It seems to me that your assumption that the author is the viewer when the node is viewed one time is plausible, but not ironclad. It probably works if the data is not really important, but in a real-life situation a race condition on viewing could have someone else view it before the node creator.