This site's design is only visible in a graphical browser that supports web standards, but its content is accessible to any browser or Internet device.

mmu | cerlim

NOVA > Final Report: RECOMMENDATIONS
or use the sitemap

13. RECOMMENDATIONS

13.1 Web page design

The design of web pages critically affects their accessibility. Even those which pass all the standard automatic checking services and adhere to the Web Accessibility Initiative Guidelines may still be difficult to access by those with visual impairments. Designers need to be much more aware of how layout and content considerations affect accessibility, and need to consider the navigational practices of users in their designs. We recommend that additional guidance be developed to encapsulate the lessons from the NoVA project and elsewhere, including the following:

Accessibility should be understood as requiring a number of different responses, such that sighted, partially-sighted and blind users can all use their choice of technology to access services.

Web pages should use a simple and logical structure, and this should be consistent throughout the site.
Web pages designed for enabling access to a variety of sources through a list of links (e.g. gateways to resources) should also provide a search facility, which should be available at the top of the page.

Such web pages should follow a logical order and the logic should be explicit (i.e. not just "A-Z" but "alphabetical by explicit criterion").

The number of links per page should be such that a user who has to listen to them being read out will be able to retain all in memory. Although we cannot recommend an exact limit (though it would be nearer 10 than 100), since this will vary from application to application, this test gives a useful rule of thumb for developers.

While WAI Guidelines advise on the use of frames, it should be understood that the critical issue tends to be the complexity of pages and the logical relationship between areas on the page. Thus a page containing frames may not itself be problematic; a page containing a number of frames which require a user to make mental links between them will be inaccessible. Designers should consider the steps needed to navigate within pages since this is the most crucial determinant of accessibility.

13.2 Assistive technology

Access to the most up-to-date software enables greater control: we recommend that libraries and museums should for this reason invest in up-to-date technology. However, not everyone has that access and many users may be using an older version or an obsolete browser, for example at home. Older versions should not be discarded without an audit of user requirements in each service. Thus providers need to be aware of the trade-off between functionality and familiarity. Web designers must also be aware of these issues and must not assume that everyone has the latest version of software available.

13.3 Staff training

We recommend that the findings of NoVA and parallel research should be used to inform accessibility training issues relating to assistive technology, for example in training library and museum staff in relation to open access computers with assistive technology. An accessible web-based tutorial, with examples of good and poor practice, might be an appropriate way forward.

13.4 User training

Experience of using assistive technology enhances performance: many of the features offered by assistive technologies provide short-cuts to help reduce the amount of time spent navigating or searching for information on the web. Not all users are aware of the features offered, and those that are may sometimes be unsure how to operate them. We recommend that greater emphasis is therefore given to the provision of training for both users and trainers.

13.5 Universal design

Some of the similarities between the two samples confirm the need for universal design rather than separate systems. The comment by a visually impaired user that text only options tend to be out of date reinforces this view. We therefore recommend that unless there are pressing reasons not to do so, all developers should adopt universal design as their underlying aim.

13.6 The appropriateness of digital approaches

Accessible gateways and portals such as those in development at the National Library for the Blind could pave the way forward for people with visual impairments. However mainstream portals also need to be addressing accessibility issues so that people with visual impairments are not forced to rely on "specialised" services alone, but are able to use the services and resources of choice.

It remains the case, however, that the Web will have to be made much more user-friendly to attract more visually impaired people to use it - for many it will still present too many barriers to be an attractive option. Service providers need to beware of making electronic access the only option. We recommend that service providers, in auditing the accessibility of their web pages, should explicitly consider whether non-Web alternatives - including human intermediaries - are needed.

13.7 Further research

We recommend that studies should be undertaken in the following areas:

Studies of how learning and cognitive styles impact on visually impaired users’ access to information resources.

Research into the effects of the usability of a web site or other online resource on the users’ motivation to continue searching, i.e. why does one user stop as soon as they find a reasonable answer, while another user continues to look further? Is this a usability or a user behaviour issue, and how does task impact on it?

The performance of screen reader technology should be assessed independently. It appeared during the NoVA tests that different screen readers performed differently on real-world web pages for reasons that were not immediately apparent. We were also surprised that screen readers do not provide effective spell-checking tools.

The feasibility of developing a software checking tool which provides a map of the stages necessary for searching by a user with screen reader technology (i.e. examining process issues) should be investigated.

A short study is needed to determine the effects of screen magnification on common character sets and to develop suitable advice both for web designers and for users, taking into account existing "smoothing" technology.

[ Previous Section: Conclusions ] [ Next Section: References ]