search
Enter at least two characters without personal information.

    LibraryArticles and FAQs Your projectRequest a quote
    LanguageEN ▾
    Methods and learning

    Quality assurance and acceptance testing: responsibilities and practical exercise

    Quality assurance turns project expectations into reproducible checks and follows discrepancies through resolution. It specifies context, data and expected outcomes before interpreting a defect. Coverage follows journey risks; a test count alone does not describe it.

    These sheets describe activities and teaching exercises. They do not constitute job offers, certifying courses or recruitment promises.

    All job descriptions
    Illustration of design and review work on a computer
    AI-generated illustration

    Explore all 28 review cases — A defect cannot be reproduced

    Missions and responsibilities

    Make defects observable and verify project requirements through scenarios, regression checks and evidence supporting acceptance decisions.

    Typical activities

    • Turn requirements into testable acceptance criteria.
    • Prepare test data, devices, states and expected outcomes.
    • Describe defects with reproduction steps and impact.

    Responsibilities to agree with the team

    Testing cannot guarantee absence of defects. Project owners decide release readiness using risks, blockers and results, with specialists for specialised checks.

    A first piece of work to demonstrate

    Deliver an acceptance matrix and a reproducible defect with expected outcome, observed result and severity.

    This is an editorial description of common activities. Job titles and scope vary across teams; it is not a certification framework.

    W3C: organising accessibility responsibilities across a team

    Practice exercise

    Establish functional validation of a form successfully, errors, duplicates and recovery.

    Work on a copy or demo environment. Define the data used and the expected result before you begin. Keep your decisions and the difficulties encountered in order to be able to explain the process.

    1. Transform needs into expected steps and outcomes.
    2. Choose test data that covers useful limit cases.
    3. Document each anomaly with context and reproduction.
    4. Retest the correction and routes that could be affected.

    Deliverable and checks

    Provide form cases with prerequisites, data, steps and expected outcomes. For each defect, record the version and actual result without real personal data. Link retesting to the fix, and distinguish cases not run from those blocked.

    Another person should be able to reproduce the defect and run the retest. Check that success, errors, duplicates and recovery have distinct expectations. A checked box without an observation, or a blocked case counted as passed, is not validation.

    Prepare website checks

    Risks and collaboration

    A long list of controls does not replace an accurate expected result. A closed fault without retest can remain present in the actual course.

    Collaboration to explore: Project management.

    Fictional deliverable excerpt : A defect record gives environment, version, entered query, returned address, expected result and observed result. Repeat the same task after correction without declaring the whole website approved.

    What passes between the roles : The reproduction record accompanies the intervention and corrected version. Separate the removal decision, operational action and post-change check, with an identifiable owner for each step.

    View the role comparison : QA / Web tester and webmaster →

    Employment, trades and training
    An example to examine

    Write a defect someone can reproduce

    Fictional teaching material: this excerpt illustrates part of a deliverable. It is not a client reference, a complete specification or evidence of professional qualification.

    Sample excerpt
    Context: three-resource list, empty filter.
    Action: enter a word absent from every title.
    Expected: no items and an explicit message.
    Observed in the defect example: empty list without explanation.
    Recovery check: clearing the filter restores all three items.
    
    Variation to replay: enter “mobile”, clear, then enter an absent term.
    Expected evidence: successive states, displayed count and return to all three items.

    What the author should be able to explain

    1. The starting state makes the check reproducible.
    2. The report separates expected and observed outcomes.
    3. Verification after correction includes recovery to a usable state.
    Is a screenshot of an empty list enough?

    It may illustrate the fault, but also provide the starting state, input, steps and test context. A screenshot alone does not show whether content is absent, filtered or still loading. Keep the original defect separate from the outcome after correction, including the tested version. Replay the steps by keyboard and verify that the result message follows the actual list. A correct message with a list still filtered remains a defect.

    Another case to examine

    Set the rule before classifying a defect

    Fictional case: someone reloads the page and finds an empty filter. That observation alone does not establish incorrect behaviour.

    Find the requirement: URL state, local persistence or an intended reset. Record version, test URL, starting state and action without copying private data.

    Compare the result with that rule. If none exists, agree the expected behaviour before marking it “fixed” or “compliant”. A tester’s preference is not an agreed acceptance criterion.

    Technical reference : W3C WAI · User Notification · . The excerpt is an original teaching example.

    Try a small filtered list

    These three fictional entries show a result, an empty state and a return to the full list. Try “mobile”, then an absent word. No network request is made.

    3 of 3 examples shown.

    • Images and alternativesPrepare an image and its description.
    • Mobile checksCheck a journey on a phone.
    • Writing a service pageClarify an offer and its scope.

    For review: use the keyboard, test an accented word and check the count after clearing. This local list demonstrates neither server loading nor network failure. Input is not saved or sent.

    Assess the quality of the work

    For the exercise, a form works with ordinary input but loses context after an error. Acceptance testing should connect the defect with the blocked task and make it reproducible. Capture or test counts are insufficient: examine risk coverage and what observations actually support.

    Scroll the table horizontally to read every column. With a keyboard, focus the table area and use the arrow keys.

    Assess the quality of the work
    SkillEvidence to examineFailure to spot
    Risk-based coverageChoose the main journey, a meaningful error and recovery after interruption. Connect each case with its risk and expected result.Many checks repeat the same successful submission while none examines an input error or return after interruption.
    Reproducible contextRetain examined version, browser, relevant dimensions, initial state and fictional data. Describe minimal steps preceding the observed result.A screenshot cannot establish how to reproduce the defect, or steps depend on an unspecified account, data item or preference.
    ConclusionSeverity and impact are distinguished from visual preferences.Automated success is presented as complete proof of accessibility.
    Reproducing a defectRecord the environment, test data, steps, expected result and observed result.Another teammate reproduces the defect; the fix is checked with the original case and a related case.

    A focused next step

    Have someone absent from the initial test reproduce a defect. After correction, replay the original case and a neighbouring one sharing the risk. Explain what remains outside scope: successful preliminary checks alone do not establish complete conformance with an accessibility standard.

    W3C WAI: preliminary accessibility checks. · Source consulted

    Documented reference points for this topic

    Open the same version with two editors, save the first and replay the second with the old validator. Also compare the stored data.

    Documented reference points for this topic
    ReferenceValue or syntaxScope
    WCAG 2.2 · 1.4.4
    200 %Enlarged text retains content and functionality; exceptions cover captions and images of text.
    WCAG 2.2 · 2.5.8
    24 × 24 CSS pxA smaller target can satisfy the criterion through spacing; other exceptions apply.
    WCAG 2.2 · 1.4.10
    320 CSS pxVertical content without two-dimensional scrolling; exceptions cover inherently two-dimensional content.
    W3C Internationalization · HTML direction
    lang="ar" dir="rtl"Language and reading direction are separate information. Declaring ar alone does not establish the base direction; dir="rtl" does. Logical CSS properties adapt spacing without hard-coding left and right throughout the layout.
    WCAG 2.2 · 3.3.7
    A · same processWithin one process, previously supplied information must be auto-populated or available to select, subject to the criterion’s exceptions. Browser autocomplete alone is insufficient; the site must offer the information already received or avoid requesting it again.
    WCAG 2.2 · 3.3.8
    AA · authenticationAn authentication step requiring recall or transcription needs assistance or an alternative route, subject to the criterion’s exceptions. Test password managers and code pasting; an audio alternative requiring transcription does not by itself remove the cognitive burden.
    HTTP · If-Match
    If-Match: "r7" → 412If-Match compares a strong validator before a write. Under the illustrated contract, a stale version receives 412; keep the draft to compare changes.

    Dates indicate when references were consulted. These facts are neither a site audit nor an overall compliance statement.

    A review case to reproduce

    A defect cannot be reproduced

    Situation to examine
    An issue report contains only a screenshot, without version, input or steps. The expected behaviour remains unclear.
    Proposed check
    Reconstruct initial state, environment, actions, expected outcome and observation. Minimise the triggering case. After correction, check that case and a neighbouring one for regressions.
    Evidence to retain
    A short report allows another person to reproduce the case. Confirmed defects, hypotheses and temporary inability to reproduce remain distinct.
    Fictional finding example
    Version DEMO-2, width 360 pixels: submitting an empty address produces no message. The expected outcome is an error associated with that field.
    Interpret this outcome
    Include starting state, input, steps and observed outcome. After correction, repeat this case and a neighbouring valid submission.

    Three complementary checks

    Fictional cases: adapt these expectations to your test contract and retain the limits of the check.

    1. Described initial state
    2. Reproduced defect
    3. Retested correction

    Scroll the table horizontally if needed; with a keyboard, focus it and use the arrow keys.

    Scenario variations and expected evidence
    Variation to testExpected result in this caseEvidence to hand over
    Restart from test dataFind the defect using the same stepsPreconditions, actions and observed result
    Change screen widthRecord where behaviour differsWidth, browser and affected element
    Retest after correctionCheck the original case and a neighbouring caseCorrected version and both test results

    Decision to hand over : Close the defect within the tested scope, retaining conditions that still need examination.

    Method reference : W3C WAI — Planning and Managing

    Learning scenario

    Extended exercise and expected deliverables

    In-depth exercise: Prepare the functional validation of a form and its confirmation with a smartphone version and a computer version. Separate nominal cases, errors and interrupts. For each defect, keep the context, the exact steps and the observed result. Use only test data. A defect does not become corrected because the code has changed: replay the scenario on the corrected version and note the effects on the nearest routes.

    What to deliver with your work

    • Functional validation scope and covered environments
    • Case with Action, Wait and Test Data
    • Reproducible anomaly reports
    • Tracking corrections and recovery results

    Acceptance criteria

    These criteria are a working basis to adapt to the context. For each, note the action taken, the expected result, the observed result and the information remaining to be confirmed.

    1. Each case has an expected result before its execution.
    2. A defect contains steps to reproduce it.
    3. Resumptions identify the controlled version and context.
    4. The coverage limits are indicated in the functional validation report.

    Understand pre-launch checks

    Ongoing management and handover

    Transmit cases, evidence and open anomalies with their explained priority. Distinguish an impossibility to verify from a compliant result. A team should know what has been replayed after correction and what remains dependent on an external environment. The balance sheet should not promise a universal absence of defects: it precisely describes the checks carried out and the risks still observed.

    Organize site maintenance

    Compare two approaches

    Verification focuses on the consequences of a defect. Exploratory checks and reproducible scenarios complement each other but meet different needs.

    Risk-guided exploration

    When to choose it
    Useful for finding problems in a new workflow. Define important functions and variations to try.
    What to plan for
    Keep steps for reproducing defects. Distinguish a one-off observation from a confirmed problem.

    Reproducible acceptance-test scenarios

    When to choose it
    Suitable for checking results after a change. Describe the context, data, action and observable outcome.
    What to plan for
    Protect important uses and maintain expectations as the product changes. Avoid checks that merely repeat the implementation.

    A risk to anticipate

    A successful test in one environment does not guarantee success elsewhere. Identify the data, permissions, browsers and dependencies tested.

    How to decide

    Classify findings by their consequences for users. Justify the release decision with blockers, known limitations and checks still required.

    Explore this decision further

    Compare scope and costs

    Setup
    Compare journeys, environments and covered edge cases. Test count does not replace the ability to check useful actions and associated failures.
    Recurring operation
    Plan retesting, fictional test-data upkeep and regression handling. Separate running a campaign from maintaining automation; outcomes must remain interpretable.
    Exit and handover
    Request scenarios, results and reproducible conditions without secrets. The next team must identify coverage gaps and repeat checks without treating a report as a guarantee.

    Compare the costs of three web solutions

    Frequently asked questions

    What makes a defect reproducible?

    State the version, page, preconditions, synthetic data, steps, expected result and observed result. Add the effect on the task and useful evidence. Avoid screenshots containing secrets. Another person should be able to reproduce the difference without reconstructing your session.

    How does a verified fix differ from full validation?

    Repeat the original case, then related uses depending on the same component. State what was and was not checked. The disappearance of one known error does not demonstrate that every site function is correct.

    What if a defect cannot always be reproduced?

    Record known conditions, observed frequency and available logs. Simplify the scenario without hiding uncertainty and distinguish what remains suspected. Describe intermittent defects by their impact and evidence.

    Further reading

    Browse other resources