Here's one way it might work, one combining the suggestions to date as well as adding a couple of new ones:

• When the node is assigned to the Reaper, replace the original text with "This node has been reaped. Reason: <Reason given when Considered>.

• I'd also add the manual link to the original text that tye suggested, but only when appropriate.

• I would also add a series of links to a set of FAQ's and useful nodes, just in case the mistake was an honest one. This doesn't have to be terribly context sensitive, if at all...just a set of documents that the true petitioner can refer to quickly to understand why the node got reaped. (More than happy to help locate those, if needed, though I suspect most of you can point to the obvious ones.)

• I like the direction of PsychoSpunk's idea, but wonder if it would be unworkable due to the increased exposure of the offending material (inlining it) and the fact that there might be multiple replies to the original node in question. Seems like it would either lead to more exposure to the offending node (e.g. multiple owners) or would leave other replies in thread limbo.

• As an attempt to avoid rewarding the obvious trolls (you can probably figure out the candidate I have in mind), I'd also add some sort of "No Mercy" switch in Consider, which would hide the original material as completely as possible. A recent ASCII artwork comes to mind, as does a short-lived attack posted this morning. This would help implement the second point.

  The idea here is to reduce the possibilities of giving certain, uh, folks any reason to think their behavior has caused a commotion. ("Just a simple edit, ma'am. No trouble at all.")

• When Reaper takes possession of a real monk's node (as opposed to one posted by AM), "he" should /msg the poster, saying something to the effect of, "I have reaped one of your nodes. For details, see <Node Title>."

  I think that could help the original poster learn from the experience, especially if we include the links of suggested reading.

I'm trying to balance the need to educate (and reduce) those making honest mistakes against simple troll prevention. As always, feedback is welcome.

Otherwise, I think it's a great idea. (Also, thanks for your continued efforts, especially the recent updates to Nodes to Consider.)