Skip to content

Ecosystem

The standard does not stand alone. This page lists what it builds on, which tools read it, and what was examined and set aside.

Every entry carries a status: proposed or implemented.

All foundations are W3C work. The format reuses them instead of inventing its own vocabulary.

  • SKOS, W3C Recommendation, 2009. Concepts, labels, definitions and relations between terms. Status: implemented.
  • JSON-LD 1.1, W3C Recommendation, 16 July 2020. Every glossary is also a linked data document. Status: implemented.
  • YAML-LD, W3C Community Group work. A YAML-LD document is any YAML document from which a conversion to JSON produces a valid JSON-LD document. Each glossary file meets this definition. Status: implemented.
  • PROV-O, W3C Recommendation. Who decided a term, and when. The format records this in its own origin and historyNote fields today. Mapping them to PROV-O is planned. Status: proposed.
  • SHACL, W3C Recommendation. Rules over the graph, for example a relation that must point to an existing concept. Status: proposed, later.

Every tool reads the same files. One <context>.glossary.yaml per bounded context, plus one atlas per domain.

The tools and projects are listed on the Adopters page. Each entry is marked as proposed or implemented.

The alternatives below were examined on 2026-09-08. They are listed here so nobody examines them twice.

Event-Sourced Domain Modeling ships a ubiquitousLanguage block in its core schema. It holds term, definition and avoid[{term, reason}], with additionalProperties: false.

One language per term. No relations between terms. No status. Extensions live in the binary. The documentation states: “not user-extensible: writing a new extension requires changes to the binary itself.”

So Rechnung cannot sit next to Invoice as a label of the same concept. And Customer in Sales cannot point to Account in Billing.

TBX is the XML exchange format of the translation industry. It has dialects TBX-Core, TBX-Min and TBX-Basic.

XML is difficult to read in a code repository. It has no relations between vocabularies. A glossary sits in a repository beside the code it describes. XML does not fit there.

A term has a name and a description inside a set.

Language tags do not exist. So Mehrwertsteuer and VAT cannot be distinguished by language. No relations between terms. No history of who decided what.

SKOS-XL models labels as resources of their own. It is not a competitor to SKOS. The JSON-LD output of the standard uses it for hidden labels.

A glossary in Markdown, TOML, or YAML with custom keys is easy to write. It is either not validated, or it has no standard behind it. Both problems are the reason this standard exists.

All retrieved on 2026-09-08.

Written and maintained by Tobias Brennecke.tobiasbrennecke.com