`documentInbox(doc)` answers "which inbox do I know for this document?" and mints
a fresh one when the answer is none. A third party therefore never reaches the
owner's inbox: they get one of their own, the owner never reads it, and the
deposit vanishes with no error.
That breaks the consumer's central act — signing up to someone else's event —
and it blocks the directory document that has to replace discovery, since
creators must deposit into a document they do not own.
The consumer tried solving it on its side: resolve the address once at creation
and publish it on the public document, so participants read it instead of
deriving it. It works, and it is still the wrong place. It puts a technical
address into a domain shape; it doubles document creation on every event, since
resolving an inbox creates one (the @data suite went from 9m37 to 21m30, with
sign-up scenarios timing out during setup on a FRESH profile — not wallet
bloat); and the library already claims this job, having stated when per-document
inboxes landed that draining them "is the library's job, not the app's".
Addressing is the other half of that same job. The app-side change is reverted.
No shape is prescribed here. Deriving the address from the document would cost
nothing and match how the rest of the model works — an overlay is derived, a
keyring is looked up, nothing is handed over out of band — with the read guard
keeping the asymmetry where it belongs. Publishing it library-side would work
too and leaves room to make it lazy.
One question decides where this note ultimately belongs: is "a document has an
inbox addressable by any holder" part of NextGraph's target model, or this
library's own invention? Upstream, it is emulation to align; otherwise it is a
design decision to take deliberately.
Sylvain Duchesne