Objective The experiment motivation is to reduce the costs of producing Real Time Engine Control systems and to therefore be more competitive in the market place. To achieve this the procedures for systematic reuse of previously developed and tested software must be put in place. Accurate measurement of the reuse process and it's effect are vitally important for effective process improvement and demonstration of cost reduction. The experiment should provide useful data in three main areas of software development. Firstly the use of the Goal Question Metric approach as a means of process measurement and improvement for a reuse process and a process that includes reuse. Secondly in showing how the resultant data is effective in demonstrating a cost reduction due to reuse. Lastly in the use of Goal Question Metrics as a driver for process improvements.THE EXPERIMENTThe experiment will use the Goal Question Metric approach to assess and improve the existing reuse measurement process. Using the revised reuse process the software development process incorporating reuse will be measured and the results assessed to infer process improvements. Current expectations of reuse on the baseline project are 50%. The aim of this experiment is to attempt to increase this reuse to around 80% or at least to provide process improvement recommendations to do so.The target project selected for PRIME is the development of an engine controller for a BR715 engine on the McDonnell Douglas MD95. This project poses greater reuse challenges than previous applications in that it features engine and airframe, as well as engine control hardware and software changes. The project will begin in May 1996 with the commencement of the first software development. This will conclude in October the same year. This phase will be followed by further software packages. RoSEC software represents a team of around 50 software engineers with a wealth of experience in the development of embedded real time systems to flight critical standards. The earlier application from which the baseline project is to be built is a software system of around 210,000 lines of source code. This figure includes 100,000 lines of ADA or assembler source code, 46,000 lines of design annotations and 64,000 lines of comments. The size of the MD95 application will be a similar order of magnitude. EXPECTED IMPACT AND EXPERIENCEThe experiment should achieve an improved reuse measurement process, an improved software development process and a good understanding of Goal Question Metrics as a tool to drive process improvement. This will result in reduced costs on future projects, a method for demonstrating the cost savings and also further benefit approved from any process improvements in the future.Software metrication is a difficult business. The very nature of it tends to introduce extra expense and a perceived "Non value added" activity. This leads to inaccurate measurements because the engineers providing the data haven't bought into the justifications. There are two important factors in making this experiment successful, firstly to demonstrate that there is value in the process, secondly in spending time ensuring that the process for collecting data is unobtrusive (Ideally invisible) and thus encouraging consistent and correct data.The results will be of interest to European software developers whose level of re-use is significant, and more generally, organisations looking to identify process improvements. Understanding of industrial reuse will be improved through a practical investigation of the mechanisms necessary both for evaluating the benefits of reuse and improving a process incorporating reuse. The work will aid European software developers to create a managed and systematic reuse programme. Fields of science engineering and technologyelectrical engineering, electronic engineering, information engineeringelectronic engineeringcontrol systemsnatural sciencescomputer and information sciencessoftwaresoftware applicationssystem softwarenatural sciencescomputer and information sciencessoftwaresoftware development Programme(s) FP4-ESSI 2 - Software Best Practice initiative within the framework of the specific research and technological development programme in the field of Information Technologies, 1994-1998 Topic(s) 1.28 - Process Improvement Experiments Call for proposal Data not available Funding Scheme ACM - Preparatory, accompanying and support measures Coordinator Rolls Smith Engine Controls EU contribution No data Address Finchley Road 765 NW11 8DS London United Kingdom See on map Total cost No data
