Requirement Error Abstraction and Classification: A Control Group Replicated Study

More testers – The effect of crowd size and time restriction in software testing

“…Second, we believe that having more "fresh eyes" helped the TR testers to perform better. With "fresh eyes," we are refereeing to phenomena when individuals become blind to defects in the areas they have been working in for a long time; e.g., to counter habituation in verification and validation is to make individuals think of the reasons behind each defect and then have them use this information to find new defects, as shown in [65]. It was also shown in [65] that simply telling the individuals that more defects exist and giving them more time for searching results in a very small increase in the number of defects found.…”

“…With "fresh eyes," we are refereeing to phenomena when individuals become blind to defects in the areas they have been working in for a long time; e.g., to counter habituation in verification and validation is to make individuals think of the reasons behind each defect and then have them use this information to find new defects, as shown in [65]. It was also shown in [65] that simply telling the individuals that more defects exist and giving them more time for searching results in a very small increase in the number of defects found. In other words, when a certain individual cannot find more defects, one can either bring in a second individual or give the first individual a mental tool that would alter their defect searching and detection process.…”

Effectiveness of Human Error Taxonomy during Requirements Inspection: An Empirical Investigation

“…Next, a series of controlled experiments [3][4][5] validated that, an error-inspection (guided with RET) significantly improves the inspectors' effectiveness when compared to fault-based inspections. It was also reported that, educating developers on errors (as opposed to faults) made them less likely to make mistakes during the requirements development [6].…”

“…The results from an independent samples t-test showed that although subjects using HET were consistently more efficient for all three teams, the improvement was not significant. This was the very first investigation of HET whereas RET has been validated and improved through a series of comprehensive empirical analysis [3][4][5]. That said, subjects using the HET not only found more faults, they found it at a much faster rate.…”

“…Next, we evaluated subjects' self-reported data (collected during post-study survey) regarding the usefulness of HET and RET on five essential attributes: (1) usability-usab., (2) orthogonality (Orth.-a lack of overlap amongst error classes, (3) usefulness of error taxonomy in locating faults -Usef., (4) confidence that taxonomy represented real RE problems -Conf., and (5) worthiness of the effort spent in using the taxonomyWorth.. A 5-point Likert scale ranging from 1-Strongly Disagree to 5-Strongly Agree was used to evaluate these attributes of both taxonomies. This comparison was performed for 23 subjects who used (and rated) HET vs. 23 subjects who used RET.…”