(BTW I just searched using terms such as "original content" or (by name, now that you know it's Corion's work) "xml" and didn't find the original.)
In any case, the line of code I offered for your Free Nodelet is literally and exactly what's in the third line of mine; viz:
Now, back to your threshhold question:
My view is that "simply to flag them" (from Eric's, above, with which I gather you concur) is NOT enough. Yes, it's nice to know something has been changed, but unless one knows "from what" it doesn't solve the potential mysteries of the now-contextless replies.
Thus, I'm not sure that the magnitude of the coding work (by the PMDev's -- esp those with a better grasp of the guts of the Monastery) -- and the extra cost of retrieving and rendering by the server) is worth the effort.
And just FTR, Eric and I may not be in total agreement, but I do agree whole-heartedly with his remark, " There is no reason not to allow some updates, its specific types of updates that are harmful." It's just that I took your ideas as a springboard for a dscussion about ways to deal with the "harmful" ones, for those values of "harmful" which seem to me to detract from the value of the information here.
Update: But that may just reflect my feeling that it's good manners for those who come seeking help make the effort to observe our standards.