Validation Methodology
Move from definition to evidence without skipping the steps in between.
The path a claim takes from a written definition to a validated result, and what each step can and cannot establish.
From method to level.
A level is earned by a kind of work, not by time passing. This table shows which work can support which level.
| Method | Level it can support | What it consists of |
|---|---|---|
| Architecture definition | E1 Architecture Defined | A reviewed specification of what the system is and how its parts relate. |
| Prototype evaluation, simulation, software in the loop | E2 Prototype Validated | Repeatable runs on prepared inputs, with the setup recorded. |
| Hardware in the loop, integration validation | E3 Integrated Validation | The parts running together, with real interfaces and recorded results. |
| Controlled field validation | E4 Controlled Field Validation | A planned trial in a real environment, with stated conditions and a record. |
| Qualification, operational evidence | E5 Qualified / Operational | A formal qualification, or records from operation within a stated scope. |
| Item | Defined | Prototype | Integrated | Field |
|---|---|---|---|---|
| Module A | Done | In progress | Not started | Not started |
| Module B | Done | Not started | Not started | Not started |
| Module C | In progress | Not started | Not started | Not started |
What each method can and cannot establish.
Simulation is not field evidence. Each method answers one kind of question and is silent on the next.
- Architecture reviewCan establish that the design is consistent and complete on paper. Cannot establish that it works.
- Simulation and software in the loopCan establish how the logic behaves on prepared cases. Cannot establish how it behaves with real signals, real clutter or real weather.
- Hardware in the loop and integrationCan establish that the parts interoperate. Cannot establish how the whole behaves in an uncontrolled setting.
- Field trial and qualificationCan establish behaviour in the conditions that were tested. Cannot establish behaviour in conditions that were not.
How a claim is written.
Every claim that rests on a record names three things, so a reader can see where it stops.
- ScopeWhat was tested, and what was not.
- ConditionsThe setting, inputs and date under which it held.
- LimitationWhat the result does not show.
Corrections and versions.
We correct in the open. A changed claim keeps its earlier wording in a visible history, and every article and evidence record carries a version and a date.
- A level can go downIf a record is withdrawn or found weaker than stated, the level falls and the page says why.
- Tell us what is wrongIf you believe a page says more than its evidence supports, write to us. That is a correction request, and we answer it.