Cognitive Factors in Software Testing

Transcript

1. Cognitive Factors in Software Testing Rahul Gopinath
2. Influential factors Comprehension Cognitive Biases Personality Gender Differences Intrinsic Motivation Group Dynamics Performance
3. Comprehension: Relevance • Code Review - Opensource • Catches more bugs • Catches more maintainability bugs • Useful for Exploratory, Structural techniques • The one barrier that makes reviews hard is expertize. Maintenance Large Teams Outsourcing Testers Expertize Simply Better OpenSource Need More Bugs Reviews
4. Comprehension Novices • Comments • Surface Detail • Single Strategy Experts • Read code • Capture Abstract meaning • Different Strategies ◦ Typical ◦ Non Typical Mapping Strategies Sufficiency Theory Hypothesis Scan Bind Information Foraging Beacons Focal Statements Color Indentation Comments Typography Influences
5. Beacons
6. Comprehension: What next • Summarize the programs ◦ slices ◦ chunks • Summarize the changes • Save contour maps while generating • Identify focal expressions • Save multiple navigational threads
7. Cognitive Biases: Relevance Judgements Test Selection Error Propagation Heuristics Release worthy Reuse Direction of Exploration Bias Relative Importance Effort • Assimilation • Confirmation • Availability • Representative • Anchoring • Automation Trusting Tool Programmer What to test Presentation of results
8. Cognitive Biases: What can we do? Tools Design Very little empirical studies Presentation of Results Groups Diversity Education Ease of use
9. Personality: Relevance • Different learning personalities for Faculty and Students • Information intake is different for each type • Preference for Abstraction • Way of thinking is different ◦ NT, comprehension • Social interaction is different Red
10. Personality: Types Types: • Introvert Extrovert • Sensing Intuitive • Thinking Feeling • Judging Perceiving Learning: • Sensing Intuitive • Visual Verbal • Inductive Deductive • Active Reflective • Sequential Global
11. Personality:Future Need for empirical studies Personality talent pool and gender Presentation of Information • Visual Verbal • Abstract Concrete (Language|Visual)
12. Gender: Relevance Very few women in Software Engineering Bias in tools is well known Situational Stress Self Efficacy
13. Gender: Difference • Software is just another tool • Internalize failures • Higher expectations • Heuristics/Comprehensive • Selective Cues/Total assimilation ◦ High tolerance/ Low tolerance • Verbal Episodic memory • Low risk tolerance • Usefulness/Ease of Use
14. Gender: What can we do? Tools: • Videos instead of tutorials • Social Presence ◦ Avatars • Lower levels of change in environment • Textual cues
15. Inrinsic Motivation: Relevance Huge variance between testers A major differentiator according to survey. A huge factor in OSS Perhaps a differentiator in gender
16. Intrinsic Motivation: Types Challenge Curiosity Control Fantasy Cooperation Competition Recognition
17. Intrinsic Motivation: Using Avatars Gender Gamification Social Dynamics
18. Group Dynamics: Relevance Large groups of developers and testers Interaction between members Brain Storming
19. Group dynamics: Problems Social Loafing Social Preassure Production Blocking
20. Group Dynamics: What can we do? Computer mediated brain storming Easily accessible histories Sub groups that can sync separately