What is the difference between a dataset and a data product manifest?

• A. A dataset is raw or transformed data; a data product manifest describes its consumption contract, metadata, and SLAs.
• B. A dataset is a processing script.
• C. A manifest is a dataset.
• D. They are identical.

Answer: (A)

In Foundry, the actual data is the dataset: raw or transformed data stored and accessible for analysis. A data product manifest, by contrast, encodes how that data should be used: the consumption contract, the metadata that describes the dataset (like schema, provenance, lineage, owners), and the service-level expectations (SLAs) for access, freshness, and quality. The manifest does not hold the data itself; it documents the rules, constraints, and metadata that govern how consumers interact with the dataset. This separation lets data products expose usable data with clear governance and expectations, while the dataset remains the underlying data resource. A processing script is not the dataset, a manifest is not the dataset, and the two are not identical, since one is the data itself and the other is the contract and metadata describing its use.