Metaphors: Realistic vs. Form-based Phone Interfaces

Introduction

    Due to recent advances in computer technology, it has been possible to provide users with interfaces that closely resemble real-world objects and actions.  Realistic metaphors are commonly used in applications where the feature set of a program mimics a "real-world" object. Many applications serving as video and media players, calendars, address books, and calculators have all been designed to look and feel like their real world counterparts.  The chief argument for using realistic metaphors is that they are easier to use due to the fact that, even if the application is unfamiliar to users, they will recognize how to perform various functions because they have had experience with the real-world object.   For example, when using a media player, most users immediately recognize that is the "Play" button and know to use it if they want to play a sound or video clip.

    One particular type of program that is often represented by realistic metaphors is a telephone-dialing program.  IBM's RealPhone is an example of this type of interface.  RealPhone is a telephone-dialing program that is modeled after a standard business telephone.  It includes most standard telephone features such as a numeric keypad, handset, flash button, and speed dial.  IBM claims that because "traditional computer-based user interface mechanisms are removed and replaced by more natural and intuitive ones," RealPhone allows "users to focus more on their tasks and less on the computer interface."

    Various studies have been conducted which support the idea that real-world metaphors are more easily interpreted by users.  One study conducted by Wendy Ark, Christopher Dryer, Ted Selker, and Shumin Zhai compared task times using 2D (iconic) vs. 3D (realistic) graphics.  Their research indicated that "subjects found objects more quickly when they had 3D realistic representations than they did when they had 2D iconic representations."  They concluded that the use of a real-world metaphor enhanced experimental task performance.  This study, however, failed to note that the measured gains in performance may not generalize to all real-world metaphors.

    The object-action interface (OAI) model provides a framework to determine why metaphors may sometimes fail.  The OAI model describes how metaphors are used to map task objects and actions into interface objects and actions (Shneiderman 61).  In other words, the metaphor is the link between the task and the interface.   The efficiency of the interface depends on the accuracy of the metaphor.   Sometimes, the metaphor is imperfect, and the interface will reflect this.  In RealPhone, for example, the receiver is an imperfect metaphor.  In the real world, it is an object that must be picked up and held to one's ear.  When it is mapped to the computer, however, it becomes an object that must be clicked on to be activated.   The confusion this generates might cancel any gains achieved by using a realistic metaphor.

    Another problem with realistic metaphors is that they sacrifice consistency with industry standards to achieve a close relation to a real-world task.   For example, RealPhone exists as an autonomous entity on the desktop--it is not contained in a form.  Because of this, it has none of the standard Windows conventions, such as a title bar, window control buttons, and a menu/tool bar.  IBM suggests that this enhances user productivity by allowing them to focus on the task.   However, users already familiar with industry standards might have a hard time accommodating an unconventional interface.

    Therefore, a metaphor may occasionally be a hindrance and need to be bent or broken to get better results.  In another study, the authors recognized the limitations of metaphors.  They "knew, for example, that double-clicking a folder icon on the Macintosh opens a window bearing no resemblance, visually or functionally, to a folder" (Moll-Carillo et al. 572).  Users were not bothered by the fact that this metaphor did not accurately represent a real-world object.   This suggests that straying from the metaphor to some degree is actually a good idea.

    To test this idea, we decided to study whether the RealPhone interface (a metaphor tied very closely to its real-world counterpart) performed worse than an industry standard, form-based interface.  We also decided to test whether or not the RealPhone interface was preferred subjectively by users.  If these suppositions provide statistically significant results, then the phone metaphor entails a tradeoff between performance and subjective satisfaction.  The poor performance would indicate that RealPhone is based on a faulty assumption--that users will easily be able to transfer their experience with real phones to the computer interface.

Department of Computer Sciences Direct questions and comments to the student editorial team