Security and test data in the Workbench’s first two days
The first work on the Journalism Workbench was not features: documentation, dependency and security housekeeping, and removing real people's names from the test data.
In this piece
The Journalism Workbench repository began on 16 September as Journalism Workbench DEV 1.24. Its first two days went on the things you do before you let other people near a tool that holds sensitive material.
A product-first README, with the developer journal moved out to a changelog. A licence, issue templates and an editor config. Dependency updates grouped, monthly, and ignoring major versions so they didn't bury real changes. Code scanning and a security policy. A patch for a vulnerable dependency set. Path-scoped CI for the front end.
Real names out of the test data
The most important change for what this tool is for: real-world investigation subjects' names were replaced with generic placeholders, and residual identifying details were scrubbed from the test fixtures.
A tool that would later hold records about real people shouldn't start life with real people's names in its test data. Doing it in the first two days meant it never had to be done under pressure.
Sources and notes (3)
- Changelog — the notes this was written from: the first two days
- README — the notes this was written from
- Security policy — the notes this was written from
Journalism WorkbenchAll 12 parts →
