This paper presents a case study of accessibility evaluation which reports the main drawbacks concerning the accessibility of Web sites of French and Belgium local administrations. We measure the accessibility of information published on Web pages, acceptance of standards of accessibility (i.e. W3C's WAI) and availability of services provided on-line. Our results allow the identification of the most frequent accessibility problems on these Web sites and measure the maturity concerning the accessibility by local administrations in France and Belgium.
In more recent years Accessibility became a legal
requirement as many countries have enacted for
Accessibility responsibility of content published on the
Web. Despite important improvements on accessibility
legislation (both Europeans and Nationals spheres) and
incentives to make e-government Web sites accessible,
there is a slow progress on the design practice of
administration Web sites. In Europe, the European
Council encourage state members to enact laws for
accessibility of public Web sites at all levels of
government. Many member states such as France [8],
Germany [
In this paper we focus on Web sites of French and Belgium public administrations. We review the current regulations and regulation mechanisms (i.e. recognized standards) in France and Belgium. By the means of a case study of accessibility evaluation we compare the maturity level concerning the accessibility of local administration Web sites in these countries. In particular, we measure the accessibility of information published on Web pages, acceptance of standards of accessibility (i.e. W3C's WAI) and availability of services provided on-line.
One of the five priorities for the new i2010 EU
eGovernment Action [
In France it started slowly in 1999 with an internal recommendation based on W3C/WAI standards and stipulating that “people in charge of Web sites should pay attention to provide accessible content for all users, in particular for impaired users”. This recommendation was not voted earlier than February 11th 2005, law number 2005-102 (“for equal rights and equal opportunities, participation and citizenship of impaired people” 7). It is important to notice that it legislates on both physical accessibility and electronic accessibility. The original proposal stated that all public Web sites should be fully accessible within 3 years. Several decrees have been published since then but the parts concerning the electronic accessibility (articles 47 and 78) have not yet been officially published and still now they are subject of revisions! The most recent decree of the law (December 13th 2006) extends the deadline for making e-government Web sites to conform with ‘international standards’ of accessibility (2 years for 7 Law N° 2005-102: “pour l’égalité des droits et des chances, la participation et la citoyenneté des personnes handicapées”. communication agencies, 3 years for administrative Web sites).
In Belgium, a law of February 2003 prevents “any kind of discrimination based on physical handicap”. This regulation clearly addresses furniture and services offered by public sites. However, there is no legislation addressing specifically the accessibility of electronic content. In April 2003, the Walloon government (one of the 3 major regions in Belgium) adopted a series of actions to make public Web sites accessible in the Wallonia region. It also has been stated that some Web sites considered critical should be accessible before 20068 according to the priorities below: • Priority 1: public Web sites devoted to social action involving any kind of impaired users; • Priority 2: Web sites devoted to information of citizens at the large or Web sites of Walloon government; • Priority 3: Web sites devoted to job search. Local administration is not addressed by these priorities.
One of the drawbacks of current legislation is the lack of official referential for accessibility. W3C/WAI is the most recognizable standard for accessibility worldwide but different EU countries have proposed their own standards. Figure 1 present the logo used to identify currently used standards in this survey. Hereafter we provide a view at glance of these standards
W3C/WAI related AccessiWeb
BlindSurfer
Anysurfer
Euracert
W3C/WAI The World Wide Web Consortium (W3C) is at the origin of the Web Content Accessibility Guidelines (WCAG 1.0) [11]. The recommendations published on May 5th 1999 contained 14 guidelines and 65 checkpoints having 3 levels of priorities. The level of conformance with these priorities levels (known under the labels A, AA and AAA) have been widely followed by other standards. WAI certification is often associated to technology compliance certification (e.g. XHTML and CSS) which can ensure cross-platform accessibility. The certification at some level is free of charges using automated tools developed for this purpose.
The WAI have been working on the development of the WCAG 2.0. However, this second version has been largely criticized by the community mostly because it introduced directives which cannot be automatically inspected (e.g. all content should be perceivable, content should be understandable, content should support current and future technologies). In some aspects these new directives are ambiguous and difficult to apply even by experts. So as far the WCAG 2.0 is under revision, only WCAG 1.0 remains the most universally recognized standard of accessibility.
The AccessiWeb label (http://www.accessiweb.org) was created by the association BrailleNet9 in France. In addition to the certification, the working group on accessibility proposes training for Web developers. AccessiWeb is based on WAI recommendations but it extends the evaluation to 92 checkpoints. Similarly to W3C/WAI, AccessiWeb certification includes three levels named Gold, Silver and Bronze. The certification process consists in 2 steps: a pre-audit performed by the applicant and an evaluation performed by AccessiWeb experts. The cost of the certification ranges from 1900 € (Bronze) to 2600 € (Silver and Gold). Although the certification is not free AccessiWeb provides tools and guidelines in order to help obtaining the certification. These resources are freely available from AccessiWeb Web site.
Recently, another initiative called RGAA (Référentiel Général d’Accessibilité pour les Administrations) has arisen from french accessibility experts and local administrations’ users. This set of guidelines is based on WCAG 1.0 guidelines and the Unified Web Evaluation Methodology (UWEM 1.0) [9, 10] which aim is to provide a set of guidelines and a standard procedure for manual and/or automated accessibility inspections. The particularity of RGAA 1.0 is that each guideline is associated with unit tests. These tests are easier to verify even for non experts and solve the problems of some high level guidelines in other existing set of guidelines. However, as this is a recent initiative the RGAA 1.0 document is still a working draft.
The label AnySurfer (http://www.anysurfer.be/) is awarded by the Belgium association “Oeuvre National des Aveugles” (ONA) after audition by certified experts on accessibility. The costs of AnySurfer certification varies according to the size of the Web site (e.g. 120 € for a Web site of 100 pages). Both French-spoken and Flemish-spoken communities of Belgium agreed to this label. From a technical point of view, AnySurfer is strongly based on the W3C/WAI recommendations. Since July 1st 2006 the label AnySurfer replaces the label BlindSurfer conceived for the same purposes.
The Euracert label (http://www.euracert.org/) is an
initiative of the European Council initiatives for
promoting the accessibility in EU countries [
Local administrations offer a great variety of services which have a huge impact on daily life of citizens. For example inscriptions on schools, reporting problems with traffic, declaration of birth/death, request of identity card and passport, and so on. In addition to these sources local administration could be considered primary sources of information about tourism, social activities, demographic data, and general public services provided to the community. Currently there is a lack of information concerning local administrations Web sites. On one hand there are few studies assessing the maturity level of online services such as information, one-way interaction (i.e. download all forms), two-way interaction (i.e. online forms), and transaction (i.e. full electronic case handling). On the other hand we don’t know how accessible these services are. Hereafter we present the preliminary results of two studies of assessment of local administration Web sites in France and in Belgium. Although the instruments used for the survey are different, both highlight the maturity level of these Web sites.
Sixty-four (64) local administration Web sites were inspected by master students on Human-Computer Interaction. Both manual inspection (i.e. human judgment of Web sites) and automated inspection by tools (i.e. CSS and HTML validator10) were employed. The inspection was limited to 10 different Web pages of each Web site, which in most case covers all the major sections available. The checklist is composed of the following main sections: i) General questions, ii) Conformance with W3C/WAI guidelines, iii) Cognitive legibility of information, iv) Adoption of labels and v) Availability of online public services. 10
http://validator.w3.org/ validator/ and http://jigsaw.w3.org/css
Checkpoints Nof. % site s General questions Is the language clearly stated in the main page? 6 9,4 Is the Web site multilingual? 5 7,8 Does the Web site propose different presentations? 1 1,6 Is there human assistance proposed to deaf users? 1 1,6
Conformance with W3C/WAI Are CSS style sheets employed for presentation? 32 50,0 Do CSS style sheets follow W3C guidelines? * 5 7,8 Are (X)HTML pages in conformance with W3C guidelines?* 2 3,1 Is alternative text provided for all images? * 6 9,4 Are tables misused for the layout? 24 37,5 Are pages equally visible with different navigators? * 54 84,4
Cognitive legibility of information Are summaries provided for long texts? 27 42,2 Does layout group similar/related information? 47 73,4 Do paragraphs present a single idea? 49 76,6 Are keywords clearly visible? 35 54,7 Is it possible to zoom in complex graphics? 16 25,0 Does links’ label reflect the content of referred page? 50 78,1 Adoption of labels Is the Web site awarded by a label? Which one? 0 0,0 Are W3C guidelines respected but not referred?* 0 0,0 Are there other standards taken into account? 3 4,7
Availability of online public services Does the Web site provide some kind of online interaction? 10 15,6 Is the number of steps in transactions always informed? 4 6,3 Do labels provide meaningful information to fill in form fields? 16 25,0 Are the form fields automatically checked? 7 10,9 * Checked with help of automated tools. The results of the survey demonstrate a huge lack of accessibility of Web sites. Half of Web sites use CSS style sheet for presentation and only 5 of them (7,8%) follow the corresponding W3C recommendations. One of the most basic guidelines concerning the use of alternative text for images is strictly (all pages on the Web site) respected by no more than 9,4%. The results are even worst when looking at Web sites that do not provide meaning labels for links (78,1% don’t). It is noteworthy that none of 64 Web sites was awarded by a label but 2 of them follow optimize the design for technology-oriented standards (i.e. FireFox and Internet Explorer 6.0) and a third one implement language of signal standards for deaf users (i.e. WebSourd LSF). One can notice a small but meaning number of Web sites moving towards providing an online service (15,6%). Some ergonomic guidelines are taken into account such as providing meaningful labels to form fields (25%) but it is clear it is not enough for Web sites intended to a very large public.
The Wallonia’s agency of Telecommunications (AWT) in Belgium started in 2004 to evaluate the 262 Web sites of Wallonia local administration. Their results shown that less than 5% considered accessible11. of
Web sites could be
Different techniques were employed in the present
survey, such as: i) Ergonomic inspection with the WAI
set of guidelines, ii) Visualization of Web sites using
the browser Lynx (for checking legibility for blind
users), iii) Automated inspection using free available
tools for accessibility inspection such as DESTINE [
The general accessibility of these Web sites is very disappointing. We have found serious problems with many labels on links not referring to the proper page. In most cases it is impossible to navigate using just the keyboard, which is a main requirement for blind users. The extensive use of JavaScript and Adobe Flash technologies only accentuates the accessibility problems with links. Surprisingly, only a few Web sites provide extensive use of multimedia documents and only 1 Web site could not be properly read using different browsers. The main results are presented in Table 2 and show the main defects found on Web sites.
Checkpoints Size of Web site in number of pages Number of sites supporting online interaction Web sites presenting problems with labels on links Content presentation in different languages Inaccessible forms Inaccessible tables Lack of support for navigation based on keyboard only Lack of alternative text for images or dynamic objects (e.g. Applets) Web sites judged quite accessible Number of sites
7 small 6 medium
2 large
Only 112 of 15 Web sites provides an average accessibility but it presents too many links on pages which reduce the legibility of content and the navigation using the keyboard. It is noteworthy that quality of the Web site has no relationship with the size of the city. Web sites of big cities such as Liege and Brussels have many serious defects.
The results of the survey demonstrate that most of Web sites (6 of 15 sites in the survey) provide at least one online service to citizens. The number of online services ranging from 1 to 14 per city. Most of the services, however, are limited to the download of forms. The main problem identified is that such forms are difficult to find in the hierarchy of the Web site. 11 By an informal communication, the AWT said that another survey is in progress but by the time of submitting this paper the results have not been published yet. 12 Sambreville, available at: http://www.sambreville.be/
The survey presented is a first step towards a comprehensive study of accessibility and maturity of online services provided by local administration in France and in Belgium. Even if Web sites are including more and more online services, the lack of accessibility is evident in most visited Web site.
None of the Web sites of local administrations in Wallonia are awarded by the label AnySurfer. The absence of labels are also remarkable on French Web sites. The poor accessibility of Web sites is an strong evidence of the lack of training of Web developers on Accessibility and Ergonomic matters.
It is also clear that current legislation and official labels for accessibility have a minor (or none) impact on the design of Web sites. French regulation is more detailed compared to Belgium but no different impact has been observed on the Web sites in terms of conformance with labels.
However, the absence of a label does not mean lack of accessibility. On the other way round, the presence of a label does not itself ensure the accessibility of the Web site as labels are awarded over a short period of time when the site is audited. Inappropriate updates on content might include defects affecting the accessibility. Although labels give a good picture of the quality of the Web site its validity is limited to a short period of time. A more detailed study should be carried out to determine why designers are so few concerned by labels.
http://www.euracert.org/