Fictional product catalog ยท Technical demonstration
HR Atlas

HUMAN RESOURCES SOFTWARE GUIDE

Data method

Data is generated deterministically. Dates represent sample content changes, not real product updates.

How is the data prepared?

The same generation rules produce the same records. Products vary by category, team size, industry, deployment model and approval steps. Initial dates represent fictional content changes, not real product announcements.

How are changes processed?

The controlled write path validates the data and the revision originally read. Content changes update the page and affected listing dates. Reimporting identical content does not refresh dates. Stale revisions are rejected, and a failed transaction rolls changes back.

Different data structures

SQLite and D1 use flat page records. The PostgreSQL proof stores entities and translations in separate tables. A shared read layer maps both structures to the same page model. Customer data still requires field mapping and a data quality review.

Content date: 2026-09-28