Making Touch-based Mobile Phones Accessible for the Visually Impaired

Dissertation Task 2 0834060

Department of Information Systems and Computing

MSc Information Systems Management

Academic Year 2009-2010

MAKING TOUCH-BASED MOBILE PHONES ACCESSIBLE FOR THE VISUALLY IMPAIRED

Alexander Dreyer Johnsen - 0834060

A Dissertation submitted in partial fulfillment of the requirement for the degree of Master of Science

Brunel University Department of Information Systems and Computing Uxbridge, Middlesex UB8 3PH United Kingdom Tel: +44 (0) 1895 203397 Fax: +44 (0) 1895 251686



ABSTRACT

The mobile phone enjoys increased popularity, providing new means of connectivity and

functionality. Today most phones comes equipped with touch-based screens, enabling

the user to interact in an easier and more efficient way, compared to standard buttons.

However, such screens require visual navigation, ruling out access for the visually

impaired; evidently, modern phones are not designed for this user group.

By analyzing the mobile market, available solutions, and current technology, it was

found that neither the society, nor manufacturers of products and services, is designing

for the visually impaired. Thus, this group is denied access to the numerous services and

possibilities sighted people enjoy.

However, it is possible to operate mobile phones through the use of applications called

screen readers; still, these applications have proven to be ineffective and less than user-

friendly on touch screens. Hence, this dissertation sets out to find an alternative

approach; to construct a solution that will make touch-based mobile phones accessible

for the visually impaired.

Design research was chosen as the methodology for the project. Design research

highlights the importance of developing a solution over the course of several iterations,

and to perform product evaluation using external participants. A total of five iterations

were carried out, resulting in several artifacts and a prototype for a user interface. The

prototype was designed to replace the phones own user interface and provide an easy and intuitive way of operating a touch-based mobile phone.

Through the process of developing the user interface, virtual prototypes and other

artifacts were created. The virtual prototypes turned out to be of great advantage,

communicating the vision and potentials of the final product to the stakeholders. In

addition, the experience from the project shows that a successful development project

should produce several iterations, have well-documented artifacts and perform external

user testing.

The new user interface was developed for the Android OS to replace the phones own user interface. Operation relies on voice and haptic feedback, where the user receives

information when tapping or dragging the finger across the screen. The proposed

solution is unique in many ways, it keeps gestures to a minimum, it does not rely on

physical keys, and it presents a menu layout similar to most Nokia mobile phones.



ACKNOWLEDGEMENTS

The process of writing this dissertation has been long and exhausting, and would not

have been possible without the involvement and support from several people; for this I

am grateful.

First, I would like to thank my supervisor Bendik Bygstad for his guidance and academic

support throughout the project. It has been indispensable and a key factor for the

accomplishment of this dissertation.

Secondly, I would like to thank Tor Ulland at Huseby Resource Centre for the blind,

for taking an interest in the project and for sharing his knowledge and insights on life

without vision. Ulland also acquired all of the participants for the evaluations,

ensuring the completion of the project

Without the involvement of May-Britt Haug, at The Norwegian Association of the Blind

and Partially Sighted, I would not have had the possibility of getting in touch with

Huseby or Tor Ulland. Thank you for believing in the project.

I would also like to thank Magne Gabrielsen, Gaby Groff-Jensen and Knut Beck at

SmartPhones Telecom for believing in the project, providing both financial and

developmental resources. Kim Ruben Teigen, whose development skills ensured the

realization of the User Interface, a key contribution to the success of the project.

A special thank you to my wife Ragnhild Eg, for her continued support throughout the

dissertation. You have stayed up countless nights proofreading, and have continued to

support me during the entire process.

I would also like to thank my family, friends, fellow students and colleagues for the

continued support during the process of writing the dissertation.

Last, but not least, I would like to thank the anonymous participants for spending their

spare time on evaluating and providing feedback on the User Interface. Not only once,

but twice! Thank you!

TOTAL NUMBER OF WORDS: 13.031

I certify that the work presented in the dissertation is my own unless referenced

Signature: .........................................

Date............31.01.2011....................



TABLE OF CONTENTS

Chapter 1: Introduction .................................................................................................. 7 1.1 Problems that people with vision loss face .................................................. 7 1.2 Back to society ............................................................................................. 8

1.3 Research aim and objectives ........................................................................ 8 1.4 Research approach ....................................................................................... 9 1.5 Dissertation outline ...................................................................................... 9

Chapter 2: Literature review ........................................................................................ 10 2.1 Equal opportunities .................................................................................... 10

2.2 The mobile market ..................................................................................... 11

2.3 Potential problems with smartphones ........................................................ 11

2.4 Improving compatibility ............................................................................ 11 2.5 Assistive technology .................................................................................. 12 2.6 Screen reader .............................................................................................. 12 2.7 Reading and input of text ........................................................................... 13 2.8 Haptics ....................................................................................................... 15

2.9 Making a better user interface .................................................................... 15 2.10 Turning the mobile phone into an assistive aid ....................................... 16

2.11 Framework ............................................................................................... 16

Chapter 3: Research Method ........................................................................................ 18

3.1 Design research .......................................................................................... 18

3.2 Awareness of problem ............................................................................... 20

3.3 Suggestion .................................................................................................. 20 3.4 Development .............................................................................................. 21

3.5 Evaluation .................................................................................................. 22 3.6 Conclusion ................................................................................................. 24

Chapter 4: Research results .......................................................................................... 25

4.1 Results in regards to development ............................................................. 25 4.2 Results in regards to testing ....................................................................... 29

Chapter 5: Discussion .................................................................................................. 34 5.1 Design as an Artifact .................................................................................. 34

5.2 Problem relevance ...................................................................................... 35 5.3 Design evaluation ....................................................................................... 35 5.4 Research contributions ............................................................................... 36 5.5 Research rigor ............................................................................................ 36 5.6 Design as a Search progress ....................................................................... 36

5.7 Communication of research ....................................................................... 36 5.8 Summary of discussion .............................................................................. 37

Chapter 6: Critical evaluation of research ................................................................... 38

Chapter 7: Conclusion .................................................................................................. 40 7.1 Summary of the dissertation ...................................................................... 40

7.2 Future research and development .............................................................. 41



References .................................................................................................................... 42

Appendix A: Ethical approval ..................................................................................... 52

Appendix B: About the author ..................................................................................... 53

Appendix C: The mobile market .................................................................................. 54 Appendix C.A - The new players .................................................................... 54

Appendix D: Using the mobile phone for assitive aid ................................................. 55 Appendix D.A - Object recognition ................................................................. 55 Appendix D.B - Navigation ............................................................................. 55

Appendix E: Interview with Tor Ulland at Huseby (Statped) ..................................... 57

Appendix F: Project plan ............................................................................................. 60

Appendix G: Project meeting reports .......................................................................... 61 Appendix G.A - Initial project workshop meeting, June 14, 2010 .................. 61

Appendix G.B - Approval of functionality, June 25, 2010 .............................. 63 Appendix G.C - Layout of UI, August 30, 2010 ............................................. 66

Appendix G.D - Layout of UI, September 13, 2010 ........................................ 68 Appendix G.E - Layout of UI, October 10, 2010 ............................................ 70

Appendix G.F - Layout of UI, October 13, 2010 ............................................. 73

Appendix H: Documentation of ui ............................................................................... 76

Appendix H.A Documentation of UI, version 0.0.5 ..................................... 76 Appendix H.B Documentation of Conceptual Design .................................. 79 Appendix H.C Documentation of Second Prototype, round 1 of

testing ................................................................................................... 84 Appendix H.D Documentation of Third Prototype, round 2 of testing ........ 87

Appendix I: Development and activity log .................................................................. 90

Appendix J: Information to project participants .......................................................... 96

Appendix K: Consent form .......................................................................................... 97

Appendix L: Participant Survey ................................................................................... 98

Appendix L.A: participant survey, round 1 ..................................................... 98 Appendix L.B: Participant survey, round 2 ................................................... 100

Appendix M: Ressults from user testing .................................................................... 102 Appendix M.A: Results from round 1, Overview of all responses ................ 102 Appendix M.B: Results from round 1, Participant 1 ..................................... 106

Appendix M.C: Results from round 1, Participant 2 ..................................... 108 Appendix M.D: Results from round 1, Participant 3 ..................................... 110 Appendix M.E: Results from round 1, Participant 4 ..................................... 112

Appendix M.F: Results from round 2, Overview of all responses ................ 114 Appendix M.G: Results from round 2, Participant 1 ..................................... 118



Appendix M.H: Results from round 2, Participant 2 ..................................... 120 Appendix M.I: Results from round 2, Participant 3 ....................................... 122 Appendix M.J: Results from round 2, Participant 4 ...................................... 124 Appendix M.K: Results from round 2, Participant 5 ..................................... 126 Appendix M.L: Results from round 1 and round 2 compared against

each other ........................................................................................... 128 Appendix M.M: T-Test and Effect sizes ........................................................ 132

LIST OF TABLES

Table 1 - Framework for the top issues people with vision loss encounter when using

mobile phones with touch screen ........................................................................................ 17

Table 2 - Design Research Guidelines by Hevner et al. (2006) .......................................... 20

Table 3 - Tasks the participants were asked to perform ........................................................ 22

Table 4 - Survey completed by participants .............................................................................. 23

Table 5 - Additional tasks presented to participants ............................................................... 24

Table 6 - Additional survey questions ........................................................................................ 24

Table 7 Summary of survey categories ................................................................................... 30 Table 8 - Table showing subjective feedback from round two ........................................... 33

Table 9 - Overview on aims and objectives .............................................................................. 38

LIST OF FIGURES

Figure 1- Design Research process model by Vaishnavi and Kuechler (2007) ............. 19

Figure 2 - Some of the menus available in the UI ................................................................... 26

Figure 3 - List navigation in the UI ............................................................................................. 26

Figure 4 - Conceptual UI prototype for people with vision loss ......................................... 27

Figure 5 - Flow map for navigation in the UI........................................................................... 28

Figure 6 - Conceptual model of application .............................................................................. 28

Figure 7 - Graph displaying mean scores from the first and second round of testing .. 30

Figure 8 - Graph displaying a comparison between results from first and second round

of testing on the menu system ............................................................................................. 31

Figure 9 - Graph displaying a comparison between results from first and second round

of testing on the complete solution .................................................................................... 32


Dissertation Task 2 0834060 Page 7 of 132

CHAPTER 1: INTRODUCTION

1.1 Problems that people with vision loss face

According to the World Health Organization (WHO) (2009), around 314 million people are

visually impaired, of which approximately 45 million are blind. Of the total number, 12

million are children, 82% are of age 50 or older; furthermore, women have the greatest risk of

becoming visually impaired. Developing countries have the largest representation of visually

impaired people, with around 87% representation. Visual impairment blindness can be

defined as the collective term vision loss, which refers to either the loss of the ability to see, or vision reduction. The terms low vision, partially sighted, or visually impaired, can be used for a person who has not lost the capability to see, but whose vision is significantly

reduced compared to normal vision (Colenbrander, 2002).

Both groups deal with daily situations that vary greatly depending on the severity of their

visual impairment. A blind person will have to rely on other senses and inputs, such as touch

and hearing, as well as assistive aids, such as the long cane and the guide dog. On the other

hand, an individual with low vision can be aided by visual enhancements, like large print,

magnifiers and illumination (Colenbrander, 2002).

Common for both groups is their encounters with situations and problems that a sighted

person would not consider an issue. This can be seen in everyday situations, where the

environment is designed for the sighted, not someone with vision loss. Accessing stores and

businesses, or simply crossing the street, can turn out to be challenging tasks. Much of our means

of communication is through the use of signs and signals; visual symbols that a person with

vision loss cannot see. For example, danger signs or signs providing direction, road blocks or

general information on public transportation. At worst, situations like these may prevent a person

from wanting to step outside, hence limiting contact with the surrounding world (Tjan et al.

2005; Joseph, 2009; Maines, 2008; Appendix E).

The same restrictions apply to other ordinary situations; TVs, radios, mobile phones, stoves and ATMs are just some examples of equipment that are becoming increasingly advanced. In the past, such items were mostly equipped with buttons; although not originally designed for people

with vision loss, they made it possible to memorize the steps in order to access a certain feature.

With todays modern design, with touch screens showing dynamic menus; it has become impossible for a person with vision loss to memorize and use such equipment. It can be argued

that companies do not consider this user group when designing new products (Picard, 2010;

Appendix E).

Another issue facing people with vision loss is the difficulty some sighted people have when

encountering blind people, making it hard for the latter to develop new relationships. Most jobs

are also designed for the sighted. Due to this, blind people do not meet the same expectations and

are often relegated to specific roles. This is an undesirable situation as employers may lose out

on valuable skills and those who lack vision are left unable to prove their potential. Society

should strive for quality and for implementing people with vision loss into the working

environment (The National Federation of the Blind of Connecticut, n.d.).



1.2 Back to society

Several ideologies and assumptions about blindness and rehabilitation exist. One of the more

accepted is the restorative approach, embraced by Father Thomas Carroll (American Printing House for the Blind, 2010). This approach works under the idea that most blind people could,

with the help of professional counseling and training, live full lives. Seven basic losses are high-

lighted, and the focus is on restoring these through training, with the white cane or a guide dog,

learning to read Braille and using assistive technology (R. A. Scott, 1995). With the correct

equipment and training, a person with vision loss can master ordinary tasks like operating a

computer, reading a book or using mobile phones (Appendix E).

People with vision loss can also participate in sports; blind Soccer is a particular example of

how well correct training in applying other senses can enable a blind person to participate in

team activities. Players are able to play soccer in almost the same way as a sighted person by

relying on shouted commands and specially designed soccer balls, containing ball bearings

(Malinowski, 2010).

1.3 Research aim and objectives

While it is clear that the technology and the will to integrate those with vision loss into society

are strongly present, society is not yet properly adapted. Several aspects can be improved,

where one is the mobile phone.

The mobile phone introduces several benefits, but perhaps the most important is the easy

access to communication with friends, family and the surrounding world. However, most new

mobile phones are designed for visual navigation, rendering the mobile phone inaccessible for

people with vision loss. Designing a phone that is inaccessible for people with vision loss can

at worst result in loss of contact with society, thus the author of this project believes in the

importance of making mobile phones not only accessible, but easy to use, for people with

vision loss.

The aim of this dissertation is to present, develop and evaluate a new User Interface (UI) for

touch-based mobile phones, which will make this type of phones available for people with

vision loss.

Based on the aim of developing a UI, the following objectives are presented:

A literature review on issues related to the operation of touch-based mobile phones

and available assistive aids and technologies

Create a framework for designing the UI

Perform an evaluation of the UI through several iterations; both internal and external

Discussion on the results from the project



1.4 Research approach

The project uses Design research method for developing and testing the UI for touch-based

mobile phones. Design research is designed with development and testing in mind,

encouraging projects to create and deliver several versions of a solution, where major releases

are tested on external participants to provide feedback (Hevner et al. 2006; Vaishnavi &

Kuechler, 2007).

Two versions of the solution were tested on external users, allowing them to provide feedback

and comments on the interfaces. Feedback from the first version was used to improve the

second version. 1.5 Dissertation outline

The dissertation report is organized as follows:

Chapter 2: Presentation of literature, research and the current mobile market relating to people

with vision loss.

Chapter 3: The research method Design research is outlined along with documentation of the

steps followed throughout the project.

Chapter 4: Research results are presented, analyzing the different between the two tested

versions.

Chapter 5: Critical discussion comparing the findings from the literature review with the

results from the evaluation of the UI.

Chapter 6: The processes and the research methodology are critically evaluated.

Chapter 7: The report concludes with a summary of the previous discussions and a highlight

of the contributions to the existing research field, with future research suggestions.



CHAPTER 2: LITERATURE REVIEW

The research field on vision loss encompasses several areas of interest, where this

Dissertation focuses on how people with vision loss can operate mobile phones with touch

screens. Hence, this chapter is divided into several subjects starting with rights of people with

vision loss, followed by a description of the current mobile market, for then to present

available assistive aids and technology. The chapter ends in a framework, used to design the

UI. 2.1 Equal opportunities

It goes without saying that people with vision loss have the same rights as sighted people; yet

the misconception that vision loss reduces work efficiency is a common prejudice (Wall,

2003). A Norwegian report on job opportunities for people with disabilities concluded that

this user group has the largest share of unemployed potential workers compared to the rest of

the population (ECON, 2003). However, a different report that looks into peoples experience with moving from school to work life, suggests that part of the problem is based on lack of

knowledge and improper administration by government organizations (Berge, 2007).

Several organizations are working towards equal rights for people with vision loss; eliminate

prejudice, integration with society, achieve equal rights and benefits and provide aid to

countries and people struggling with diseases causing blindness.

World Blind union (n.d.) (WBU) is an internationally recognized organization that represents

160 million people with vision loss in 177 member countries; as a universal voice, it aims to

achieve equal rights and opportunities in all aspects of society. However, WBU does not

provide direct services such as training, guidance and access to assistive technology (AFB,

n.d.; NABP, n.d.). These services are designated to local and national organizations, like the

American Foundation for the Blind and The Norwegian Association of the Blind and Partially

Sighted (NABP).

Vision 2020 is a joint cooperation between the WHO, the International Agency for the

Prevention of Blindness (IAPB) and professionals within the field. The main focus is to

eliminate the main causes of avoidable blindness by the year 2020 ant to prevent 100 million

people from becoming blind (Vision 2020, 2009; International Agency for the Prevention of

Blindness, 2010).

Organizations are not the only means for support for people with vision loss. Several

countries have also implemented rules to protect the rights of people with vision loss. In the

United Kingdom, the Disability Discrimination Act states that a person should not be treated

less favorable because of his or her disability. The European Union has introduced a similar

act, which directs its member countries to introduce measures to reduce discrimination against

people with disabilities (Wall, 2003). Norway has taken it a step further with the introduction

of a new discrimination and accessibility law, which state that all new products marketed

towards the general public should be equally accessible to all, regardless of personal

limitations or handicaps (Steria, 2009).



2.2 The mobile market

The mobile phone market enjoys increased popularity, with market shares expanding every

day (IDC 2010), it is now considered larger than the PC market (Neset, 2010). The mobile

market consists mainly of two types of mobile phones; feature phones and smartphones.

Where feature phones are considered low-end phones, with limited features and computer

power (Nusca, 2010; Wikipedia, 2010a), smartphones offer more advanced computing power

and operating systems, as well as greater connectivity; they are essentially small mobile

computers (Wikipedia, 2010b).

During the first quarter of 2010, a total of 314.7 million mobile phones were sold to end users

worldwide, with smartphones making up 54.3 million of the sales (Gartner, 2010b); this is an

increase of 17% compared to the same period in 2009. Although, feature phones are still

considered the platform with the largest sales volume, smartphones have had a positive

increase and sales are expected to increase further in the coming years (IDC, 2010).

According to Gartner (2010) the top five OS for smartphones are; Symbian (44,3%),

Blackberry (19.4%), Apple iOS (15.4%), Google Android (9.6%) and Windows Phone

(6.8%). Apple iOS and Google Android are currently the fastest growing (The Nielsen

Company, 2010) and it is estimates that the Android OS will be the second largest OS

worldwide in 2010 and it will challenge Symbians top ranking position in 2014 (Gartner, 2010a). The numbers show that smartphones sales are increasing, suggesting that the market

wants more advanced phones, with more functionality and better connectivity.

A more detailed view on the mobile market and its actors is available in Appendix C. 2.3 Potential problems with smartphones

While smartphones come with a range of additional features, there are drawbacks. Studies

have shown that the majority of users only take advantage of a small portion of the

functionalities available (Gomns, 2005). In addition, the touch screen can make it more

difficult to type, as no physical feedback is provided. Nevertheless, the increasing sales

indicate that most sighted people do not consider these factors as show stoppers. However,

this is not the case for those with vision loss.

The majority of information on a mobile phone is presented through graphical means,

rendering access almost impossible for a person with vision loss. A touch-based mobile phone

becomes even harder to access with a touch-based visual design and navigation that contains

no physical representation of where or how to press buttons and icons. With these limitations,

it seems likely that a large user group will be unable to use these types of mobile phones,

segregating them further from the general population. This is also in contradiction with laws

set by the government (Meyers, 2008). 2.4 Improving compatibility

Studies on mobile phone accessibility for elderly and people with disabilities; all conclude

that mobile phones are not designed for these user groups. They point out several aspects that

could improve mobile phone accessibility. Mobile phones should be of adequate size and

shape and have texture with good grip. The screen should be large and the buttons should

have a logical placement for easier memorization of layout. Voice feedback should provide

information and confirm execution of commands. The phone should also provide easy access

to emergency numbers. Finally, the number of features should be kept to a minimum to ensure

ease of use (Abascal & Civit, 2001; Plos & Buisine, 2006; Smith-Jackson et al. 2003; S.K.

Kane et al. 2009). These suggestions are all in accordance with the guidelines on how mobile



phones should be designed, provided by the NABP (n d).

Creators are recommended to follow good practices and standards to ensure that their

applications or web pages are accessible to all user groups (Perea et al. 2006). However, this

recommendation is not always followed; not for mobile phones nor for applications or web

sites. A recent study on disableds relationship to social media, like Facebook and Twitter, concludes that these and similar services are designed for visual navigation, forcing those who

are visually impaired to use mobile versions of the media that are better formatted and have

less content (Tollefsen et al. 2011; Rossen, 2011).

Several guidelines have been presented in recent years to show developers how to design

accessible applications and web sites. For instance, the W3C organization has formed such an

initiative, which develops accessibility guidelines (W3C, 2008). These guidelines are in line

with those published by NAPB (n.d.). Furthermore, Fukuda et al. (2005) proposes to

introduce navigability and listenability metrics for designing and developing applications or

web sites, ensuring that screen readers can navigate and present the information in a clear and

consistent way. Similar suggestions were presented by Calabr et al. (2009) to ensure that e-

books are compatible with screen readers.

With a range of different web browsers, both on computers and mobile phones, compatibility

cannot always be guaranteed. In light of this, The Mobile Web Initiative Device Description

Working Group has proposed the creation of a Device Description Repository; a database

containing information about mobile devices that can be queried by applications to present

data in the most efficient way (K. Smith & Sanders, 2007).

2.5 Assistive technology

Although mobile phones are not designed for use by people with vision loss, there are solutions

that can compensate or improve such use. These are defined by the general term assistive

technology; they include any product, instrument, equipment or technical system designed for or

used by a person with disabilities, which prevents, compensates, supervise, alleviates or

neutralize the effects of the disability (Perea et al. 2006). Products may range from physical and

living objects like the long cane or the guide dog, to software in a device like a screen reader. For

an individual who is blind, it is impossible to read the content on a screen; hence the aid of

assistive technology is a great benefit in their personal and professional life. This section will

look at current solutions and relevant research that addresses the issue of assistive technology

on mobile phones. 2.6 Screen reader

It is possible to operate a mobile phone or computer without being able to see what is on the

screen. Software applications called screen readers can convert text to speech, enabling the

user to read and navigate the content of a screen through hearing. By listening to voice

communication provided by the screen reader application, users are able to perceive and

navigate the content on the screen, making it possible to perform tasks like word processing,

e-mailing, listening to music and surfing the web (Wikipedia, 2010b).

A screen reader is generally made out of two components; the application which monitors the

content on screen, and a synthesizer, which provides the spoken feedback. This is produced

through text-to-speech, where the text input is provided by the screen reader and the synthetic

voice is produced by the synthesizer. The synthesizer works with different languages and

supports phonemes and grammatical rules (AFB, n.d.). For a screen reader to work efficiently,



applications need to follow common standards so that the content can be interpreted and

presented correctly (Babinszki, 2010).

Screen readers are available for both computers and mobile phones, ranging from products

that are free of charge to those that cost closer to $1.000,-. One example of a screen reader for

computers is JAWS, one of the most feature rich products, but also the most expensive

(Freedom Scientific, n.d.). Screen readers can also be run in a web browser, allowing a person

to use almost any computer (Bigham et al. 2008). On mobile phones, screen readers have

been most common for phones with physical buttons, but are becoming available for phones

with touch screens (Babinszki, 2010).

However, a screen reader must not be confused with the voice feedback often built into

modern mobile phones. Although they are capable of representing menu and application

content, these functions are less advanced when it comes to conveying accurate information.

Moreover, they are limited to phone menus; and so unable to access information inside

applications (Theofanos & Redish, 2003).

Mobile Speak is a commercial screen reader designed for mobile phones that run the

Windows Mobile or Symbian OS. It is considered one of the best solutions for mobile phones,

and it supports phones both with physical buttons and with touch screens (Code Factory, n.d.).

TalkBack and Spiel are two open source screen readers designed for the Android OS; both

enable developers to incorporate screen reader functionality into their applications. Although,

not as advanced as commercial versions, they provide developers with an easy and convenient

way making their applications more accessible (C. Chen & Ganov, 2009; Darilek, 2010).

2.7 Reading and input of text

Braille was created to enable blind people to read text. It is not a new language, but a text

system with dots laid out in an organized manner enabling a blind person to read and write

text in the same manner as a sighted person. By moving the finger across the dots, a trained

person is able to understand the letters that the dots represent, and thus able to read the text

just like a sighted person (AFB, n.d.). Reading text from a computer can also be achieved

through Braille, through the use of refreshable Braille displays, Braille printers and Braille

note takers (AFB, n.d.).

Although Braille is a great tool, its reliance on finger sensitivity can exclude the elderly that

may have reduced feeling in the fingertips (AFB, n.d.). For people who have lowered vision,

Braille might not be needed; magnifiers, loupes and special glasses are just some of the

equipment that can assist a person in reading text NABP (n.d.).

However, Braille is not suited for mobile phone use. Instead, users will have to rely on

different means of typing, depending on the phone in use. For a person with vision loss,

mobile phones equipped with physical keyboards are the preferred means of interaction; this

relies on the feedback provided by buttons. In conjunction with a screen reader, operating a

mobile phone is feasible with a keyboard (Babinszki, 2010; Buxton et al. 2008).

Physical keyboards for mobile phones come either with a T9 (text on 9 keys) or a QWERTY

keyboard layout. A T9 layout consists of a phone pad with numbers ranging from 1 to 9 and

three associated letters on each button; while a QWERTY layout represents the same keys as

a keyboard attached to a computer. Just like a computer keyboard, it is possible to memorize



the layout of these buttons and thus write without looking at the keyboard. In addition,

predictive text is often used together with these layouts, enabling the phone to predict what is

being typed, completing the words faster and correcting spelling mistakes (Wikipedia,

2010d).

Mobile phones equipped with touch screens do not always have a physical keyboard; instead

text is entered on a virtual keyboard presented on the screen. This can render their use a

challenge, as the virtual buttons do not provide any tactile feedback (Yfantidis & Evreinov,

2005; Buxton et al. 2008). Several solutions have been presented to improve the typing of text

on touch screens, ranging from physical equipment that work in cooperation with the mobile

phone to software keyboards installed on the phone.

HumanWare (n.d.) has created an overlay for touch screens that encompass the phone screen

and enables the user to interact with the phone and gain access to the most common functions.

The overlay can also communicate with a separate Braille device, making it possible for the

user to read and write using Braille (HumanWare, n.d.). The advantage with this solution is

that the user can carry only one device and interact directly with it; however, the downside is

that it is only compatible with resistive touch screens (Wikipedia, 2010b); conversely, most

modern phones come with capacitive screens that are only capable of interacting with

conductive materials like a human finger (Wikipedia, 2010a).

As mentioned, mobile phones with touch screens come with virtual keyboards. While they are

handy, experience shows that these keyboards can be troublesome when typing. Hence,

several third party software keyboards have been created. A keyboard named ThickButtons

provides a virtual QWERTY layout on the screen; ThickButtons differentiate itself from a

normal keyboard by anticipating the coming letters and making those letters larger and others

smaller (Wells, 2010b). SlideIT and Swype are two other solutions, where the use of a

dictionary enables users to write words by sliding a finger along the letters without removing

it from the screen (Wells, 2010a; Sorrel, 2010a). Yet another option called SwiftKey

anticipates the following word; developers claim that 1/3 of all words will be correctly

anticipated, resulting in up to 50% faster typing (TouchType, 2010).

However, all of these keyboards require some degree of visual navigation, which may be

suitable for people with low vision but not for blind users. A keyboard called BlindType

(Anon, n.d.) might be suitable for blind users, although this was not its original intention.

BlindType is a QWERTY virtual keyboard, which perceives the users typing style. Through an assumption that the user is familiar with the QWERTY layout, the keyboard is able to

recognize where and or what is being typed on screen without the user having to focus on the

screen. Essentially, a word can be typed at any place on the screen without a keyboard

displayed. In theory, this keyboard could also work for blind users, since it does not require

text to be entered in a defined area and it can automatically correct typing errors. Recently,

the company behind the solution was acquired by Google (Noyes, 2010). Google has also

released a technology called Google Scribe, which predicts what the user is planning to type

(Lofts, 2010).

Furthermore, Yfantidis & Evreinov (2005) have designed a gesture based keyboard specifically

for blind users. It works by tapping the screen to make a square appear, which represents eight

directions; where each direction holds a separate letter. By moving the finger towards one of the

directions, the letter is read out loud and proceeding to remove the finger confirms the selection

for typing.



2.8 Haptics

Haptics is a technology that provides tactical feedback, resulting in a more intuitive and less

visually reliant interface. However, haptics can never replace hearing or vision; rather, it

provides an additional sense. For touch-based interfaces, haptics can make a user feel and

visualize the shape of an item without looking at the screen, or it can provide feedback when the

finger reaches the border of an element or a button. Haptics is anticipated to become more

advanced in the feature, with an expected ability to provide even more fine-grained details, such

as the fur of an animal (Buxton et al. 2008; Gemperle et al. 2001; Rassmus-Grhn, 2006).

Yu & Brewster (2003) have developed a solution for presenting graphs and tables to users with

vision loss. Users can explore virtual graphs and tables through a force feedback device that

creates haptic feedback, while audio is used to present detailed and abstract information.

In a study on user reactions to haptics, Rassmus-Grhn (2006) found that some rely more on the

audio feedback than the haptics feedback. This suggests that, although haptics can be a valuable

additional modality, some users depend more on the senses they are used to. Hence, solutions

that incorporate haptics will likely be greatly improved if they also implement sound.

2.9 Making a better user interface

Mobile phone manufacturers have noticed the inadequacies in usability of the UIs native to

various OS. Most manufacturers develop their own set of UIs for the mobile phones,

replacing the OSs own UI (HTC, n.d.; Topolsky, 2008). In addition, Nokia, Intel and the University of Oulu have established a dedicated research center, which focuses on improving

the UI on mobile devices (Darling, 2010).

Raman and Chen (2008) argue that the UI is only a means to an end and should blend

seamlessly into the users way of operating the mobile phone. In contradiction to a computer OS, the mobile OS should provide access to what the user needs and remain unattended when

not in use. According to Karampelas and Akoumianakis (2003), the layout for mobile phones

should include the following, consistent presentations, alternative navigation tools,

accessibility to common options and functions and self-explanatory navigation.

Hence, several solutions have been put forward to improve the navigation of touch screens for

users with vision loss, essentially providing eyes-free navigation.

One of this was presented by Amy K Karlson et al. (2005) in a study on a user-interface

designed for one-hand thumb use. The interface created a 3x3 square-grid; with each square

containing a link to an application or a sub-menu. Selection is done by tapping one of the

squares, while gestures were introduced for zooming and to navigate between menus and.

Results showed that the participants liked the way the navigation and selection of applications

worked, although they were hesitant with the gestures.

Kane et al. (2008) have developed Slide Rule; a UI that solely operates through the use of

gestures and auditory feedback. By using several combinations of gestures, Slide Rule allows

the user to open applications and perform commands. Results revealed that the proposed

solution performed significantly faster than button based menus. However, due to the gesture

commands, more errors were produced.



Amar et al. (2003) built a prototype handheld device that used tactile and auditory feedback to

convey a menu structure that provided easy access to common functions. The menu could be

navigated through one hand access, where each finger would perform a specific task. Rotating

a dial would move through various options and menus, a separate button would return to a

previous menu, and four buttons performed navigation and application functions. Most

participants were satisfied with the solution; however, it was noted that users had problems

forming a conceptual model of the hierarchical menu. This was improved by adding the

feature of pressing a button halfway down, which would state the functionality of the selected

button.

With their application, Strumillo et al. (2009) have abandoned the default phone menu on the

Symbian OS and offer a new, simpler menu that provides audible feedback and access to

common functions and features. While navigating through menus, the user is informed of

what is a displayed and possible action to take. The solution has received good feedback;

however, it was developed for an older version of Symbian, making it incompatible with

todays Symbian phones.

Eyes-free Shell is an open-source project for the Android platform created by Chen and

Raman (2009); it replaces the phones existing UI with a new menu system, providing easy access to applications, as well as information on time and current location. Eyes-free Shell

uses auditory feedback and gestures to navigate between menus. The user can search for

applications by using the same type of technique presented earlier (Yfantidis & Evreinov,

2005). The application is part of a larger open-source initiative, sponsored by Google, which

aims to make touch-based mobile phones more accessible. However, some features require

physical buttons, rendering the proposed solution inaccessible to certain phones. The reliance

on gestures can also be problematic, as indicated from the findings in other projects

mentioned above (Helft, 2009).

2.10 Turning the mobile phone into an assistive aid

Extending on the common functionalities of the mobile phone, several solutions exist today

that turn the phone into an assistive aid. Appendix D presents some of the functions available,

such as object recognition and navigation.

2.11 Framework

The literature review has pointed out several aspects that hinder the effective use of mobile

phones with touch screen for people with vision loss. Although several factors could warrant

further scrutiny, the aspect that stood out as the main obstacle is the operation of the touch

screen itself.

Based on the literature, a framework pointing out the most important factors for the

development of a new UI for touch-based mobile phones are summarized in table 1.



Table 1 - Framework for the top issues people with vision loss encounter when using mobile phones with touch screen



CHAPTER 3: RESEARCH METHOD

Design research was chosen as the research method for the development project in this

Dissertation. This chapter outlines the selected method along with a documentation of the

steps followed throughout the project.

3.1 Design research

Design research aims to provide solutions for human purposes by creating an artifact; the

artifact is tested to provide feedback on improvements (March & Smith, 1995). It is a design

solving process, with several cycles of iterations, where new and improved artifacts are

constantly put forward. Each artifact is evaluated and new artifacts build on the knowledge of

previous ones; thus it is an approach that encourages innovation (Hevner et al. 2006; Grnli &

Bygstad, 2009).

The research field arose as a result of several failed projects; the failures highlighted the

importance of closely studying the design of a product during development (Vaishnavi &

Kuechler, 2008). The same approach is used in engineering, where a product is developed and

tested to further improve the coming versions (Peffers et al. 2006; Vaishnavi & Kuechler,

2007).

Frederick R. Brooks Jr., best known for his book The Mythical Man-Month (Brooks, 1995),

argues that constant incremental iterations should replace the practice of building complete

product versions. Development teams should produce quick and effective prototypes that can

be tested by external users, which provide valuable feedback for further development (Kelly,

2010).

According to Grnli and Bygstad (2009), design research is particularly relevant for

understanding mobile services and innovation. For a project like the current, where the

developers face a previously unfamiliar condition, best guess is not enough (Nielsen, 1993).

In these situations, feedback from the intended target group becomes vital to the projects success. Hence, Design research is the process in which it is uncovered what works and what

does not work (Vaishnavi & Kuechler, 2007).

Several models for successful completion of a Design research project have been presented.

One, presented by Grnli and Bygstad (2009), leverages on two widely accepted approaches

to Design research. The method recommends Vaishnavi and Kuechler's (2007) model for the

overall project framework and the model from Hevner et al. (2006) to evaluate the results.

Vaishnavi and Kuechler's (2007) approach is a five step model, designed to guide a project

from startup to the finished solution, as shown in figure 1. The first step is the Awareness of problem, which focuses on highlighting or improving the results of an existing problem. This is followed by the Suggestions step, where the problem is further analyzed using existing knowledge and theory, culminating in a tentative design. Development is the next step, where the project attempts to implement an artifact according to the suggested solution, which

is then evaluated in the fourth step. The whole project is completed with a conclusion where the project results are evaluated (Grnli & Bygstad, 2009).



Development, evaluation and further suggestions are often performed iteratively, allowing the

researcher to step back and evaluate and make changes as necessary. Stepping back achieves

an understanding that could only be gathered from the specific act of construction, the model

refers to this process as circumscription (Vaishnavi & Kuechler, 2007).

Figure 1- Design Research process model by Vaishnavi and Kuechler (2007)

Vaishnavi & Kuechler's (2007) model has been criticized for being too dependent on

traditional development methods. The suggestion step is the only stage that allows creative

input and is also the only stage separating the model from the mentioned development

methods. However, it is noted that the steps are only suggestions and not absolute

requirements (Vaishnavi & Kuechler, 2007; Grnli & Bygstad, 2009), thus allowing the

project to make changes as deemed necessary.

Hevner et al. (2006) have developed seven guidelines, presented in table 2, to help provide

better understanding and measurements for effective design research projects. Although quite

general, these guidelines are implemented in the same way as Vaishnavi and Kuechler's

(2007) model, allowing the project freedom to decide on implementation. However, it is

advised to address each of the guidelines during the projects course (Grnli & Bygstad, 2009; Hevner et al. 2006).



Table 2 - Design Research Guidelines by Hevner et al. (2006)

As mentioned earlier, Vaishnavi and Kuechler's (2007) model were used for the overall

project framework and are described in the following section. Hevner et al.'s (2006) seven

guidelines will be used to evaluate and discuss the results of the project, as described in

chapter 5.

3.2 Awareness of problem

The first stage of Vaishnavi & Kuechler's (2007) model was initiated by a general

assumption, that the majority of todays mobile phones are touch-based and designed for those with no visual impairment. The assumption was based on the authors previous experience from the mobile industry; former projects had demonstrated that the UI is not

always up to standards and hinder effective use of mobile phones.

Once the general assumptions were in place, these needed to be verified by more thorough

research. The initial literature review focused on the mobile market and recent phone releases,

using resources found mainly through Google and later moving on to the major market

analytics companies. With no prior experience related to vision loss, a literature review on

users with vision loss was needed, not only to identify existing solutions for both daily life

and mobile phone use, but to understand more of the experience if vision loss. Through

resources, such as Google and NABP, literature on the most common problem areas was

collected, along with information regarding manufacturers of assistive technology.

3.3 Suggestion

A more thorough literature review was carried out in the suggestion phase, focusing on all

elements and issues regarding the development of assistive technology for people with vision

loss. The literature review was completed in four iterations, each step contributing with more

information regarding issues facing those with vision loss, how people with vision loss use

mobile phones, application compatibility, operating systems to develop for, screen readers

and haptics, and implementation of such technology. The end result was a framework of

guidelines that worked as a source of reference throughout the project.

The work that came out of the literature review supported the earlier assumptions,

highlighting the need for a UI for touch-based phones, designed specifically for those with

visual impairment. The new design would require operation using touch and sound, menu

elements placed in a logical arrangement, and well-defined categories for assistive



functionalities and standard mobile phone functionality.

It was important for the project to provide new functions and improved access for people with

vision loss. From the literature review, several projects with similar solutions came to the

authors attention. However, the projects relied mainly on gestures and hardware buttons, functionalities that could make the solutions harder to learn, and that are certainly

incompatible with mobile phone that are not equipped with buttons. Based on the feedback

from these projects, it became evident that the use of gestures should be kept to a minimum

and that the dependency on hardware buttons should be removed.

3.4 Development

This stage involved the development of the solution in addition to documentation and

planning of the project. The Rational Unified Process (RUP) was chosen as the framework for

the development process due to its emphasis on multiple iterations and its clear and defined

guidelines for developing applications (Gornik, 2004; Pollice, 2002). However, time

constraints necessitated a lightweight implementation of RUP, including only core activities.

Nevertheless, this approach is still in accordance with the framework (Pollice, 2002).

RUP requires all objects to start with an inception, where the project vision is defined and

presented to the stakeholders (Pollice, 2002). Stakeholders for the current project include

representatives NABP, Statped and SmartPhones Telecom; the latter is the company

responsible for providing developer resources. To present the vision and the idea in an

intuitive manner, a virtual prototype was created to show a possible solution for replacing a

touch-based mobiles GUI. The virtual prototype was created in Microsoft Expression Blend 4 with SketchFlow (Microsoft Corporation, 2009), a tool for creating working prototypes of

applications without writing any code.

Further on, RUP dictates an elaboration on the design and a definition of the baseline,

including functionality requirements and possible applications (Pollice, 2002). The virtual

prototype served as a foundation for the outcome of the project and was used to communicate

the idea and the vision to the developers. During several project meetings, discussions

centered on what would be realistic outcomes from the allocated time. The decision was to

create a working UI prototype for people with vision loss, while ignoring the server side

functionality. Hence, the UI would work and behave as the final product, but all data

presented would be static, not dynamic. An early prototype of the application was developed

at this stage, to test core functionality. Complete logs of the project meetings are available in

Appendix G.

Construction is the third step in RUP; this is where the main development takes place (Pollice,

2002). The application was developed in Java for Android, to support Android version 1.6

onwards. Functionalities for the screen reader and haptics were not developed during this

project. Instead, existing modules were implemented with API integration. During the

construction phase, a total of five iterations of development were completed, with a working

prototype coming from each one. Three iterations were for internal use, while two were tested

externally. With each iteration new virtual prototypes, associated documentation, and added

functionalities evolved. Guidelines and best practices for Android development, defined by

Google, were followed through (Google Inc, 2010b; Google Inc, 2010a). Documentation on

the UIs is available in Appendix H in addition to complete development logs in Appendix I.



3.5 Evaluation

The final phase of RUP is the transition phase, where the developed product is tested on real

users (Pollice, 2002). Each version of the UI was tested in three different scenarios, in a

virtual environment, by the developer on a real device, and by the author, who made sure that

all functionalities were implemented according to specifications. In addition, evaluations of

the prototype were performed through two iterations of user experiments, where external

participants tested and provided feedback on the prototypes.

The participants were recruited with assistance from Statped, a Norwegian agency, which

provides guidance to municipalities for supporting people with learning disabilities and vision

loss (Statped, n.d.). All participants had previous experience from volunteering to other

projects aimed at people with vision loss; moreover, before commencing, they were informed

of their rights and of relevant ethical regulations.

Participants all performed the same steps, starting with a general introduction to the project

with its goals and aims, followed by ten minutes of training and some time to become familiar

with the UI, finally performing a couple of exercises and providing feedback and completing

a survey. Each test was performed on a HTC Hero (HTC, n.d.) device, with an average

duration of one hour. Participants were monitored throughout the test. The introduction to the

project emphasized to the participants that the UI was only a prototype and that the aim was to

improve the navigation on touch-based mobile phones, they could therefor expect limitations

and fewer available options.

During the first iteration, participants were trained in operating the UI. They then had to

choose a language for operating the system, either English or Norwegian. They were informed

that the English version would be more advanced and sound more natural than the Norwegian

version. All menus and feedback would reflect the chosen language.

Table 3summarizes the exercise that the participants were asked to perform. The exercise was

designed to evaluate whether the participants were able to understand and navigate the UI on

their own. Instructions were to perform one task at the time and sufficient time was allowed

for each task.

Table 3 - Tasks the participants were asked to perform

During the exercise, the participants behavior, responses, reactions and verbal feedback were documented. The participants were also providing verbal feedback on functions. Feedback

provided included both comments on current functions and suggestions for improvements.



After the exercise, participants were asked to complete a survey. This collected background

information and presented a set of statements for participants to mark their level of agreement

with. From 1, strongly disagree to 5 strongly agree. Statements were grouped in categories according to the assessed function of the UI. Table 4 presents a summary of the

most important questions from the survey. The complete survey is available in Appendix L.

Table 4 - Survey completed by participants

Data gathered and analyzed in the first user experiment resulted in a new version where the

major issues and concerns uncovered in the first iteration were improved upon. In the first

iteration, a single tap of the finger would state the function of the selected function, while a

second tap would select the function. Although, this served the purpose, the solution caused

frustration among the participants; they pointed to difficulties with tapping the correct area

and mistakenly opening elements. To overcome these issues, the navigation of the UI was

changed so that a single tap would still state the function of an element, but dragging the

finger to the right would open it.

The second iteration also introduced numbering and made changes to the naming of elements.

One set of numbers reflected the elements; current position in a list, while another set

presented the total number of elements in a list. A new list view, for lists with several

elements, was also introduced. The new list introduced a second row of letters, ranging from

A to Z, on the right side of the screen.



With the introduction of the new functions, the participants were asked to perform the same

exercises and survey once more. Additionally, they were asked to test the new list view and

answer a couple of new questions related to the added functionality, as seen in table 5 and 6.

Table 5 - Additional tasks presented to participants

Table 6 - Additional survey questions

3.6 Conclusion

To conclude the research, the data gathered from the two user iterations were categorized and

further reviewed. Important findings from the surveys and the evaluations were analyzed and

documented for further development of the prototype. The results are also presented in this

report; providing new knowledge to the research field of assistive technology and mobile

phone user access for people with vision loss.



CHAPTER 4: RESEARCH RESULTS

The previous chapter presented the selected research method and the steps taken to produce

the results. This chapter presents and evaluates the research results.

4.1 Results in regards to development

Following the Design research model of Vaishnavi and Kuechler (2007), the current project

has gone through the recommended stages in developing an alternative mobile phone UI. The

new UI provides users with a different way of interacting with the phone, compared to the

mobile phone manufacturers original indentation. Through the UI, any function or application can be utilized in an easier and more efficient way, making the phone more

accessible for people with vision loss. However, it is important to note that the applications

need to be compatible with a screen reader to be useful for people with vision loss.

Interaction with the new UI is made possible through two sensory systems, using sound and

tactile feedback. An organized and intuitive menu system provides the user with access to

relevant information and functionalities. The UI is designed in accordance with Nielsen's

(1993) usability principles, and adapted based on the findings from the framework in the

literature review. It also follows Brook's (1995) recommendation in keeping the number of

available functions to a minimum, avoiding potential confusion.

Figure 2 portrays screenshots from the UI running on a mobile phone; the key words represent

the different menus and functions available. The intended audience is people who are either

blind or that have strongly reduced vision; the text is for the benefit of the latter group. The

key words can also serve as reference points for external users, for instance if customer

support is needed. The font size is formatted according to guidelines from (NABP, n.d.).



Figure 2 - Some of the menus available in the UI

As mentioned, the UI is operated through sound and haptics. A single tap on the screen

initiates the UIs screen reader, which states the function of the selected element, while the gesture tap and drag finger executes the element. The gesture was implemented to simplify the execution of elements, and to reduce the number of incorrect selections. Moreover,

dragging a finger across the screen commands the UI to read out functions and at the same

time makes the phone vibrate when a new item is available. A back element is placed at the

bottom of every screen, with the exception of the main menu, allowing the user to easily

navigate to the previous menu.

To help the user create a mental map of the different menus and elements, the UI will also

state the level it is currently displaying and give feedback when successfully executing an

element. Furthermore, the UI can be customized to read out the selected elements position on a list, for example, number three of five.

Longer lists are displayed in the same way as regular lists, only with an added ribbon on the

screens right side, where an alphabetical list of letters is displayed. Figure 3 shows an

illustration of such a list. The letters refer to the initial letter of available elements, which

could be anything from applications to contacts. Dragging a finger across the ribbon causes

the screen reader to name the letters as the finger crosses them.

Figure 3 - List navigation in the UI



In addition to improved accessibility on touch phones, the UI is intended to work as an

assistive aid that can provide useful information, such as time, date and the weather forecast.

The UI is also designed, to provide navigation and to analyze objects or text using the phones camera; however, the features were not implemented in the current version.

The UI was tailored specifically for people with vision loss and it does not depend on existing

functionality or layout. This approach differs greatly from other solutions that typically rely

on a screen reader and the manufacturers UI implementation. By creating a new set of menu elements, the project did not have to focus so much on compatibility issues and could instead

focus on creating new functionality. At the same time, this allowed freedom to customize the

UI and optimize solutions for the visually impaired.

Design features that were of particular importance in the development of the new UI included,

the reduction of required gestures, a minimal learning curve and no reliance on hardware

buttons. The use of gestures was decided to be reduced since they can be hard to master (Amy

K. Karlson et al. 2005; Shaun K. Kane et al. 2008). In addition, by organizing menu elements

in a vertical layout, the UI operation would be similar to that of the Nokia phones, common

among people with vision loss. Finally, the dependency on hardware buttons was

circumvented, as this could result in compatibility issues with certain phone models.

Figure 4 shows one of several virtual prototypes created. Developing a virtual prototype

turned out to be of great advantage, working as an aid in communicating the vision and as a

tool for developers in finalizing and trying out the product. The virtual prototype resided in a

web browser and was operated with a computer mouse. It behaved like the UI developed for

the mobile phone, with fully operating menus and sound. In addition to the virtual prototype,

figure 5 shows a flow map that was created to visualize the different elements and the

relations between menus.

Figure 4 - Conceptual UI prototype for people with vision loss



Figure 5 - Flow map for navigation in the UI

The Android OS was chosen as the developmental platform for the UI, a decision based on

several factors. The Android platform is adopted by several phone manufacturers and is

currently the fastest growing mobile OS in the market (The Nielsen Company, 2010; Gartner,

2010a; Comscore, 2010). Hence, products developed for the Android OS have the potential of

reaching out to a large customer base. The Android OS was also a natural choice for the

developers, since it allows anyone to create a new UI for the phone. Furthermore, it is

considered a very open and flexible OS for application development (Ableson, 2009; Google

Inc, 2010a; Google Inc, 2010b).

Figure 6 presents a conceptual overview of the application, developed in Java for Android,

with support for Android version 1.6 and onwards.

Code Behind

Base Activity

Modules TTS Extended

Haptics

Activity 1 Activity XActivity 2 Activity 3

OnTouchListener 1 OnTouchListener 2 OnTouchListener 3 OnTouchListener X

UI - ViewLanguage packs English Norwegian Etc.

Application

Figure 6 - Conceptual model of application



Base Activity contains standard functionality and provides communication with the external

modules TTS Extended (Eyes-Free Project, n.d.) and Kickback (Ganov, n.d.); both are

modules used for creating voice feedback (screen reader) and vibration (haptics). All feedback

to the user, menus, voice and haptics are provided through functions called Activities. Each

Activity has an OnTouchlistener, which measures where and how the user is tapping on the

screen. All menu elements reside in XML format, which can easily be translated to different

languages. The current version supports English and Norwegian, but other languages can be

added as appropriate.

4.2 Results in regards to testing

Two iterations of user testing were carried out. A total of four people participated in the first

iteration, three women and one man with an average age of 50 years; a total of five people

took part in the second iteration, the same four with an extra man, making the average age

49.4. As mentioned, statistics from World Health Organization (2009), show that women over

50 make up the largest share of the worlds blind population; thus the projects sample is representative of this population. Of the participants, three were blind, one was diagnosed

with very low vision, although her interaction could indicate complete vision loss, the final

participant had low vision and could see objects in very close proximity.

According to Nielsen and Landauer (1993), only a small number of people are required for

usability testing; in fact five people should be sufficient to obtain adequate results (Richtel,

1998). Nielsen (2000) states that such a small number of users will provide better results than

a larger number; largely due to limited project budgets, the advantage of running several

smaller compared to one large test. This notion is in accordance with Design research, which

recommends several iterations of user testing.

The user tests were planned according to usability guidelines from Nielsen (1993), with

usability measured relative to certain users and certain tasks, and where every test should

define a representative and measurable set of tasks relevant to the users. For consistent

measurements, a survey should be completed subsequently. The mean value of the measured

attributes should exceed a specific minimum; on a Likert scale ranging from 1 to 5, the mean

value should be at least 4 (Nielsen & Levy, 1994) and 50% of participants should provide the

score of 5 (Nielsen, 1993). With two iterations of user testing, the ratings from the two

surveys can be compared to evaluate whether the improvements affected usability ratings and

if further improvements are still required (Nielsen, 1993).

Survey questions were categorized to the features they were set to evaluate, these are listed in

Table 7. Menu system and The solution are considered the two most important categories, as they measure how well the UI works and the success of the solution. The participants

received the same set of questions in both rounds.



Table 7 Summary of survey categories

Figure 7 presents the results from round one and two, with an overview of the mean scores for

each functionality category. Several questions make up a category, thus the mean scores are

averaged across questions as well as participants.

Figure 7 - Graph displaying mean scores from the first and second round of testing

The difference between mean scores from round one and two clearly shows responses were

more positive in round two, where mean were higher than four for all categories. The two

categories of particular interest, Menu system and The solution both presented fairly poor feedback in the first round, with initial mean scores as low as 2.3 for The solution and 3.9 for Menu system. According to Nielsen and Levy's (1994) criteria, these low scores would constitute a project failure. However, round two of testing provided a substantial increase in

positive feedback; thus, it can be concluded that the solution tested in round two is more

efficient and better received among the participants.

Menusystem

Phone Navigation Applications Information The solution

Round 1 3,9 4,4 4,1 4,0 4,3 2,3

Round 2 4,5 4,7 4,7 4,7 4,8 4,3

1,0

2,0

3,0

4,0

5,0

Ave

rag

e s

co

re

Comparing components in the UI



Figure 8 shows the results from the individual questions for the Menu system category. The mean scores illustrate the changes in the participants reaction between the first and second

iteration.

Figure 8 - Graph displaying a comparison between results from first and second round of testing on the menu system

The results show an increase in scores for all but one question, indicating that the second

menu system received a better acceptance than the first version. The reduction in scores for

the questions relating to the vibration function was clarified through verbal feedback from

participants. In the first round of testing, a greater reliance on haptic feedback was required to

navigation; however, in the second round of testing, the improvements facilitated the auditory

feedback and participants chose to rely more on sound than touch. These results are consistent

with findings from Rassmus-Grhn (2006), who found that people tend to rely more on the

sense they are accustomed to.

The menusystem waseasy to use

The systemmade the

touch phoneeasier to use

The spokenoptions were

easy tounderstand

It was easy tonavigatebetween

menus andsub-menus

The vibrationmade it

easier tochange

betweenmenus

Thecategorizatio

n of sub-menus were

logical

Round 1 4,0 3,7 4,3 2,5 4,3 4,5

Round 2 4,6 4,6 4,6 4,6 3,6 4,8

1,0

2,0

3,0

4,0

5,0

Ave

rag

e s

co

re

Menu system



Figure 9 presents results from the individual questions for the solution category. The mean

scores portray participants acceptance of the solution for rounds one and two.

*1 Three participants did not respond due to no prior experience with solutions for touch screens *2 One participant did not respond due to no prior experience with touch screens to compare against

Figure 9 - Graph displaying a comparison between results from first and second round of testing on the complete solution

The results show an increase in mean scores for all questions, indicating that the whole

solution received a better acceptance in the second round of testing. It should be noted that

some participants decided not to answer the second and third question, feeling that they

lacked the experience to give an informed response. However, all participants responded to

the most important question, I personally felt that the menu system enabled and improved my use of the mobile phone, with a great increase in positive feedback.

While the Design Research method does not require statistical analyses, these were

nevertheless carried out to shed more light on the changes in scoring that were evident

between the two versions of the UI. Effect sizes and differences between means were

evaluated across the categories previously outlined. Due to the small sample size, a certain

measure of leniency with respect to level of significance was judged necessary; this was

therefore set to 0.1. Thus mean scores, available in figure 7, were found to be significantly

higher for the second round of testing, compared to the first round, for the navigation

functions (t(3) = 2.45, p>0.1), for the information functions (t(3) = 2.45, p>0.1), and for the

solution functions (t(3) = 5.44, p>0.05). Effect sizes were judged to be high for all three

functions (r = 0.41; r = 0.40; r = 0.89; respectively). Effect sizes were similarly high for the

menu functions (r = 0.37) and the application functions (r = 0.41), but the differences between

means were non-significant (t(3) = 1.47, ns; t(3) = 1.48, ns; respectively). The remainder of

the analyses is included in Appendix M.M.

I personally felt that themenu system enabled and

improved my use of themobile phone

I prefer this solution aboveothers*1

I would use this solution inmy daily life*2

Round 1 2,3 2,5 2,3

Round 2 4,6 4,0 4,3

1,0

2,0

3,0

4,0

5,0

Ave

rag

e s

co

re

Acceptance of the solution



Nielsen (1993) suggests that participants should be asked for their subjective opinions, to

better provide an understanding of their satisfaction. Participants were asked to state their

thoughts and suggestions to improvements, following each round this feedback was used to

improve the application further. Feedback from round one is available in Appendix M.A,

while a complete list of feedback from round two is available in Appendix M.F. The most

informative comments from round two are summarized in Table 8.

Table 8 - Table showing subjective feedback from round two

The positive feedback from participants shows how much they appreciated the second

iteration of the solution; this is consistent with the survey results. The negative feedback

mainly focuses on the new functionality introduced in the second iteration, indicating that the

issues encountered in the first iteration had been solved and that a third iteration should aim at

solving the new problems. Comments pointing to the inadequacy of the vibration result from

the over-all improvements to the solution causing participants to shift their attention to the

voice, where the participants paid more attention. Participants suggestions for improvements are reasonable and well thought through and should certainly be considered in a third

iteration.

MAKING TO

Documents

Making Touch-based Mobile Phones Accessible for the Visually Impaired