Archive
37 guides on automation frameworks, API testing, CI pipelines, and the process habits that make quality teams trusted.
2026 · updated entries marked
The four black-box techniques in the ISTQB Foundation syllabus, applied to one checkout on this site's practice shop, with every result checked by a test. Each technique asks a different question, and the state transition model found something none of the others could: a URL that skips the details form.
test cases · tutorial · manual testing
8 MIN
A two-test Playwright run against Sauce Demo, the JUnit XML it produced, and what PublishTestResults@2 does with that file on its documented defaults — including the pattern that matches nothing and the switch that lets failing tests pass. Then the other half: why publishing results never creates a Test Case, and the CSV format that does.
azure devops · ci cd · playwright
10 MIN
A catalogue of API test cases where every case was sent to a real API. 34 requests to Restful Booker, grouped by endpoint, with what a well-behaved API should return and what this one did, followed by a Playwright spec that checks the contract and records the known defects.
api testing · test cases · playwright
12 MIN
Updated
The four common defect metrics that compare bugs found before and after a release are one fraction under different names. This guide gives the formulas with their sources, shows how a different denominator changes the answer, explains the time window, puts the often-quoted 85% average in context, and shows when a change between releases is real.
quality metrics · release · process
7 MIN
Should you use page objects, fixtures, or plain hooks in Playwright? We wrote the same three Sauce Demo tests all three ways, ran them, and measured the differences: how much code each needs, how many places a locator lives, where a failure points when a locator breaks, and what a beforeEach does to a test that doesn't need it.
playwright · automation · tutorial
8 MIN
A hands-on guide to checking a database the way a tester checks a screen. Load a small shop database with seeded data bugs, write a query that returns the rows breaking each rule, and see why the first version of each query often finds nothing.
sql · test data · tutorial
8 MIN
A login page test case list where every case was run. 22 cases against Sauce Demo, grouped by what they protect, with the error text the site actually showed, the cases a practice site can't answer, and a data-driven Playwright spec with its real output.
test cases · playwright · tutorial
11 MIN
Updated
The two are constantly confused, and teams pay for it twice: quality rules copied into every story, or skipped entirely because everyone assumed they were somewhere else. This guide separates them, shows what belongs in each, and turns a Definition of Done into checks a pipeline enforces.
acceptance criteria · definition of done · process
8 MIN
Updated
A hands-on introduction to accessibility testing. Scan a page with axe, read a violation properly, see the same bug in the accessibility tree and in your locators, then find out why a keyboard pass misses it entirely. Includes an exercise with an answer key.
accessibility · playwright · automation
9 MIN
The measured quality gap in AI-generated code is real but narrower than the headline, and it is not the interesting part. The interesting part is that review, testing, and traceability all assume the author understood what they wrote — and a product owner is the person who can put a name against each story before that assumption stops holding.
ai · code quality · product owner
6 MIN
A Definition of Ready stops half-formed work entering a sprint, and it also turns collaboration into a hand-off with an acceptance gate. Both are true. This guide covers when it helps, the failure mode that makes practitioners call it an anti-pattern, and the diagnostics that tell you which one you have.
definition of ready · process · product owner
6 MIN
How to run a go/no-go decision that can genuinely say no: one named owner instead of a group vote, evidence gathered before the meeting, criteria with thresholds written in advance, every accepted risk carrying an owner and an expiry, and escaped defects measured afterwards to correct the criteria.
release · product owner · process
8 MIN
Updated
Acceptance criteria decide what gets built and what gets tested. This guide covers the eight properties that make a criterion checkable, with a real story rewritten from a score of 48 to 100 in our acceptance criteria grader, and the questions each rewrite answered before the sprint instead of during it.
acceptance criteria · requirements · product owner
11 MIN
Updated
Support data is the earliest external signal that a release went wrong, and most of it gets read as a support-team scorecard instead. This guide separates the metrics that say something about the product from the ones that say something about the queue, and gives a weekly routine for turning volume into defects.
quality metrics · product owner · support
7 MIN
Accessibility stopped being a nice-to-have in June 2025. This guide covers who the European Accessibility Act applies to, which standard it points at, what enforcement has actually looked like in France, why overlay widgets do not count, and five checks a product owner can run on their own checkout in fifteen minutes.
accessibility · compliance · product owner
7 MIN
A hands-on introduction to load testing with k6. Start a tiny shop API locally, check it works for one user, prove it meets a response-time goal under steady load, then add users until it doesn't. Includes an exercise with an answer key.
performance testing · k6 · api testing
15 MIN
A pairwise model of the Sauce Demo checkout, 31 generated cases run as a data-driven Playwright test on Chromium, Firefox, and WebKit, and a run of every account, product, and form combination to check what those cases found and what they could have missed.
test cases · playwright · automation
16 MIN
A measured look at Playwright's toHaveScreenshot: how many pixels a real regression changes, why maxDiffPixels: 100 and maxDiffPixelRatio: 0.01 can hide it, what animations: 'disabled' does not freeze, and how to handle baselines in CI.
playwright · visual testing · ci cd
8 MIN
What the keyv and Bitwarden CLI attacks did to CI runners and AI coding agent config files, which defenses we verified on real npm versions (including one that npm 10 silently ignores), and a pull request check for files that run code on folder open.
ci cd · security · github actions
9 MIN
Updated
What a retried flaky test actually looks like in Playwright's list, JUnit, and JSON reporters, how we diagnosed a real one on this site, and a quarantine policy with owners and expiry dates that CI enforces.
playwright · flaky tests · ci cd
11 MIN
Updated
A file-by-file look at what Playwright's planner, generator, and healer agents are allowed to do, plus the review rules we recommend before letting a healer edit your tests.
playwright · ai · automation
6 MIN
Real defects from this site's own codebase: what the symptom was, which check caught it, the root cause, the fix, and the regression test we added.
case study · playwright · accessibility
6 MIN
Updated
2025 · updated entries marked
A hands-on introduction to API testing: explore an API with curl, write a Postman collection and run it with Newman, then write the same checks in Playwright. Uses a public practice API whose real responses include a few surprises.
api testing · postman · rest
9 MIN
Updated
A working Cucumber project from feature file to green build: Gherkin scenarios and a Scenario Outline, step definitions with hooks, a JUnit Platform suite runner, and the dependency mismatch that breaks it if you add JUnit 6.
cucumber · bdd · gherkin
7 MIN
Updated
One test (log in, add a product, check the cart badge) written in Cypress, Playwright, and Selenium against the same public demo shop. How the code differs, how long each run took, what broke during setup, and which framework fits which team.
cypress · playwright · selenium
7 MIN
Updated
A complete, working Selenium WebDriver project: Maven dependencies, a page object, explicit waits, JUnit tests, and a troubleshooting section built from errors we hit while running it.
selenium · automation · java
7 MIN
Updated
A practical path to tests in CI: the GitHub Actions workflow Playwright generates and what each line does, the additions a real project needs (taken from this site's pipeline), and how to shard a slow suite and merge the results into one report.
ci cd · automation · github actions
9 MIN
Updated
A practical guide to test cases: what to include, three techniques for choosing inputs, ten real cases for a checkout form with the results we observed, and how the cases become an automated test. Three of the ten raised questions the form's requirements don't answer.
qa · test cases · process
8 MIN
Updated
We found a real bug in a practice shop and wrote it up two ways. The difference between 30 and 100 points is steps, expected versus actual, environment, reproducibility, and evidence. Each part is explained with the exact lines from the report.
qa · bug reporting · process
8 MIN
Updated
A concrete comparison instead of a debate: which bugs on a deliberately broken demo shop an automated suite caught, which ones a person caught, and a short set of questions for deciding what to automate next.
qa · manual testing · automation
5 MIN
Updated
Create a Playwright project, write login and checkout tests for a real demo shop, see what a failing assertion looks like, and use fixtures and the trace viewer. Copy the code and it runs.
playwright · automation · javascript
8 MIN
Updated
Ten skills for testers, in the order they build on each other: test design, bug reporting, exploratory testing, API testing, one UI framework, Git and CI, test data and SQL, debugging failures, accessibility, and working with AI assistants. Each comes with a free project that proves you can do it.
qa career · skills · automation
7 MIN
Updated
Three rules for test data (own it, make it unique, clean it up) and the code that enforces them: a builder with safe defaults and a Playwright fixture whose cleanup runs even when a test fails. Includes the strategies for when you can't create data through an API.
test data · automation · strategy
7 MIN
Updated
Testing tools grouped by job: browser automation, API testing, BDD, accessibility, CI, test management, device clouds, and reporting. Each entry says what the tool is for, when you don't need it, and whether we've used it ourselves.
tools · qa · automation
7 MIN
Updated
A fixed wait that added up to five seconds to a test, an assertion that passed for text that doesn't exist, a click on the wrong product, and a dialog listener that froze the page. Each mistake is reproduced with real output, and each has a fix.
ui automation · best practices · cypress
8 MIN
Updated
Where the test pyramid comes from, the only widely cited ratios and what their authors said about them, the ice cream cone and hourglass antipatterns, and a rule for choosing the layer of each new test.
test pyramid · strategy · qa
7 MIN
Updated
A plain-language introduction for new testers: what testing is and isn't, seven principles every tester should know, the main levels and types, and a practice session on a deliberately buggy demo shop with a list of real bugs to check your findings against.
qa · software testing · beginner
6 MIN
Updated