CarSpotter

The clue is the primary object on the play screen. The choices need to stay readable, and the reveal needs to lead clearly to the next car. I treat those as connected states of one round, rather than separate screens to decorate.

The sequence below is taken from one recorded Daily. The question and reveal belong to the same Toyota Camry round; the result is the completed five-car run. Scores and car images are preserved from the capture.

A Saturn wagon is a more personal signal for this game than a generic sports-car silhouette. It reflects the kind of enthusiasm behind CarSpotter: everyday cars can be recognisable and interesting too.

The cream wagon, electric lime and dark grey surfaces give the identity continuity across the home screen and marketing artwork. I kept the photographic game content distinct from the illustrated brand character, so the identity does not suggest a driving or racing mode.

The earlier result card used letters tied to types of detail. That made the summary depend on a category the player did not need to know, and that the production content could not consistently support.

The redesign uses Q1–Q5 with explicit check and miss symbols. It answers a more useful question: what happened in each round? The same outcome language carries into the exported card, while the car names and clue photos stay out of it.

These are different recorded runs; the comparison documents the change in visual language. Comprehension and conversion remain questions for further testing.

The work extends beyond the visible screens. Game rules, saved progress and media references need to remain dependable across interruptions and updates. I separate game logic from presentation so changes to the interface can be checked without rewriting the scoring rules.

One earlier failure involved media references tied to an installation-specific file location. The design response was to use stable references and validate media before trusting it. A file existing at a hash-shaped path is not enough: the bytes must be checked against the expected identity.

For the latest interface work, the practical checks are concrete: complete answer labels at larger text sizes, reachable reveal actions, a usable static state with Reduce Motion, and a run that can resume without losing its score. Simulator evidence and device checks need to be reported separately.