The revision tree try to display the whole history of an item so user may get a better feeling about the history of an item. It has to scan the complete log of the repository 'cause it requires some information subversion does not (and I think will not) give. Because this produces a lot of traffic the revision tree uses always the internal log cache.