I have a great relationship with my wife, my parents, and my brothers. I also have a good relationship with my daily supervisor (work). But fortunately, a different one with all of them. I have different expectations of a relationship with my wife compared to one with my supervisor. But, if the quality of my relationship with my wife is determined only by our expectations and our ability to fulfill those, then life would be quite boring. As a matter of fact, I enjoy the unexpected as well. Also, I don't often know what to expect, I cannot predict the future.
Hence, the quality of a relationship is determined by the expectations + all unexpected that happens between the two persons. Does this also apply to objects in this site? When can we say a relation between two objects is of high quality? What are expected relationships between objects? And which "unexpected" relationships could be valuable?
How do pre-defined relationships compare with random or serendipitous relationships? I met my wife not through a dating site but by chance after a party we both went. I think that if a person is purposeful, we could take the purpose of the person in order to define expected relationships: a person who wants to make a collection of news-items from his birth-year will be helped if we set the fixed relationships "object=news-item" and "year-published=year-of-birth".
Leaving aside the totally random relationships, we can (pre-)define some common relationships that could be concerned relevant. We notice that a relationship is defined by a common concept, such as 2 brothers have the same parents. Let's try to think of some relevant relationships on this site for information objects: same author, same reader, same interested people, same topic(s), same set, same 'age', same type(?)... etc. In addition to "similar to" relationships, there are "contrary/opposing to" relationships, and more.
How do we get the most important relationships from the users, because we do not want to show them all relationships possible. How can we enable them to define their own 'relevant' relationships? How can we derive relevant relationships just by looking at how people use the objects? And how do we improve the serendipitous relationships or is this utopian?
