AT

FLUID vs. ODCS vs. ODPS: Data Product Specs Compared

If you are standardising data products, you will meet three acronyms quickly: ODCS, ODPS and FLUID. They sound like competitors. They are mostly complementary. They describe data products at different layers.

ODCS: the Open Data Contract Standard

ODCS describes the contract of a dataset: its schema, data types, quality expectations, SLAs, and the terms between a producer and a consumer. It answers “what is this data, and what can I rely on?”

ODCS is focused and well-scoped. It is about the agreement, not the surrounding product.

ODPS: the Open Data Product Specification

ODPS describes the product around the data: its description, pricing, licensing, the data holder, and the datasets it bundles. It answers “what is this product, who offers it, and on what terms?”

ODPS sits a layer up from ODCS. It is the packaging and commercial framing, where ODCS is the technical agreement.

FLUID: one contract, compiled

FLUID is an open standard with a different emphasis: it is designed to be compiled. A FLUID contract is a single file that captures a data product’s structure, governance and platform binding, and a compiler turns it into real, deployed infrastructure on a specific cloud.

The point of FLUID is not to out-describe ODCS or ODPS. It is to be executable, and to interoperate. forge-cli, the FLUID compiler, emits ODCS and ODPS from a FLUID contract, and imports them too. You model a data product once and stay compatible with the wider ecosystem.

How to think about them

A useful way to hold all three at once:

  • ODCS: the technical contract. What the data is.
  • ODPS: the product wrapper. What the product is.
  • FLUID: the compilable definition. How the data product gets built.

They are layers, not rivals. Adopt FLUID and you do not give up ODCS or ODPS. You generate them. Already invested in ODCS or ODPS? FLUID can read them. The honest advice for a team standardising today is to avoid betting against any of the three: pick the layer you need first, and choose tools that interoperate across all of them.

Learn more about the FLUID standard, or request the introductory field guide.

← All resources Map your data products