FAQ

Is this a new ATL?

No. ATL is a cryptographic substrate: append-only log, receipts, inclusion proofs. APL is a semantic layer that runs on top of a substrate like ATL.

ATL proves existence and history; APL specifies meaning under frame.

Is APL just metadata?

No. A piece of metadata alongside a claim is annotation. APL is a protocol: it defines a binding between the claim and a content-addressed frame, resolution and verification rules for that frame, and validity semantics for the claim under that frame.

The difference matters in practice: annotations can be silently changed or ignored; APL’s frame binding is cryptographically enforceable.

Why not C2PA?

C2PA focuses on media-lineage provenance with a specific trust model and a specific set of assertions. APL is broader and more neutral: it is a general protocol for frame-bound claims in any domain (AI evaluation, scientific measurement, document attestation, photojournalism, etc.), and it is explicit about separating cryptographic substrate from semantic frame.

APL can coexist with C2PA. They solve overlapping but distinct problems: C2PA answers “what is this artifact and how did it get here”; APL answers “under which declared observational frame was this claim made, and is the claim valid under that frame”. A photograph can carry a C2PA provenance chain AND an APL claim (observer, instrument, aspect = scene-light-capture, explicit exclusions about denoised output) in the same receipt; the two layers do not compete.

Who decides which frame is “right”?

No one. A verifier checks that a claim is valid under the frame it declares. Two claims can both be valid, under different frames, and disagree — and APL makes that disagreement inspectable rather than hidden.

Choosing which frame to trust is a matter of domain policy, not protocol.

Can I use APL without ATL?

Yes. APL requires some carrier that meets the minimal contract defined in The Protocol § 5.1. ATL is the canonical substrate (and the reference profile APL-on-ATL wires it in), but APL is deliberately substrate-agnostic.

Why do I need a registry?

You don’t, strictly. frame_ref resolution works from any source that serves the correct bytes — including a local cache.

A registry is a discovery convenience: it maps human-readable aliases to hashes and hosts frame documents. A verifier’s correctness must not depend on any single registry, because frame identity is the content hash.