7. Comparison and choices 1

The choice of which implementation strategy to use will depend on the nature of the system. Each strategy has merit and all of them have disadvantages. These are summarised over the next couple of pages:

Parallel approach

Advantages of a parallel approach

• Minimises risk as the old system can continue to be used should a problem arise in the new one.
• The new system can be optimised whilst the old system is still available
• New data created by the old system can be used to check the consistency of the same data in the new system

Disadvantages of a parallel approach

• Duplication of effort and system hardware adds cost to this approach
• There may be a need to keep synchronised copies of database records. One set is used by the old system and the other by the new system

Phased approach

Advantages of a phased approach

• Minimises risk as the old system is replaced one bit at a time.
• Only having to find bugs and errors in one part of the system makes it easier fix compared to having to solving problems in the complete system. This is the approach often taken in computer programming where the code is made up of separate functions and modules.

Disadvantages of a phased approach

• The system must be able to be partitioned into sub systems for this approach to work
• It does assume that a problem in one module has no effect on another module. This is called 'loose coupling' in computer programming and is a good thing. A tightly coupled system means that modules strongly affect one another.
• Can take a long time to roll out the whole system, especially if there are a large number of complicated sub systems
• Expensive in terms of the time taken up by staff to evaluate each module and loss of productivity