Design Considerations for Spatial Metaphors
reflections on the evolution of viewpoint transportation systems
- / -
Position Paper by Heiner Benking and Anthony Judge fro acm-ECHT Conference, Edinburgh 18-23 September 1994
Invited Workshop: Spatial Metaphors for Information Systems
This note assumes that the future evolution of the use of spatial metaphors in
interface design will not be focused on any single metaphor. For a variety of reasons
there is a need for environments permitting user-choice, or AI-selection, of metaphors
through which information is communicated in either direction. The following remarks
highlight some of the issues relating to this evolution.
1. Classes of metaphors
There is a need to distinguish classes of metaphors offering different advantages and
disadvantages. Typically they would include: geometric forms (cube, sphere, polyhedra in
general), artificial forms (townscapes, house, room), natural forms (landscape, trees,
etc.), systemic structures (highway systems, pathways, flow systems), dynamic systems
(atomic, molecular, planetary, galactic systems), traditional symbol systems (mandalas,
sand paint ings, etc.).
Of special interest are those sets of metaphors which permit inter-transformation with
minimal loss of conceptual integrity (in terms of maintaining relationships between data
referents). This function is important with respect to user learning, user preferences for
information organisation and cultural biases and preferences in information organisation.
It is however equally important as a way of handling different levels of complexity.
With information beyond a certain degree of complexity it is questionable whether any
single metaphor is adequate as an interface for adequate comprehension. This is best
exemplified by the wave/particle metaphors used to comprehend fundamental physical
systems. There is therefore the interesting possibility, that some complex systems may
only be adequately comprehended through three or more complementary metaphors. This
consideration also relates to psychological type-theory by which the number of types
distinguished may correspond to the number of metaphors appropriate for comprehension (The
astrological types are perhaps an extreme example).
4. Relevance of morphing techniques
It is interesting that the debate on metaphors has proceeded independently from the
development of morphing. The reason may lie in the emphasis on the necessarily visual
(superficial) features of the images which are morphed. These have not yet been considered
as ways of carrying information as is the case with the metaphors considered earlier.
There is no reason why users should not be free to distort any initial metaphor into some
preferred form. Morphing can also be used by the user to reconfigure information in terms
of complementary metaphors.
5. Alternation and perspective
There are learnings from the use of two eyes to obtain a sense of perspective. It is
equally desirable to understand the use of metaphors in terms of the alternation between
perspectives which provide a sense of depth that would otherwise be unavailable. Such
'depth' is distinct from that obtainable from any 3-D metaphor which although it
offers depth is cognitively not as significant as that offered from the cognitive
integration of two contrasting metaphors. Such 'depth is only achievable by
alternation between metaphorical interfaces (as the wave/particle example suggests).
6. Metaphoric library
There is merit in considering the possibility of a library of metaphors analogous to
the 'libraries' available in software packages. Such a library might contain
many of the metaphors noted above. A stage may be envisaged in which metaphor libraries
might become equivalent to figures as in drawing packages. These could then be seen as
archetype libraries to trigger ideas in brain-(quickstorming) processes.
7. Pattern fitting
Of great interest is the possibility of using algorithms to fit data to spatial
metaphors. Through this function the user would be presented with a selected sub-set of
metaphors to which the information can be usefully fitted.In its simplest form this would
be of value in interrelating key concepts in the outcome of conferences. Typically the
intellectual product of such events (e.g. UN Conference on Environment and Development,
Rio de Janeiro, 1992) takes the form of complex declarations and programmes (e.g. Agenda
21). The challenge is to configure the conceptual elements into a globally comprehensible
form. This is necessary to counteract the tendency to generate an asymmetric agglomeration
of elements (of which Agenda 21 is a typical example). In addition to its mnemonic
function, a metaphoric framework can then highlight possible missing elements as well as
suggesting ways of understanding inter esting functional relationships between such
elements. In this sense a metaphoric framework can be used to 'challenge' any
conventional legalistic structuring of academic or policy conferences. Such feedback can
suggest the use of alternative metaphors.
8. Sterile metaphors
It is useful to treat the use of metaphors 'self-reflexively'. This can best
be illustrated by the many proposed architectural and landscape metaphors. Some of them
appear to suffer from inadequacies analogous to those noted by critics of sterile
architecture and city-scapes (sterile or criminal downtown areas) - to which
environmentalists have proposed many alternatives.
9. Knowledge spaces and navigation
An optimal metaphor for storing and exploring data and knowledge are scapes (deep
structured spaces) as found in Nature. We can define knowledge topologies, even a
panorama, ranging from particles, areas, bodies, scapes to a Gestalt. We gain by expanding
the perspective by agreeing on views in given frames (superstructures) by the mind's eye.
In such a virtual space, previously only available in an aperspective mode, we can make
sense by locating and interpreting virtual entities and objects. To immerse or
'dive' into such a topology of information, maybe seen as a multi-story
information warehouse, needs making use of teh third dimension and an internal deep order.
We need to navigate in and between bodies of expertise (domains) and not restrict
ourselves to the surface, avoiding depth and hidden or invisible structures on other
levels. Learning to explore the virtual space like children explore the physical space
makes fun and is the real challenge. The challenge to make use of cyberspace to escape,
but to make sense by exploring new domains.
The metaphor should than take the form of Nature Space, which has features of infolded
(or fractal) ordering to carry other levels of information. Much thought has been given
tin the 1960s to such a Conceptual Superstructure concept by Douglas Engelbart (
Augmenting the human intellect: A conceptual framework).
10. Piggy-backing on drawing packages
Aspects of what is advocated above can be envisaged as an adaptation of certain kinds
of drawing packages. The requirement is that it be possible to attach user information to
any feature of a drawn object. This function is to be distinguished from conventional
labelling. Rather each feature effectively allows the user to open a box into which
information may be put. The box can than be hidden until clicked open. Any object drawn or
retrieved from a library are then effectively, clusters of information, boxes or drawers
which may or may not be filled. If the package is used to select a metaphor object to
cluster and hold a set and subsets of user information elements (statements and
relationships between them, images, etc.), the features might be differently coloured if
filled in this way. Equally if information was independently acquired and particular
categories (e.g. accumulation of transactions entered into the information databases by
other external parties) the colour might be used to signal the quantity of information
attached to any feature of the metaphoric view on that database.
Within such a context conventional zooming could be used to expose finer levels of
structure that could be used in the same way as hangars for more detailed clusters of
information. The emphasis here is that a drawing package (or some drawing packages) could
probably made 'intelligent' with relatively little effort. One approach might
simply be to attach a record ID field to every such drawn feature. The ID would then point
to user information in a related datafile. Refinements might enable a number of different
figures from the library to have the same set of record IDs plugged into them. The user
then has access to different visual configurations of teh same information. Things become
features, possibly to the size of a palace. Alternatively a user could select a
'palace' from the library and furnish/decorate it with information (as pointers
to an appropriate file). The palace could be 'remodelled' at any time, gardens,
and animated features (animals, servants, etc.) could possibly be added. Such an approach,
whether applied to personal interest information or to that relating to an official
function, would then allow different users to 'introduce' themselves through
opening selected parts of the customised metaphor to each other, namely
This approach is clearly related to the potential or virtual reality techniques. A
workgroup might then take thevisual form of a cluster of houses linked by pathways (with
gates, secret doors, etc.) which those involved might be constantly redesigning. Some
might prefer the additional feature of 'gateways' to alternative realities or
spaces where the same information was configured otherwise.
11. Mnemonic advantages
It is appropriate to note that mnemonics was a highly developed art prior to the
widespread availability of paper - and especially to facilitate the task of orators of
renown. A major feature of this art was the distribution of information onto visual
surfaces which could subsequently, be walked through and interrogated for the next points
in an argument. Typical surfaces were so-called 'memory palaces' and garden
scapes (mandalas should also be seen in this light). It can now be argued that there is an
analogous need for such visual metaphors to enable individuals to handle information
overload and retain some control over the information they endeavour to possess. Spatial
metaphors can be seen as vital to retaining posses sion of information and avoiding
'memory leakage' or the effective 'dismemberment' of one's information
12. Intellectual property
It is possible that intellectual 'property' may come to be usefully defined
in terms of spatial metaphors. Access to them, in the case of corporations, may be
jealously guarded since they would then effectively define the corporations self-identity.