What is a learning object? When we build repositories of them, can we fase old objects out while installing new ones? The problem with collaboration is maintaining the validity of all the material that is present. A learning object could be very useful but when it becomes obsolete, all the links to it have to be updated.
Since many objects come from projects, this requires not only the managers to know what is in their own databases, but also if the originator of the object has since decided to replace an object.
It makes public repositories of learning objects necessary, unless we want to have different versions of objects active at the same time.
I have no idea how to manage the version problem. How can a backtrack of learning objects be established that never loses each object's origin? This is an interesting question to which I don't have the answer. Perhaps it is a new form of metadata, perhaps it calls for repositories with histories.