|
Abstract
|
|
|---|---|
| Software verification and validation might be improved by gaining knowledge of the variables influencing the effectiveness of testing techniques. Unless automated, testing is a human-intensive activity. It is not unreasonable to suppose then that the person applying a technique is a variable that affects effectiveness. We have conducted an empirical study on the extent to which tester and technique contribute to the effectiveness of two testing techniques (equivalence partitioning and branch testing). We seed three programs with several faults. Next we examine the theoretical effectiveness of the technique (measured as how likely a subject applying the technique strictly as prescribed is to generate a test case that exercises the fault). To measure the observed effectiveness, we then conduct an observational study where the techniques are applied by master?s students. The difference between theoretical and observed effectiveness suggests the tester contribution. To determine the nature of this contribution, we conduct a qualitative empirical study in which we ask the students to explain why they do or do not detect each seeded fault. We have found that the human component can reduce or increase testing technique effectiveness. We have also found that testers do not contribute equally to the different techniques. | |
|
International
|
Si |
|
Congress
|
Seminario impartido en Swedish Research School in Software Verification and Validation |
|
Entity
|
SWELL -- Swedish Research School in Software Verification and Validation |
|
Entity Nationality
|
Sin nacionalidad |
|
Place
|
Västerås (Sweden) |
|
Start Date
|
07/11/2011 |
|
End Date
|
07/11/2011 |