There are extensive implications for this that we have never found good methods for implementing. How would metadata or field merging happen, for example? Which document wins out? There are a lot of questions like that. When coming in from windows, the new entry doesn't have any content that would need to be merged.

It is something we've looked at for each release, but it has very major implementation implications. Part of the problem is that everyone we've talked to seems to have a different idea how exactly they would use it and how exactly handling of potentially conflicting information should get handled.