Current study page: Interaction, Interface & Service Design

The earlier reference below is retained for continuity.

Design judgment studio¶

Canonical target: DR08; interfaces with DR03, DR04, DR06, DR11, DR12, EX-M04, ML02 and TS-F22.

Task and reference¶

Design a low-stakes reservation or equipment-loan interaction for actual users. Start with paper and observation. Identify the user's goal, permissions, possible errors, recovery route and accessibility needs before choosing a visual style.

Read Jakob Nielsen's 10 usability heuristics, source SRC-DR-NIELSEN. They are professional rules of thumb, not universal laws or a substitute for user research. Reconstruct the distinction between a system state, a user's mental model and the visible sign that connects them.

Compare and articulate¶

Collect three authorized examples of the same task: one clear, one difficult, and one visually impressive but misleading. Remove brand cues where practical. Rank them against declared criteria before discussing reputation. Explain tradeoffs in information hierarchy, affordance, error prevention, feedback, reversibility, typography and operational cost. Preference is evidence about a person, not by itself a justification of the design.

Draw the state transitions, including cancellation, invalid input, unavailable resources, partial completion and loss of connectivity. Compare a diagram with a paragraph describing the same system. State which mistakes become visible in each representation and where either representation hides timing or responsibility.

Make, observe and revise¶

Build a paper prototype and observe consented participants attempting two realistic tasks. Do not coach them toward the expected answer. Separate observed behavior from your interpretation, and record where your own language misled them. Use synthetic identities and no real financial or health transactions.

Revise one consequential failure. Show the rejected alternative and why it was rejected. Then test a changed context: limited vision, different language, low bandwidth, intermittent service or a user unfamiliar with the institution. Do not simulate a disability and claim to have validated accessibility for that population.

Compression test¶

Attempt the task using only technical correctness and rhetorical explanation. Record whether spatial hierarchy, affordance, embodied use or perceptual comparison still causes recurrent failures. This is an internal DR deletion/compression probe. It does not prove that DR deserves permanent trunk status; an independent matched comparison is needed to distinguish a missing capability from poor instruction.

AI and assessment¶

Before using generation, internalize task states and high-cost failure conditions. Reconstruct the interaction without the prototype. AI may generate alternatives, but the learner owns selection, accessibility, permissions and evaluation.

Receive an unseen change request after the first review. Revise the interaction, explain what the change invalidated and demonstrate recovery from a seeded error. Defend the final choice through the sequence: perceive, compare, contextualize, articulate, curate, make, revise, defend. A beautiful screen that conceals failure or removes recourse does not pass. Nor does a correct but unusable interaction.

Use examiner calibration. Preserve the observation log, versions and disagreement. Expert preference and small user samples are evidence with limits, not numeric guarantees of quality.