Structural testing is not enough to identify all the defects

Added on -2019-09-21

Test Report The test results clearly show that the structural generation technique used (combination of Branch and Condition/Decision technique) was not sufficient to identify all the defects. While technique provides an easy full code coverage, it does not easily allow to identify and understand the edge cases for the control structures, neither does it guarantee the absence of defects (Yang, et al 2006). However, while test cases show that this technique is not too good in finding unexpected bugs outside the requirement scope, it is arguably the
| 2 pages
| 527 words
| 175 views

Trusted by 2+ million users,
1000+ happy students everyday

Showing pages 1 to 1 of 2 pages

Test ReportThe test results clearly show that the structural generation technique used (combination of Branch and Condition/Decision technique) was not sufficient to identify all the defects. In fact, it helped finding a single ambiguous error that had to be tested further to understand its impact. The chart (Appendix A) shows the correlation between the number of bugs found for every technique used.Considering these findings I would suggest that the structural testing is only useful when the actual (final) source code is available. While technique provides an easy full code coverage, it does not easily allow to identify and understand the edge cases for the control structures, neither does it guarantee the absence of defects (Yang, et al 2006). However, while test cases show that this technique is not too good in finding unexpected bugs outside the requirement scope, it is arguably the best technique to check if a program conforms to its specification.Error Guessing technique proved to be much more effective in this task, mainly because it leveraged human intuition and experience. Goal of the tests generated using this technique was to tests program requirement/documentation flaws and logic in extreme cases. While the success of this technique in this particular case is more of an exception, or luck as Kuckis would put it (2013), I would argue that when time is not a constraint, this technique allows to generate very creative tests that might be quite useful to have in the test suite.Lastly, Boundary Value Analysis has proven to be satisfactory in testing edge input cases. It allowed to find a new defect which was not identified with any other techniques, by executingjust 3 tests. Technique is very easy to apply and its error-finding capacity is high (Müller 2005). I would say that boundary value analysis is very important in any test suite and should probably be used as the first option when generating the tests.In conclusion I am very disappointed in how unsuccessfully the structural testing was appliedin this task. However my understanding is that given the circumstances it was a pure luck that the error guessing performed much better. The Boundary Analysis effectiveness, on the other hand, was predictable, and I am completely satisfied with its result.Word Count: 376ReferencesKuckis, T. 2013. Error Guessing. Available at: www.elen.ktu.lt/studentai/lib/exe/fetch.php?media=error_guessing.ppt (Accessed: 08 December 2015)Müller, T., et al. 2005 Certified Tester. Foundation Level Syllabus (2005). Available at: http://www.bcs.org/upload/pdf/istqbsyll.pdf (Accessed: 08 December 2015)

Found this document preview useful?

You are reading a preview
Upload your documents to download
or
Become a Desklib member to get accesss

Students who viewed this