Saturday 17 August 2013

How to Find Possible Number of Test Cases

It absolutely depends on your approach you take(test case generation strategy), but you should use Pair Wise (Orthogonal Array) to reduce the number of TCs. Personally I don’t think there is a method to know the exact number of test cases before writing that. It depends on the approach which you use. Some folks write functional level and some write unit level test cases. According to me test coverage matters rather than no. of test cases.
Also, traceability matrix should be maintained and checked regularly, in order to ensure your test coverage.
The estimation can be done using the following data/method
Historical data
===============
If you have a similar application, then you can get the historical data, and find out how much your current project varies from that. Accordingly, give a rough guess.
Function points
===============
Estimate the number of function points that could be generated for your requirements.
This is done by the development team also. So, you may borrow the data from them.
Types of testing
================
Functional, security, Performance/stress, Interoperability etc…
Testing techniques
==================
BVA, EP, Decision Tables, State Transition etc.
The methods could be something like below:
1. Identify the types of testing to be done.
2. Identify which testing type is applied on whICH function point(FP).
3. Determine which testing technique is applied to test which FP (while doing this, remember what kind of testing you wanted to do on the FP)
4. For each testing technique, give a rough estimate on the number of test cases that can be generated, example:
FP1 – BVA – 8 test cases
FP2 – DT – 15 test cases
5. Compare this final data with the Historical data.

No comments:

Post a Comment