The idea that the page *is* the database has some conveniences, but is often awkward. There ultimately is no perfect way to create a single statement that does it all–you often do have to repeat yourself, e.g., having a version that is really clearly human readable, and another that is really clearly machine readable.

This used to be the recommendation with CC–in your HTML, you’d put a human readable communication in visible text, and then also a separate machine readable chunk of RDF/XML hidden in the comments. But, without an easy tool to generate AND maintain the appropriate code, it’s not really viable to maintain over time.

If you look at BMI as an example, they have a web form for entering songs in their rights database. One way to think about it is: the real data standard for the web is web forms. IMHO, the form is more important than the format.