# Editing responses that have perceived errors

I don't think that the edit to this response [1] was ideal. If a response is incorrect, we can point that out in comments, or vote it down. But editing someone else's non-community-wiki answer should be limited to obvious typos, math fixes, and similar "invisible" fixes. This edit seems exactly like the sort of remark that should be in a comment. Otherwise we will run into all sorts of "disagreements" over whether some proof is "correct". I would not appreciate someone editing one of my posts in this way, and I have much more more mathematical self-esteem than an a student or hobbyist would likely have.

Added: I don't want to imply that the person who made the response is an amateur; his profile shows that he isn't. That's even more reason to let him fix his own answer and respond they way he likes.

• While I agree with Bill, that this alleged "proof" is totally bogus, completely altering the original reply as he has done is unacceptable. It shows no respect to the original author, and is likely to start an "editing war". The best response is to comment on the post and to provide one's own reply as well. – Robin Chapman Sep 13 '10 at 6:35
• Bill, please do not tell me that I am confused. You inserted new material into Douglas's answer; this new material was longer than his original posting and completely antithetical to it. As I said, it was totally disrespectful to him. – Robin Chapman Sep 13 '10 at 18:52
• @Robin: Totally bogus? Really? If possible, can you please explain. Maybe I will post a new question on the parent site for that. – Aryabhata Sep 13 '10 at 19:38
• (Note: I have removed the last 3 comments which adds nothing but flames.) – kennytm Sep 14 '10 at 9:23
• @BD: I see that you have deleted your answer. Does this mean that you no longer stand by what you said in it? If not, why did you delete it? – Pete L. Clark Sep 14 '10 at 21:45
• @Pete: If you must know the reason is as follows: I went to great lengths here and elsewhere to explain the seriousness of this error. I was extremely disappointed by the lack of any indication whatsoever from the OP that he made any attempt to understand this explanation. Indeed, instead of that, he shifted the discussion to one of my proofs (one which received much praise elsewhere) and made nitpicking criticisms of it. After that I decided that the discussion was no longer productive and that it would be better to remove my remarks till they could later be elaborated much more precisely. – Bill Dubuque Sep 15 '10 at 3:50
• @BD: I didn't need to know, but I wanted to. Thank you for your explanation. Deleting an answer that 3 others have commented on and 10 others have voted upon while leaving your many comments on others' answers may leave a negative impression on some of this site's users. I am reminded of a certain proverb contrasting a person's capacity for dishing it out to the same person's capacity for taking it... – Pete L. Clark Sep 15 '10 at 4:26
• As the OP, I did "make an attempt" to understand the explanation. However: (1) I still do not agree that there was any "error" in the original post, because I don't believe that the original post was intended to be a proof. The person who wrote the original post commented here and confirmed it was intended to be a hint, not a proof. (2) Even if the original post was completely and utterly erroneous, the way to handle that is to vote it down and leave a comment. We get bad posts here all the time. So the edit was inappropriate even if the post was completely flawed. – Carl Mummert Sep 15 '10 at 11:12
• Also, the issue here is not a lack of mathematical expertise: the person who posted the original response has a PhD, I have a PhD, probably the majority of people who have commented here have PhDs. If there are difficulties communicating the issue, they are not because the audience is simply unable to understand basic undergraduate number theory. However, discussing any actual errors in the post misses the point. The issue at hand is not whether the original post was flawed. The issue was the unprofessional, heavy-handed edit to the original post. I see no support for that edit here. – Carl Mummert Sep 15 '10 at 11:27
• @Bill: You're saying that appeal to authority is irrelevant, but your own individual expertise is relevant for assessing this situation? That we will encounter a diverse group of mathematicians online, but if they consciously choose to handle a question differently than you would, it's a severe error? The diversity of participants is a motivation to treat answers with more respect, not less. The point here is not about correctness at all, it's about respect for fellow participants on the site. – Carl Mummert Sep 15 '10 at 13:39
• I don't see how studying number theory would help anything. The issue here is not the beauty of number theory or reverse mathematics. It was only your edit to someone's answer that led me to post here. I will note (again) that nobody here has written in support of that edit, and that your own explanation you gave as an answer here was significantly downvoted before you deleted it. – Carl Mummert Sep 15 '10 at 15:28
• @Carl: If you think that random "downvotes" in an amateur mathematics forum are reason enough for me to alter my deep-seated convictions - founded upon over three decades of experience working in related number theory and algebra - then I'm afraid that our view of what is important in mathematics is so far apart that it is hopeless to continue the discussion any further. – Bill Dubuque Sep 15 '10 at 15:41
• @Bill: Downvotes and upvotes from members of the site are extremely relevant to the question of whether your edits were in line with the norms of this site. That is the issue here, and it is not a mathematical issue. I think any professor with decades of experience would know that sometimes deep-seated convictions have to be moderated in professional discourse. – Carl Mummert Sep 15 '10 at 16:09
• Apart from that, I'll point out that if you want to be able to rely on your "over three decades of experience working in related number theory and algebra" then you will need to add some information to your profile to allow others to actually verify your background. Otherwise such claims might appear to be only bluster. However, you have argued above that "Mathematical proof correctness is not by appeal to authority", which presumably includes appeal to your own authority as well. – Carl Mummert Sep 15 '10 at 16:13
• Non sequitur. But I agree with Qiaochu Yuan that the issue here is not about mathematics. There's nothing about mathematics in my original post above, for example. If the question was about mathematics I would not have asked it on meta. The only reason I can see to claim the question is about mathematics is to dodge the actual issue, which is how to respond to perceived errors in responses. – Carl Mummert Sep 15 '10 at 17:08

The guidelines of editing is already in the editing page, and also have been blogged before:

As it says on the sidebar of every edit page, here’s what makes up good editing practice as we see it on Stack Overflow:

• Fix grammatical or spelling errors.

• Clarify meaning without changing it.

• Correct minor mistakes.

• Always respect the original author.

You edit to make things better, clearer, more effective — never to change meaning.

While Bill's addendum does clarify mistakes of the poster, I would not encourage this kind of editing as this does alter the original theme of the answer.

• @KennyTM: Consensus has not been reached that the poster has made "major mistakes". Douglas S. Stones has a PhD in mathematics. I think he deserves more respect than this. – Pete L. Clark Sep 13 '10 at 9:40
• As much as I'm not fond of wading in meta, I have to say this: there's a reason the "always" in "always respect the original author" was emphasized like this. Maybe they should make it bold for the sake of the forgetful? – J. M. isn't a mathematician Sep 13 '10 at 11:29
• I explained my reasons for my edit in this case and I stand firmly behind them. The history of number theory is full of errors of precisely this sort. Indeed, statements equivalent to unique factorization were implicitly assumed even by leading mathematicians until Gauss brought these issues to the fore in Disq. Arith. Because of this it is especially important that errors of this nature be nipped in the bud. – Bill Dubuque Sep 13 '10 at 14:19
• @Kenny: Please do explain your opinion that pointing out an error "alters the original theme of the answer". Try as I may, the only way I can make any sense of your remark is that the while the original answer is incorrect, the remark is correct, so it has altered the theme of being incorrect! Surely that is not your intent, but I can't think of any other reasonable interpretation of your remark. – Bill Dubuque Sep 13 '10 at 14:59
• @Pete: It is not that uncommon for someone with a PhD in one field of mathematics to make naive mistakes in other fields outside their expertise. Another expert (Robert Chapman) has also opined above that the 'alleged "proof" is totally bogus', so I am not alone in thinking that the original version of the proof was invalid. I highly doubt that any expert would think that the original proof was valid. – Bill Dubuque Sep 13 '10 at 15:08
• @BD: Are you referring to the post of Robin Chapman at 06:35UTC in which he describes your edit as "unacceptable"? I do not see any support here for editing an answer as you did. – Carl Mummert Sep 13 '10 at 16:24
• @Pete: I removed the word "major". This should sound fairer. – kennytm Sep 13 '10 at 17:42
• @Bill: Right. Originally it was a (false) proof, then it becomes some sort of discussion. Also, if you find the post is wrong, it is nicer post a comment... Like, imagine a journal editor appends a "this article is completely bogus because ..." section in your paper and publish it, instead of just rejecting it. – kennytm Sep 13 '10 at 17:49
• @kenny: Since this forum is far from a journal, I do not find that argument convincing in the least. I still await your explanation of your puzzling remark about "altering the theme" of the post. – Bill Dubuque Sep 13 '10 at 19:43
• @Bill: While the analogy may not be an exact match, what I see as the key point of KennyTM's and Carl's analogy is that we have a case of some writing ostensibly authored by one person (the paper's author/whoever's profile appears below and to the right) which has been edited to include things most people agree that person would not write himself. – Larry Wang Sep 14 '10 at 0:25
• @Bill Dubuque: As I see it, the fuss was made when you edited the response instead of just posting a comment as the site protocol dictates. That would have allowed the author to fix it equally well. I posted here because (1) your edit seemed out of line to me and (2) it is somewhat embarrassing for the original answerer to have to point that out himself. It appears that the consensus here agrees with me that the edit was inappropriate. – Carl Mummert Sep 14 '10 at 2:17
• @KennyTM: muad suggested in his/her answer that perhaps the FAQ should be edited. Another option would be to edit the page that appears when a comment is edited. I don't know whether moderators have the ability to edit these pages on math.stackexchange, or whether you have any desire to do so, but I agree with muad it might help to point out that comments about a post belong in the comments section below it. That would also give a concrete resolution to this thread. – Carl Mummert Sep 14 '10 at 2:28
• Yes, I don't appreciate the seriousness: there was no objective "error". The response was obviously just a hint about a possible approach. It appears you misinterpeted it as a proof, which it clearly was not intended to be, then decided that you should edit the response itself to claim that the entire method was invalid. That edit went against the norms of this site, the instructions on the edit page, and common professional courtesy. Meanwhile someone else has had to apologize for your edit here on meta. I think everyone expects much more from a professional mathematician. – Carl Mummert Sep 14 '10 at 3:23
• @Carl: (1) See meta.math.stackexchange.com/q/774/171 for what section of FAQ we can edit. (2) The "How to edit" section is shared by the whole SE network. So that change should be raised on meta.SO instead. – kennytm Sep 14 '10 at 6:27
• @KennyTM: thanks for the response. I was under the impression the mods here had slightly more ability to customize things that that. However, based on your answer, it seems like we'll have to rely on the "always respect the original author" language. The SE philosophy in general seems to be to have few documented rules and assume that the users will be able to get the point. The present situation highlights the main difficulty with that: it leads to lengthy discussions when someone doesn't get the point. – Carl Mummert Sep 14 '10 at 11:00

Personally, I found it quite rude that this opinion was expressed so prominently... (thanks for jumping to my defense Carl). Bill: if you really feel it's that important, you could have just appended a proof of the auxiliary result.

Also, when I read the question, I got the feeling that the OP was stuck, wanted some help, but didn't want the complete proof, writing: "How do we prove..." rather than "Can you give me a proof of..."

[Snip previous remarks (perhaps I overreacted)]

Addendum: Actually, now that I think back, I remember being presented with a fairly long finite descent proof of the irrationality of sqrt(2) in undergraduate analysis. I guess I can understand the skepticism when I claim a one-line proof of a generalisation of this result.

Despite several claims, has anyone actually found an error in the proof yet? (aside from me not including a proof of a fairly obvious result)

• @Douglas: I agree that you were treated less than respectfully, and I'm sorry about that. Please feel free to edit out Bill Dubuque's comments, especially now that you've edited your post to fill in the details of the proof. – Pete L. Clark Sep 13 '10 at 9:30
• @Douglas: I'm afraid that even after your edit, the answer is still misleading. In particular, uniqueness of prime factorization is not only necessary for the first sentence that you invoke it in, but also in the following sentence "Therefore...". Could you please clarify that. – Bill Dubuque Sep 13 '10 at 14:43
• @Douglas: I'm happy to see that you deleted the erroneous and/or misguided remarks in your first version of this answer. – Bill Dubuque Sep 13 '10 at 14:48
• @BD: I thought the comments were apropos. For example, you did not prove in your answer that 1/(c/d) = d/c. If you wish to criticize others for incomplete proofs, it opens you up to the same criticism. It's a common courtesy to assume that an answerer with a PhD would be able to fill in these details if asked, and to gently ask them to do so rather than claiming their intentionally-incomplete proof is "invalid". – Carl Mummert Sep 13 '10 at 16:23
• @Douglas: I like to do the same thing, leaving an incomplete answer when I think the person who asked will benefit from working through the details more than I will. In fact the answer I left for the question under discussion was written that way, and is equally vulnerable to the claim that it is incomplete as a proof. – Carl Mummert Sep 13 '10 at 16:28
• @Carl: Don't be ridiculous. No one expects a proof of 1/(c/d) = d/c at this level. If an answer was meant to be a hint or a sketch of a proof then that should have been explicitly mentioned. Anyone who has the faintest clue about the teaching of number theory would know that presenting the proof that way reflects a major misunderstanding - one that has permeated number theory throughout history - in various guises about the "obviousness" of unique factorization and related results at the foundation of factorization theory. – Bill Dubuque Sep 13 '10 at 17:33
• As the original (now edited) post stated, there are numerous other omissions in your original answer. For example, $M$ may be empty in one of your lemmas, in which case it has no least element, so the lemma is ill posed. Also, the phrase "Let $r = \frac{c}{d}$, $\min d > 0$." is nonsense: a number does not have a minimum. You expressed an opinion here that responses should convey a good impression to first-time readers, but that post is written in a cryptic non-prose style that would be unacceptable for homework from a number theory student, much less a professional publication. – Carl Mummert Sep 14 '10 at 3:35
• @Carl: That is complete and utter nonsense. – Bill Dubuque Sep 14 '10 at 4:03
• So the empty set does have a least element? – Carl Mummert Sep 14 '10 at 4:04
• @Carl: being a universal statement, it's vacuously true (and if you look at the original version you will see that I explictly removed "nonempty" due to space constraints, knowing that it was safe to do so) – Bill Dubuque Sep 14 '10 at 4:04
• @Carl: What you call "cryptic" is usually called elegant by most people. That post received much praise elsewhere, so your attempts to pick it apart will make no impression on me other than to continue you emphasize just how ridiculous your remarks are. – Bill Dubuque Sep 14 '10 at 4:08
• I said that lemma is ill-posed: the statement of the lemma itself refers to a least element that may not exist. The empty set does vacuously satisfy the hypothesis of the lemma, but the conclusion does not even make sense in that case. – Carl Mummert Sep 14 '10 at 4:10
• @Carl: I have no interest in continuing this discussion. – Bill Dubuque Sep 14 '10 at 4:11
• @Carl: fyi, I know quite a bit of logic, including some reverse mathematics, so don't waste your time trying to attack my logic. It is is quite sound. – Bill Dubuque Sep 14 '10 at 4:23
• Trimmed to remove some inappropriate comments. – Larry Wang Sep 14 '10 at 16:31

editing someone else's non-community-wiki answer should be limited to obvious typos, math fixes, and similar "invisible" fixes

Agreed, this should probably be added to the FAQ

This edit seems exactly like the sort of remark that should be in a comment.

Yes - infact it IS a comment. It is even signed.

Otherwise we will run into all sorts of "disagreements" over whether some proof is "correct".

That's a good point - acting this way can only cause friction.

I would not appreciate someone editing one of my posts in this way

No I don't think it is correct to add a signed comment onto someone's actual post - Here is an example of an edit which I thought was necessary.

It is easy to see (e.g. here) that Bill Dubuque appreciates a very high standard of rigour and that is one reason why his answers to other questions are so illuminating to read but I don't think it is necessary to get frustrated at non-rigorous, partial or false arguments - these can also be very useful and lead to insight too.

Of course it is important to be able to recognize these and differentiate them from formal arguments and this is probably a difficult thing for a beginner but I don't see it as a real problem because for any mathematical argument one reads - they should put enough thought into it to understand it and thus notice mistakes.. without doing so what would be the point of reading it at all?

• I think the edit you made is qualitatively different because (1) it fits in seamlessly with the existing post and (2) it just neutrally points out where the extra assumption was used. People can then read the comments to see that the assumption in question is the heart of the proof. Also, the original answerer had already marked the post as "community wiki", which is why you could edit it with under 1000 rep. – Carl Mummert Sep 13 '10 at 12:19