<blockquote>
Following the start of a discussion on the Fedora User list about content models and how they might be used, we offer the following.
At the University of Hull we intend to use appropriate content models widely and to use them as the basis for much of our repository security (see the second document in the list).
We are still at the proof-of-concept stage and the documents here are very much drafts offered without any sort of warranty! If nothing else, they show the direction in which our thinking is going. That said, everything described here is working and the process may give you some ideas.
<p/>
Image object model v0.3 (pdf)