that appropriate CASE technology and standards offer significant potential to improve the production of software artefacts, to facilitate coordination and communication amongst individuals, and to provide an improved organisational infrastructure. Domain-specificity contributes strongly to the acceptance and ease-of-use of CASE by the users, enables organisational changes that improve the effectiveness of individuals, and offers the potential to improve the structure of an organisation. Model-based analysis offers the potential to significantly improve the quality of the specification and the software developed from it. For design maintenance and automatic code generation to be successful throughout the life cycle, the study revealed that additional CASE tool functionality is required that supports the testing and maintenance of the software in a domain-specific manner. In the domain of manufacturing control, support is required for incremental code generation and on-line modification of operational PLC software. The investigation is guided by an approach called action research, wherein the researcher and a client organisation that exhibits characteristics of the problem of interest collaborate to set up an experiment to bring about improvements, carry it out, and then assess its consequences. The action research process proved effective in facilitating the development of a CASE tool that is likely to be adopted in practice by permitting practitioners to influence the design of the CASE tool, as well as to participate in its assessment. This collaboration in the research process brought about changes in their perceptions of how CASE technology may be used to improve the process and products of the software development life cycle. In conclusion, domain-specific CASE methodology companions have the potential to significantly improve quality in the specification and coding phases of the life cycle and can contribute to increased effectiveness of both individuals and the overall organisation. Design maintenance and automatic code generation are problematic and require additional automated aid to be accepted in practice. The action research approach is an effective method for investigating the potential and limitations of CASE technology in a field-based, industry setting.