Increase Quality¶
Enforce 100% test coverage based on real business scenarios and requirements.
Quality in most organizations is an afterthought. Testing is something developers do when they have time. Test coverage is a metric nobody really believes. "It works on my machine" is still an acceptable answer in too many places.
This doesn't work for an Enterprise Knowledge Graph. An EKG that serves strategic use cases across the organization cannot afford "mostly works" or "works in the demo." It has to work. Every time. For everyone.
The Testing Problem¶
Traditional testing approaches fail for several reasons:
- Tests are disconnected from requirements - Developers write tests based on what they built, not what the business asked for
- Coverage is meaningless - 80% code coverage says nothing about whether the system does what users need
- Tests rot - Requirements change, tests don't, and nobody notices until production breaks
- Testing happens too late - Problems found in QA are expensive; problems found in production are disasters
100% Coverage That Actually Means Something¶
The Use Case Tree enforces, requires, and enables 100% test coverage—but coverage that's based on real business scenarios and requirements, not arbitrary code metrics.
Every story in a use case defines what the system must do, for whom, and why. Each story has explicit acceptance criteria. Those criteria are the tests. If a story passes its criteria, it works. If it doesn't, it doesn't ship.
No story without criteria. No criteria without tests. No deployment without all tests passing.
The Use Case as Living Specification¶
A use case starts with just a name and a description. Over its lifecycle, it grows into a fully detailed model:
- Concepts - What things are we talking about?
- Personas - Who uses this?
- Outcomes - What value does it deliver?
- Stories - What must the system do?
- Workflows - How does work flow?
- Dependencies - What else does this need?
This model is captured in the EKG itself. The specification is the system. When the model changes, the tests change. When the tests change, you know immediately if something broke.
Build Time and Run Time¶
Most testing happens during development. Ship it, and hope for the best.
The Use Case Tree approach tests continuously—not just during build, but in production:
- Build phase - Every commit triggers full test suites
- Run phase - Production monitoring validates that the system continues to behave as specified
This is the quality level required for enterprise strategic use cases. Anything less is hoping. Hope is not a strategy.