This report is also available as an Acrobat file.

Contents
The Design of Virtual Environments with particular reference to VRML

The uses of spatiality

The active eye and the layout of information

It is clear that perception is a process which involves action on the part of the observer in relation to the observed. Even something as apparently trivial as decorative pattern engages the observer s mind in an active way. Gombrich (1979) argues that the faculties of expectation and extrapolation are essential to an appreciation of pattern, so that much of the pleasure of pleasing patterns lies in their touching the interplay between utter predictability on the one hand and chaotic unpredictability on the other. It is difficult to see how this could work unless, as Gombrich suggests, the mind is always predicting, and directing the eye to seek what it expects to find. The decorative artefact balances between redundancy, where no new information is introduced to the scanning eye and which the mind soon tires of attending to, and complete variety, which seems to the mind equally unrewarding of attention.

Gombrich is dealing with decorative surfaces, and uses the language of information theory to do so. However our need is often to deal with the conveying of information in the more limited everyday sense the transfer of facts from author to observer. The structuring of information on a two-dimensional surface conforms to such obvious rules that we tend to take it for granted. Nevertheless such structuring can be regarded as providing useful meta-information which enhances the understanding of the content-information. So many design procedures make use of the spatial organisation of information in order to enhance understanding that only one or two examples need be cited.

Miles (1987) succinctly puts the case for spatial organisation in graphic design: Typography and layout, thoughtfully used, can provide signposts which reinforce the significance of each passage and help the readers to find their way about in a document. A gamut of conventional techniques used in laying out a page relies on our sure interpretation of juxtaposition, separation, alignment and non-alignment, relative sizes and so forth. It includes indents, paragraph spacing, columns, tabular layouts, and of course white space . The findings of the gestalt psychologists described above seem to suggest that this works at a basic precognitive level, though no doubt strongly reinforced by the learning of conventions.

Tufte devotes a whole chapter of Envisioning Information (1990) to another spatial artefact: layering and separation. Most of the book is devoted to the representation of multivariate information on a two-dimensional surface. In discussing layering, none of his examples depicts a third dimension in an obvious pictorial way. However, many of his examples allude to the missing dimension by means of tonal and colour differences. For example, in discussing the relation between signal and background marks providing substantive data and those providing the framework in which it can be understood he draws attention to the importance of suppressing the tone and saturation of the background. Perhaps our tendency to give attention to the marks of highest contrast as against those of lowest derives from our experience of the landscape, where tones and hues diminish with distance and we must attend for reasons of survival to those things which are nearest.

Interface design guidelines depend on the use of position on a surface for much of their effect. Examples (taken from Brown 1985) include:

Use invariant fields on each screen: If an item of information has a similar function on similar screens, do not move it around. This advice denotes a use of space informed by restraint rather than virtuosity. Effort may be required by the designer to ensure that this consistency is achieved.
Present lists in useful orders: Simply to lay out information in a linear array, so that items are sequenced in a way which makes sense in relation to their individual meanings, uses space importantly, if so simply that we often take it for granted.
Arrange information in logical groups: Clustering (and aligning) objects so that those which belong together are seen together enhances the meaning of the components. Such spatial organisation has the status of a grammar, like that which gives words their significance in a sentence. An example familiar to all computer uses is the GUI convention of clustering radio-buttons : those within a set are mutually exclusive while there is no exclusivity between sets. The user can only make use of this knowledge if it is clear which set each button belongs to.
Minimise pointer movement: Many early hypermedia products imitated the conventions of turning the pages of a book, by placing the next page and previous page triggers respectively at the right and left of the screen. This was a blunder significant for any discussion of Virtual Reality, since it was based on an ill-considered transfer of some attributes of real paper documents to a virtual environment. It involved users in wasteful pointer movement whenever they wanted to flip pages back and forth. Soon afterwards designs began to appear which still borrowed the idea of the right button denoting next and the left button previous but which brought them together in a tight cluster of controls. Taking liberties with reality had proved the better solution than following it slavishly.

The significance of these guidelines is that they are about constructing meaning or at the least, amplifying meaning using space. Of course, information structures of this kind can be enhanced by the availability of an additional, third, dimension. Even so, it is worth remembering that when using three dimensions in this way, three will often not be enough , just as two was not (Information spaces: from database to the Virtual Museum page 69).

Contents

Graphics Multimedia Virtual Environments Visualisation Contents