I wrote a little DTD of my own over the weekend to use for biblliographic information I'm starting to collect for my book. I needed something to structure the data, and emacs/psgml/DTD seems like a good starting point, since it's a tool chain I already use and am comfortable with. Perhaps this is a wheel I'm reinventing, but my structure is parsimonious (author, journal, date, title and room for open-ended notes), so it should be easy enough to move it into some other data storage bin if and when I decided on making a change. And thanks to python/libxml/libxslt, dealing with the data is relatively easy.

Suggestions for bibliographic dialects of XML and related software are welcome.

The underlying issue I'm grappling with is a need to adapt my work style. Writing newspaper stories, I can just blunder ahead. Both literaly and metaphorically, I can easily keep up to 10 bits of knowledge in a heap (either in my brain or physically, via notebooks or research papers on my desk) and find whatever I need. Beyond 10, though, it gets increasingly difficult to rummage around and find stuff in a way that's quick enough to keep up with the writing. I regularly bang up against the limitations of this technique when writing longer newspaper stories. It obviously will fail completely when trying to write a book.