System Requirements

SYSTEM IMPLEMENTATION

In this article we will mainly deal with the main strategies involved in system implementation. “Implementation is a project in its own right and it will benefit from controls that are applicable to any project” (Keen, 1981). Implementation tasks вЂ" testing, file conversion, documentation and training вЂ" require lots of planning and co-ordination. As an overview, testing for example represents the last opportunity for preparing the system modules before their exposure to users and the harshness of the real world. File conversion creates important technical and operational requirements which have to be planed for by the designer. Poor planning may lead to delayed or erroneous file creation. System documentation on the other hand ensures that the system does actually reflect the proposed physical system. Finally training covers the retraining of current staffs and the recruitment of new personnel.

After the approval and test of the system, a changeover from the old to the new system is eminent. This may be done by use of three possible strategies: parallel running, pilot implementation and direct changeover. Parallel running involves running the old and the new systems concurrently for an agreed period of time and the results from the two systems compared. Results have to be cross checked and the sources of errors located. This leads to system modification if problems are discovered in the new system. When the system user has complete confidence with the new system the older one is abandoned. This strategy has the characteristics of being the safest, costliest and most time consuming.

In pilot implementation strategy, instead of all the transactions being passed through the new system, as in parallel running, only a limited number are entered into the computer system. This reduces the overheads of duplicated entries and also gives practice in live processing. Since a limited set of transactions are used this strategy is less rigorous in its testing than parallel running. However, it has always been found that most of those un-entered transactions are the ones in most cases that cause the system to crash, a fact that is only realized when the existing system is abandoned and full live running commences.

In direct changeover, implementation of the new system involves complete withdrawal of the old one without any kind