Interdependent Components for the Development of Accessible XUL
Applications for Screen Reader Users
Xabier Valencia
1
, Myriam Arrue
1
, Halena Rojas-Valduciel
2
and Lourdes Moreno
2
1
EGOKITUZ: Laboratory of HCI for Special Needs, University of the Basque Country (UPV/EHU),
Informatika Fakultatea 20018, Donostia, Spain
2
Computer Science deparment, Universidad Carlos III de Madrid, 28911, Leganés, Spain
Keywords: Accessibility, User Testing, Expert Review, XUL, Mozilla Firefox Browser, Screen Reader.
Abstract: Web applications based on XUL technology have reached great development. This technology enables
developers to easily create extensions and add-ons of Mozilla Firefox browser. It is essential to keep in
mind accessibility in the development of such applications in order to not discriminate user groups. In this
sense, standards and good practices have to be considered. Furthermore, User-Centred Design and Inclusive
Design approaches should be followed as they involve users with disabilities in the development process.
This paper presents an analysis of XUL accessibility guidelines created by Mozilla Foundation. An
accessible XUL application has been designed and developed based on the guidelines. User testing has been
conducted by two blind users revealing several important accessibility barriers. In addition, an expert review
process was carried on by a blind accessibility consultant. They all used JAWS screen reader. The results
obtained show that the existing guidelines conformance is not enough for ensuring accessibility of the
application. There are other factors dependent on assistive technologies and user agent that have to be
considered in the development of accessible XUL applications.
1 INTRODUCTION
The use of Internet has experienced a vertiginous
growth in the last few years. Users access the Web
employing diverse devices, modalities and
technologies. Due to this diversity, inclusion
approaches are necessary in order to provide full
accessibility to Web contents and avoid the
exclusion of some user groups.
Users with disabilities are the most affected by
accessibility barriers on the Web. They access the
Web using assistive technologies, for example, a
screen reader that relates content in audio to the
visually impaired. It is essential to develop
accessible web applications to ensure appropriate
assistive technology support.
Currently, research work regarding web browser
functionality augmentation is gaining attention.
Some examples of Mozilla Firefox add-ons are
(Greasmonky, 2012), (Stylish, 2013) and (Turn off
the Lights, 2013).
These augmented functionalities could be
utilized by all users only if accessibility aspects are
considered in their development process. Thus, a
comprehensive inclusive design paradigm for
augmented browser functionalities should integrate
User-Centred Design (UCD) methods and Inclusive
Design approaches in addition to accessibility
guidelines compliance (Lawton, 2007) (Newell,
2000).
The interest of “design for all” paradigm is
rapidly increasing in the community and several
efforts have been made in this way. In fact, there are
several organizations concerned with web
accessibility. They develop and maintain support
resources for complying with accessibility standards
such as guidelines. This is the the case of Mozilla
Foundation (XUL, 2013b). In addition, it provides
extension mechanisms to augment browser
functionality and develop application add-ons for
Mozilla Firefox through one specific technology
such as XUL (XUL, 2013a).
The objective of this paper is to analyse the
appropriateness of the set of accessibility guidelines
defined for XUL technology. For achieving this
objective, an accessible add-on for augmenting
Mozilla Firefox browser functionalities has been
developed. In the development process, a
65
Valencia X., Arrue M., Rojas-Valduciel H. and Moreno L..
Interdependent Components for the Development of Accessible XUL Applications for Screen Reader Users.
DOI: 10.5220/0004855200650073
In Proceedings of the 10th International Conference on Web Information Systems and Technologies (WEBIST-2014), pages 65-73
ISBN: 978-989-758-024-6
Copyright
c
2014 SCITEPRESS (Science and Technology Publications, Lda.)
comprehensive inclusive design paradigm has been
applied, including UCD approach and user testing.
User testing was performed by two screen reader
users. Several accessibility barriers were observed in
the testing which were later on evaluated by an
expert screen reader user who works as an
accessibility consultant. As a result, a review of the
XUL accessibility guidelines and conclusions of the
development process carried on are presented.
2 OVERVIEW
2.1 XUL
XUL (XML User Interface Language) is a XML
based language to create User Interfaces (UIs), in the
Mozilla platform. The main goal of the language is
to allow easy development of cross-platform add-on
applications which run on any Mozilla integrated
platform.
It separates the program logic from the user
interface components, facilitating the work of the
designers and programmers. This approach is also
applied in languages like (QML, 2013) or (XAML,
2013).
XUL is based on existing standards such as
XML, HTML, CSS, DOM and Javascript. In
addition, the Cross-Platform Component Object
Model (XPCOM) technology can be applied
(XPCOM, 2013) when operating system
functionalities are required.
2.2 Related Work
In the development processes, technological, human
and legislative aspects must be considered in order
to manage accessibility issues. Consequently, related
work from numerous disciplines should be taken
into account. In the standardization field, the W3C
ought to be highlighted along with the Web
Accessibility Initiative (WAI) (WAI, 2013). The
Web Content Accessibility Guidelines (WCAG) 2.0
(WCAG 2.0, 2008) is one of the most important
components, and is viewed as the official standard.
The ISO 9241-210:2010 standard provides a
framework for following and incorporating a UCD
approach into a particular context of accessibility.
Following methods that integrate usability and
accessibility in products design processes will
ensure that users with and without disabilities could
be able to access. This is the distinguishing
characteristic that User Sensitive Inclusive Design
(Abascal et al., 2007) has; the user with disabilities
is in mind. This work is focused on carrying out user
testing technique in order to validate the
accessibility included in a web application.
Several works related to XUL and accessibility
has been found in the literature. These research
works are related to developments of browser
extensions for Mozilla Firefox. All of them are
oriented to people with disabilities (Mirri et al.,
2011) (Hanson et al., 2005).
Nevertheless, very few articles have been found
that directly address the question of how to model
accessibility according to the WCAG (Moreno et al,
2013), (Martin et al, 2010). An interesting attempt
meriting particular mention is the Dante approach
integrating the Web Authoring for Accessibility
(WAfA) ontology (Yesilada et al, 2004) (Harper and
Yesilada, 2007) for the visually impaired into
WSDM (Plessers et al, 2005).
2.3 XUL Accessibility Guidelines
XUL language is based on web standards so its
accessibility guidelines do not differ too much from
previously published web accessibility guidelines.
The XUL accessibility guidelines are divided in
six different sections: Keyboard Access, Assistive
Information, Display, Human Computer Interaction,
Media and Custom Widgets (XUL, 2013b). Each of
one has a set of checkpoints to verify. For instance,
the guideline Keyboard Access defines eight
checkpoints that should be considered: one related
to tab order, another one to keyboard shortcuts and
so on.
All the guidelines and the related checkpoints
can be seen as an accessibility checklist to be
considered in order to evaluate the accessibility of a
developed add-on application. The XUL
accessibility guidelines document states a pass/fail
statement to each checkpoint which could be applied
in order to elaborate a checklist easy to verify by
developers at design time. Table 1 presents the
checklist for evaluating XUL accessibility
guidelines.
3 EXPERIMENTAL DESIGN
3.1 Object of Study
The object of study is to analyse the appropriateness
of XUL Accessibility Guidelines for the
development of accessible XUL-based applications.
WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies
66
Table 1: XUL accessibility checklist.
1.1 Logical tab order is provided
1.2
Keyboard functionality is provided for
inaccessible features such as the column picker
or added features such as column sorting
1.3
Keyboard alternatives are provided for
toolbarbutton functionality
1.4
Keyboard shortcuts are present for important
functionality
1.5
Context menus are triggered by the
oncontextmenu event handler
1.6
All mouse operations have keyboard accessible
equivalents
1.7
All scrollable elements are controllable with the
keyboard
1.8
Keyboard focus is maintained and does not move
unexpectedly
2.1
Alternative text is provided for meaningful
images
2.2
All windows, including dialogs and wizards,
have a descriptive title
2.3
Every form element has an associated label and
radiobuttons are encapsulated in a groupbox
3.1 System settings are maintained
3.2
Color alone is not used to convey meaning and
sufficient contrast exists between font color and
background color
3.3 Visual elements and containers resize gracefully
4.1
Help documentation is provided including a
description of keyboard shortcuts
4.2
Alerts are displayed using the alert scripting
function or the notification box element
4.3
Interactive elements are sufficiently large and
visible
4.4
Alerts are presented when the user initiates an
error. The user has the opportunity and
instruction to fix the error
4.5
User is informed of time limits and has control of
response time when appropriate
5.1 Transcripts are provided for audio tracks
5.2 Video is captioned and a transcript is provided
5.3
User has control over animation and is warned
about flashing content
6.1 Custom widgets provide accessible functionality
3.2 Experiment Context
Following XUL accessibility guidelines, a XUL-
based application has been developed. Its
accessibility has been evaluated using the XUL
Accessibility Checklist (see Table 1). The developed
application is an add-on for augmenting Mozilla
Firefox browser functionalities.
Figure 1: Demographic information form.
3.3 Sample
The add-on application for Mozilla Firefox includes
several pages with different type of web content.
For this study, two web pages containing forms
have been selected. The selection of such type of
application was due to the following reasons: the
diversity of elements included in it, importance of
accessibility in order to get to each question of the
forms and fill it in and the high interaction it
requires from users (Lazar et al., 2007).
The questions that users are required to fill in are
about demographic information, emotional aspects
and issues related to the design of the visited web
pages. These forms are presented to the user after
some specific time interval browsing in a website.
Figure 1 shows one the forms developed based on
XUL.
Forms were implemented using the following
XUL elements: Window, Radiogroup/Radio,
Textbox, Label/Description, Image, Button,
Hbox/Vbox/Box, DialogHeader and Spacer.
The developed forms share a similar structure.
We can resume this XUL structure in the following
way:
Each form is a Window element which has a Box
as a container of the form.
The title of the form is defined with a
DialogHeader element.
Description elements have been included for
providing explanations.
The input elements are one of the following: text
inputs, number inputs and radio inputs. In Figure 2
an extract of a XUL document is shown.
InterdependentComponentsfortheDevelopmentofAccessibleXULApplicationsforScreenReaderUsers
67
As it can be appreciated in Figure 2, a label has
been attached to the radiogroup through the control
attribute. This mechanism ensures that assistive
technologies would adequately present the form to
the user.
Figure 2: XUL extract for demographic information form.
The other input elements have been implemented
similarly, ensuring that all title values are unique and
that all labels were attached to the corresponding
input element
The implemented application was verified by
developers using the checklist presented in Table 1.
Not all the checkpoints are relevant to the developed
application as some of them are considered content
not usually present in common forms. The results
obtained in this initial evaluation are presented in
Table 2.
All checkpoints in the checklist were fulfilled
with the exception of item 4.1, the help function. It
was decided not to include this function in this
preliminary version. However, for upcoming
development iterations, the help function will be
incorporated. In conclusion, we would not expect
interaction problems with the application as almost
checkpoints were fulfilled. Therefore, accessibility
of the application was ensured.
3.4 Participants
Two screen reader users with more than 6 years of
expertise browsing the web were recruited for the
user testing: a woman (User 1) and a man (User 2)
whose ages were 30 and 40 respectively. User 1
considered herself as an intermediate Web browser
user whereas User 2 considered himself as an
advanced user.
They both use JAWS, but one of them (User 1)
uses it infrequently as she prefers to use VoiceOver
screen reader on Mac OS operating system. User 2
uses Windows and JAWS, but he does not usually
use Firefox. Even though, both are Mozilla Firefox
sporadic users.
The experimental sessions were carried out in the
same lab. They were asked to bring their laptop so
they used JAWS configured with their personal
preferences. The platform used was similar for both
users: Windows operating system (User 1 used
Windows XP and User 2 used Windows 7), JAWS
12 and Mozilla Firefox 22. Users were encouraged
to report any barrier they detect when interacting
with the XUL application. The sessions were
recorded with a camera located behind the user in
order to obtain information about the interaction.
The interviews were taped with a voice recorder.
Table 2: The XUL accessibility checklist applied.
1.1 The tab order works correctly
1.4 Buttons have shortcuts
1.6
All actions are accessible from keyboard and
mouse
1.7 The scroll can be done with the keyboard
1.8 The focus works as expected
2.1
The images have no alt text since are targeted to
other users and it has nothing meaningful for
blind
2.2 All windows have different titles
2.3 Labels are connected with their input element
3.1 Elements size has been set using “em” units
3.2
Elements colour have not meaning and the fonts
and background has enough contrast
3.3
Flex elements has been used to avoid unexpected
UI behaviours.
4.1 It is not provided in this draft version
4.2
Alerts are displayed using the alert scripting
function
4.3 Form is clearly differentiated
4.4 Uncompleted form or errors are advised
In addition, an expert evaluation was performed by
an expert screen reader user who has been working
as ICT accessibility consultant at least for the last
five years. All the evaluation was carried on at her
usual working setting, and reported her findings by
email.
3.5 Procedure
First, the XUL application was installed on users’
laptop. Then, all users were asked to perform two
tasks. The first one consisted on freely navigate
during five minutes in a concrete website
(www.discapnet.com). Then, the application
presented the first form to complete. The questions
in this form were related to their navigation
experience. The second task consisted on a search
task on the same website with a limited time interval
of ten minutes. Finally, the application presented the
second form containing questions related to
demographic data.
WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies
68
4 RESULTS
4.1 User Testing
4.1.1 User 1
This user experienced several barriers when filling
in the forms developed with XUL. The barriers
reported by the user are the following. Besides,
related XUL guidelines are indicated in each case:
Barrier 1.1: She was unable to know which answer
option was checked in the multiple-choice type
questions. This occurred when the user wanted to
verify that the selected answer was the correct one.
She navigated with the virtual cursor of JAWS and
it read the labels correctly. However, it only read
the value of the option and informed that it was not
checked even if it really was checked. Therefore,
we had to tell her which option was selected in
order to ensure that it was the desired one. (Related
guideline: 2.3)
Barrier 1.2: She experienced navigation problems.
Surprisingly, when she filled in a multiple-choice
question, JAWS focus was moved to the first
question of the form. Then, she had to navigate to
the next question from the beginning of the form.
This happened even though the program focus was
at the correct position. (Related guidelines: 1.1,
1.8)
Barrier 1.3: JAWS shortcuts navigation feature did
not work properly. Due to the difficulties she was
having, she tried to navigate through the form
using the JAWS shortcuts, like for instance, forms
or headings shortcuts but it did not worked as
expected. (Related guideline: 2.3)
Barrier 1.4: Problems for clicking on a button. She
was unable to find the button to continue. The
button was located at the end of the form and was
accessible using tab or arrows. However, JAWS
did not correctly detect this element. (Related
guideline: 2.3)
4.1.2 User 2
This user experienced similar problems reported by
User 1 except of Barrier 1.3 (he did not try this
mode of navigation). In addition, he reported other
barriers:
Barrier 2.1: Problems with text input questions.
JAWS was unable to detect a text area element
even if the focus was on one. Sometimes, he
reported listening a label that was not the correct
one. (Related guideline: 2.3)
Barrier 2.2: Alert messages were not adequately
presented. JAWS detected the alert windows but
not the containing text. Therefore, he only could
hear the default sound of the alert and “OK” button
but he missed the alert message. (Related
guideline: 4.2)
Barrier 2.3: Problems with numeric type inputs.
Firefox adds special controls for this type of
inputs. These controls are for entering the numeric
value using two small buttons inside the element,
one of them for increasing the value and the other
for reducing it. These controls entered into conflict
with his navigation controls. They are activated
with keyboard arrows which were the navigation
mode used by this user. Consequently, he was
unable to correctly enter his age. (Related
guidelines: 1.6, 2.3)
Barrier 2.4: JAWS read not existing options in the
UI. He reported us that JAWS sometimes read text
elements that were not in the UI. (Related
guideline: 2.3)
4.1.3 Discussion of Results
The user testing carried on indicates that there are
quite accessibility barriers in the developed XUL
application, even if accessibility guidelines have
been considered.
Some of the detected barriers are high impact
ones as users could not complete the tasks without
any assistance, for instance, barrier 1.4 (Problems
for clicking on a button) experienced by both users.
This barrier is related to activating the submit button
of the forms. Users could not get to these buttons so
they could not complete the tasks on their own.
These types of barriers are accessibility problems,
which should be documented in the accessibility
guidelines. In the group of accessibility barriers
should be also included the following ones: Barrier
1.1, Barrier 2.1, Barrier 2.2 and Barrier 2.3.
Other group of barriers detected in the user
testing were of moderate impact, as they do not
compromise the accessibility of the application.
However, they make the application less usable and
users can be disappointed inducing negatively in
their accessibility perception. These barriers should
be erased as well for ensuring a satisfactory user
interaction. For instance, Barrier 1.3 (JAWS
shortcuts navigation feature did not work properly)
user has other alternatives of navigation with the
screen reader so this problem does not compromise
the correct completion of tasks. Even though, the
inexistence of this barrier makes the application
more usable and user experience could be more
satisfactory. In the group of usability barriers should
InterdependentComponentsfortheDevelopmentofAccessibleXULApplicationsforScreenReaderUsers
69
be also included the following ones: Barrier 1.2 and
Barrier 2.4.
The detected barriers influenced negatively in the
user satisfaction when interacting with the XUL
application and they both needed around 15 minutes
to fill in each form. Considering that each form
consisted of ten short questions the time spent is not
acceptable.
There are remarkable differences between the
results obtained by each user. User 1 detected only
two high impact barriers (Barrier 1.1 and Barrier
1.4) whereas User 2 experienced more barriers of
this type (Barrier 1.1, Barrier 1.4, Barrier 2.1,
Barrier 2.2 and Barrier 2.3). They both used the
same version of the screen reader but the navigation
strategies applied can differ a lot from user to user.
In addition, our opinion is that User 1 is a more
experienced user. Our observation in the
experimental sessions revealed that User 1 has a
wide range of knowledge about functionalities of
screen readers. This observation differs from the
perception of their own expertise as User 1 defines
herself as an intermediate user and User 2 as an
advanced one.
All in all, accessibility guidelines should
consider all potential barriers independently of
assistive technology version used, navigation
strategies applied and user expertise level.
4.1.4 Improvements to XUL Guidelines
As can be seen most of the barriers are related to
those guidelines regarding form elements, like the
guideline 2.3 or the keyboard related issues 1.1 or
1.8.
Tagging labels with the corresponding control
seems to be not enough. It is essential to correctly
identify all questions and provide mechanisms in
order to alert user and assistive technology about the
existence of a list of questions or choices. Adding a
new XUL element to tag the whole form would
allow assistive technologies to handle better the
information and also ensure the correct behaviour of
the screen reader cursor.
Orientation of screen reader users would
considerably improve by applying a simple good
practice: informing at the beginning of the form
about the total number of questions in it and
numbering each question.
Regarding the keyboard, the added controls and
shortcuts could create conflicts with browser
controls or shortcuts as well as with assistive
technologies controls. In this sense, information
about the shortcuts available in the most used
assistive technologies would be helpful. This would
allow a better and more efficient navigation to the
user and would avoid unexpected technology
behaviour.
Finally, for the alert message issues, instead of
using the standard alert element, the
“notificationbox” should be used. In our preliminary
tests, this element seemed to work better with
JAWS. However, it may cause inconveniences to
other type of users such as those using magnifiers.
Another alternative solution could be to apply alert
functions on the active window. This issue requires
more investigation to carry on.
4.2 Expert Evaluation
Due to the distinct results obtained by the users and
in order to obtain a factual knowledge of the
situation; a review process was conducted by an
expert blind screen reader user. She is a consultant
on ICT accessibility.
She was asked to conduct a test with different
versions of JAWS, Firefox and operating systems. A
summary of the testing can be found in Table 3.
Results of the testing show the strong dependency
that XUL accessibility features have with the user
agent, the operating system and version of the screen
reader. For instance, checkpoint 2.2 “All windows
have different titles” only can be correctly detected
if the user interacts with the application on a
Windows 7, JAWS 13 and Firefox 23 platform. The
fulfilment of checkpoint 2.2 is not so essential for
user interaction (the user could complete a task even
if the title of windows are not correctly presented).
However, the same platform is required for
checkpoint 2.3 “Labels are connected with their
input element”. This checkpoint is essential if the
user is supposed to access and fill in questions
presented in a form.
The expert noted that Firefox generates
inaccessible HTML elements for the application
interface. Therefore, JAWS does not correctly read
them to the user. This is due to the way JavaScript
implements and interprets the DOM. Visually, the
element appears in the interface, but it is as if the
element was non-existent in the DOM. The screen
reader just ignores it.
In conclusion, the latest version of Firefox and
JAWS (at least versions JAWS 13 and Firefox 23)
seems to be the best combination for using the
developed XUL application. Nevertheless, it is
unusual that users have the latest versions of screen
readers. The same occurs with browsers. Most of
screen reader users use outdated browser versions
WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies
70
since the cursor mode of JAWS does not work with
the latest versions. Sometimes this mode is
necessary in order to access user interface elements
that cannot be read as usually.
Table 3: Correspondence of XUL accessibility checklist
with the expert review results.
1.1 Only in Firefox 23.
1.4 Better using Firefox 23
1.6 Only using Firefox 23
1.7 Only using Firefox 23
1.8 JAWS 13 and Firefox 23
2.1
When running the extension on Windows 7,
using Firefox 23
2.2 Windows 7, JAWS 13 and Firefox 23
2.3 Only using JAWS 13 and Firefox 23
3.1 Good feature for low vision using a magnifier
3.2 Good feature for low vision
3.3 Better using Firefox 23 and JAWS 13, JAWS 14
4.1 -
4.2 Better using Firefox 23
4.3 When refreshing the virtual buffer of JAWS,
refresh the page or when the form is being read
automatically
4.4 Only using Firefox 23 JAWS 13 and Windows 7
Some suggestions for improving the current version
of the developed XUL application were indicated by
the expert:
The text for the questions should be defined as a
header element.
The multiple-choice questions should define the
possible answers as a list element.
Users should be provided by an input text element
for introducing the answer. This suggestion would
increase the form response time, therefore it would
be less efficient and usable. However, it would be
accessible for screen reader users.
Implementing these suggestions would overcome
some of the most significant problems that users will
have probably to face to. Mainly when they do not
use the latest versions of browser and screen reader.
5 LESSONS LEARNED
From the results obtained in user testing and expert
evaluation, it is clear that there are extremely
important accessibility barriers in the developed
XUL application. The user testing can bring up
problems that have not been detected previously.
Findings reported in section 4 also highlight the
importance of testing applications with different
versions of Firefox and screen reader. There are too
many differences between versions and possible
combinations. Each combination has its strengths
and weaknesses. Using the latest versions of each
one seems to be the best solution. Even though,
many people do not update their software,
principally, due to the price of these updates. But
sometimes there are some functionalities that the
user is used to, that he cannot leave behind. For
these reasons, it is crucial to ensure that the
developed applications can be used in the wider
range of versions as possible.
User agent developers should consider the
accessibility barriers detected in this work. Universal
access to existing augmented browser applications
can be guaranteed only if inclusion design
paradigms are adequately defined and applied.
In the mean time, a transitory solution could be
to transform the application into a browser XUL
element to display the forms coded in HTML inside
the XUL code. As HTML accessibility issues have
been more analysed and considered in the last
decades. A great amount of documentation, tools
and methods exist for making accessible HTML
code. It would avoid the creation of the keyboard
scripts, because the keyboard behaviour would work
as expected and also reduces the workload. But the
main advantage is that it would make possible a
higher compatibility between different Firefox
versions and screen readers.
Anyway, XUL accessibility guidelines should be
reviewed in order to update issues regarding
keyboard navigation, assistive technologies
compatibility, forms elements tagging, etc. Not only
to ensure that all elements are accessible but
guarantee also the HTML-like behaviour so users do
not get confused or disoriented during the
interaction.
6 CONCLUSIONS
Web access to all users should be ensured, including
people with disabilities who use assistive technology
to access ICTs. Many organizations are concerned
about this issue, and they work towards accessibility
compliance. This is the case of the Mozilla
Foundation that provides accessibility guidelines to
apply when developers use their technologies like
XUL.
This paper presents a study of XUL and its
accessibility guidelines. We have developed an
application for augmenting browser functionalities
in XUL. The development process considered UCD
approach with the aim of creating an accessible
InterdependentComponentsfortheDevelopmentofAccessibleXULApplicationsforScreenReaderUsers
71
application with form content type. User testing with
2 legally blind users was carried on in addition to
accessibility guidelines conformance evaluation. The
developed application seemed to be accessible
according to XUL accessibility guidelines. However,
the results gathered in the user testing indicated
important accessibility barriers. Some of the
detected barriers make the application not operable
for screen reader users.
An expert evaluation has also been considered in
this paper. A screen reader user with more than five
years of experience as a consultant on ICT
accessibility has conducted the review. The results
reveal a strong dependence between platform used
(versions of user agent, operating system and screen
reader) and the accessibility barriers experienced by
users.
The findings of this paper should be considered
in next versions of XUL accessibility guidelines.
Some of them could be included as guidelines
whereas others could be considered as best practices.
This research was oriented to screen reader users
and XUL applications containing forms. In the near
future, there is a need of performing evaluation
studies with more content type, other groups of users
and other assistive technologies. Therefore, future
work will be motivated to the evaluation and
improvement of other XUL accessibility checkpoints
not considered in this work. This research work will
lead to ensure universal access to Mozilla Firefox
add-on applications.
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the support of
the Spanish Ministry of Science and Innovation
through Project ModelAccess (TIN2010-15549) and
the MULTIMEDICA project (TIN2010-20644-C03-
01).
EGOKITUZ is supported by the Department of
Education, Universities and Research of the Basque
Government (Eusko Jaurlaritza/Gobierno Vasco)
through Grant# IT-395-10.
REFERENCES
Abascal, J and Azevedo, L. Fundamentals of Inclusive
HCI. Design. (2007) Universal Access in Human
Computer Interaction, 4th International Conference on
Universal Access in Human-Computer Interaction,
UAHCI 2007, Held as Part of HCI International 2007,
Beijing, China, July 22-27.
Greasmonky (2012) http://wiki.greasespot.net/Main_
Page.
Hanson, V. L., Brezin, J., Crayne, S., Keates, S.,
Kjeldsen, R., Richards, J. T., Swart, C., & Trewin, S.
(2005). Improving Web accessibility through an
enhanced open-source browser. IBM Systems Journal,
44 (3), 573 - 588.
Hanson V. L., Richards. J. T., and Swart. C., (2008)
Browser augmentation, Harper S. and Yesilada Y,
Web Accessibility, Springer London, pp. 215-229.
Harper, S. and Yesilada, Y., (2007). Web Authoring for
Accessibility (WAfA). Web Semantics: Science,
Services and Agents on the World Wide Web. 5, 3, pp.
175-179.
Lawton S, H. (2007) Just Ask: Integrating Accessibility
Throughout Design. Madison,: ET\Lawton, available
at www.uiAccess.com/justask/
Lazar J, Allen A, Kleinman J, Malarkey C. (2007) What
Frustrates Screen Reader Users on the Web: A Study
of 100 Blind Users, International Journal of human-
computer interaction, Taylor & Francis, 22 (3), pp.
247-269.
Martín, A, Rossi, G, Cechich, A and Gordillo, S. (2010)
Engineering Accessible Web Applications. An Aspect-
Oriented Approach. World Wide Web, Springer US
13 (4), pp. 419-440,
Mirri, S, Salomoni, P, and Prandi, C. 2011. Augment
browsing and standard profiling for enhancing web
accessibility. In Proceedings of the International
Cross-Disciplinary Conference on Web Accessibility.
ACM, New York, NY, USA, Article 5 , 10 pages.
Moreno L, Valverde F, Martínez P, Pastor O, (2013).
Supporting accessibility in Web engineering methods:
a methodological approach, January, 2013, Journal of
Web Engineering , RINTON PRESS, INC, ISSN: 12
(3&4), pp. 1540-9589.
Newell, A.F.; Gregor, P. (2000) User Sensitive Inclusive
Design: in search of a new paradigm. Proceedings on
the 2000 conference on Universal Usability, pp. 39-44.
Plessers P, Casteleyn S, Yesilada Y, De Troyer O, Stevens
R, Harper S, and Goble C (2005) Accessibility: A Web
Engineering Approach. In Proceedings of the 14th
International Conference on World Wide Web (WWW
’05) ACM, New York, NY, USA, pp. 353-362.
QML (2013) http://qt-project.org/doc/qt-5.0/qtquick/
qtquick-applicationdevelopers.html.
Stylish http://userstyles.org/
Turn off the Lights (2013) http://www.stefanvd.net/
project/turnoffthelights.htm.
WAI (2013) http://www.w3.org/WAI/
WCAG 2.0 (2008). http://www.w3.org/TR/WCAG20/
XAML (2013) http://msdn.microsoft.com/en-us/
library/ms752059.aspx.
XPCOM (2013) https://developer.mozilla.org/en-US/
docs/XPCOM.
XUL (2013a) https://developer.mozilla.org/en-US/
docs/XUL.
XUL Accessibility Guidelines (2013b)
https://developer.mozilla.org/en-US /docs/ XUL_
accessibility_guidelines.
WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies
72
Yesilada,Y., Harper, S., Goble, C. and Stevens. (2004) R.
Dante annotation and transformation of web pages for
visually impaired users. In The Thirteenth
International World Wide Web Conference. ACM,
New York, NY, USA, pp. 490-491.
InterdependentComponentsfortheDevelopmentofAccessibleXULApplicationsforScreenReaderUsers
73