Computers are not conventionally "clever". They perform calculations very quickly but they don't understand the problem that they are trying to solve. This limitation is particularly noticable when attempting to process language and images; although significant research has been undertaken in this area, the holy grail of a computer which is able to understand content and perform any cognitive process based on it remains quite distant.
Meanwhile as the web has exploded in popularity, the number of ways of accessing the information it contains has diversified. Originally (and still in the most part) people use graphical web browsers which display the content on a screen. But people also access the web on handheld devices like mobile phones or PDAs, and some people use software which reads the text of a web page out loud. Some people use web browsers which don't display images, colours or fonts.
Because there remains no way of intelligently, automatically converting the content which I see with my eyes into a form that someone else might listen to, for example, it is important to ensure that the content we wish to convey is available in a form can be easily adapted for each medium. It is a trivial task to add visual information to purely semantic content because the intended understanding of the content has been provided by the author. Similarly, audible information such as different voices, pauses and tones can also be added based on the semantic information.
The process of providing that semantic information is the basis of accessibility; additionally there are some purely pragmatic constraints related to particular disabilities.
The Disability Discrimination Act 1995 (part III) enforces access to goods and services for people with disabilities. Websites which are inaccessible to those with disabilities fall foul of this act.
The most often mentioned deficiency is failure to accomodate blind or partially sighted people who browse the internet with text-to-speech software which reads webpages aloud. As an example, screen reader software is unable to describe an image. If text is contained within that image it will be inaccessible to people using such software. In some cases this can mean that navigating your site becomes impossible.
Search engines can be an effective way for potential customers to reach your website. However, it must be remembered that search engines are automated pieces of software and must be able to read your page in order to properly determine what content it contains and therefore in which searches it shows up.
Search engine robots (also known as crawlers or spiders) traverse your site and millions of others by jumping from link to link. They do not fetch images. If your site were to use a glossy graphical logo for the company name with no alternative form available, then a search engine would not be able to return your site even if someone were searching for your company name.
Accessible websites cannot place restrictions on what people can do with the content upon them. They can be printed, copied and quoted from freely. This is often extremely useful for people who are looking to purchase from you, and in other cases attributed reference to your website may persuade people to visit it.
Note that this does not detract from the copyright of the content on your website. This remains your intellectual property and copyright infringment claims can and should still be persued.
The W3C maintain the internet standard for web accessibility, the WCAG. It consists of a number of tests broken into three levels according to importance. If a website passes all the tests for a given level and the previous levels, it can claim that level of conformance.
The levels are as follows:
The tests cannot generally be performed automatically, for the reasons previously stated: they would require a computer to have an understanding of the content, which is only a very distant possibility.
