The European Internet Accessibility Observatory (EIAO) - Aspects of Internationalisation

M.H. Snaprud, Agder University College/Faculty of Engineering and Science, Grimstad, Norway

F. Aslaksen, VISTA Utredning AS, C. J. Hambros plass 5, 0164 OSLO, Norway

Main Points

A European project to design the machinery of an Internet Accessibility Observatory has been outlined. ¹
Accessibility is a design goal recognised by international organisations and national authorities, to secure an inclusive design of the Information Society.
A frequently updated overview of the accessibility status may trigger competition among counties and business sectors leading more rapid improvement.
The measurement results will be made freely available on the web to enable different groups including policy makers and users with special needs, to analyse the status and the impact of measures taken.
Open Source Software has been chosen as the basis for developing the accessibility assessment tools. This will enable participation from developed nations and from the majority world to co-operate on more equal footing.
Considerable language barriers, limit the use of electronic information. In Europe, English is the most frequently used foreign language. However, less than half of the EU population are able to speak English.
The Open Source license and the planned design for localisation, and distribution will enable the inclusion of additional sites, dissemination of measurement results and facilitate the forming of a global network with distributed responsibilities for accessibility monitoring.

Background

As the amount of information and the number of services available on the web increase, the "digital divide" between those who can access online content and those who cannot is growing. Appropriate use and distribution of new technologies can bridge the divide and enable new user groups to participate in the online community. In particular, people with disabilities and people in developing countries may also benefit from the development.

The use of the World Wide Web has in less than 10 years become an important part of many peoples life. The web has changed how we work, how we study, communicate, buy and sell goods and service etc.

Standards and guidelines such as Web Content Accessibility Guidelines[1] (WCAG) are needed to build an information society. These guidelines should be followed to ensure that we build an information society for all. In this paper we present some early results from quantitative web accessibility assessment of public web sites covering some European sites representing, central government, and newspapers.

Unlike the built environment, where barriers for groups of users are difficult to remove, the web is undergoing a rapid persistent change where new sites are created and existing ones are redesigned continuously. Therefore the web has a great potential for being inclusive and contribute to a more inclusive society.

However, today a lot of web sites are inaccessible for large groups of users, including blind people, partially sighted people, people with learning difficulties, physically impaired people and people who are deaf. Bert Massie, chairman of the Disability Rights Commission, says in the foreword of The Web, Access to Inclusion for Disabled People ², "This report demonstrates that most websites are inaccessible to many disabled people and fail to satisfy even the most basic standards for accessibility recommended by the World Wide Web Consortium."

To assess the accessibility of Web sites and web services, manual checks are frequently carried out by means of a check list or a talk-aloud protocol.

Some European countries conduct manual checks of public of Web sites. Due to the cost the checks are not carried out more often than yearly and with a limited scope of assessments. Moreover, the criteria for the assessment are not coherent among different states and there are even examples (e.g. Norway) where the criteria are changed from one assessment to another making any trend analysis impossible. The current practice lead to substantial problems with obtaining coherent results.

To overcome some of these problems a number of automatic benchmarking tools for analysing Web pages have been developed such as the HTML validator ³ Bobby ⁴, Dr Watson ⁵, A-Prompt ⁶, Wave ⁷, or Validate ⁸. Many of the services are available on the web and can be used for free, whereas some will require payment. These tools typically produce a report, listing deviations from standard or indicate potential barriers for accessing the content or services of a web page.

These tools typically do not support storage of data in a database for further analysis, although stored data is needed both to compare different sites effectively, and to monitor development over time.

The percentage of central government web sites that conform to the WAI ⁹ accessibility guidelines (level A) is listed as one of the indicators for eEurope Benchmarking. Still, there is no systematic overview of the actual use of accessibility guidelines in Europe.

The EIAO Project Outline

The overall objective of the EIAO project is to contribute to better e-accessibility for all citizens and to increase use of standards for representing on-line resources.

The project will develop a prototype technical machinery for a European Internet Accessibility Observatory (EIAO) consisting of:

A set of web accessibility metrics.
An Internet robot for automatic and frequent assessment of web accessibility and deviations from Web Content Accessibility Guidelines (WCAG).
A data warehouse [1] providing on-line access to collected accessibility data.

The tools for collecting, assessing and disseminating data will be continuously improved, based on feedback from end users (e.g. policy makers, associations of disabled people etc.) and user tests to sharpen the relevance of the automatically collected data.

In this way, the EIAO can provide benchmarking of websites. Such benchmarking will hopefully lead to competition between countries, organisations, developers etc. It may also be used to provide background for political action.

The project proposal partners are universities and private companies from six European countries. There is a reference group where NGO's and public bodies working for better e-accessibility are represented. The project will be open to the scientific society by information, invitations to discussions and active use of the projects Internet site. This is encouraged as the project is based on Open Source Software and the project as a whole intends to be open to users, scientist etc. to discuss development and get input to the project. Also the project's objective is to provide the public and decision makers with quantifiable information for promoting eInclusion and for design of policies for a more including Information Society. To achieve this impact the project is designed to be as visible as possible.

Open Source Based Development and Accessibility

All software developed in the project will be released under an Open Source license. The software will be the technical machinery used to drive the prototype Observatory. The prototype will only validate open standards. To allow for validation of vendor owned formats the LGPL ¹⁰ license is proposed, which can use plug-ins released also under other licenses.

The Open Source license ensures access to source code and seem very natural for software developed for assessing accessibility:

access to code for verification of assessment method. Transparent validation methods seem to be important to raise the credibility in the result.
freedom to improve and use the validation methods. Further installations of the system will enable additional teams to experiment and to contribute.
localisation (web interface, development guidelines...). Local language important for community building and for acceptance in additional countries.

Considerable language barriers limit the use of electronic information. In Europe, English is the most frequently used foreign language. However, less than half of the EU population are able to speak English[2]. Automatic detection of user preferences language preferences is available. However this feature is seldom used.

Local language support is crucial for involvement and for the outreach of the accessibility effort. Therefore the EIAO tools will be designed for handling multiple languages and to support translation. Some successful examples showing how localisation can be organised include both of the two major Linux desktop application GNOME ^{11, 12} and KDE ¹³ are being translated into 87 different languages by translations teams. The web browser Mozilla has 104 ¹⁴ translation teams.

We plan to design a localisation package containing documentation for translation of the web interface to the datawarehouse and the guidelines for developing Web Accessibility Metrics (WAM) plug-ins.

The translation could technically be handled as a web interface for entering the keywords used in the web interface (eg. By listing the needed keywords in the already available languages and requesting the translator to enter the translation to a new language).

Online feedback forms will be developed to allow for feedback from users, encouraging them to comment on additional sites to monitor and on possible improvements in particular regarding usefulness of the data and on cultural issues.

Selection of What to Assess

All kinds of web sites can be assessed. In the EIAO project we plan to assess only sites using open standards, including HTML, XML, SVG etc., and disregarding sites based on e.g. Macromedia Flash.

Examples of quantifiable indicators include the share of web sites declaring what version HTML/XML they are using. On a correct web page this will be the first line in the document.

Further quantifiable indicators include the number of deviations from the declared document type, share of images with an informative alternative text for the images on Web site, and the deviations from the HTML standard and from the WAI guidelines. All this will be collected and stored in the datawarehouse.

Selection of Sites to Assess

For a first preliminary study we used a selection compiled by European Union representing the European Governments on-line ¹⁵. More about the results from this initial survey was presented at the ETSI conference Access for All ¹⁶ and at the EMBEC 2002 [3].

For the selection of sites we intend to focus on public services, governmental and local authorities, and some private service providers like banks, public transport etc. Clearly, the organisations responsible for a site need to be informed about the monitoring before results are published on-line.

The project will also rely on local contributors in different countries to provide and maintain links to relevant websites to monitor. The EIAO localisation package should also include an outline of how to organise local maintenance of a directory of websites to monitor for a country (or smaller geographical unit). Lessons learned from the Open Directory Project (ODP) ¹⁷ project will be valuable here.

Call for Volunteers

To improve the impact and enhance the scope of the project feedback and participation is encouraged. You may consider to apply for local funding based on the potential benefits from participating in the project. In particular we welcome contributors to:

website directory maintenance
volunteers to translation teams
governmental representatives providing input on useful features and user interface design
people with special needs providing feedback on user interface design, and on measurements

In this early stage of the project there will be more room for exchange of ideas to improve the current concept.

Conclusions

An Internet Observatory could provide a vehicle to enable more inclusive and sustainable design of assessment tools and thereby create more awareness of Web accessibility and universal design.

The outlined research project will develop prototypes of the basic components of an Internet Accessibility Observatory, and not a full-scale implementation. The technical development into a full scale Observatory will then mainly focus on streamlining, and adjustments for scaling up the number of sites. The next challenges will be related to organisation and financing and legal issues. The tools will be freely available and built for easy localisation, allowing use in countries also beyond Europe. It is designed to be a general tool and can cover websites all over the world. It will be based on Open Source license ensuring the free use and distribution of the tools.

Resources

[1] T. B. Pedersen, C. S. Jensen: Multidimensional Database Technology. IEEE Computer 34(12): pp 40-46. 2001.

[2] European Parliament, Multilinguality and the Internet, Briefing Note N° 2/2001, PE nr. 297.557, 2001.

[3] M.H. Snaprud, C.S. Jensen, N. Ulltveit-Moe, J.P. Nytun, M.E. Rafoshei-Klev, A. Sawicka, and Ø. Hanssen, Towards a Web accessibility monitor, Proceedings of IFMBE, EMBEC '02 Vienna, pp 1718-1719, 2002.

Notes

Provided negotiation with the European Commission is finalised as planned, the kick-off is scheduled for the end of 2004.
The Disability Rights Commission: The Web, Access and Inclusion for Disabled People, April 2004
http://validator.w3.org/
http://bobby.watchfire.com/bobby/html/en/index.jsp
http://watson.addy.com/
http://aprompt.snow.utoronto.ca/
http://www.wave.webaim.org/index.jsp
http://www.htmlhelp.com/tools/validator/
WAI: http://www.w3.org/WAI
http://www.gnu.org/copyleft/lesser.html
http://developer.gnome.org/projects/gtp/teams.html
http://developer.gnome.org/projects/gtp/teams.html
http://i18n.kde.org/teams/
http://www.mozilla.org/projects/l10n/mlp_status.html
http://europa.eu.int/abc/governments (as accessed 2003-02-15)
http://www.etsi.org/cce/proceedings/7_3.htm
http://dmoz.org/

This research has been partly funded by the Research Council of Norway under the contract: 160564 - Pilot project - An Internet robot for Assessing the Accessibility of on-line content.

Top