ZJX Documentation

Home | Docs | Who / Why / Better | llms.txt | Sley

1. Thesis

ZJX is the native compressed structural transport and cache envelope for Sley. It carries typed graph shards, graft bundles, diagnostics, trace receipts, seals, and repeated compiler-agent snapshots.

ZJX is not the Sley semantic identity contract yet. Sley semantic identity should come from deterministic canonical graph bytes. ZJX wraps those bytes for movement and storage.

2. Correct Boundary

canonical_graph_bytes -> semantic hash
canonical_graph_bytes -> ZJX envelope

Avoid treating compressed envelope bytes as the semantic hash until both the Sley graph encoding and ZJX canonical encoding are frozen.

3. Initial Payloads

PayloadPurpose
shardA bounded Sley graph slice for agent inspection or edit planning.
symbolsSymbol and binding tables needed to understand a shard.
diagnosticCompiler, checker, authority, lifetime, or graft rejection details.
graftA structural patch bundle submitted to the Loom.
receiptThe accepted or rejected result of a checked graft.
traceAccepted provenance chain for graph changes.
sealImmutable content-addressed artifact metadata and bytes.
snapshotRepeated module or arena state used for compiler-agent caching.

4. Sley Integration

Sley uses human-readable .sley source as the review projection. The canonical program is a typed graph. ZJX carries graph fragments and edit artifacts through the AI-facing loop. 

.sley       human-readable source projection
.sley.graph canonical graph bytes, if exposed
.zjx        graph shard, graft, trace, receipt, seal, or snapshot envelope

5. Agent / Compiler Loop

  1. The Loom parses Sley source into canonical typed graph bytes.
  2. An agent requests a bounded shard.
  3. The shard, local symbols, diagnostics, and authority context move as ZJX.
  4. The agent returns a graft bundle as ZJX.
  5. The Loom validates parse, type, effect, authority, lifetime, and provenance checks.
  6. The accepted or rejected result returns as a trace receipt.
sley graph src/app.sley --slice task:get_user --wire=zjx
sley graft patch.zjx
sley trace src/app.sley --wire=zjx
sley seal src/app.sley --out seal.zjx

6. What ZJX Should Not Claim Yet

  • No universal compression dominance claim.
  • No semantic identity claim while graph and envelope encodings are still moving.
  • No token-window reduction claim after decompression.
  • No replacement for encrypted secret storage.
  • No production-stable binary standard claim during active v0 work.

7. Why Not Plain JSON?

JSON remains useful as the inspectable baseline for tests and early debugging. ZJX becomes useful when Sley graph payloads repeat, grow, and move often enough that compact deterministic transport and caching matter.

The benchmark target is not ideology. The target is real Sley graph payloads: shards, grafts, diagnostics, traces, and repeated snapshots.

8. Current Limits

  • The Sley graph encoding must stabilize before ZJX can be a hard semantic contract.
  • JSON remains the inspectable test baseline.
  • The first real ZJX milestone is deterministic graph byte encoding, then measured transport savings.
  • Public docs describe architecture direction, not a frozen external package.

9. Related Product

Read Sley Documentation for the language model that ZJX carries. Sley and ZJX are designed as one structural programming stack: language graph plus transport envelope.

Return to ZJX home