The assumption to test
A team wants documents to remain usable without a mandatory cloud service. It is tempting to assume that removing someone from sharing also withdraws every document already delivered. A local replica exposes the gap: permission governs future exchange, while an existing copy creates continuing storage, protection and recovery responsibilities.
A copy beyond the switch
Consider the proposed lifecycle test: add a device to a collection, disconnect it for a period, then remove it from sharing. Ask what remains readable and what happens when it reconnects. If a document was copied outside the application, synchronization cannot guarantee its deletion. This counterexample limits the promise of remote administration; it does not establish that every application handles revocation identically.
Make ownership inspectable
A proposed settings view should identify stored collections, distinguish originals from metadata, and explain recovery arrangements. Authorized local removal should target the chosen collection while preserving unrelated work. Device and departure policies should specify where complete copies are allowed. These choices make control understandable without implying that an offline label or encrypted network traffic proves protection of stored copies.
A hypothetical authority boundary
Suppose an engineer has an offline maintenance document but needs a manager's approval before authorizing a repair. The document remains useful for reading and drafting, yet local possession does not supply the missing decision authority. This is a hypothetical boundary case, not a reported deployment. A practical rule is to pause binding approval when required authority is unavailable, while allowing permitted preparatory work.
Check behavior, limit conclusions
Run the lifecycle test and record local availability, future exchange after reconnection, and the explanation shown to the owner. Separately check whether an external copy persists and whether local removal preserves unrelated work. Compare observations with the stated policy. The research basis is limited to abstract and introductory material from "Local-first software: you own your data, in spite of the cloud." Applying it through the direct-sync plan is an engineering hypothesis. ABOLTUS is in development; these are proposed checks, not benchmark results or claims of available features.