Software Testing Glossary
More than a hundred terms a working tester meets, including the vocabulary of AI detection and content authenticity, each defined in one or two sentences. Where a guide on this site explains the idea in depth, the term links to it. Definitions follow ISTQB usage where the term is defined there.
Last updated Editorial method
A
- Acceptance criteria
- The conditions a specific story or feature must satisfy to be accepted by the product owner or customer. Per-item, unlike the definition of done.Read the guide
- Acceptance test driven development (ATDD)
- A collaborative practice in which the team writes concrete acceptance examples before development and turns them into automated acceptance tests.Read the guide
- Acceptance testing
- Testing against business needs and user requirements, usually by or for the customer, to decide whether a system is fit to release.Read the guide
- Agile inspection
- A sampling-based review of a specification in which checkers count major defects on a few pages against clear rules and extrapolate the defect density.Read the guide
- Agile testing
- Testing practised inside an agile team: continuous, collaborative, example-driven, with no separate test phase.Read the guide
- All-pairs testing
- A combinatorial technique that covers every pair of parameter values at least once, cutting case counts sharply against full combination.Read the guide
- Automation pyramid
- A model recommending many fast unit tests, fewer service or API tests and few end-to-end UI tests.Read the guide
B
- Base rate
- The proportion of a population that actually belongs to the positive class, for example the share of submitted essays that are machine-written. It sets what a positive verdict is worth: at a low base rate, most flags from a detector with a modest false positive rate are wrong.Read the guide
- Black-box testing
- Testing based on the specified behaviour of a component or system without knowledge of its internal structure.Read the guide
- Boundary value analysis
- A design technique that tests at the edges of equivalence partitions, where defects concentrate.Read the guide
- Branch coverage
- The percentage of decision outcomes in the code exercised by a test suite.Read the guide
- Bug bash
- A time-boxed session in which many people, often from outside the test team, test a product at once to find defects quickly.Read the guide
C
- C2PA manifest
- The data structure the C2PA specification defines to carry provenance: a set of assertions (ingredients, actions, creator), a claim that hashes them, and a signature over the claim. Manifests nest, so an edited file carries its parent manifest as an ingredient.Read the guide
- Charter
- A one- to three-line mission for an exploratory testing session: what to explore, with what resources, to find what kind of information.Read the guide
- Code coverage
- A measure of how much of the source code a test suite executes, expressed as statement, branch, condition or path coverage.Read the guide
- Condition coverage
- The percentage of boolean sub-conditions evaluated to both true and false by a test suite.Read the guide
- Content Credentials
- The public name for provenance attached to a file under the C2PA specification: a signed manifest recording who made the content, with what tool, and which edits followed. Valid credentials show a signer made a claim and the file is unchanged since, and nothing about whether the claim is true.Read the guide
- Continuous testing
- Running automated tests at every stage of a delivery pipeline so that feedback arrives within minutes of a change.Read the guide
- Contract testing
- Verifying that a service and its consumers agree on the shape and behaviour of their interface, without running both together.Read the guide
- Cost of quality
- The total cost of preventing, appraising and failing on quality: prevention plus appraisal plus internal and external failure costs.Read the guide
D
- Decision table
- A tabular technique listing combinations of conditions and the actions each combination should produce.Read the guide
- Defect
- A flaw in a component or system that can cause it to fail to perform its required function. Also called a bug or fault.
- Defect density
- The number of defects per unit of size, typically per thousand lines of code, per function point or per requirement page.Read the guide
- Defect leakage
- The proportion of defects found after a phase or release that should have been found before it.Read the guide
- Defect removal efficiency
- Defects found before release divided by all defects found before and after release, as a percentage.Read the guide
- Definition of done (DoD)
- The team-wide checklist an increment must satisfy to be considered complete, usually layered by story, sprint and release.Read the guide
- Disclosure label
- A visible or machine-readable statement attached to content that it was generated or altered by AI, whether declared by the uploader, derived from signed provenance, or inferred by a detector. Tested as a feature: it must display on every surface and survive share, embed and re-upload.Read the guide
- Dynamic analysis
- Examining a running program for defects such as memory leaks, race conditions or security weaknesses.Read the guide
E
- Entry criteria
- The conditions that must hold before a test activity may start, such as a stable build and an available environment.Read the guide
- Equivalence partitioning
- Dividing inputs into classes expected to behave the same so that one value tests the whole class.Read the guide
- Error guessing
- Designing tests from experience of where defects typically hide.Read the guide
- Escaped defect
- A defect found in production that testing did not catch.Read the guide
- Evidence-grade record
- A test record complete enough to be relied on in a dispute or an audit: who ran the test, inputs identified by hash, the exact version of the system under test, criteria fixed before the run, raw outputs, timestamps and environment, stored write-once and reproducible by a second tester.Read the guide
- Exit criteria
- The conditions that must hold for a test activity to be considered complete, such as coverage reached and no open critical defects.Read the guide
- Exploratory testing
- Simultaneous learning, test design and execution, structured by charters and time-boxed sessions.Read the guide
F
- False positive rate
- False positives divided by all true negatives: the share of genuinely human-written documents a detector flags as machine-written. Measured on a labelled human slice with a confidence interval, and the rate whose ceiling should be fixed before a threshold is chosen.Read the guide
- Flaky test
- An automated test that passes and fails on the same code without any change, because of timing, state, environment or order.Read the guide
- Functional testing
- Testing what the system does against its specified functions.
G
- Grey-box testing
- Testing with partial knowledge of internal structure, typically using design documents or logs to shape black-box tests.Read the guide
H
- Hard binding
- A content binding that ties a C2PA manifest to the exact bytes of an asset by cryptographic hash. Any change to the covered bytes, including a re-encode, breaks the hash and the manifest fails validation, which is the signal a pipeline test looks for.Read the guide
I
- Integration testing
- Testing the interfaces and interactions between integrated components or systems.Read the guide
- ISO/IEC/IEEE 29119
- The international series of standards for software testing: concepts, processes, documentation, techniques and keyword-driven testing.Read the guide
- ISTQB
- The International Software Testing Qualifications Board, whose syllabi and glossary define widely shared testing terminology.Read the guide
L
- Load testing
- Performance testing at the expected production load to check response times and throughput.Read the guide
M
- MC/DC
- Modified condition/decision coverage: every condition is shown to independently affect the decision outcome. Required in safety-critical standards.Read the guide
- Metadata (EXIF, XMP, IPTC)
- The descriptive layers embedded in a media file: EXIF written by the camera (device, exposure, time), XMP written by editing software (edit history, rights), and IPTC fields for captions and credits. Each layer is stripped, rewritten or preserved differently by each upload channel.Read the guide
- Metric
- A measurement of a testing attribute, useful only when paired with the question it answers and the behaviour it might distort.Read the guide
- Mobile app testing
- Testing applications on phones and tablets across devices, platform versions, networks, interruptions and store rules.Read the guide
- Model-based testing (MBT)
- Generating test cases from a model of the system's behaviour, such as a state machine.Read the guide
- Mutation testing
- Introducing small faults into code to measure whether the test suite detects them.Read the guide
N
- Non-functional testing
- Testing how well the system works: performance, security, usability, reliability, maintainability, portability.Read the guide
O
- OSSTMM
- The Open Source Security Testing Methodology Manual, a methodology for measuring operational security by testing.Read the guide
- OWASP Top Ten
- A regularly updated list of the most critical web application security risks, used as a shared vocabulary.Read the guide
P
- Pairwise testing
- See all-pairs testing.Read the guide
- Penetration testing
- Simulated attack on a system by skilled testers to find exploitable weaknesses.Read the guide
- Percentile (p95, p99)
- The response time below which 95 or 99 percent of requests complete. More informative than the mean for performance targets.Read the guide
- Performance testing
- Testing to determine how a system behaves under load, including response time, throughput and resource use.Read the guide
- Product risk
- A risk that the product will fail to satisfy stakeholders, scored by likelihood and impact to steer test depth.Read the guide
- Provenance
- The recorded origin and history of a piece of content: who created it, with what device or software, and which edits followed. Provenance systems record claims and sign them; a tester checks that the record survives every pipeline stage and validates at the end.Read the guide
Q
- Quarantine
- Removing a flaky test from the gating suite for a fixed period while it is fixed, without deleting it.Read the guide
R
- Regression testing
- Re-testing after a change to confirm that previously working behaviour still works.Read the guide
- Requirements coverage
- The percentage of requirements traced to at least one executed test.Read the guide
- Risk-based testing
- Prioritizing and sizing test effort by the assessed product risk of each area.Read the guide
S
- Security testing
- Testing to find weaknesses that could allow unauthorized access, data loss or misuse.Read the guide
- Session-based test management (SBTM)
- A way of structuring exploratory testing into charters, time-boxed sessions, notes and debriefs so that it can be measured and reported.Read the guide
- Shift left
- Moving testing activities earlier in the lifecycle, into requirements, design and development.Read the guide
- Smoke test
- A shallow, fast set of tests that confirms a build is stable enough for further testing.Read the guide
- Soak test
- A performance test at normal load over a long period to find leaks and degradation.Read the guide
- Soft binding
- A binding that links a manifest to content by a perceptual fingerprint or an invisible watermark rather than an exact hash, so the manifest can be found again after metadata is stripped. It tolerates re-encoding and resizing and gives weaker evidence than a hard binding.Read the guide
- Specification by example
- Using concrete examples, agreed by the whole team, as the specification and the acceptance tests.Read the guide
- Spike test
- A performance test with sudden large increases in load to see how the system copes and recovers.Read the guide
- State transition testing
- Designing tests from the states of a system and the events that move it between them.Read the guide
- Statement coverage
- The percentage of executable statements exercised by a test suite.Read the guide
- Static analysis
- Examining code or documents without executing them, by tool or by review.Read the guide
- Stress testing
- Performance testing beyond expected load to find the breaking point and the failure mode.Read the guide
- System testing
- Testing the integrated system as a whole against its requirements.Read the guide
T
- Test automation strategy
- The decisions about what to automate, in which layer, owned by whom, in which environments, and how results are reported.Read the guide
- Test basis
- The documents and knowledge from which test conditions are derived: requirements, designs, risks, user stories.Read the guide
- Test case
- A set of preconditions, inputs, actions and expected results developed for a specific objective.
- Test condition
- An aspect of the test basis that could be verified by one or more tests.Read the guide
- Test coverage
- The degree to which a specified coverage item (requirement, risk, code element, model element) has been exercised by tests.Read the guide
- Test data
- The data needed to execute tests: inputs, existing state and expected results.Read the guide
- Test environment
- The hardware, software, network and data configuration in which tests run.Read the guide
- Test estimation
- Predicting the effort, duration and cost of testing using techniques such as work breakdown, three-point or test point analysis.Read the guide
- Test level
- A group of test activities organized together, such as unit, integration, system and acceptance.Read the guide
- Test management
- Planning, estimating, monitoring and controlling test activities and reporting their results.Read the guide
- Test oracle
- A source of the expected result for a test: a specification, a comparable system, a formula or a person.
- Test plan
- A document describing the scope, approach, resources and schedule of intended test activities.Read the guide
- Test point analysis
- An estimation technique that sizes testing from function points adjusted for quality requirements and productivity.Read the guide
- Test process
- The set of activities from planning through completion by which testing is carried out.Read the guide
- Test process improvement
- Systematic improvement of how an organisation tests, often assessed against a maturity model such as TMMi.Read the guide
- Test strategy
- A high-level description of the test levels, types and approach for an organisation or programme, from which plans are derived.Read the guide
- Test type
- A group of test activities aimed at a particular quality characteristic: functional, performance, security, usability, regression.Read the guide
- Three amigos
- A short conversation between a business, a development and a testing perspective to agree examples before work starts.Read the guide
- Three-point estimation
- An estimate combining optimistic, most likely and pessimistic values, commonly weighted (O + 4M + P) / 6.Read the guide
- TMMi
- The Test Maturity Model integration: five levels of test process maturity from Initial to Optimization.Read the guide
- Traceability
- The ability to relate tests to the requirements and risks they cover, and defects to the tests that found them.Read the guide
- Trust list
- A published list of the certificate authorities or signers whose certificates a C2PA validator accepts. A manifest whose signature verifies but whose certificate chains to no trusted list validates as signed by an unknown party, which is one of the four verification outcomes.Read the guide
U
- Unit testing
- Testing individual components in isolation, usually by the developer who wrote them.Read the guide
- Usability testing
- Testing how easily users can learn and use a system to achieve their goals.Read the guide
V
- Volume testing
- Testing with large amounts of data to check capacity and behaviour at scale.Read the guide
W
- Watermark (statistical text)
- A hidden pattern a cooperating text generator embeds by nudging word choices towards a secret subset of the vocabulary. A detector holding the key counts how often the subset appears and runs a significance test. Evidence weakens with editing and short length, and absence proves nothing.Read the guide
- White-box testing
- Testing based on knowledge of the internal structure of the code.Read the guide
- Workload model
- A description of the mix, rate and pattern of transactions a performance test should reproduce, derived from production data.Read the guide
No term matches. Clear the search or choose All letters.
