PDF is a common format used to distribute content online, and provides more of a "digital paper" feel than more flexible formats like HTML. This determinism is why tools like LaTeX work well with PDFs. Unfortunately, the PDF specs have evolved over the years, incorperating more of HTMLs "features" like JavaScript (PDF 1.3, 2000), and as the specs get more complicated, PDF readers get larger and more complicated, and there are more possiblities for security holes.

For a quick, quasi-technical introduction, see Didier Stevens's articles on the structure of PDF files and malicious variants as well as his list of parsing tools. If you want a bit more detail, Didier wrote up a 23 page summary.