How to Measure the Progress of Testing?
There is no hard & fast strategy to measure the progress of testing process. The target of discovering the errors is not clearly visible.
Measurement of test progress is extremely important area, although it is ignored in quite a large number of organizations.
The testers are required to provide a convincing reply to their managers when asked about the status of testing. However generally the testers are not well prepared to answer it.
The reason of being not adequately equipped to answer this tricky question is that as a general practice in test measurement we usually count the things. Our mind is tuned to count various numbers of inputs applied by us, the percentage of code covered by us and the number of times we happen to invoke the application. We count how many times, the application was terminated by us successfully, the number of failures detected by us and so on.
Now the question comes is – How should we interpret these counts?
If we get more failures, should we take it as good news or as a bad one? The answer could be either.
A high bug count can imply that the testing had been thorough and rigorous and possibility of left over bugs is remote now, alternatively it could imply that the software application had large number of bugs, while some of them have been exposed. Thus such counts may be misleading and may not help us in measuring the progress of testing.
Next question that comes is – when to release the software?
Time of releasing the software is a tricky decision and must be thoughtfully taken based on the status of testing. Since no testing standards are available as of now, “Economic considerations”, “Target of time to market” and finally “Gut feeling of the decision makers” have become vital factors that tend to override the technical considerations while releasing the software application in the market.