Consistency, standards, and formal approaches to interface development and evaluation: a note on Wiecha, Bennett, Boies, Gould, and Greene

Wiecha et al. [5] describe ITS, a system for interface construction. Application developers use ITS to compose style rules that are applied during compilation. These rules result in a certain interface conformity, described as “consistency,” and are said to reduce the development effort considerably. I have no reason to dispute the authors’ central claim, that their interfaces are useful, usable, and easy to develop. This is a response to their secondary claim, that their interfaces owe their virtue to consistency, which is incorrect. In [2], I illustrated a range of problems with the concept of interface consistency. In some circumstances it is difficult to define consistency or to determine the appropriate features or dimensions on which to be consistent. Consistency is a trade-off against other goals: at times, it is not the best design strategy. Random or careless design is to be avoided, but the best overall strategy is generally to use all available means to obtain the best possible understanding of the prospective users ancl their work situations. Consistency and other design rules are best seen as guidelines that may have to be violated for the benefit of users. ITS generates interfaces from design rules. Wiechla et al. argue for establishing a set of design rules to cover all allowable interaction techniques and then prohibiting exceptions (p. 233). They claim that resulting interfaces will be both good and consistent or perceived to be consistent. The generality of this claim must be evaluated carefully, with an eye to possible limitations. I