form for this type of interface, illustrating its suitability also for information systems whose data are spatial in nature. The following section describes an experimental study to compare various techniques to provide a better understanding of these types of interfaces from the human-computer interaction perspective.
3 . 3
A Comparative Study of Graphical User Interfaces for
Topological Maps
A comparative evaluation of the presentation techniques described in this section will help to identify their merits and shortcomings. The main aims of this study were to uncover users' experiences and preferences with these interfaces and to measure task completion times in the performance of various tasks typically faced by users of such systems. More specifically, the objectives were:
• to observe how users used these interfaces and what difficulties were encountered
• to uncover what station name searching strategies were used on various interfaces
• to ascertain the order in which subjects preferred to use the interfaces
• to find out if there was any significant difference in users' performance of various tasks with the interfaces
CHAPTER 3: THE BIFOCAL DISPLAY
3.3.1 An Integrated Approach to Interface Evaluation
45
Burger & Apperley (1991) consider a list of problems associated with current interface evaluation techniques and propose a multi-dimensional evaluation approach which entails three major evaluation types, namely, critical evaluation, subjective evaluation and objective evaluation. They argue that "using one, or even two, of ihe techniques in isolation can result in a biased or incomplete view of the interface. Some aspects of the interface will be highlighted by one technique and not by the others and it is therefore essential that none of the three be left out in the evaluation process."
This integrated approach is particularly applicable to graphical user interface evaluation as complex human behaviour exhibited in the course of user interaction with this type of interface cannot be simply summarised by a set of numbers in terms of task completion times and error counts. In order to get a richer picture of user interaction and a more accurate and reliable basis to compare these interfaces, this integrated evaluation approach has been adopted.
In this study, the London Underground wall map together with four map-based interfaces implemented on a workstation were used. A detailed explanation of some of these map-based interfaces and their implementations has been discussed in Section 3.2. The following subsection describes in brief their general features and characteristics.
3 . 3 . 2 Map-based Interfaces
The five interfaces used in this experiment were, a paper wall map, Bifocal Display (point and shoot), Bifocal Display (scrolling), simple scrolling view and split screen (scrolling) systems. The fish eye view system used by Hollands et al ( 1989) had not been considered in this experiment as the complexity of the London Underground map would incur excessively long processing time required to generate the fisheye view on the Sun IPC Workstation; the slow response time would render the fisheye view interface unusable.
Paper Wall Map
The paper wall map, which measured 900mm x 600mm, was the official London Underground map produced by the London Regional Transport Authority. Being one of the world's oldest and largest mass transit systems, the London Underground railway system comprises more than 10 different lines and over 250 stations. The map is printed in colour, and a specific colour is assigned to each line on the map. Stations are normally represented by small rectangles along the line and interchange stations by circles.
CHAPTER 3: TilE BIFOCAL DISPLAY Bifocal Display - Point and Slwot
46
Figure 3.6a shows an implementation of the London Underground map using this technique. With the point and shoot interface, the user moves the central focus region to another location by first positioning the cursor using a mouse to the point of interest and then clicking a mouse button. The new Bifocal image will then be presented; on a Sun SP ARC IPC workstation the response time was about a quarter of a second. Bifocal Display - Scrolling
The static presentation of the Bifocal Display is identical to the one discussed earlier. However, with a scrolling interface, the user moves the central focus area by scrolling a mouse. The system detects the direction of the mouse movement and updates the Bifocal image in real time; the amount of movement of the central focus area is directly proportional to the scrolling action on the mouse made by the user.
Fulhu e;..O.doNy Porsons Gr-oen Putney Bridge Figure 3.7 � St Poul'a Uno Euex Road Old Street London 8ridgo
CHAPTER 3 : 1HE BIFOCAL DISPLAY
Simple Scrolling View
47
The scrolling operation for this interface is similar to that described earlier. With simple scrolling view, a section of the map will be displayed in detail at any one time. Although this interface accommodates a larger detailed display area on the screen, it provides no global information about the position of the current display in relation to the entire map. If the current display is at one extreme of the map, further scrolling in that direction will not result in a change of view. Figure 3.7 illustrates an implementation of this presentation technique.
Split Screen Display - Scrolling
The split screen display as the name suggests, partitions the display area into two sections with one section displaying a small area of the map in detail and the other a reduced size of the entire map. On the demagnified map the area which the detailed display is currently presenting is highlighted to enable the user to maintain an overall perspective of the map. This technique is akin to an inset on a geographical map. Figure
3.4 shows an implementation of this presentation technique.
As the user scrolls the mouse, the two sections of the screen will be updated in real time according to the direction of and proportional to the scroll movement.