What I *try* to do is write TITLE text so that it makes sense jointly and severaly -so it makes sense as a standalone item read out of context, but also so that it sits perfectly with the surrounding text as though it were the text by default.

Here’s a simple example:

Visit the
home page

That’s the ideal anyway - I confess that it’s rare to get it exactly right, and I’ll err on the side of self-contained rather than contextually-replaceable!

Lately I try not to be redundant of text content. In brothercake’s example above that is a bit of the case, so I am trying to not do that as often, whereas I might be more inclined to do something like this example (as used on my portfolio site):

Accessibility

I would do this to better clarify what the single word “Accessibility” means in the context in which it’s used. “Home” on the other hand has such a universal understanding it may be redundant to title it. It may be err, on the side of caution albeit.

I’m just trying to be more careful about the usage is all.

Straight document reading, no problem (the titles did not read aloud.
Line reading (by hovering mouse over each line or element was impossible. the title tag cut in after two words, and the line couldn’t be read.
Tabbing to links, both link and title attribute were read aloud, which is only a benefit if the title provides additional information, and doesn’t just repeat link text.

Great test, it’ll be interesting to see what the consensus view is.

Just wondering in assistive tools if their is a default setting that the the alt would be read before the title? Or is it just which one comes first in the code? In the case of text links the link would always be read first followed by the title then?

I havde also tried to make links and titles blend together into a coherent sentance… but never really stopped to ask if their is a set order to what is read first.