Session 2.2 – Principles of Verification

Module 2: Verification & Validation | Duration: 1 hour

Learning Objectives

Verification = “Are we building the product right?” — a discipline focused on reviewing artifacts before and during implementation to prevent defects from reaching validation.

  • Define verification and articulate why it complements validation.
  • Walk through the verification workflow phases with concrete deliverables.
  • Apply (and measure) structured techniques such as inspections, static analysis, and contract testing.

Principles

Verification is a proactive quality activity. It does not ask “does the user like it?” (validation) but “is this artifact technically sound and aligned with its specification?”

Prevention over detection
Find issues in the cheapest artifact (requirements & design) before coding begins.
Independence
Fresh eyes (peer reviewers, QA, architecture guilds) reduce confirmation bias.
Evidence trail
Reviews recorded in tools (Jira, Confluence, Git) with action items and waivers logged.
Repeatable criteria
Entry/exit checklists ensure quality gates are objective.

Verification Goals, Checkpoints & Artifacts

Goals
  • Prove compliance with specs, policies, and regulations.
  • Surface ambiguity/risks early so they don’t escape to validation.
  • Create traceable evidence for go/no-go decisions.
Checkpoints (Entry/Exit)
  • Entry: artifact shared, roles assigned, checklist agreed.
  • Exit: all action items closed or waived, metrics captured, stakeholders sign-off.
  • Escalation path defined for unresolved high-severity items.
Artifacts Produced
  • Review minutes & defect log (Jira, Azure DevOps).
  • Updated traceability matrix linking requirement → verification evidence.
  • Automation reports (static analysis, contract tests) stored with release notes.

Verification Workflow

Use the following loop for any artifact (requirements, design, code). Each phase should answer a specific question and produce tangible evidence.

Plan & Prep Inspect Record & Share Close

Always capture who attended, what was found, and when actions will be closed.

Plan & Prepare
Define scope, roles, checklist, artifacts, and success criteria. Share materials ahead of time.
Inspect / Review
Walkthrough, inspection, or tool run. Focus on objective findings, not ownership blame.
Record & Share
Log defects in Jira, note waivers, capture metrics, update traceability matrix.
Close Actions
Track fixes, rerun tools/checklists, obtain sign-off. Communicate learnings for the next iteration.

Techniques

Walkthrough (early drafts)
  1. Author summarizes scope and risks.
  2. Reviewers ask “what if” questions and note ambiguities.
  3. Outcomes become backlog items or acceptance criteria refinements.
Formal Inspection (mature artifact)
  1. Moderator schedules session, assigns roles (reader, recorder, experts).
  2. Each participant reviews offline against checklist.
  3. Meeting logs defects, severity, owner, ETA. Metrics captured (defects/page, prep time).
Static Analysis & Linters
  1. Integrate tools (SonarQube, ESLint, Bandit) into CI.
  2. Establish baseline severity threshold (e.g., fail build on new high severity issues).
  3. Review trend dashboard weekly—treat regression as action items.
Contract/API Verification
  1. Document schemas with OpenAPI/GraphQL SDL.
  2. Automate schema validation and consumer-driven contract tests.
  3. Use versioning strategy (semver) and backward compatibility checks.

Checklists & Artifacts

Requirement review checklist
  • Clear actors, inputs, outputs
  • No ambiguous terms (“fast”, “user-friendly”)
  • Acceptance criteria traceable
Design review checklist
  • Error handling defined
  • Logging/observability strategy
  • Non-functional requirements addressed
Code review checklist
  • Tests included/updated
  • Security best practices (input validation, secrets)
  • Performance considerations (N+1, caching)
Artifacts to file
  • Review minutes and action items
  • Updated traceability matrix
  • Waiver log with owner & expiry

Metrics & Reporting

Defect Removal Efficiency (DRE)
DRE = Defects found in verification / (Defects found in verification + escaped defects)
Example: 24 defects found during inspections, 6 escape to system testing → DRE = 24/(24+6)=80%.
Defects per Review Hour
Example: 4 engineers spend 45 minutes inspecting a design (3 hours total) and find 9 defects → 3 defects/hour. Use this to calibrate whether reviews are too shallow or too long.
Action Closure Time
Track mean/median time from discovery to fix confirmation. Flag items older than SLA (e.g., 3 days for critical design issues).
Static Analysis Trend
Plot “new issues vs resolved issues” each sprint. Scenario: Sprint 5 introduced 12 new critical warnings, only 5 resolved → add remediation task before Sprint 6 coding.

Report metrics alongside validation KPIs so leadership sees the full quality story.

Summary & Assignment

Verification is most powerful when it is systematic, evidence-driven, and collaborative. By pairing process checklists with automated gates, teams reduce rework and build confidence before validation even begins.

Assignment: Select an upcoming feature and create a verification plan that lists review participants, checklists, automation tools, and the metrics (DRE, closure time) you will report to stakeholders.
Course Home