Now that the puzzle in question has been solved, it seems to me that it really was just a gimmick without much value. As far as I can tell this puzzle could exist entirely as-is on the main question page. I think this is a poor choice because:
- It doesn't add anything of value to the real puzzle.
- It requires that a user understand revision histories and how to access them, which is certainly not true of all users.
And, if one more edit is made to this puzzle at some point, the "real" puzzle will fall entirely into the collapsed edits at the bottom of the revision history, further obscuring it.
At the same time, I could envision a situation where having to look carefully at the revision history would indeed provide some clues to a puzzle and thus be integral to solving it. Clues could be deleted bits, or timestamps of edits, or edit notes, or something like that. I'm sure that there would be a clever way to do this.
BUT I still think this should be strongly discouraged (i.e., downvoted) because of concern #2 above, as well as the fact that a puzzle maker does not have entire control over the revision history of his/her puzzle. Since others can edit the puzzle as well, it seems risky to put critical information in a revision history.
If you want a good model of what kind of information can be hidden in edits/revision, click the "edit" button for this puzzle.