# how to correct an erroneous edit

"Black adder" edited my answer to a question asking for a proof of noether normalization. the edit was nicely done, except in the next to last sentence a change was made from "≤ d" to "less than d", which introduces an error. I do not know how to either fix it or message the editor of the question. Or perhaps, in some languages "less than" means ≤?

• You must mean here: mathoverflow.net/a/92434/2926 I fixed it for you, but please check. (BlackAdder evidently meant to be helpful and also asked in the edit summary to check that no errors were introduced in his/her edit.) – Todd Trimble Dec 15 '15 at 17:16
• I see now however someone has also edited the original question until it is nonsense when combined with my answer! I.e. Miles Reid wrote two books, one with a difficult proof of NN, his commutative algebra book, and another one with an easier proof, his UAG book. The original question complained about the one in the algebra book. The answer I gave is the easier proof in the geometry book. So now the question and answer are rather puzzling! I.e. I now seem to be providing the same argument that was objected to! subsequent editing poses an interesting challenge to historical accuracy. – roy smith Dec 15 '15 at 17:34
• There was no recent edit, and all edits after the original were minor typo corrections. Evidently your answer was helpful anyway even if you think you merely reproduced Reid's UAG proof, so maybe different words were all it took. The OP by the way seems to have a deleted account and I don't think will be returning, so all puzzlement aside, let's let the answer stand. Your efforts were and are appreciated. :-) – Todd Trimble Dec 15 '15 at 17:43
• I am not sure what you refer to. The question was not edited at all after you gave the answer. All edits happened before and in any case they seem minimal. All version of every post are available in the revision history I mentioned [except in very very rare cases, or when only present for less than five minutes]. For the specific case see: mathoverflow.net/posts/92354/revisions – user9072 Dec 15 '15 at 17:43
• you are right. i misread the date of the last edit of the question. now i am puzzled as to how my answer helped. perhaps as you say, just the fact my notes were written to be especially clear to my class helped. – roy smith Dec 15 '15 at 17:57

## 1 Answer

To make a minor fix to an edit that took effect one proceeds as for a general edit of the post, that is one selects "edit" just below the post, makes the necessary change, and saves it.

It is possible to message editors of a post using the usual @user-syntax for comments; that is one writes a comment starting (or at least containing) @username where username is the the editor's displayname (without space). In fact, it suffices to use the first three (or more) characters too, but it needs to be the start of the displayname, so I could notify you with @roy or @roysmith or @roysm yet not with @smith).

For completeness let me add it is also possible to undo an edit completely. To this end go to the revision history of the post (clicking the timestamp in the middle at the end of the post; note if the post was not edited it is not there so for example on this meta post it is not present [indeed on your meta-post it is meanwhile present]). There you see all old versions. Look for the one you want to go to and click "rollback" there.

• I thought there was a lower bound on the amount of changes the system accepts as an edit. If so, just fixing $<$ to $\le$ may well be impossible. – Emil Jeřábek Dec 16 '15 at 9:36
• There is only a lower bound for suggested edits, not for edits. Thus, a user wishing to correct an edit to their post will never face any restriction of this form. (A third user without the edit-privilege indeed might face this problem, but there is no difference between fixing an edit and any other minor edit, which was the information I meant to convey.) – user9072 Dec 16 '15 at 10:21
• All right, I stand corrected. – Emil Jeřábek Dec 16 '15 at 10:56