Sunday, April 16, 2006

Usability requirements

Different usability attributes are usually weighted differently. With games, satisfaction is probably the most important criterion. However, measuring usability requirements can pose problems. Some attributes are somewhat subjective, such as satisfaction, but their target levels can be set and evaluated nonetheless. The table below lists possible metrics for defining and evaluating the usability standard.

Satisfaction - Anonymous questionnaire
Efficiency - Time to perform certain task(s)
Learnability -Time to reach certain stage(s)
Errors - Number of errors, while using a system, in certain task(s)
Memorability - Number of correct answers (a memory test)

How to set the requirements
Different levels of usability can be specified for each attribute evaluated. For example, let’s assume satisfaction is measured on a scale from 1 to 5. Optimal performance is 5, but the target level may be set at 4, which is still difficult to achieve. The minimum acceptable level could be set at 3 and the current level assigned a value of 2.5.

Other attributes, such as the number of errors, are more straightforward to evaluate. Identifying them may, however, require experience and expertise. A fairly simple method is to evaluate an existing user interface with real users and use that as a basis for new requirements.
When no competing products exist to use as a benchmark, usability goals are very hard to establish. However, projects rarely start from scratch without any kind of prior knowledge or experience. The minimum level of usability should be set to what it was on a previous project.

When to set the requirements
Usability requirements should be defined when the other requirements of the mobile game have been established. At that point, it is not usually possible to define the exact numerical values for the evaluation, especially since the questionnaires have not been drafted and the final mobile game user experience of the game has not been implemented. However, the requirements can be established in a less detailed way and the weight of each usability attribute can be assessed. Typically the usability requirements are defined after the technical requirements have been set.

If the usability requirements are defined too late in the process, it may be too expensive to implement them, even if it turns out that the requirements are not being met Nokia N110. Therefore, the requirements must be laid out before programming begins, and usability should be evaluated against the requirements at different stages of the development cycle.