También disponible en Español


The digital magazine of InfoVis.net

Visual Metaphors
by Juan C. Dürsteler [message nº 91]

The use of Visual Metaphors is one of the most common ways of elaborating interfaces and visual representations. However not all the developers think the same way.

Recently I gave a seminar on Information Visualisation in Turin and Barcelona. The objective was to present some of the more interesting visualisations in order to debate with the participants about their interest and utility.

During the Barcelona session an interesting discussion arose about the convenience that visual representations are guided by the so called Visual Metaphor. 

The visual metaphor can be defined as the representation of a new system by means of visual attributes corresponding to a different system, familiar to the user, that behaves in a similar way. An archetypical case is the desktop metaphor. In it the traditional hierarchical tree of directories and subdirectories is substituted by the graphical interface of folders and files.

It’s worth noting that metaphors (be them visual or not) have become common in this technological world in which we live. For example, the directory structure mentioned before is, mathematically speaking, an acyclic connected graph whose properties make it especially suitable to store and sort information in an easily retrievable way. IT professionals call it a tree because its representation looks a lot like a tree and probably because if they call them acyclic connected graphs, nobody would understand them. 

Connected Acyclic Graph. This mathematical construction represents conveniently hierarchies and data structures that in information technology are called "trees". Obviously only the connectivity properties are shared with an actual tree. The abstraction, however, simplifies the process of assimilation.  Drawing of a tree. The hierarchy of branches, stems and leafs can be described by a Connected Acyclic Graph.

This way, non trivial mathematical properties are substituted by a closer to experience metaphor. That of the tree is a common experience that allows the layperson to anticipate many of the properties that belong to this data structure. Everybody can distinguish the stem branches and leafs. It is easy to understand, for example, that by isolating one branch of the tree you isolate all that depends on it. The same occurs with many other properties that we know about trees, which allows us to assimilate, understand and foresee many of the properties of data structures that expressed mathematically would seem alien to us.

One of the participants was worried about the importance that visual metaphors and 3D representation have, preferring clearly the use of conventions instead of visual metaphors, somewhat in Alan Cooper’s line.

Cooper in his 1994 article titled “The Myth_of_Metaphor” , considers that the search for a good visual metaphor is essentially counterproductive, even though many of the software developers were using it then (and 7 years after the article, things haven’t changed so much).

Cooper differentiates 3 paradigms used for building user interfaces: 

  • Technology Paradigm. The interface is expressed in terms of the technology used to build it showing the inner workings of the software. The user has to understand how it works in order to interact with it. It’s not the software that adapts to the user but the user that adapts to software.

  • Metaphor paradigm. The interface is expressed through a metaphor that hides the inner workings of the program and refers to terms more or less familiar to the user. For example: the above mentioned desktop metaphor or the “organic design” of Ben Fry’s anemone (see the final part of the article number 67 "Monitorising the web" ) 

  • Idiomatic paradigm: According to Cooper this method solves the problems of the preceding ones since it’s based on the way that we humans learn to perform operations, by using instructions expressed by means of a language. For example the mouse isn’t the metaphor of anything but it’s unbelievably easy to use in an effective way. Scroll bars are another example of the same, they don’t represent anything in particular but can be easily apprehended.

This interesting rationale deserves a debate which we don’t have space for here. The fact is that many current programs have interfaces that combine, to a greater or lesser extent, the three paradigms, probably because each of them solves different problems in different situations. 

Reality doesn’t appear to be approachable in a unique and uniform way in all situations.

Thanks to Marc Noguera from multiplica.com for the link to Alan Cooper’s article and for raising the debate.

Links of this issue:

© Copyright InfoVis.net 2000-2014