When "Better" is Worse

From Wired for Speech: How Voice Activates and Advances the Human-Computer Relationship by Clifford Nass and Scott Brave:

p. 155: The word of technology is inspired by the desire to do “better and better”. There has long been a tendency to define better in terms of a technology’s distance between its objective performance and the objective characteristics of the physical and social world that it models. The definition has led to the development of a set of principles that urge marketers to deploy the most advanced technology that are available at a given time. This strategy, which is blind to the physiology of users, has led to the creation of less usable systems in the name of “improvement”. [...] Although interface builders should certainly cheer for each technological advance, mindful manifestation demands that, to paraphrase Alexander Pope, the proper judges of interfaces be users.

Nass and Brave provide warnings on the design principles that consider the “best” interfaces as the ones that are as “accurate” and “realistic” as possible. Indeed, frequently accuracy burdens people. For instance, the physical world is often less convenient to navigate than is its abstracted counterpart. They quote Henri Matisse who taught use that “Exactitude is not truth“: Sometimes a less accurate mirroring of the world can in fact be more effective.

In addition, the risk for using the most advanced technology is that the contrast highlights failure:

Consider users who encounter an interface in which one dimension is clearly superior to the other. Initially, they may simply feel uncomfortable at the incongruity. In a (potentially unconscious) search to resolve the discomfort, they realize that these two technological creations are very different in that one is high in quality while the other is low in quality. Having a clear category for contrast (quality), the brain accentuates the difference and labels on very low quality and one very high quality.
If human psychology were simple, this accentuation should not be important: very hight quality and very low quality should balance each other out in the same way that high and low would. Unfortunately, two cognitive biases prevent this from happening. First, negative experiences are more arousing, memorable, and noticeable thant the best part, leading to more negative overall judgments. Second, poor quality in a interface generally leads to unpredictability, which, especially in a task-oriented situation, is very worrisome. These biases lead to overly harsh judgments of interfaces of mixed quality.

Relation to my thesis: These thoughts (based on Nass and Brave’s experiments) go in the direction of my current work with the analysis of accuracy used to geo-reference photos in Flickr. One goal is to provide arguments for location-aware systems designers to reject the obsession with veridicality in all aspects of an interface because this goal is impossible in the near term and lead to high levels of user distress and failure.