Accessibility findings vs recommendations

This article on newspaper site accessibility dropped into my inbox, from the author hoping to get some coverage from accessibility related sites. Whilst the aim is good (testing sites to increase the awareness of accessibility), I wouldn't be happy with the recommendations.

The meat of the article is the findings from John (a retired research worker from Sussex) and his use of the Independent Newspaper site. For example:

The text-heavy appearance of Independent.co.uk and its lack of a left-hand navigation bar aided John's ability to tab through the content, as he said he would rather be tabbing through individual links to articles than an excess of section headings.

This is a little confusing to a web developer, as in the tabbing order, there are 10 section heading before the content. In the source code, there are almost 200 links in the top navigation, one of which is the 'Front pages' link referred to later in the article. That doesn't really match the conclusion drawn about accessibility, and it's just a red herring to infer from this that the text nature of the site is a good thing.

What would be really useful for that situation is a skip link, or a better method of drop-down navigation.

There are some valid points on pop-ups and consistent navigation. The search issue (there being two) was especially good because visually, the two searches are very separate. However, in the source they are next to each other, with no means of distinguishing them. However, you might take from the article that one should be removed, whereas a hidden heading or two would be more appropriate here.

I think one of the bigger concerns for me is that John apparently uses JAWs 3.7, which has been out of date for at least 5 years! The impact of this is that many functions standard since version 6 (and improved in 7 & 8 ) aren't available, so John didn't know to complain that many of the headings are not marked up as such, or ask why the navigation items are wrapped in <pre> elements.

As well using an old screen reader, it's problematic to just use one screen reader. I wouldn't expect many organisations to test with more than one screen reader, but you can't then generalise the results without reference to the accessibility standards. It is notable that the expert opinion has to try and put John's finding in perspective, bringing in comments about non-screen reader issues.

Don't get me wrong, getting a real user to test a site is a good thing, and helps to hammer home why it's important.

However, to work out how to improve, I would want a much broader view. Usability testing of any kind is essentially a case study approach, and the findings have to be taken in a wider context that encompasses of the all the site's users and goals.

(For a much more in-depth look at this type of issue, try reporting accessibility issues. You can also find all the accessibility articles on journalism.co.uk)