Yesterday, I was working on PRAjax. The UpdatePanel did not work completely as I wanted it to work: in the background, the whole page was still fetched and updated. A cleaner way would be to just fetch updated content and not the whole page.
In my search for a PHP HTML parsing class, I found a lot of libraries, but all with disadvantages: one was too big in file size, another only parsed XHTML, ... Luckily, I stumbled on SAX parser! So if you ever want to parse HTML and read out specific tags and attributes, try this one.