There is aleady editor_vote - (patch) that would enable janitorial reaping as you've described. What I'm trying to point out is that unreap on its own doesn't do much good unless there is a way to make it stick.

Another idea that occurs to me is that whenever a reaped node is unreaped, we would record that in an unreaped table. Then we make a superdoc that shows any nodes in reapednode that were once unreaped, so janitors can unreap again.

What do you think?