Lecture 2.1 Direct Manipulation

Direct Manipulation

The lecture explains how good interface design makes it clear both what users can do and what happens when they do it. Examples involving measuring cups, thermostats, and a confusing voting ballot show how observation, feedback, and users’ mental models affect usability.

Main Ideas and Lessons

Observe what people do, not just what they say

People using a conventional measuring cup repeatedly poured, checked the level, and poured again. When asked how to improve the cup, they did not mention the difficulty of reading the measurement.

Observing their behavior revealed an opportunity: a cup designed to be read from above, allowing users to check the level while pouring. Some versions also display English or metric units depending on how the cup is held.

Use prototypes in field research

A prototype can change how people interact with an object and help uncover needs that are difficult to express in an interview. The measuring-cup example illustrates the value of combining behavioral observation with prototypes rather than relying only on stated preferences.

Design for two basic user questions

1. “What can I do?” The interface should make available actions apparent.
2. “What happened?” The interface should provide feedback that makes the result of an action understandable.

Make actions and results easy to understand

Direct manipulation helps connect a user’s actions to their results. The classic Honeywell thermostat makes its central action—turning a dial to set the temperature—visible and straightforward. Its immediate feedback connects the action directly to the displayed setting.

Interfaces that rely on more indirect controls, such as confusing button systems, can be harder to use. The lecture notes a renewed interest in directness in devices such as the Nest thermostat.

Account for users’ mental models

People form internal representations of how things work, often using analogies from prior experience. For example, they may think of a new thermostat as working like an older one, or explain electricity as flowing like water.

These models can be incomplete or inaccurate, and people may hold several models at once. Learning involves building and revising these representations. Designers should not assume that users share the knowledge and mental models designers develop through expertise. A mismatch can lead to slow performance, errors, and frustration.

Distinguish slips from mistakes

• A slip occurs when someone has the right goal or understanding but accidentally performs the wrong action—for example, hitting Delete instead of the equals key. Possible design responses include separating keys, enlarging targets, or adding corrective features.
• A mistake occurs when someone performs the intended action but acts on an incorrect understanding of how the interface works.

Learn from ambiguous interfaces: the butterfly ballot

The lecture describes a Florida ballot from the 2000 election in which candidate names and punch holes did not align clearly. Voters could punch the hole they believed corresponded to their chosen candidate, while the ballot recorded a different selection.

The lecture cites unusually high support for Pat Buchanan in Palm Beach County—0.85%, compared with 0.23% in a comparison set—as a clue that the interface may have contributed to errors. The example shows how election data can help identify possible design problems.

Use consistency to prevent errors

Ambiguous designs can result when engineering constraints take priority over clarity for users. If interfaces vary between counties or elections, people cannot rely on what they learned previously, and errors become harder to diagnose. Consistent interfaces let users build on a familiar mental model.

Speakers and Sources

• Lecturer: Not identified in the subtitles.
• Alex Lee: Credited for the OXO measuring-cup example, presented at a 2008 conference.
• Bill Verplank: Cited for a drawing about the two fundamental interface questions.
• Henry Dreyfuss: Identified as the designer of the classic Honeywell thermostat.
• Don Norman and Jakob Nielsen: Cited for questions designers can use to evaluate whether an interface will feel natural and easy to use.