Testing vs. Development

Lecture 37 examines the psychological and organizational dichotomy separating software development from software testing. Developers operate under a constructive mindset—striving to prove that code works and delivers required features. In contrast, testing requires a destructive, critical mindset. The lecture cites Glenford Myers' classical axiom: 'Testing is the process of executing a program with the intent of finding errors.' A successful test is one that uncovers a previously undiscovered bug, not one that passes trivially. Students analyze why programmers should not be the sole testers of their own code due to blind spots and subconscious bias. The economics of testing—balancing exhaustive testing impossibility against diminishing defect discovery returns—are evaluated. The curriculum systematically aligns theoretical software engineering models with professional industry practices and modern Agile engineering workflows. Reviewing the accompanying lecture handouts and conceptual diagrams reinforces understanding of object-oriented analysis, UML modeling, and modular design. Consistent evaluation of requirements specifications, architectural trade-offs, and quality assurance metrics ensures robust preparation for academic assessments. Students investigate the role of software requirements specifications in aligning stakeholder expectations and preventing expensive late-lifecycle defects. The discussion highlights how functional and non-functional requirements dictate architectural style selection, performance thresholds, and security controls. Key concepts such as source and sink analysis, context boundaries, and external data interfaces are rigorously analyzed in system modeling scenarios. Use case modeling techniques demonstrate how primary actors, preconditions, postconditions, and alternative execution flows document user interactions. State transition diagrams illustrate how event-driven systems transition between operational states in response to external triggers and guard conditions. The lecture contrasts traditional sequential lifecycle models like the Waterfall model against iterative, prototyping, and risk-driven evolutionary approaches. Software design principles emphasize high cohesion and low coupling to minimize ripple effects and enhance long-term system maintainability. Coupling classifications, spanning content, common, control, stamp, and data coupling, are critically evaluated to eliminate structural dependencies. Cohesion levels, from coincidental and logical to communicational and functional cohesion, provide objective criteria for evaluating module quality.