Datasets:
Describe access_events accurately: grants only, revocations replayed
Browse files
README.md
CHANGED
|
@@ -54,10 +54,16 @@ index memory.
|
|
| 54 |
| `documents` | 4,000 | Documents with ACL (`allow_tokens`, `min_level`, embargo window) |
|
| 55 |
| `chunks` | 14,067 | Indexed units; `canary` is a unique marker for exact leak detection |
|
| 56 |
| `principals` | 600 | Departments, projects, clearance levels, capability tokens |
|
| 57 |
-
| `access_events` | 72 |
|
| 58 |
| `queries` | 240 | Evaluation queries; `kind` is `normal` or `probe` |
|
| 59 |
| `revocation_cases` | 50 | Triples where a revocation provably removes retrieved content |
|
| 60 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 61 |
## Access semantics
|
| 62 |
|
| 63 |
A principal may read a chunk when **all** of these hold:
|
|
|
|
| 54 |
| `documents` | 4,000 | Documents with ACL (`allow_tokens`, `min_level`, embargo window) |
|
| 55 |
| `chunks` | 14,067 | Indexed units; `canary` is a unique marker for exact leak detection |
|
| 56 |
| `principals` | 600 | Departments, projects, clearance levels, capability tokens |
|
| 57 |
+
| `access_events` | 72 | Authority timeline as generated: time-limited secondment grants that expire on their own |
|
| 58 |
| `queries` | 240 | Evaluation queries; `kind` is `normal` or `probe` |
|
| 59 |
| `revocation_cases` | 50 | Triples where a revocation provably removes retrieved content |
|
| 60 |
|
| 61 |
+
`access_events` holds only `grant` rows by design. Revocations are *replayed* at
|
| 62 |
+
evaluation time from `revocation_cases`, so each system under test is measured
|
| 63 |
+
against the same withdrawal on a directory it has not already mutated. Rebuilding
|
| 64 |
+
a directory per condition is what keeps a later condition from "passing" merely
|
| 65 |
+
because an earlier one already expired the grant.
|
| 66 |
+
|
| 67 |
## Access semantics
|
| 68 |
|
| 69 |
A principal may read a chunk when **all** of these hold:
|