The chapter is about the running of this project, the problems encountered and how they were solved.
The project started with a discussion with the Managing Director of Source Computing Ltd. with regards to the progression of "Forecourt Plus" and whether resources would be available for development.
The resources were cleared for use and the project proposal was written and entered for review. The proposal was accepted and Ian Bray was chosen as the project tutor.
The first problem occurred with the monitoring of the data format used between a PC and the Gilbarco TMS-15. The original plan was to link a PC and a Gilbarco TMS-15 with another PC eaves dropping on the data being passed between them. This was abandoned when it became apparent that the second PC to be used did not have a functional Comm Port. This problem was solved by using a link program written by Alan Usher of Valid Link Ltd., which displayed the data being sent to and from the till.
The Requirements Specifications for both the Till_Link and the back office systems were written. These were reviewed by Ian Bray, rewritten and then submitted to Source Computing Ltd. These Requirements Specifications were finally signed off by the Colin Simm three months after the software had been written. Alan Usher had verbally agreed the Requirements Specifications when they had arrived at Source Computing Ltd., which enabled the software to fit the specifications.
The Till_Link software was based on the SerialDemo software by Sun and supplied with the Comm Port API. The link was developed incrementally, by testing every change immediately it had occurred. Each problem that arose was solved before the next change. This took longer than was originally envisioned due to the new area being worked in, and the unforeseen problems that arose.
The Till_Link software was tested using a null modem cable and a testing program also based on the SerialDemo software. This worked well and then the software had to be altered to allow another program using networking to send data through this link to the test program.
The back office was then developed and tested with the Till_Link software and the test program again using the null modem cable. This testing was completed successfully.
The Gilbarco TMS-15 was then connected to the Comm Port and the finished Till_Link and TMS-15 software activated. An attempt was made to log onto the Gilbarco TMS-15, but this failed. All further attempts also failed, but the Till_Link software does work, the problem is with the data being sent to the Till_Link and then to the Gilbarco TMS-15. This problem was not solved due to lack of time and the pressing need to write the dissertation.
A sample of both the original QuickBasic code and the Java code was given to 6 Software Engineers who answered a questionnaire. This questionnaire was used in statistical analysis of how well each paradigm produced easy to understand and maintain software. The progress of this analysis was hindered due to the disagreements between various members of the university staff with regards to the validity of these tests in using subjective data. Once an agreement was made, these tests were finished.
The project ran smoothly overall with a few problems, most of which were solved. The unsolved problem remained that way because of the limited time of the project. This problem theoretically should be easily solved as the Till_Link software works, if the correct "log on" data was sent to the Gilbarco TMS-15, it should log on.