After some reflection, I think the metaphor of structured, semi-structured and unstructured unities is incorrect. For example, there are structured, semi-structured and unstructured data, however the examples Steve Hoffman provides are incorrect. A relational database is structured data, however both a spreadsheet and a text document are both semi-structured data. They are just at either end of a spectrum. Unstructured data is data which is a complete exception–there exists no format to make it readable. Ultimately, it is simply ignored by the system because, though it may have value it cannot be captured.
If my hypothesis is correct, then there is more semi-structured examples of each of the unities than any other type.