»In which our hero dives into the xml

Frustrated with the forever-limited export options available in Delicious Monster's Delicious Library media-management software, I decided to wade through the single XML file the program uses to store data. Hey, at least it's not a plist.
The file does not contain timestamps, but it does contain links, however inconvenient, to the hack-y "shelf" concept in the software. The xml file swells to about 12400 bytes per entry

Using perl's XML::Simple on my late-model PowerBook on the 4M XML file, a simple run of the parser takes several minutes: 106.13s user 1.62s system 57% cpu 3:07.57 total. I want to extract title, author, and ISBN information, as well as the date I added the item; but the XML output by Delicious Library does not conform comfortably to a structure that I can parse, and reading through the source to DeliciousExporter shows that the authors had to include plenty of special cases and output-munging in order to build the HTML pages from the source XML.