Scheman

Scheman is a term used in data architecture to describe a family of approaches for describing, validating, and evolving data schemas across diverse data stores and services. It refers to both a schema declaration language and the surrounding tooling that manages schema lifecycles, including versions, migrations, and governance. Scheman aims to provide a unified way to express data structures — such as records, objects, and relationships — and to enforce those structures at boundaries such as APIs, message buses, and storage layers.

Core concepts in Scheman include a formal schema language, a central registry, and runtime validators. Schemas can be defined in human-readable formats and compiled into machine-enforceable rules, supporting primitive and composite types, constraints, and references. The schema registry stores versions, enables discovery, and supports compatibility checks between producers and consumers. A migration or evolution component can generate or apply transformations to move existing data to a new schema while preserving backward compatibility where possible.

Scheman is architecture-neutral and designed for distributed systems, including microservices, event-driven architectures, and data lakes. It seeks interoperability with related standards such as JSON Schema, Avro, OpenAPI, or XML Schema, while prioritizing cross-store portability and lifecycle governance over vendor-specific serialization formats. Adoption typically centers on improving data quality, contract stability, and traceability of schema changes across teams and systems.

In practice, Scheman supports use cases such as API contract enforcement, event schema validation, data quality checks upon ingestion, and automated data lineage. Critics note that it can introduce complexity and require cultural alignment around governance and versioning.