Glossary
Software-delivery terms, defined crisply
The vocabulary Stride works with (Plan, Design, Optimize, and Verify modules) plus the cross-cutting concepts that underlie all four. Crisp definitions, deeper context where it helps, and links to the relevant product surface.
Plan · 73 terms
- Acceptance criteria
- Agile release train (ART)
- Backlog refinement
- Built-in quality
- Calibration error
- Capacity planning
- Capacity vs velocity
- Carryover
- Cone of Uncertainty
- Confidence vote
- Continuous exploration
- Cost of delay
- Daily standup
- Definition of Done
- Definition of ready
- Discovery vs delivery
- Dual-track agile
- Enabler epic
- Enabler story
- Epic
- Gold plating
- Guardrails
- Hardening sprint
- House of Lean
- Inspect & adapt workshop
- INVEST criteria
- Iteration planning
- Iteration retrospective
- Iteration review
- Jobs-to-be-done
- Kanban
- Kano model
- Lean budget
- Lean portfolio management
- Minimum marketable feature (MMF)
- Minimum viable product (MVP)
- Mob programming
- North Star metric
- OKRs
- Pair programming
- PI planning
- Planning fallacy
- Planning poker
- Portfolio kanban
- Premortem
- Program board
- Program increment (PI)
- Relative estimation
- Release on demand
- Release train engineer (RTE)
- RICE scoring
- Risk ROAMing
- SAFe (Scaled Agile Framework)
- Scope creep
- Shape Up
- Spike
- Sprint burndown
- Sprint goals
- Sprint spillover
- Sprint zero
- Story points
- Story splitting
- Swarm pattern
- System demo
- T-shirt sizing
- Three-point estimation
- Timeboxing
- User-story mapping
- Velocity
- Velocity stability
- Vertical slicing
- Weighted shortest job first (WSJF)
- WIP limit
Design · 56 terms
- Accessibility tree
- Accidental vs essential complexity
- ADR
- Anti-corruption layer
- API gateway
- Architecture fitness function
- Backend for frontend (BFF)
- Big ball of mud
- Bounded context
- Brownfield vs greenfield
- C4 model
- Circuit breaker
- Clean architecture
- Client component
- CQRS
- Defense in depth
- Dependency injection
- Domain-driven design
- Edge function
- Event sourcing
- Event storming
- Event-driven architecture
- Evolutionary architecture
- Function as a service (FaaS)
- Hexagonal architecture
- Hydration
- Idempotency
- Incremental static regeneration (ISR)
- Inversion of control
- Island architecture
- Kappa architecture
- Lambda architecture
- Layered architecture
- Micro-frontend
- Microservices
- Modular monolith
- Monorepo
- N-tier architecture
- Onion architecture
- OWASP Top 10
- Polyrepo
- Ports and adapters
- Principle of least privilege
- Progressive enhancement
- Saga pattern
- Semantic HTML
- Server component
- Server-side rendering (SSR)
- Serverless
- Service mesh
- Service-oriented architecture (SOA)
- Static site generation (SSG)
- Strangler fig pattern
- System of systems
- Ubiquitous language
- Zero trust architecture
Optimize · 57 terms
- Artifact repository
- BPMN
- Branded type
- Breaking change
- Canary analysis
- Case duration
- Case throughput
- CDN edge
- Chaos engineering
- Code churn
- Conformance checking
- Cumulative flow diagram (CFD)
- Cycle time
- Decision mining
- Dependency pinning
- Deployment pipeline
- Deprecation window
- Deviation analysis
- Discriminated union
- DORA metrics
- Dotted chart
- Error budget
- Event log
- Exception path
- Feature branch
- Five whys
- Flow efficiency
- Gitflow
- GitOps
- Happy path
- Hotfix
- Infrastructure as code
- Lead time
- Lockfile
- Petri net
- Postmortem
- Process bottleneck
- Process conformance
- Process discovery
- Rebase vs merge
- Release candidate
- Rollback strategy
- Semantic release
- Semantic versioning (semver)
- Shift-left security
- SLO
- Software bill of materials (SBOM)
- Squash merge
- Takt time
- Theory of constraints
- Throughput
- Type narrowing
- TypeScript strict mode
- Value stream
- Value stream mapping
- Variant analysis
- Webhook
Verify · 73 terms
- AI test generation
- Apdex score
- ATDD (acceptance-test-driven development)
- Backpressure
- BDD
- Bug bash
- Change failure rate
- Code coverage
- Configuration drift
- Contract testing
- Defect density
- Deployment frequency
- Disaster recovery
- Distributed tracing
- End-to-end testing
- Escaped defect
- Exploratory testing
- Fault tolerance
- Flaky test
- Four golden signals
- Four key metrics
- Game day
- Gherkin
- Graceful degradation
- High availability
- Horizontal autoscaling
- Immutable infrastructure
- Incident commander
- Ingress controller
- Integration test
- Latency percentile
- Mean time to restore
- Mock vs stub
- MTTR
- Mutation testing
- Observability
- On-call rotation
- OpenTelemetry
- Property-based testing
- Quality gate
- Rate limiting
- Real user monitoring (RUM)
- Regression suite
- Regression test
- RTO and RPO
- Runbook
- Saturation
- Severity vs priority
- Shift-right testing
- Site reliability engineering
- SLA
- SLI
- Smoke test
- Snapshot testing
- Synthetic monitoring
- TDD
- Test case vs test scenario
- Test coverage
- Test data management
- Test double
- Test environment
- Test fixture
- Test flakiness rate
- Test impact analysis
- Test isolation
- Test pyramid
- Test quarantine
- Three pillars of observability
- Toil
- Traceability matrix
- Triage
- Vertical autoscaling
- Visual regression
Cross-cutting · 75 terms
- Agent loop
- Agentic RAG
- Agentic workflow
- AI citation (cross-LLM)
- AI code review
- AI credits
- AI Overview (Google SERP)
- AI pair programming
- AI-native delivery
- Autonomous agent
- Blameless postmortem
- Blue-green deploy
- Bus factor
- Canary release
- Chain-of-thought (CoT)
- CI/CD pipeline
- Code review
- Cognitive load
- Connected delivery graph
- Context window
- Context-switching cost
- Continuous deployment
- Conway's Law
- Dark launch
- Deep work
- Delivery debt
- Depersonalization
- Distillation
- Dunbar's number
- Embedding vector
- Emotional exhaustion
- Engineering ladder
- Feature flag
- Few-shot prompting
- Fine-tuning
- Function calling
- Function spec (tool definition)
- Grounding (LLM)
- Hallucination
- Inference cost
- JSON mode
- Karasek demand-control model
- Large language model (LLM)
- LoRA adapter
- Maslach Burnout Inventory
- Model Context Protocol (MCP)
- Multi-agent system
- Orchestrator agent
- Process debt
- Prompt caching
- Prompt engineering for software teams
- Pull request
- Quantization
- Refactor
- Reinforcement learning from human feedback (RLHF)
- Retention cohort
- Retrieval-augmented generation (RAG)
- Semantic search
- Skip-level 1:1
- Software delivery operating system
- Specialist agent
- Spotify model
- Streaming response
- Structured output
- System prompt
- Team Topologies
- Technical debt
- Temperature (sampling)
- Token budget
- Tool use (LLM agent)
- Top-p sampling
- Tree-of-thought (ToT)
- Trunk-based development
- Vector database
- XML-tagged prompt
About this glossary
- What is the Stride glossary?
- A curated reference for software-delivery terms (agile, DevOps, SRE, process mining, AI agents, architecture) written with crisp definitions of 200-400 characters so AI search engines can cite them verbatim. Every term is its own URL with DefinedTerm JSON-LD schema.
- How are the terms organised?
- By category aligned to Stride's product modules: Plan (sprint planning, capacity, estimation), Design (architecture, ADRs, domain modelling), Optimize (process mining, flow), Verify (testing, observability, SRE), and Cross-cutting (AI-native delivery, prompt engineering, agentic workflows).
- How can I contribute a term?
- The glossary is editorial. We maintain it as part of our research and content workflow. If you spot an error or want to suggest a term, file a GitHub issue against stride.page; we triage suggestions quarterly.
- Why are definitions so short?
- Definitions are 200-400 characters because that's the size AI systems (Google AI Overview, ChatGPT, Perplexity, Claude) can cite as a self-contained answer. Deeper context lives in the optional body section; the definition is the citation target.
- Are these definitions consistent with industry standards?
- Where industry standards exist (ISO 19510 for BPMN, Maslach Burnout Inventory for burnout, DORA metrics for delivery performance) we cite them. Where the field disagrees or is evolving (AI-native delivery, prompt engineering), we make our editorial position explicit.