Editorial Policy
How Testpartout writes about its tests
The written content on this site follows one rule above all: it must match the product as it actually behaves. These are the working principles behind that rule.
Editorial principles
Content describes implemented behaviour
What a page says a test does must be what the shipped code actually does. Methodology sections are written against the implementation, not against an aspiration or a roadmap.
No unsupported averages, ranks or percentiles
Testpartout does not publish population averages, global rankings or percentiles that are not derived from real aggregated measurements. Where that data does not exist, the comparison simply is not shown.
Factual claims must match the product
Any factual statement about features, storage, data handling or measurement behaviour must be verifiable in the product as it exists today. Planned features are described as planned — never as available.
Uncertainty and limits are disclosed
Browser measurements carry inherent limits: hardware latency, refresh rate, scheduling and input pipelines. Pages state these limits openly instead of presenting numbers as laboratory-grade.
Pages are updated when the implementation changes
When a test's behaviour changes, its content is updated to match. Stale methodology text is treated as a defect, not as an acceptable leftover.
Errors are corrected
When an inaccuracy is identified in published content, it is corrected. Corrections take priority over preserving wording.
Content exists to explain, not to fill pages
Editorial content exists to explain the tests, help interpret results and support troubleshooting. It is not produced to inflate page counts for search engines, and thin pages with no real information are not published.