As the system was developed, my partner and I held frequent meetings to determine the exact specifications of each section of Smart Ticket, in order that our two sections of code would merge together into a working whole with as little trouble as possible. It was through these meetings that the abstract class detailed in Appendix B was developed, and allowed us both to create independent code, that would fit together when finished. Specifications were always agreed before any code was written.
There were a total of five code merges, although the last contained no new code or fixes, and was merely cosmetic. Even from the first merge, there were no errors in compilation, and thanks to the firmly agreed specification, interaction between the two halves was as expected. Each merge resulted in bugs being reported to the author of the offending code, and was followed by agreements on any new specifications that appeared necessary. This model of working was found to be very efficient and effective.
The emphasis on the London Underground in the project has mainly been lost, with the entire system being designed with generality in mind. The route-engine has been tested with a full implementation of the network and works with it. It was decided that nothing was to be gained by testing the complete system on it, as it would take a great deal of time, and prove little. Instead, a small test network has been used.
The final version of Smart Ticket performs largely as specified. Through a graphical interface it is possible to describe a network and weight it, as well as request routes to be found across it. The returned route will always be the one of lowest weight. The system appears to run on a variety of platforms functionally, albeit with some cosmetic differences. The only area really falling short of the specification is the route-engine's inability to process routes including a request to journey via a particular station. This, however, could be fixed merely by alteration to the engine, as the rest of the system is designed to be able to do this.
Smart Ticket has been a successful project, and I am proud of the programs. It is a prototype of a system that it demonstrates to be suitable for a variety of network traversal applications, at a range of scales.
Smart Ticket can be found at http://smack2.sthughs.ox.ac.uk/smart/ or from http://www.simonmack.com/, although sections may be password-protected to prevent damage to the system.