Wednesday, August 02, 2006

Testing for Accessibility

This is an article from the August edition of the EuroSTAR Newsletter - STARTester, from Ruth Loebl. You can view the complete newsletter by clicking here and don't forget to subscribe to receive future issues.

Testing software for accessibility involves little more than imagination and common sense, but you have to pick the right standards, and then get to know some users.

• Lots of disabled people use computers, even people whom you might at first assume could not possibly use one.
• Many disabled people are not "disabled". My mother simply can't see as well as she used to, and has a bit of arthritis in her hands.
• Research commissioned by Microsoft indicated that in the United States, 60% (101.4 million) of adults from 18 to 64 years old "are likely or very likely to benefit from the use of accessible technology due to difficulties and impairments that may impact computer use.

"Even excluding people who are 65 or over, that's more than half the population – this isn't a niche market. So there ought to be a demand for accessible interfaces, although it's sometimes hard to detect. I'm encouraged by the improving legal situation" – check out the Code of Practice on the Disability Equality Duty in the Disability Discrimination Act 2005, example in para 3.46.

So how do all these disabled people use computers? For most, the answer is: in the same way that non-disabled people use computers, with a standard keyboard, mouse and screen.
As an example, one of the most important features of Windows is the ability to change the colour scheme and system fonts. While some of us just like a bit of variety in the colours we look at on the screen all day, for quite a few people choosing the right font and colour scheme is what enables them to read the screen at all. A few pre-set font and colour schemes are offered through the Accessibility Wizard (Programs, Accessories, Accessibility). More can be achieved through the Control Panel (Display, Appearance tab, Advanced button).

Accessibility testing should highlight when systems interfere with or disable these features that are provided through the operating system. It would be most annoying if your choice of colour scheme were ignored by a system that you have to use. All too often, oh dear, it is.

Access technology
Access technology is used where the effect of an impairment is such that an intermediary tool is needed to enable someone to use a computer.

• Partially sighted people who can't get by with an alternative font or colour scheme often use screen magnification software, which enlarges some or all of the screen contents, and provides other powerful features such as image smoothing and colour manipulation.
• People who have a reading impairment, poor literacy or dyslexia can use 'text-to-speech' software, where text highlighted with the mouse is spoken out loud. Sometimes, voice input is useful too.
• People who have a problem with their hands or arms will need adjustments or alternatives to the standard keyboard and mouse. This might simply be different hardware (a one-handed keyboard, a trackball, joystick or mouse pad) or a full speech recognition system.
• People who are blind use a standard keyboard but cannot operate a mouse. Speech output software conveys the contents of the screen, sometimes complemented by electronic braille output on hardware called a braille display. A full keyboard interface without reliance on the mouse is essential for effective access.
These access technologies are very powerful, but are often helpless when faced with really poorly designed software.

Software testing
As testers, you may want to know more about disability and access technology, and to meet some disabled people – I would certainly encourage this. For effective accessibility testing, it is important to involve as many different types of real users as possible, with different abilities, background, experience and so on. We often criticise software and web designers who don’t include people with disabilities in their testing processes.
Real user testing does demand a functional interface, but at this late stage it may be too late to change some aspects of the underlying design without compromising the viability of the whole software development project.

So as well as getting to know your users, we advocate the use of 'inclusive design' standards and guidelines at the earliest stages of interface design. The testable statements focus on the software itself, in isolation, independent of any particular user. They are intended to minimise barriers to both accessibility and usability, and to address many of the requirements of disabled users.

Standards and guidelines
Which standards and guidelines are most applicable for testing software? Three suggestions are below, and for our latest thinking on software and web accessibility, visit the RNIB technology site and follow the link to Software Accessibility. This information is due to be refreshed in early August.

ISO 9241-171 (formerly ISO/TS 16071)
The full title is "Ergonomics of human-system interaction – Guidance on software accessibility". It is in final draft, but as an internationally recognised standard written by professional standards-makers, we hope it will become a reference point alongside the guidelines and checklists that exist.
Within RNIB, we have adopted ISO 9241-171 as the basis for our software acceptance procedures, but it is actually too wide-ranging to implement in its entirety! For each software development project so far, we have had to extract a more workable subset of the full range of standards, tailored to the particular development and delivery platforms for that project.

IBM software accessibility checklist
The IBM Checklists are available for various technologies, including software in general as well as Web, Java and hardware. Each key point is explained clearly, and some information about implementation and testing is also given. They are much more usable and user-friendly than ISO, and free, but less comprehensive.

Section 508
Section 508 is US legislation to ensure that “electronic and information technologies” which Federal agencies develop, procure, maintain, or use, conform to standards designed to provide comparable access for people with and without disabilities. If you want to sell to Federal government in the US, these are the standards to apply. They don't include enough to make your system fully accessible, though.

Some messages to end with
• Accessibility is ultimately subjective, like usability. Effective testing depends on having a wide variety of users to test products in the later stages of the design process.
• Inclusive design is more objective, and applies to the software itself, independent of users. It can be tested from the earliest stages of a software design project.
• Inclusive design does not stifle creativity: good design for people with disabilities results in good design for all.

In my presentation at EuroSTAR 2006, you'll have the chance to see access technology in action, and some examples of accessibility standards and testing in the real world. See you there!

"Ruth Loebl has been with the Royal National Institute of the Blind (RNIB) for 13 years, working in the area of sight loss and technology."

No comments: