Afterberry Vault
Status & Revision History
Revision 2.0 — 20 September 2026. What exists, what has been rehearsed, what is planned.
§T1 How to read this page
The rest of the manual describes a design. This page says how much of the design exists. An item is done when there is evidence a stranger could check; declared when it rests on the maker’s word alone; rehearsed when it has been carried out at small scale with everything published; planned otherwise. Dates are added when things happen, not before.
| Milestone | State | Date | Evidence |
|---|---|---|---|
| Founding collection of 8,888 photographs | done | 2 Sep 2026 | The 8 Museum register |
| First seal: locations manifest published | done | Mar 2026, declared | file and digest; the date awaits an independent timestamp |
| First seal: key destroyed | declared | Mar 2026 | the file’s own statement |
| Architecture, revision 2.0 | done | 20 Sep 2026 | this manual |
| Reference implementation | done | 20 Sep 2026 | §T3 |
| Lock, both doors, relay proofs | rehearsed | 20 Sep 2026 | §T3 |
| Capsule encryption and sealed keys | rehearsed | 20 Sep 2026 | §T3 |
| Carrier embedding and extraction | rehearsed | 20 Sep 2026 | §T3 |
| Erasure coding, 60 of 88 | rehearsed | 20 Sep 2026 | known-answer vector in selftest; independent check in Go, §T3 |
| Independent timestamps for this revision | planned | ||
| Canonical carriers for all 8,888 photographs | planned | ||
| Assembly of the 101 capsules | planned | ||
| Ceremony script published for comment | planned | ||
| The ceremony | planned | ||
| First replica with a fixity report | planned |
§T2 Deposit register
A copy is entered here when its custodian has returned a fixity report (Preservation §R4). No capsule has been sealed, so nothing has been deposited, and the register is empty. It will not be padded.
| Id | Custody class | Custodian | Contents | Placed | Last fixity |
|---|---|---|---|---|---|
| no entries | |||||
§T3 The rehearsal set
SHA-256 of each published file. The set is described in Sealing §S10 and Photographs §E9.
The rehearsal’s primes are in the lock file. They are meant to be.
§T4 Open questions
- Who holds the primes, even briefly. A jointly generated modulus would mean that no single machine ever knew p and q. The protocols exist; whether they can be run by a handful of witnesses in a room, to safe-prime standard, is unsettled.
- The reference rate. One nanosecond per 8,192-bit squaring is a judgement. ASIC designs of 2020 already expected a few nanoseconds at 2,048 bits. It will be reviewed against the hardware literature before the exponents are fixed at the ceremony; if it moves, it will move towards a faster rate and larger exponents.
- A post box. If each ring also published an encryption key whose private half is derived from its waypoint, anyone could seal a letter to the same horizon, forever, without the maker. It would make the lock a public utility. It would also add a third door — the security of that encryption scheme — and the trade has not been made.
- Relay governance. Who publishes the log, how forks are resolved, and how a runner proves a leg without doubling its cost.
§T5 Revision history
| Revision | Date | Summary |
|---|---|---|
| 2.0 | 20 Sep 2026 | Architecture revised throughout. Introduces the two-door lock (AB-L1), rings and the Remainder, chunked capsule encryption with sealed keys (AB-E1), 101 × 88 coding (AB-S1), the 88-byte photographic record (AB-R1) and keyed carrier embedding (AB-C1), a single signing identity created and destroyed at the ceremony, and the rehearsal set. Licence revised to ATL 2.0. Supersedes all 1.x documents as the description of the design. |
| 1.0 | 25–26 Mar 2026 | First publication: manual, ATL 1.0, resilience, locations, reconstruction and embedding documents. Locations manifest sealed. |