Skip to content

What DTCS is (and is not)

Short design philosophy for evaluators and new users. Normative detail: SPEC.md. Limits: limits.md.

What it is

  • A draft standard for vendor-neutral data transformation contracts (YAML/JSON)
  • A way to capture semantics: schemas, actions, expressions, rules, lineage, guarantees
  • A reference toolkit to parse, validate, analyze, plan, and (for small fixtures) run contracts
  • A CI / governance surface: fail builds on invalid contracts; compare revisions; declare capabilities

What it is not

Not this Instead
A Spark / Polars / DuckDB / SQL engine Contracts describe meaning; your engine executes
A production ETL runtime dtcs run is an in-memory teaching interpreter
A storage or orchestration system Out of scope by design
A finished, ratified standard Spec 3.0.0 is draft; tools are alpha
A dual-compiler certified portable suite for all A.9 profiles Advanced profiles are Experimental; /2 is Candidate
A full Node/WASM analytics stack npm bindings: parse / validate / declare only

Mental model

Author contract  →  validate / analyze in CI  →  match engine capability
                                              →  execute in your platform

Use DTCS so teams agree on what a transform means. Implement how in the engine you already trust.

Next