Someone mentioned XML/XSL/FO. Don't try to write your content in XSL-FO. You'll hate every minute of it.
I'd look in to using DITA (Darwin Information Typing Architecture). It's a set of canned XML structures, plus a specification for how to process and customize those structures. It includes tags for stuff like footnotes...I bet it covers a lot of your use cases. There are some good intros to how these XML structures work here: http://dita.xml.org/book/dita-wiki-knowledgebase
As DITA is XML, you can convert it to HTML and whatever else you feel like, pretty easily. There's an open-source implementation of the DITA spec called the DITA Open Toolkit (http://sourceforge.net/projects/dita-ot/). The DITA Open Toolkit includes stylesheets/scripts to publish HTML and PDF, among other things. PDFs are published via XSL-FO. Just like HTML needs a web browser to render something useful, XSL-FO requires a FO processor to create a PDF. So, in the end you write DITA, XSLT and other scripts transform that DITA to XSL-FO, the a FO processor consumes the XSL-FO and spits out a PDF. The DITA Open Toolkit comes with an open-source FO processor (Apache FOP). FOP doesn't fulfill everyone's needs, but it might work very well for you.
Unfortunately, working with the Open Toolkit and customizing its output can be a bit unwieldy. http://groups.yahoo.com/search?query=dita+users is a pretty good place to look for help.
Time for a sequel to Animal Farm. At some point in the story, the paranoid pigs will order a culling of all humans.
Great idea...Bet our energy future on something that's succeptible to FoxDie.
Intel CPUs are not defective, they just act that way. -- Henry Spencer