I’ve commented on Roger’s article (which he’s responded already – that was quick) to on the usability of the implementation. To put it into context…

When a user clicks on a typical navigation link, they expect to be taken somewhere, i.e. another page. With that expectation, the focus of the user then changes to the content. But the content won’t have changed.

Likewise, for a screenreader like JAWS, the focus (of the cursor) stays on the link rather than switching to the content.

All and all I’d urge caution with this kind of implementation of navigation.

What I’d much prefer to see is “static” navigation and “gallery” pages (where navigation appears in the content, not just the side-bar).

All said though, Roger has taken a mish-mash of code and elegantly web-standardised, rather than make it “accessible” per se*

* note I’m intentionally trying to start a bit of constructive argument/dialog here