Validating dental software is not the same as testing it
Testing asks whether the software does what it was built to do. Validation asks whether what it was built to do is the right thing, and whether you can demonstrate that to somebody else.
For clinical and research tools the second question is the one that matters. A tool that produces plausible output and cannot show its reasoning will not be accepted by the people who have to rely on it.
Validation means a defined intended use, a defined population, and evidence that performance holds within both. It is closer to study design than to software engineering, which is why it usually needs both skill sets in the same room.
Working on something like this?
Get help from an expertWhat an evidence plan actually needs before you run the first test
Most testing programmes are designed backwards. Decide what claim you need to defend, then work out what data defends it.
Mateus RochaThe aims page decides the grant
Reviewers form a view in the first half page. Three aims that depend on each other read very differently from three that do not.
Dayane OliveiraShort color science training measurably improves shade matching
What our 2025 study found, and what it means for how restorative dentistry is taught.
Dayane Oliveira