NOTE: Older notes are deprecated. The whole redesign may be an epic. This is about coming up with an "as optimal as we can make it in the allocated time" relational model for Zanata's database. It does not include implementing it. Verification: includes prototype implementation with production-sized sample data requirement: sample queries run in a reasonable time stats for document, project (all locales), project (single locale) find copytrans candidates look up locales for a version (or equivalent) get text flows in a document matching a search phrase Some requirements for the model: copytrans query can work efficiently stats can work efficiently must be consistent in use of inheritance or templating between different levels of hierarchies (such as group->project->version) may: revise the concept of how projects and versions work, may not want to keep versions at all. file blobs still should not be stored in the database
Migrated; check JIRA for bug status: http://zanata.atlassian.net/browse/ZNTA-238