This preview shows page 1. Sign up to view the full content.
Unformatted text preview: ss would lead to a disaster.
As shown in Figure 13.2(b), in this method, both the old and the new systems are
parallely operated with the same data for the initial three or four cycles. During
this overlapping period of complete operation of the two systems, the results
produced by the new system are compared with those produced by the old system
to develop confidence in the new system. Some discrepancies may be discovered.
Often these are due to inaccuracies in the old system that were not recognized
before as inaccuracies. Some discrepancies may also be due to missing program
logic for which no programming was provided or due to mistakes in the
programming itself. These must be corrected by further debugging before the
conversion is complete. This method is considered to be one of the safest ways to deploy a new system
because of the availability of old system as a backup. There is no interruption of
work if there are problems with the new system because the old system is still in
operation and the problems found in the new system can be corrected while the old
system is still being used.
This method, however, is very expensive because additional manpower is needed
during the overlapping period for the operation of two systems in parallel. Due to
the requirement of additional manpower, the organization is under considerable
strain during the period of parallel operation, and organizational breakdowns tend
to occur if the period of parallel operation is long. Hence parallel operation must
not be carried on any longer than needed to establish confidence in the new
system. Continuing them too long is a sign of weakness in the new system.
Due to the reason mentioned above, this method is not preferred in situations
where manpower resource is scarce. It is also not used in situations where the new
system differs to a great extent from the old system in the functions that it
performs and its input and output.
As shown in Figure 13.2(c), in this method, the complete changeover...
View Full Document
- Spring '14