Suggested by Xinju Wang:

(1) Always use at least discount usability.

• Explanation: From this paper, they got much better result by using discount usability engineering technique than using no usability engineering at all. The discount usability engineering does not cost a fortune to get good result. These discoun t techniques are designed to use quickly and cheaply.
• Example: Use scenarios (reduce both the level of functionality and the number of features, simple for test and change), simplified thinking aloud (small test group, take note instead of videotape subjects), and simplified guidelines (use only 9 basic guidelines).
• Exception: The discount usability technique cannot find all the problems. Some discount technique, like the simplified guidelines, can be hard to use for nonexperts.

• Explanation: When developing critical software or big project, the budget is not a big problem. The traditional methods can find more problems. Even though traditional methods can be costly, we can get benefit in the long run.
• Example: In a big project, assign some user interface specialists to use all the usability engineering techniques.
• Exception: Only apply for critical software or big project. Too expensive and time consuming for small companies or projects.

Suggested by Michael Chisholm:

(3) Doing simple heuristic (in the sense of this paper) evaluation with a few important guidelines will help find many of the biggest problems with a user interface.

• Explanation: A heuristic evaluation in this paper refers to looking over and evaluating a UI with respect to a set of guidelines. A few important guidelines will be easy for a developer with little experience in UI evaluation to apply - there may not be a need to hire an expensive UI expert to do an extensive evaluation of the interface. You can catch many of the biggest flaws in an interface by using a few simple rules (the paper suggests 9). It is also a good idea to have several developers do the heuristic evaluation, as different people might discover different problems.
• Example: The "Provide clearly marked exits" guideline is an easy one for a developer to apply. He may just sit down and make his way through the interface and look: for each dialogue box or screen, is there an easy way to back up? An easy way to cancel your previous choice? Or to start over again? To quit (if applicable)?
• Exceptions: It has been experimentally shown that simple heuristic evaluations can uncover many problems with a UI; nevertheless, a UI expert will almost certainly uncover many more problems than an inexperienced developer would. And this simplified technique shouldn't be relied upon with larger projects. A large important project should invest in a full-blown usability study.