Kristina Höök Swedish Institute of Computer Science kia@sics.se

Catriona Macaulay Napier University catriona@dcs.napier.ac.uk

Whenever we present the problem of navigation in electronic worlds, the so-called "lost in hyperspace" problem, there will always be one (male) engineer in the audience who suggests that users would be helped by an orientation map in the interface over the space. We always reply that this would be all very well for those users who are already quite good at finding their way through the information spaces, as they have the cognitive abilities required to understand maps. But what about those users who have problems with orienting themselves in abstract information space as well as problems understanding maps? In our view, this points at the core problem for designers of navigational tools: there are large individual differences that cannot be evened out through adding overviews or maps. Instead we believe that we have to rethink our whole approach to the design of information spaces.

So what can we add that would make navigation easier for those users who find it hard to navigate and understand maps and other overviews? If we look at how users navigate in the real world - in cities, in finding information in libraries, in finding their way through buildings, in choosing which television program to see, in deciding which medical doctor to see, etc., we see that this is often done through talking to other people. For example, Streeter and Vitello (1985), discovered that travellers found their way to a destination by driving to the nearest gas station and asking for help. Jon O'Brien (pers.comm.) discovered that employees in an organisation found it more useful to search for persons than documents, despite having a nicely designed three-dimensional information space. Andreas Dieberger (1997) analyses behaviour on the WWW in terms of following other peoples' advice and choices, for example through studying their bookmarks and assembled links. We name this strategy "social navigation", since it does not have to involve building a (mental) spatial model. Instead, social navigation relies on interactions with other people. Sometimes it means interacting with another person, sometimes it means studying what a large group of users does.

There are several reasons why social navigation may be preferred over more spatially demanding methods. When we talk to someone else, the information we get back can be personalised to our needs. We are perhaps told a little bit more than exactly the information we asked for, or if the information provider knows us, the instructions may be adapted to fit our knowledge or assumed reasons for going to a particular place. The instructions are also adapted to the user in another sense in that they start from the point where the user is at, and is given in a sequential form (first go there, then go there). Social navigation also has another quality, in that we can judge to what extent the directions given can be trusted depending upon the credibility of the information provider. If I am provided with information on how to find a particularly good definition of some concept from a well-known researcher in my field, I will probably follow the advice, while if the information is given by someone not in my field, I might not even bother to look it up. Yet another aspect of social navigation is the form in which instructions are given; frequently, it will be given in a verbal, dialogue-driven form, rather than an abstracted spatial form.

Following groups of people, for example, going through an airport, may also be considered as a form of social navigation, even if it does not encompass all of the aspects mentioned above. It does involve the matter of trust. It also provides the seeker with a sense of security: if all these people have chosen this route, it must be the right one. In general, navigation is a scary activity. Most people are scared of getting lost, and employ various methods to feel more secure. Talking to other people seems to be one of the more reassuring methods.

Social navigation will not replace the need for well-designed information spaces and navigational tools that assist users in forming models of the space. It would be stupid to replace alphabetical order in a library with chaos just because there is a librarian who knows the way around the chaos, who is available for a book borrower to talk to. Furthermore, for some applications, the main task of the user might in fact be to understand the layout and relationships in the space. What our goal is, instead, is to broaden our view on the design space and to include social navigation as one tool in the repertoire.

From a sociocultural perspective we might also consider the view that there is no such stable entity as 'the individual' . Rathe,r we might consider that individuals, cultural groups and differences are defined by context. Using this lens, human behaviour can only be understood in relation to its context (Cole 1996). Part of that context will arise from the socio-cultural-historical setting within which 'we' are situated. For example, feminist geographers have long considered the complex relationships between gender and space/place (Massey 1994). Part of our project therefore will concern itself with investigating the implications for information space/social navigation design of taking just such a socio-cultural perspective. In so doing we will be recognising that design is not just about cognitive processes or technology, but about politics as well.

In the PERSONA project we are attempting to take a grip on the overall design problem of information spaces and navigational tools. We do so through two main lines of research. The first is concerned with finding a method by which a designer can analyse a targeted domain, its users, their tasks, and so forth and determine whether there are navigational issues that need to be resolved. We believe that navigation comes into play in many applications - not only in hypermedia, for example, but also in hierarchical file systems, in augmented realities (such as interactive museums), and other such structures. The method will include making suggestions for good design possibilities.

The second line of research investigates the social navigation concept. Our goal is to define it in such a way that we can understand when and where it comes into play, and why it works better for certain individuals.

For both these research directions, we need to find evaluation criteria by which we can determine what signifies a successful navigation. Rather than relying on the traditional usability metrics, such as time spent, number of errors, etc., we would like to get at the underlying cognitive processes involved, as well as more qualitative measurements such as whether the user feels secure or not.

Cole, M. (1996). Cultural Psychology. A Once and Future Discipline. Cambridge, MA, The Belknap Press of Harvard University Press. Massey, D. (1994). Space, Place and Gender. Cambridge, Polity Press.

Streeter, Lynn A. , Diane Vitello, and Susan A. Wonsiewicz (1985) 'How to Tell People Where to Go: Comparing Navigational Aids', Int. J. Man-Machine Studies, 22: 549-562.

© 1998,1999,2000 i3net	-- Search i3net --	About the i3net web pages
Revised: 15 December, 1997.		Mail to webmaster@i3net.org