@jens admittedly this is interesting but what i am really getting at is how to layer metadata from different sources onto the same "content" -- http headers, rdf statements, and so on, what is the atomic unit of content, that kind of thing. the content's embedded metadata might conflict with the external metadata, so which one do you trust? and how do you deal with potentially multiple nested layers of container formats? (plaintext message in an html document in an http message in a text file)