Better detection also means preserving content
The next quality questions for RiffBlock are fewer false positives, understandable protection states and measurable work instead of broad promises.
More rules alone do not establish that a blocker is improving. An ordinary control with an advertising-like name must remain usable. A repair should help a specific problem. An enabled extension does not establish that every required permission is available.
Understandable protection checks
The development version’s protection check brings together extensions, applied filters, filter age and the local Safari page test. Unknown, paused and confirmed remain different states. An old test cannot confirm a new filter generation.
Less repeated work
Concurrent configuration requests now share certain local reads. Imported learning rules notify affected open tabs, while outdated responses are discarded after a protection change. Controlled comparisons show less repeated work. This is not yet a measured battery improvement on a physical iPhone.
The next quality round
I want a more systematic collection of real missed advertisements and ordinary counterexamples. Detection, false positives and usability issues need separate assessment. Longer Safari and YouTube sessions, device checks and energy measurements remain on the list.
Privacy also needs clear boundaries: reducing tracking does not automatically anonymise a connection. A password-protected settings export is a possible next step I am evaluating, not a delivered feature. The 1.1 notes remain in preparation until further acceptance.