A lot of information technology seems to be around the dynamic of having your cake and eating it too: you want your documents to be databases, or your databases to be documents. Either way, you want both.

RDBMS is document. A database is a document, or a collection of documents. XML is a database, or a collection of data. But, RDBMS has only a limited document-interface, and XML has only a limited database structure.

XML in general, and RSS 2.0 and Atom as concrete examples, seem put upon to squeeze more and more database features out of a document-interface structure. XML, taken to a database-like extreme, becomes RDF, which then is stuck with the same old RDBMS problem where the concern shifts to how to provide document-interface features on the database structure.

I keep trying to write a bunch more about this, but keep getting somewhat tongue-tied. Mostly, I am frustrated with RDBMS and XML—I can't quite get either to do what I want, and putting them together, while maybe working (I hope!), also suggests that in some cases, the limitations of both might compound each other.

Anyway, I am coding the iCite net prototype which should bring some more light on this subject (perhaps through utter failure!, but I hope through some success).