Design backward from the evidence

“Visit an asteroid” names a destination. “Does its surface vary between regions in a way consistent with different materials?” starts to specify an investigation. A question becomes useful for design when competing explanations predict observably different results. Without that difference, a mission may collect attractive data that cannot decide what its team wanted to know.

Start by stating the possible explanations and what would count against each. This does not guarantee a decisive result. It makes clear why particular observations are worth obtaining and where the investigation may remain ambiguous.

An observable — open definition is a quantity the investigation can measure or derive from measurements. For a fictional surface study, the team might compare reflectance in selected wavelength intervals across multiple regions. That choice still requires calibration, viewing geometry and enough signal to distinguish a physical difference from instrument variation.

The desired conclusion and the recorded detector output are not the same thing. Write the chain from detector values through calibration to the physical inference. Each link reveals a possible requirement — open definition on the instrument, navigation, timing or ground processing.

Suppose two models predict seasonal changes on a hypothetical world: one predicts a signal that varies, and the other predicts an approximately stable signal. A single snapshot cannot directly test the time pattern. The design needs repeat observations, useful seasonal separation and comparable measurement conditions.

A higher-resolution camera does not compensate for missing the required observing times. The most important design choice may be the orbit or mission duration. Selecting hardware first can hide that problem until the mission is expensive to change.

Work backward from the difference to detect

The observing plan should say how many locations or times are needed, which controls matter and how the data reach the analysis team. A measurement stored aboard an inaccessible spacecraft has not completed the evidence chain. Communications and archive documentation belong to the science design.

NASA’s system design process connects stakeholder expectations with requirements and candidate solutions. In this course, the fictional examples use that general reasoning to keep choices tied to evidence. They are exercises in design logic, not specifications for an actual flight project.