My references look like a haiku
Intuition, Interaction, Metaphors and Musings: An Analysis of the Korean Air Website
Systems Analysis Part II
Michael Sean Gallagher
October 24, 2005
This paper serves a continuation of a thorough systems analysis of the Korean Air website begun in the first installment, which detailed the effects of memory and practice on successful systems navigation and the incorporation of those principles in design methodology, specifically as they applied to the Korean Air website. This installment will address the concepts of intuition, interaction styles and the use of metaphors as design tools. Concepts addressed in this analysis will be derived from “The Anti-Mac Interface” (Gentner, Nielson, 1996), “Interaction styles and input/output devices” (Jacob et al, 1993), “Intuitive vs. Familiar” (Tognazzini, 2001) and “Ten Usability Heuristics” (Nielson, 1994). Intuition, interaction styles and the use of metaphors are fairly ubiquitous in the field of design, yet certain pitfalls exist when relied on excessively. Design has recognized the limitations of exclusive, often inconsistent, metaphor use as a means of establishing intuition through a thorough investigation of the interaction styles of users. In certain respects, the Korean Air website abides by these industry standards, yet other aspects fail to convince the user of its usability.
“The Anti-Mac Interface”, in particular, serves as a primer on the development, use, and overuse of the metaphor as a learning tool in systems design. The original Macintosh, along with its subsequent design methodology, revolved around a few principles, including the notion that it “needed to sell to naïve users, that is users without any previous experience” (Gentner, Nielson, 1996, 71). Therefore, design needed a means by which this inexperienced user could access relatively complicated platforms. A metaphor is a mere extension of that desire to incorporate the learning needs of the inexperienced user. Metaphors are rife throughout the realm of systems design, especially in the field of web design. The website for Korean Air is no exception. For example, on the homepage for Korean Air (http://www.koreanair.com/
), the metaphor for arrivals and departures is an airplane, fittingly, while the metaphor for mileage inquiries is a bag of money with a dollar sign on the front, somewhat perplexingly. This mileage inquiry metaphor suffers from the target domain having features not located in the source domain, an inverse of a situation outlined in “The Anti-Mac Interface” (73). The actual function of the mileage inquiry has one aspect related to the money as illustrated by the metaphor; registered members can exchange mileage for cash conditional to certain limitations. However, this is an exception to the concept of the mileage inquiry, rather than the norm. The primary function of the mileage inquiry is to allow the user to review their accumulated mileage and to determine what rewards they have qualified for, most of which are not cash related.
On a related note of interest to the Korean Air web designers, the use of metaphors as visual guides and learning tools are inherently limiting because they involve the matching of real world symbols for elements on the interface. This is an especially daunting task considering that often interface functions are designed precisely because they do not exist in the real world; they are naturally innovative. The services offered by Korean Air via their mileage inquiry are not faithfully represented in the mileage inquiry (cashbag) metaphor, nor can they logically be expected to be. In essence, the metaphor has been thrust on the user and, similar to the desktop metaphor mentioned in “The Anti-Mac Interface”, its limitations on the interface do not serve the customers’ needs (73). In fact, the cashbag metaphor does not serve the needs of Korean Air because it distorts the nature of the services provided. Please see Screenshot 1-A for a visual representation.
Although the mileage inquiry metaphor fails to elicit an instantly recognizable real world equivalent, the Korean Air website does abide by a few of the design principles as illustrated in “Intuitive vs. Familiar”. For example, the mantra of building a design on a few, simple driving concepts is manifested in the Korean Air website (Tognazzini, 1998). Along with the mileage inquiry there are few other metaphors the user would be inclined to grapple with. The most notable, aside from the mileage inquiry, would be the airplane metaphor, which unfortunately is inconsistently labeled depending on the screen from which it is accessed. For example, the homepage labels the airplane image as Arrivals and Departures, while many of the subsequent screens label it as Flight Schedules. This inconsistency violates one of the ten usability heuristics, which states specifically that “users should not have to wonder whether different words, situations, or actions mean the same thing” (Nielson, 1994). This logic can be extended to user having to wonder whether different metaphors or symbolic representations have the same meaning, as is the case with the inconsistently titled Arrivals and Departures or Flight Schedules, depending. Please see Screenshot 1-B for further illustration.
Inconsistency naturally destroys a good deal of the intuition necessary to navigate such a website. Intuition, albeit inherent, can be stimulated by accessible, logical design, which adheres to a preconceived mental model. Inconsistency leads the user to not trust their intuition, forcing them to look for more recall oriented tools to navigate a website. Furthermore, the lack of consistency does not promote a learning grid for the user to follow. If one went from Point A to Point B and the labels were inconsistent, one would not mentally correlate Part A with Part B, storing the as fractured entities within memory. A thorough round of testing should have detected this inconsistency. This leads to a related point that testing allows the prototype to be evaluated according to the extent in which the user is “getting it” (Tognazzini, 1998). Prototype testing should be able to recognize the existence of inconsistent labeling and its subsequent effect on efficient user navigation. In a nutshell, the design team should have noticed the inconsistent labeling on the Korean Air website.
Aside from the use of metaphoric visual language to allow for greater accessibility on the part of the user, there are other design considerations that attempt to reduce the cognitive learning necessary to perform a systems task. Many of these design considerations are outlined or alluded to in “Interaction styles and input/output devices” in theoretical form. By theoretical form, the point is implied that many of these design implications are represented as areas demanding further research. However, the practicality of many of the design considerations is worth mentioning, if only to illustrate how common they have become in modern design. For example, as illustrated by the Korean Air website, menu options serve to access a user’s sense of recognition rather than recall (Jacob et al, 72, 1993). Drop down menus have less flexibility than command language functions, but they prove to be less mentally taxing and more appropriate to the casual interaction a user often experiences with a commercial website. Although these web interfaces have yet to convert to voice recognition input devices as mentioned in the article, the logic of menu driven interfaces is sound in its understanding of human-computer interaction.
The Korean Air website employs a pseudo-natural language that is neither technical jargon nor completely natural interface language. This proves especially effective in offering accessibility to the system by avoiding jargon. The user is afforded the opportunity to interact with the system based on familiar terms, rather than technical terms relevant only to the airline industry.
Despite this accessible language, the Korean Air website does not offer natural language processing in any of the menu interfaces. This proves especially effective in this regard, because of the precision necessary for the desired interaction. For example, a passenger wishing to travel from Seoul to San Francisco would not always have the recall necessary to express which airports are represented in these two cities. A menu option, by limiting possible choices, serves to keep the transaction free of errors. In essence, the lack of choice keeps the transaction valid. As mentioned in the article, “the result is a computer interface consisting of a set of objects and operations that closely resemble their counterparts in the physical world” (Jacob et al, 72, 1993). The familiarity of the Korean Air website with a real world model greatly contributes to its effectiveness. Please see Screenshot 1-C for an example.
Although the menu functions do enable users to navigate especially sensitive information entries with minimal strain and lessened error rates, the inconsistent labeling as well as the often misrepresented metaphor serve to limit the effectiveness of the Korean Air website. Further investigation should be conducted in this regard to replace these wieldy and mislabeled representations with something which serves to stimulate intuition by firmly matching a real world model.
Gentner, D. & Nielson, J. (1996). The Anti-Mac Interface. Communications
of the ACM, 39(8), 70-82. Retrieved October 15, 2005, from http:// drexel.blackboard.com/courses/1/INFO608-900-200515/ content/_581003_1/antimac.pdf
Jacob, R.J., Leggett, J.J., Myers, B.A. & Pausch, R. (1993). Interaction styles
and input/output devices. Behavior & Information Technology, 12(2),
69-79. Retrieved October 15, 2005, from http://drexel.blackboard.com/
Nielson, J. (1994). Ten Usability Heuristics. Retrieved October 15, 2005, fromhttp://drexel.blackboard.com/webapps/portal/
Tognazzini, B. (1998). Intuitive vs. Familiar. Retrieved October 15, 2005, fromhttp://drexel.blackboard.com/webapps/portal/