Yesterday at work I got stuck retyping a handful of PDF files. They were nothing too fancy: black text in a garden-variety font on a white background, with a logo (image) in the upper-right corner. Nothing I couldn't do in a word processor, but I didn't want to retype all that junk (I have no idea what became of the original documents--I just had the PDFs).
xpdf was a big help to me. xpdf is an open-source PDF viewer, and it comes with several command-line programs. Yesterday I was able to use two of these programs to help me recreate the documents: pdfimages extracted the logo image (in ppm format) from one of the PDFs, and pdftotext converted each PDF file to text. So I was able to use oowriter (the word processor of openoffice.org) to create the new documents, and I could just copy-and-paste from the text files generated by pdftotext. Still tedious and annoying, but better than typing from scratch (less error-prone, too).
CentOS 5 doesn't have xpdf. This experience highlighted to me how important that package is to me (I use xpdf to view PDFs all the time). So I spent some time this morning building xpdf on CentOS 5 from source, and I came up with an SRPM. I sent it to the CentOS Extras site--maybe they'll add it to their package list. If you want the spec file, leave a comment.