Re: PDFs on Kindle
Going back through a document would be a pain in the butt, no argument there. However, tagging as you go really is trivial. I believe Andy is using InDesign, which makes it super simple as of CS3, but I'll admit I don't know about prior versions.
Afterthought: I wouldn't suggest going back through the PDF archive to tag for reflow. Not with 6e on the distant horizon. But 6e would be a nice place to start.
I believe PDFs are the industry standard for file formats sent to a printer (my last job dealing with print publishing ended in 2000), not the relatively low fidelity files we consumers use, but massive files with very high resolution pics and other graphical data. InDesign and Quark, the industry standards for print document creation, output both from a master native file. That native file is where the tagging is done. Even if the printer uses something other than PDF, InDesign and Quark would still output the file format called for, either including or leaving out the reflow tagging, since the print version should be 100% unaffected.
The reason to do this is that print media, while nowhere near dead, has only a limited remaining reign as the King of mass publishing. Portable display devices are the future. I currently read my Hero books on a 4" 480x640 PDA screen using PocketXpdf, which at least can re-render a PDF page into plaintext, though it's a little hit or miss about where to place floating elements that aren't reflow tagged. It would be great to read it with the floating elements logically placed and pictures appearing at all.