# Real-Use Testing and Constructed Fixtures Catch Different Bug Classes ## Claim Exploratory use of a product in its ordinary workflow surfaces different defects than deliberately constructed edge-case fixtures do, so a QA practice needs both rather than treating one as a substitute for the other. <!-- atomicity: 1 idea ✓ --> ## Scope & conditions - Holds when the constructed fixtures target states (high volume, rare configuration, long-tenure data) that ordinary product use wouldn't naturally reach within a normal testing window. - The claim is about coverage breadth, not about which method is more efficient per bug found — McCarthy doesn't compare cost-effectiveness, only failure-class difference. - Doesn't establish the actual overlap or ratio between the two bug classes — only that they're not identical, which is a qualitative claim in the source, not a measured one. ## Evidence Anne McCarthy runs real WordPress 7.1 sites for ordinary use alongside deliberately built fixtures (revision-heavy posts, deeply nested comment trees, titleless posts for excerpt-fallback testing) for behaviors her real sites don't naturally reach. Her stated reason: "you simply run into different bugs and points of confusion when you're in the flow of using a product over doing artificial testing" — treating the two modes as complementary rather than interchangeable. ## Links - [[DevRel MOC]] — why both testing modes are needed, and what each one structurally cannot see - [[3. Resources/Permanent Notes/Generated Test Fixtures Are Not Evidence Until Independently Validated]] — because applies-in; both are about disciplined QA practice around the same WordPress 7.1 testing workflow, one on coverage strategy, one on evidence quality. - [[3. Resources/MOCs/Open Source Governance MOC]] — because applies-in; a concrete testing-methodology practice for open-source release QA. ## Tests / falsifiers - Track which WordPress 7.1 confirmed Core/Gutenberg issues originated from real-site use versus from a constructed fixture, over a full release cycle — would quantify the claimed difference instead of asserting it. - Find a defect class caught by neither method, which would bound the claim's coverage rather than falsify it directly. ## Maintenance - Last tended: 2026-08-16 - Next review: 2026-11-16 ## Source - [[3. Resources/Literature Notes/Using WordPress Studio to speed up testing 7.1]] — states the real-use-vs-constructed-fixture distinction directly and lists the fixtures built to reach otherwise-unreachable states.