Arne Harder & Rainer Michel

The Target-Route Map
Evaluating its Usability for Visually Impaired Persons

Abstract and Homepage Index

Reference Of the Original Publication.
Harder, A. & Michel, R. (2002). The target-route map: Evaluating its usability for visually impaired persons. Journal of Visual Impairment and Blindness, 96, pp. 511-523.

Abstract. Two experiments demonstrated the usability of the target-route map in locating and identifying unfamiliar routes by persons who are visually impaired. Compared to the orientation map and the mobility map of the same area, the target-route map was read faster, and more of its details were recalled correctly.

Keywords. Blind orientation, tactile map, Target-route Map, usability.

German Summary (Deutsche Zusammenschau)

  1. Introductory Remarks
  2. General method
  3. Experiment 1 Sighted participants
  4. Experiment 2 Visually impaired participants
  5. General discussion
  6. Apendix

I Introductory Remarks

§1 Basic Facts on Tactile Maps

Visually impaired people (persons who are blind or have low vision) who prepare themselves for traveling an unfamiliar route in an unknown environment by reading and memorizing a tactile map travel more safely and efficiently through that route than do those who either use an audiotaped description or travel the route accompanied and advised by a sighted observer (Brambring & Weber, 1981; Espinosa et al., 1998). With a tactile map, one can perceive the spatial layout of the represented area without having to conduct any distracting navigational activities, such as avoiding obstacles.

On the other hand, anyone who reads a tactile map will have to build up a mental spatial structure of the represented area while successively touching parts of the map (Révész, 1938). Many persons who are visually impaired have difficulty accomplishing this task (Ungar, Blades, & Spencer, 1995a, 1996). They generally read a tactile map slower and understand it less than do sighted people who see a visual print of the same map (Ungar, Blades & Spencer, 1995b). However, after fewer than 10 hours of systematic training of appropriate tactile scanning strategies, blind children improve their ability to orient themselves on a tactile map (Ungar, Blades & Spencer, 1997) or to locate certain distinctive features on a tactile map (Berla & Butterfield, 1977).

Physical parameters can also affect a user's reading or understanding of a tactile map. For example, the minimal distance required for symbol distinction between two points or lines on a tactile map depends on the sizes of the symbols; the absolute minimum size seems to be around 3 mm (about 0.12 in.; Berla, 1982).

§2 Types of Tactile Maps

One can make maps more convenient to use by adapting their features to the needs of special users or the requirements of specific tasks. Tactile maps for people who are visually impaired and maplike representations of routes for people who are cognitively impaired are addressed to certain populations of users (LaDuke & LaGrow, 1984). Road maps and maps for hiking each refer to the requirements of different tasks.

The process of adapting maps for specific tasks may involve the scale, the types of environmental features being represented, the number of features represented, and the symbol coding system.

One can distinguish between two types of tactile maps as follows:

  1. Orientation maps help their users get an "overview" of a large area, such as an urban area or a small town (James, 1982). They represent few distinctive features of the environment and use a comparatively small scale.
  2. Mobility maps, however, are designed to help users determine how to travel safely in unfamiliar environments (James, 1982). They represent the spatial arrangement of all streets that will be encountered in the designated area, together with features that are considered crucial for mobility purposes, by using a comparatively larger scale.

Symbols that represent features that are necessary for enhancing mobility, such as bus stops or traffic lights, are referred to as mobility symbols.

§3 The Target-Route Map - An Individual Map

At the University of Magdeburg, the second author, a computer scientist, was engaged in the problem of individual maps for his doctoral thesis. Individual maps are designed to aid special users to solve particular spatial tasks (Michel, 2000, p. 2). Michel decided to create a map that would help people who are blind locate a particular unfamiliar route.

This target-route map presents both the desired route (target route) with all its relevant details and the distinctive features of the surroundings of the target route, so that the user may relocate the route when lost. To avoid overloading the sense of touch, the target route should be comparatively larger than its surroundings (Michel, 2000, pp. 36-38).

Thus, the target-route map can be characterized as an orientation map of an area in which the mobility map of the target route is embeded. This type of map represents a radical example of the term "individual map", for two maps can each represent the same urban area but emphasize different routes by enlarging the scale and enriching the number of details of those routes, and thus each will look different.

§4 The Task-Dependence Hypothesis

Anyone who wants to establish a new type of map will have to prove its usefulness. The following considerations formed the basis for the experiments reported in this article.

  1. The target-route map shall help its users easily locate routes in maps of unfamiliar areas. When a specific area is shown on both an orientation map and a target-route map, those who use the target-route map should find the target route faster and should have more space for their fingers to identify the mobility symbols in that route than should the users of the orientation map. Thus, the group who uses the target-route map should read the map faster and remember more of its details than should the group who uses the orientation map.
  2. The results of the comparison between the target-route map and the mobility map, both of which represent the same area, are considered to be task dependent.
    1. If the task is primarily to become familiar with a specific route (that is, if the target route was announced before the map was read - see Espinosa et al., 1998), the target-route map, which emphasizes that target route, should have an advantage for its users.
    2. If, on the other hand, the task is to gain an understanding of the entire area and the target route was not announced before the map was read (see Brambring & Weber, 1981), the mobility map may be more convenient for its users, since it displays the whole area using a uniformly large scale.

Two experiments were conducted to test the first prediction of this task-dependence hypothesis. The experimental tasks should favor the target-route map, and the experiments should clarify how much better would be the outcomes for the users of that type of map.

The first experiment was conducted with sighted persons who were blindfolded while exploring the tactile maps. Because most sighted persons are unfamiliar with tactile maps, the results are unlikely to have been influenced by any preferences for certain types of maps, which may have occurred if the participants were blind and had experience with tactile maps.

The second experiment was conducted with visually impaired persons because they are the real test population for any new type of tactile map.

II General Method

§5 Experimental Material and its Production

Three types of tactile maps were produced, each of which represents a certain geographic area. Within this area, the target route is emphasized by a raised line.

The orientation map shows the whole geographic area by using a uniformly small scale, the mobility map shows the same geographic area by using a uniformly large scale, and the target-route map represents a specific route within that area by using a large scale for the target route and a small scale for the surrounding area. The target-route map is the only one that emphasizes the target route twice: once by a raised line and again by a representation thatr is enlarged in comparison to its surroundings.

The experimental maps were manufactured with the help of the computer-based tactile map production system "Map Wizard" (Michel, 1998, 2000; Michel & Hamel, 1998 - see Box 1 for details).

To produce the orientation map or the mobility map, the data with the basic geographic information is represented on the screen using either a small or a large scale. After some changes in the resulting map layouts, these maps were printed in black lines on white paper, which was copied onto swell paper (lines on swell paper can be raised so they are tactile).

To produce the target-route map, the layout of the orientation map was used as the basic data. The target route was then manually marked within this layout. After that, the focus line algorithm, the main component of Map Wizard, was used to enlarge the marked route. The area surrounding the marked target route was not enlarged, so that the target route and surrounding areas still fit within the boundaries of a single map sheet.

In simpler terms, imagine the map with the target route laid out on a twodimensional balloon. The focus line algorithm "pumps up" or enlarges the target route, but "releases air" out of the symbols within the surrounding area so that they will appear smaller when compared to the orientation map. The greater the distance from a symbol to the target route, the smaller it will be presented on the target-route map.

Map Wizard holds the difference in scale between the target route and its surroundings as large as possible and those within the surrounding area as small as possible. Thus, the user of the target-route map experiences a map with two scales: a large scale for the target route and a small scale for the rest of the map.

Two geographic areas were shown on the maps. The same symbol code was used for all the maps (see Figure 1).

Area A represents a 5.62 square km (about 3.47 square miles) residential part of Berlin, and Area B represents a 4.66 square km (about 2.88 square miles) part of Magdeburg, the capital of the German state Saxony-Anhalt.

Each map was shown on at least one sheet of swell paper. Every map was 42 cm long and 29.7 cm wide (about 16.54 x 11.69 in.).

Both orientation maps - one for each geographic area - had a scale of 1:7500. Each orientation map filled one swell sheet.

The mobility map - only for Area A - had a scale of 1:3000. Because Area A was too large to be presented as a mobility map on one swell sheet, this map was divided into nine submaps by cutting it three times in both orthogonal directions. The target route was presented on three of those nine partial maps (maps 4, 7, and 9; see Figure 2 for an illustration of the subdivision process).

The target-route maps - one for each area - showed the target route in a scale of 1:300 and its surroundings in a scale of 1:7500 (see Figure 3 and Figure 4). Each target-route map filled one sheet.

Additional skeleton maps were designed to test the participants' ability to recall the routes. They were the same size as the orientation maps, but represented only the streets and rivers of the corresponding area (see Figure 5).

§6 Dependent Variables

Two dependent variables - reading time and recall ability - were used in both subsequent experiments. Each variable was measured with each map for each participant.

  1. Reading time was defined as the number of minutes a participant took to explore the map tactually. The participants knew that they would have to recall the details of the map after they read the map.
  2. Recall ability was defined as the percentage of items a participant recalled by filling in the missing details of the skeleton map (see Table 1).
    Each participant had to mark the target route by drawing a central line through the correct streets; recall each building or mobility symbol, regardless of its position on the corresponding map, and correctly mark the position of each building or mobility symbol. A point for recall ability was assigned if one of the following happened: the participant:
    1. drew the "central line" in such a way that the marked and the correct target route were identical to each other for one city block to either side of the streets of the target route;
    2. marked a building or mobility symbol on the skeleton map that existed somewhere on the experimental map; or
    3. located the correct position of the building or mobility symbol (this task was judged to be correct, if a building or mobility symbol was marked on the skeleton map within half a city block of where it was located on the experimental map).

Table 1 Recall Ability Units per Area

Area Feature Description Occurrence Units
(A) city-blocks along both sides of the target route 26 26
  buildings outside the target route 4 8
  mobility symbols within the target route 12 24
  maximal number of features and recall units 42 58
(B) city-blocks along both sides of the target route 16 16
  buildings outside the target route 2 4
  mobility symbols within the target route 5 10
  maximal number of features and recall units 23 30

§7 Procedure

Each experiment consisted of two phases: a preparation phase and a testing phase.

During the preparation phase, the participant learned the symbol codes for the tactile maps. Each symbol was presented to the participant on a sheet of swell paper measuring 5 x 5 cm (about 1.97 x 1.97 in.). During the first trial, the meaning of each symbol was explained to the participants; the presentation order was constant for all participants for this trial. In subsequent trials, the symbols were randomly presented to the participants, and they had to recall their meanings. As soon as all symbols had been correctly identified during two subsequent trials, the preparation phase ended.

Before the testing phase began, the participants were told that they would have to explore a tactile map by actively touching it; this map would contain a route of major interest marked by a central line. In addition, they were told that they would have to recall the elements of the map, particularly those found along the route of major interest.

Each participant tactually explored the experimental map. As soon as the participants had been exposed to the map, a stopwatch was started; the watch was stopped when the participants verbally indicated that they knew the map. Then the participants explored the corresponding skeleton map and filled in the missing features from the experimental map. No time limit was given for reading the map or completing the skeleton map.

Figure 1 Symbol Code for All Tactile Experimental Maps.

Figure 2 Illustration of the Sub-Division Process for the Experimental Mobility Map.
The geographic area covered by the orientation map of area A was divided into a matrix of nine elements - three rows by three columns (see part a). To yield the mobility maps, each of the nine elements of the matrix was represented on a separate sheet of swell paper using a unique scale of 1:3000. Part b shows the mobility map of matrix element 3. This map represents a part of the target route (emphasised by an additional line) and its nearest surroundings.

Figure 3 Orientation and Target-Route Map for Area A.

Figure 4 Orientation and Target-Route Map for Area B.

III Experiment 1 Sighted Participants

§8 Method

The participants in this experiment were 24 sighted students of the University of Magdeburg. They all reported to have no experience with tactile maps prior to this experiment.

The participants were randomly assigned to the three types of maps: 8 to the orientation map, 8 to the mobility map, and 8 to the target-route map. All these maps represented area A (see Figure 3).

All the participants were blindfolded when they actively touched the tactile map. Afterward, they completed a printed version of the skeleton map with full vision.

The data for reading time and recall ability were statistically analyzed (Kruskal-Wallis-Test). For each analysis, the value of statistical significance was set at p < 0.05 (see Table 2).

§9 Results and Discussion

Both analyses demonstrated a statistically significant effect (see Table 2).

The users of the target-route map read the map significantly faster than did the users of the orientation map. The reading-time values for the target-route map tended to be lower than those for the mobility map, but the difference was not significant.

The users of the target-route map also recalled more of its details than did those of either the orientation or the mobility map.

Thus, people who had never "been in touch" with tactile maps seemed to benefit from the target-route map presentation. But since the participants of this experiment were sighted, this result does not necessarily demonstrate the usability of the target-route map for people with visual impairments. Therefore, an experiment with visually impaired participants was conducted.

Table 2 Results of Experiment 1

Variable Map Type m (sd)  
reading time orientation map 51.8 (4.4)  
  mobility map 40.9 (22.9)  
  target-route map 31.1 (12.5) sig.
recall ability orientation map 59.8 (20.0)  
  mobility map 56.6 (10.2)  
  target-route map 78.3 (16.1) sig.
Comment. sig.: statistical significance (Kruskal-Wallis-Test, two-tailed testing, p < 0.05).

IV Experiment 2 Visually Impaired Participants

§10 Method

Only 4 persons participated in this experiment, all of whom were adventitiously visually impaired. They were members of a rehabilitation course at the Berufsförderungswerk Halle/Saale (Germany), which includes braille reading and other blindness-related topics.

No participant reported any formal education with tactile maps. The only participant who reported to be totally blind stated that he had experienced "tactile pictures" before this experiment.

Each participant worked with two maps consecutively. One map always represented Area A, and the other always represented Area B (see Figure 3 and Figure 4). Each participant explored a target-route map and an orientation map. The geographic area and the type of map for the first map were randomly chosen for each participant.

All the participants were blindfolded throughout the experiment, so the skeleton maps were tactile. Each participant completed a skeleton map of each area by drawing the missing features on the map with a pen (see Figure 5).

Table 3 presents the descriptive results of this experiment. Because of the small sample, no statistical analyses were conducted.

§11 Results and Discussion

The four participants achieved better recall when they used a target-route map than when they used an orientation map (see Table 3). In addition, they seemed to read their second map faster than their first map, and the values for recall ability were smaller for the second map than for the first one.

The first result may reflect an advantage of the target-route map. The second and third results may be due to a fatigue effect. All the participants began their experimental work immediately after they had passed the day's 10 lessons. They all reported to be tired when they received their second map. Maybe the fact that they speeded up their reading of their second map and were working with its skeleton map explains why they did not recall as many features as they did when they worked with their first skeleton map.

Table 3 Results of Experiment 2

Factors Reading Time Recall ability
  m (sd) m (sd)
map type orientation map 20.25 (9.54) 65.55 (34.12)
  target-route map 25.75 (11.95) 90.95 (14.10)
Area (A) 23.50 (10.66) 83.85 (23.36)
  (B) 22.50 (11.82) 72.65 (30.30)
sequence first 27.00 (7.07) 89.88 (14.01)
  second 19.00 (12.70) 66.63 (35.16)
total 23.00 (10.43) 78.25 (27.72)

Figure 5 Example of Filled-in Skeleton Map. This skeleton map was filled in by a visually impaired person whose residual vision had been occluded.

V General Discussion

§12 Methodological Considerations

According to a strict significance definition (Bonferroni correction), neither of the two dependent variables in Experiment 1 reached statistical significance. But since both of them reached the uncorrected .05 significance level, the results of Experiment 1 were considered to be interpretable.

The first experiment demonstrated an advantage of the target-route map: The users of the target-route map read their map faster and recalled its details better than did the users of either the orientation or the mobility map of the same geographic area. But the advantage for each variable was small, or the effect would have matched the Bonferroni-corrected significance criterion.

The results of Experiment 2 are not inconsistent with those of Experiment 1. The four visually impaired participants seemed to recall the details of the target-route map better than they recalled those of the orientation map. Much more data are needed for a statistically valid analysis.

§13; Task Appropriateness of the Target-Route Map

The target-route map integrates the "mobility map" of the target route into the "orientation map" of the whole area. By considering only this construction principle, one may expect that its users should read and recall the route as fast and accurately as people using a mobility map. In Experiment 1, the users of the target-route map tended to read their map even faster than did those who used the mobility map. As predicted by the hypothesis presented earlier, more features of the target-route map than of the orientation map were recalled. But the mobility map seemed to be the most difficult to recall for the participants of Experiment 1.

The tactile mobility map used in Experiment 1 consisted of nine separate parts. The participants who used this map reported the need to integrate the cognitive "images" of these map parts mentally into a single representation of the whole area; they reported that this task was difficult.

The target-route map, however, presented the whole area on a single map sheet, and the target route was emphasized by its enlarged presentation. This type of spatial representation seems to be appropriate when the task is to gain an accurate concept of the route and its differentiated parts in relation to its position in the environment. Because the target routes in both geographic areas contained the highest number of decisive informational elements (see Figure 3 and Figure 4), the participants' route-understanding was the crucial task of both experiments.

Amplifying the route in the target-route map may have drawn the participants' attention to enable them to memorize the route elements. This possibility may explain why the best results at recall of map details in both experiments were achieved with the target-route map.

However, one cannot exclude the possibility that the superiority of the target-route map is just due to the enlarged route, which has more space for the fingers to detect the symbols, at least for the visually impaired participants in the second experiment. If this symbol-distance hypothesis is accepted, the participants should have read the target-route map faster than the orientation map. This result was not empirically found in Experiment 2.

If the distance between the symbols is the only factor, the mobility map in Experiment 1 should have been the easiest for the participants to read and recall, since there is the same large space for finger movement everywhere in the map. The empirical data of Experiment 1, however, do not confirm this expectation at all.

One may even consider the superiority of the target-route map in both experiments to be a result of the chosen dimensions of the geographic area. If one chooses an area so small that its mobility map fills one map sheet only, perhaps one would expect the mobility map to be superior to both the orientation and the target-route map.

But according to the task-dependence hypothesis (see §4), one should see a shift in favor of the mobility map only when the detailed recall of the environment becomes more important than the understanding of the target route. The conditions of this experiment, however, favored the learning of the target route, and the results are consistent with the predictions of the task-dependence hypothesis.

§14 Conclusions

This study demonstrated that the target-route map provided an advantage for memorizing an unknown route, compared to conventional types of tactile maps. This finding was repeated in two experiments, one with blindfolded sighted participants and one with visually impaired participants.

For a final judgment about the target-route map, further and more differentiated data must be collected. Meanwhile, orientation and mobility specialists may test the practicability of the target-route map principle with tactile maps they make themselves to demonstrate difficult spatial layouts to their clients who are visually impaired.

Future research could consider person-, environment-, and task-related variables that influence the mobility of people who are visually impaired. It may include the personal conditions of the participants, such as the degree of their experience with tactile maps; spatial features, such as the presence or absence of a characteristic sound network (Mansfeld, 1940); or task differences, such as spatial orientation compared with walking independently through an environment after consulting the relevant tactile map.

VI Apendix

§15 References

  1. Berla, E. P. (1982). Haptic Perception in Tangible Graphic Displays. In: W. Shiff & E. Foulke (eds.). Tactual Perception: A Source-Book (pp. 3643-386). New York: Cambridge University Press.
  2. Berla, E. P. & Butterfield, L. H. (1977). Tactual Distinctive Feature Analysis: Training Blind Students in Shape Recognition and in Locating Shapes on a Map. Journal of Special Education, 11, pp. 335-346.
  3. Brambring, M. & Weber, C. (1981). Taktile, verbale und motorische Informationen zur geographischen Orientierung Blinder (tactile, verbal and motoric informations for a geographic orientation of the blind). Zeitschrift für experimentelle und angewandte Psychologie, 28, pp. 23-37.
  4. Espinosa, M. A., Ungar, S., Ochaita, E., Blades, M. & Spencer, C. (1998). Comparing Methods for Introducing Blind and Visually Impaired People to Unfamiliar Urban Environments. Journal of Environmental Psychology, 18, pp. 277-287.
  5. James, G., A. (1982). Mobility maps. In: W. Shiff & E. Foulke (eds.). Tactual perception: A Source-Book (pp. 334-363). New York: Cambridge University Press.
  6. LaDuke, R. O. & LaGrow, S. J. (1984). Photo-Bus-Route-Map: An Intervention to Produce Independence in Bus Travel for Mentally Retarded Adults. The Mental Retardation and Learning Disability Bulletin, 12, pp. 71-75.
  7. Mansfeld, F. (1940). Die Verdunklung und die Blinden (the darkening and the blind). Arch. f. d. ges. Psychologie, 107, pp. 411-436.
  8. Michel, R. (1998). Tactile Maps for Blind People. In: Th. Strothotte (ed.). Computational Visualisation: Graphics, Abstraction and Interactivity (pp. 339-355). Berlin-Heidelberg-New York: Springer.
  9. Michel, R. (2000). Interaktiver Layoutentwurf für individuelle taktile Karten (interactive layout design for individual tactile maps). Dissertation, Otto-von-Guericke-Universität Magdeburg (Germany), Aachen: Shaeker.
  10. Michel, R. & Hamel, J. (1998). Distortions and Displacements in 2D. In: Th. Strothotte (ed.). Computational Visualisation: Graphics, Abstraction and Interactivity (pp. 139-150). Berlin-Heidelberg-New York: Springer.
  11. Révész G. (1938). Die Formenwelt des Tastsinnes (the world of shapes for tactile perception). The Hague (The Netherlands): Martinus Nijhoff.
  12. Ungar, S., Blades, M. & Spencer, C. (1995a). Visually Impaired Children's Strategies for Memorising a Map. British Journal of Visual Impairment, 13, pp. 27-32.
  13. Ungar, S., Blades, M. & Spencer, C. (1995b). Mental Rotation of a Tactile Layout by Young Visually Impaired Children. Perception, 24, pp. 891-900.
  14. Ungar, S., Blades, M. & Spencer, C. (1996). The Ability of Visually Impaired Children to Locate Themselves on a Tactile Map. Journal of Visual Impairment and Blindness, 90, pp. 526-535.
  15. Ungar, S., Blades, M. & Spencer, C. (1997). Teaching visually impaired children to make distance judgements from a tactile map. Journal of Visual Impairment and Blindness, 91, pp. 526-535.

§16 Box 1: Description of Map Wizard

Map Wizard is a software prototype for creating the layout of tactile maps. It has been written in C++ for Microsoft Windows operating systems. On the basis of geographical data in vector format (e.g. GDF, NTF) map symbols for streets, bus stops, taxi stands, bus lines and pedestrian crossings are placed on the map area. The program creates the symbols in a size that allows tactile perception of the symbols. The thus created map layout will be printed on swell paper and heated afterwards until the lines are tactually perceivable.

The displayed area may be chosen arbitrarily. It is scaled such that it fits on the available paper sheet. It is ensured that the symbols are not scaled down to a size below the threshold of tactile perception. Hence choosing a large area and a small paper may lead to symbol cluttering.

Map Wizard detects and analyses symbol cluttering. A user may thereupon run the automatic clutter correction or displace the symbols interactively.

For symbol displacement the Focus Line algorithm is used. It supports symbol displacement with a linear origin (like streets or rivers). It moves the symbols in the vicinity until they have at least a distance such that they are Perceivable as different symbols.

The force that is applied to each symbol moving it away from the center of cluttering is decreased with growing distance to the center. At the map border the movement is zero to prevent the map size. You may think of this procedure as a balloon surface that is pulled away at one place and thus stretching the area around it.

To create a tactile route map a user may specify the start, and end and optionally intermediate points. Using Dijkstra's algorithm for route search in graphs the shortest route is thus determined. This route may now be enlarged up to three times of its former size. To achieve this Focus Line algorithm is applied again with a predefined set of parameters.

§17 Acknowledgements and Affiliations

We would like to thank Jochen Schneider and Thomas Wagner (students at the University of Magdeburg, Germany) for observing the sighted experimental participants, and Ine Langer (student at the Fachhochschule Harz/Wernigerode, Germany) for observing the visually impaired experimental participants. We also thank the members of the Berufsförderungswerk Halle/Saale (Germany) for their friendly cooperation. Finally we want to thank Stephanie Santel and Frauke Krämer (students at the University of Magdeburg) for reviewing, Dona Sauerburger (Orientation and Mobility Specialist at Maryland, USA) for helping to edit, and Hans-Jürgen Warmbold (member of the staff at the Clinic for Neurology II at the University of Magdeburg) for preparing the figures.

Dr. Arne Harder
Luxemburger Str.124-136 Whg.2208
D-50939 Köln, Germany
e-mail: nc-harderar@netcologne.de
Dr. Rainer Michel
General Electric Network Solutions
Engelbergerstraße 21
D-79106 Freiburg i. Br., Germany
e-mail: rainer.michel@ps.ge.com

To Homepage Index.
to the Index of Harder's Homepage: Mobility.