The current architecture might not support this, and there are probably a bunch of questions to consider, but it was what I thought of when I encountered the EO glossary. Some questions suggested by @do-me :
- What language is authoritative (eg for cross-linking terms)?
- What if language nuances lead to a different graph?
- What if a language has multiple words for the same concept?
- What about versioning in different languages?
- Do we need a separate translation approval strategy?
The current architecture might not support this, and there are probably a bunch of questions to consider, but it was what I thought of when I encountered the EO glossary. Some questions suggested by @do-me :