AppTheory
AppTheory is the serverless application framework that sits directly on top of TableTheory in the Theory Cloud stack. Every AppTheory construct that needs durable state — event workloads, jobs ledger, idempotency, sanitization audit, logging profiles — persists through TableTheory.
AppTheory (serverless runtime)
│
┌─────────┴──────────┐
│ │
TableTheory (data layer) KnowledgeTheory
What AppTheory depends on
| AppTheory feature | TableTheory primitive used |
|---|---|
| Event workloads | Single-table state machines + version locking |
| Jobs ledger | Sort-key ordering + TTL |
| Idempotency state | Conditional puts + TTL |
| Lambda configuration cache | Lifecycle timestamps |
| Sanitization audit | Append-only writes, time-ordered keys |
| Logging profile bindings | Per-tenant partition keys |
The AppTheory team owns the shape of these tables; TableTheory owns how the shape is expressed and persisted.
Why this matters for TableTheory stewardship
A breaking change to TableTheory’s marshaling, locking, or encryption semantics breaks AppTheory’s idempotency guarantees. The blast radius is not “AppTheory needs a code update” — it’s “every AppTheory workload silently misbehaves.” That’s why the contract scenarios gate every commit and why breaking changes are coordinated explicitly with the AppTheory steward.
Stack diagram
AppTheory ──depends on──▶ TableTheory
│ │
│ ▼
└──depends on──▶ DynamoDB
No reverse arrow. TableTheory does not consume AppTheory.
Related
- Architecture Patterns — the Theory Cloud stack and TableTheory’s place in it
- Features · Optimistic Locking — the primitive AppTheory’s event workloads compose on
- AppTheory repository — AppTheory’s own documentation site