Icons at the interface: their usefulness by Yvonne Rogers (1989) is one of the earliest studies on the taxonomy of interface icons. Through her research, Rogers hopes to discover primitive syntax and semantics for icons, which could be used as a basis for designing a set of icons for a given application. This is a summary of some of the main points of the paper.
Icons should reduce the complexity of a system and make it easier to learn. The user is given the impression that the system is easy to use, and this can have a positive effect on the user. The icons can also be universally meaningful without being restricted to language. However, icons can fall short of this goal. Many icons have been used to represent very ambiguous system information. If the user has to look up the meaning of an icon in a manual, then this type of communication is inadequate. There is always a risk that the icons become arbitrary symbols.
A common problem with iconic interfacing is the different meanings that can be attributed to a single icon. Verbal languages have syntactic and semantic rules which help to disambiguate their meaning. Pictorial languages don’t yet have equivalent rule sets. Paradoxically while pictorial communication could potentially be universally understood, it lacks the rules to guide this process.
Context is key
Context is a key factor in disambiguating the meanings of icons. Rogers presents Hungarian toilet signs as an example.There a stiletto shoe is used to represent women and a typical man’s shoe represents men. Because of the user’s prior knowledge of where to find a toilet in a restaurant, and the fact that there are generally separate toilets for men and women, this kind of icons are immediately recognisable.
Experienced computer users some times prefer command line tools over graphic interfaces. Having learned to use command based interfaces, they see no reason for the use of icon-based interaction, which in many cases can also be slower. For new users it is much easier to recognise and select icons, than it is to learn and remember all the commands and function keys necessary for operating the software. Rogers still points out, that there is potential for using interface icons for all types of users. They are not only valuable to new users.
Rogers recognizes four ways in which interface icons can refer to their referent.
- Resemblance icons – depict the referent through an analogous image.
- Exemplar icons – serve as a typical example for a general class of objects. One stands for many, just like the knife and fork represent a restaurant.
- Symbolic icons – the relation between the referent and referrer is more abstract or metaphoric. For example an image of a wine glass with a fracture conveys the concept of fragility.
- Arbitrary – Bear no resemblance to the referent and the association has to be learned.
The effectiveness of an icon also depends on to what extent the visual appearance of an icon infers the underlying referent. This effectiveness can also be called articulatory distance. Icons that are easiest to understand are considered to have more direct mapping, while the ones that are more difficult to understand have a less direct mapping.
When a function or a system object is so abstract, that there is no way of designing an icon with a logical or isomorphic connection to it, a common solution is to select a completely arbitrary and abstract design. The idea is, that such forms have no prior association connected to them. Once the meaning of the icon has been learned, there is no risk of it being confused with another function or object. This approach can be effective when the user needs to learn one or two icons, but it is less useful when a user has to learn a whole set of icons.
In her user studies, Rogers also discovered that icons based on concrete analogies can also perform very badly, if the link between the referent and the referrer is tenuous and not made explicit, such as using an image of a wheelbarrow with bricks in it to represent the move operation in a word processing environment.
An important aspect of interface icons is, that they almost always have to function as a set, while other real-world pictorial signs can more often function as individual items. The challenge of a functional set is, that not only do the icons need to be able to represent individual referents, they also need to discriminate between various connected referents. A system of combining common and distinct graphical elements can be effective in such cases. This system is a kind of a simple grammar for the interface design.
Rogers also point out the multidimensional nature of pictorial communication as an advantage compared to verbal language. For instance variations of form, shape and size can be used effectively to communicate different parts of command sets.
In a study comparing icon-based and name-based interfaces for a word processing environment, Rogers discovered that users remembered the meaning of icons better than names, over time. This might be due to the fact that pictorial signs are more likely to be stored in both visual and verbal memory, according to Allan Paivio’s (1971) dual coding theory.
In her summary Rogers recognizes the usefulness of icon-based interfaces. To overcome the challenges of choosing when to use icons and how to design them, Rogers calls for the development of a taxonomy and grammar for icons. Following such a systematic approach can be especially beneficial for designing complete icon sets.
- Rogers, Y. (1989). Icons at interface: their usefulness. Interacting with computers.1 (1), 105117.
- Paivio, A (1971). Imagery and verbal processes. New York: Holt, Rinehart, and Winston.