There's a pretty good article about how a web application that utilizes AJAX should work in the face of adversity. Adversity, as in not having the ability to use AJAX.

What?I found most interesting about the article was the approach that's suggested. I would guess that most people think about building the functionality that they want their applications to have, and add in support for alternate uses later. And no, I'm not suggesting that you don't plan for the alternate cases. I'm just saying that it's usually built after the main use case is satisfied.

But the approach in the article says to build a non-interactive website first and get that completely working, and then add in the AJAX stuff later. Once it's out there, it seems much more logical than the other way around. And, coincidentally, it makes porting current sites much less of a daunting task.