Systematic approaches to software development have long been recognised as important. In conventional development environments, structured analysis methods have been adopted. The design, analysis and implementation of object-oriented systems is a comparatively new approach in software engineering, and initial attempts to apply structured analysis to these developments failed as a result of mismatches between the structured and object-based paradigms implicit in each environment. First-generation object-oriented methods improved this situation, but did not address all aspects. Weaknesses persisted at various points in the development process and software life cycle, for example project management progress metrics and code generation. The Fusion methodology has been developed by Hewlett-Packard in response to these perceived weaknesses. It utilises best practice from a number of object-oriented methods, for example Rumbaugh, Booch, Wirfs-Brock, etc. and extends some areas with original contributions. As such it is classed as a second-generation object-oriented method.
Paradigm is a CASE tool, offering an environment in which a number of methods can be used to model system developments. Fusion is one such method supported. The tool is graphically based, making extensive use of iconic controls and utilising a wide range of output capabilities. Code generation, reverse engineering, and links to other applications including spreadsheet, word processing, etc. are among the features provided.
The objective of this application experiment is to use and assess the above tools and methods when applied to application development in a software integration platform (SIP) environment. A SIP provides standard 'housekeeping' functions commonly used by applications, such as database, networking, configuration, graphical user interface (GUI), etc., so that standardised applications can be seamlessly integrated into a standards based, common environment.
Specifically, the following main tasks will be carried out during the experiment:
* Investigation of the potential benefits attributable to second-generation object-oriented methods when applied to software engineering projects which are targeted at implementation on SlP's, and to identify the impact on development team structure and approach given this development paradigm. The experiment will be based around the development of two example software modules for process applications (environmental consents monitoring and statistical process control). These will be compared to baseline developments of the same applications.
* Verification of the efficacy of software applications developed using one specific second-generation method (Hewlett-Packard's Fusion) when deployed on a 'best in class' object-oriented SIP (Prism's OpenBase), thus identifying the potential for enhanced automatic code generation within this paradigm, and any gaps and weaknesses in the process
* Where gaps are identified, proposal of additional process steps to correct the problems, and elaboration of how the approach can be applied in the generic case of other object-oriented methods and SlP's, Prior to introduction into Prism's standard software development process.