Just to follow on from pixeldiva’s comments about hiding links offscreen. She mentions sighted non-screenreading users but some screen readers users can see very well. Some dyslexics use screen readers to support their comprehension of web pages.

Given that this is a group that may be easily disorientated, I’ve always been a bit concerned that offscreen positioning of page content could create problems for these users as the screen reader output would no longer match the visual display. I’m not suggesting that offscreen positioning should be abandoned - merely that it should be used judiciously. More than a couple of instances on a page may well create barriers for this group.

I think it’s also worth remembering that only about 5% of those who are registered blind (in the UK) are thought to have no usable sight. With that in mind, it becomes harder to asses what extent (if any) individual visually-impaired screen reader users rely on the visual display. Again, any significant disparity between the visual display and the screen reader output could create confusion.

In short, screen reader users aren’t necessarily blind. They just have problems reading. :-)

Jacks comment: “Any accessibility developers should ‘use’ it!” is good. Maybe I am naive but after becoming the law in 1999 that all websites should be accessible (in the UK) I would assume that all web developers take accessibility very seriously (removes tongue from cheek).

I’m not sure whether its the screen reader that’s figuring out what to display or the browser. I guess using a MSAA debugger/watcher will reveal which one is doing what.

My explanation of the heuristics is based on the first table of test cases - JAWS7.10 and Internet Explorer 6, basically since I have the test results for a normal link, a titled empty link, and an empty link, I deduced what was going on based on the test results.