OpenDtect AttributeEngine Process Flow

In making attribute computation program using OpenDtect we should follow the right flow so that our algorithm will be run nicely.

Process steps:

1. prepareForComputeData First to be called, and only once
2. getInputData
3. computeData a const (called n times; depends on our processor core number ? and if we set allowParallelComputation to yield true)
   
4. finalizeCalculation

Again. This is a quite simple problem, but it takes some times for me to figure it out (d'oh)
NOTE: getInputData(), computeData(), and finalizeCalculation() are called n times, where n = number of traces in a line. (Called inside doWork() method of Provider class)