1. While dealing with software testing strategy for object-oriented architectures, the following are a few significant changes required to be normal testing strategies,
2. Here, new testing techniques favoring the error discovery procedures should be launched in order to actively test various analyses as well as the design models.
3. The initial phases of testing unit testing possess only little value in this aspect, also significant changes are needed to be integration testing.
4. While the object-oriented representation is under progress, its completeness and consistency should be checked as and when required, etc.
5. Testing object-oriented software begins by considering only the small parts of code and later extending to large software.
6. The small parts here refer to classes which inturn possess attributes and operations and are capable of implementing communications and collaborations.
7. Now, comes the matter of integration of these classes.
8. Once the integration is completed, the uncovering of errors caused due to communications and collaborations with other classes is accomplished.
9. This is achieved by performing several regression tests in these classes.
10. Now, as all classes are bounded to form one complete system, hence, this system is tested as a single entity.
11. Due to this test, if there exists any errors prevailing in requirements can be easily traced.
12. In general, it is quite impossible to say that we have done testing.
13. This is because, software testing can never be completed but this doesn't mean that the software product can never be delivered to the client, rather there are certain issues that can address the actual truth behind the complexity of termination of software testing.
14. Software testing can never be finished, rather the responsibility of testing gets shifted from the software engineer to the end-users implementing the system.
15. Also, it is often referred that software testing can be said as complete when the given software engineer is lagging in terms of time and money.
16. The above two considerations look quite low-level issues, but they bring out the true issues behind testing.
17. But experts believe that they are not yet concluded with this issue.
18. Rather research is still in progress to arrive at a meaningful conclusion.
19. A good testing strategy also assesses other quality characteristics such as portability, maintainability, and usability.
20. These should be specified in a way that is measurable so that testing results are unambiguous.
21.Tests effectiveness, test coverage, mean time to failure, the cost to find and fix defects, remaining defect density or frequency of occurrence and test work hours per regression test all these measures should be stated within the test plan.