The design phase in this project consisted mainly of the integration of the virtual hardware and documenting the implementations. The integration was successful and the documents have been produced.
The original aim was that the group would meet once a week, but this has been impossible sometimes due to the tight schedule of some of the group members. This is clearly a subject to improve.
In this phase, there have no ambiguities about the roles of the group members. Also, the schedule, and the amount of tasks per member have proven to be sufficiently well planned -- only few tasks have been out of schedule, and every member have been equally loaded.
We have reached the internal milestone we set -- the core system has been implemented and integrated.
At this point in time the project is on schedule; the integration of the units is done (meaning that the virtual hardware is operational), testing has begun and the documentation is close to the state planned for this phase.
The coding started at the early stage of the project. This has proven to be the right choice; we have both been able to implement the core system and gather a fair amount information and experience for designing the rest of the system.
| 8.12.1997 | Huima | Holländer | Kirma | Herva | total | Difference |
|---|---|---|---|---|---|---|
| Studying | 2 | 2 | 1 | 1 | 6 | -10 |
| Project management | 2 | 4 | 1 | 2 | 9 | -3 |
| Specificating | 4 | 2 | 1 | 2 | 9 | -6 |
| Designing | 4 | 2 | 2 | 4 | 12 | -4 |
| Coding | 16 | 13 | 10 | 11 | 50 | -10 |
| Testing | 6 | 7 | 7 | 6 | 26 | -6 |
| Meetings | 8 | 6 | 6 | 6 | 26 | 0 |
| Documenting | 12 | 9 | 8 | 15 | 44 | +12 |
| Host maintenance | 1 | 1 | 0 | 2 | 4 | +1 |
| TOTAL | 55 | 46 | 36 | 49 | 186 | -26 |
| Individual differences | -1 | -9 | -14 | -2 | -26 |
As a summary it can be stated that the overall scheduling in this phase was a true success; there was no particular rush towards the deadline of the phase. The time allocations were off just by a bit, and the total amount of work done was a bit less than expected.
While the designing and planning were dominant tasks in the previous phase, this phase has mainly involved implementing and documenting. The emphasis on the implementation has moved from the per module work to the integration.
So far, the core system is specified, implemented and documented. We have also written testing plan that describes the testing in the system level as well as module level. The module integration testing has also been considered.
We have considered the implementation of the core system as the most important task in this phase. Most of the module implementation took place already in the previous phase, and the emphasis in this phase has been in debugging, code refinement and interface documentation. We can say that the schedule for every module has proven to be sufficiently loose. There were, however, some difficulties with the heap module implementation and integration.
The most interesting and exciting phase forehand was the integration. So far, our coding discipline and interface specifications have proven to be good enough -- no major difficulties have arisen. The current status of the system is that we can compile a scheme program (for example a program that calculates the Fibonacci numbers), assemble it, run it as an addressed agent in the virtual machine synchronized by the scheduler. We are also able to send messages between agents.
Technical specification in our case consists of several per module documents that specify the modules and describe their implementations. Adopting the style of distributed working in this task was necessary because the design -- not to mention the implementation -- was a highly solitary task for each module. We just specified the rough interface of each module in forehand, the person responsible for the design and implementation was then supposed to complete the interface and implement the module. Although this has not made the project as much team oriented as the project could have been, the solution of ours has spread responsibility to everyone and enabled very individual working schedules.
The result is a fair amount of documentation summarized below:
The testing manager and the process manager together produced the guidelines for the module testing. They were also responsible for planning the integration testing and system testing. The module level test plans were produced -- as for the technical specification -- by the authors of the modules.
The resulting documents can be seen below:
The risk control of the project has worked quite well for this phase of the project; all the work that was planned have been done. There were few tasks that required some additional work (concerning the integration and some parts of the documentation) but this was observed in time and resources were realloced to adjust to the new situation. The schedule also had a bit of extra time that was justified when the project manager was unable to participate in the documentation for a while due to a hand injury.