Hi. I've finaly got some monograph code in some sort of shape. Sory for how slow this has been. It's very close to the proposal from Michael with differences as follows:
* Not solely for structured vocabularies - can be used for other graphical data like database schemas, object models, frames or sematic networks. * Classes have different names - search/replace can fix this - I'd prefer to keep the names closer to that of graphs than of vocabularies. * Collections not iterators used (Michael says that he has no firm view about this). * The Visitor.Direction concept is expanded into the Link entities. These label the role of a node in a particular arc. * The constraints allow you to follow from an arc to a node with a namespace which tells you more about that relationship. It could give extra constraints, such as saying that a relation is 1-many. It could also be used to generate relationship types or instances from data, for example, auto-generating java beans from vocabulary types. Anyway, I've attached source for some interfaces which you can use to build javadocs and classes. There's an in-memory and an SQL-backed implementation you can have if you ask nicely (I didn't want to send too big a file via e-mail). Matthew
namespace.jar
Description: application/java-archive
