Concepts of Interface Quality

Key ideas#

Everyday words don’t work#

“Convenient”, “simple” and “usable” are how people judge interfaces, but they are too vague for professional work. “Simple” may not even be achievable: an aircraft cockpit is more complex than a hang-glider’s controls without being badly designed. Golovach’s advice is not to use these words in your own work, but to expect users to judge in exactly these terms.

Shneiderman’s measures#

Ben Shneiderman proposed five measurable characteristics: speed of performance, rate of user errors, subjective satisfaction, time to learn, and retention of skills over time. Their strength is precision, since “faster” is easier to work on than “more convenient”. Golovach lists the weaknesses. The measures are incomplete (operator fatigue isn’t covered) and they conflict (speed against learning). In practice you can usually improve only two at once, and the concept doesn’t say which two. Learning time and retention are also hard to measure and often irrelevant for daily-use systems.

Three design philosophies#

• User-centered design (UCD): study the audience and optimize for it; the developer’s opinion doesn’t count. Golovach’s objections: where users are paid and easily replaced, as at a cashier’s terminal, productivity outweighs their opinion. People also adapt, so complaints about a new version fade after a few weeks. And the research is expensive and open-ended.
• Task-centered design (TCD): an interface is good if users’ tasks get done efficiently. This allows economic estimates and keeps projects manageable, but it has no stopping rule. More tasks always looks better, which feeds feature creep.
• Goal-centered design (GCD), which Golovach prefers to call motive-centered: people perform tasks to satisfy personal motives. His example is a secretary entering a meeting dictated over the phone. The task view yields “support creating calendar events”. The motive view (not getting blamed for a wrong entry) reveals that accuracy matters more than speed. A practical tip: list the audience’s motives and compare them with the task list. Mismatches point to gaps in your task analysis.

Usability (ISO 9241-11)#

The standard combines the earlier ideas (specified users, goals, and measures of effectiveness, efficiency and satisfaction) and adds context of use: equipment, noise, interruptions, monitoring of staff, required speed. Golovach finds the standard’s terms vague. He suggests replacing them with Shneiderman’s measures and replacing “context of use” with the more concrete “environment”.

Commercial success#

Golovach considers this the most practical concept: a good interface helps the product sell. Measures are not equal here, and satisfaction matters more commercially than error count. Related points:

• Users call “simple” what is familiar, and “convenient” what needs no unproductive actions. So look for extra steps and remove them.
• Improvements should be explainable in a few words and should look good in screenshots.
• When buyer and user differ, choosing whom to serve is a real dilemma. The common mistake is to forget one side entirely.
• The designer is paid by the client and should work for the client’s success. An unsuccessful product doesn’t help users either.

He adds one more test, deliberately provocative in wording: is the interface “sexy”, meaning something people want to have?