# Maybe anonymous users should not be allowed edits?

This suggested edit erased the first line in my answer and replaced it with

I think that gravity has an effect on any particle in our universe. Even the photons are attracted by the back holes (despite the fact that some physicists consider photons as having no mass !) But in quantic physics, particles may adopt one these two behaviours: being "real objects" i.e with a mass or being a wave!

So what is the mass ? Does the mass really exist ??

Edit suggested by anonymous user. It was approved by one reviewer and I just managed to catch it and reject it.

Why are anonymous users allowed to mess up things? At least if an anonymous user does the editing the person who wrote the entry should be primarily responsible for approval or rejection.

I am asking, a feature request, for a veto from the original person who wrote the question/answer that was edited by an anonymous user, as in this case already there was an approval! It would have made the answer really stupid.

Edit in response to comment by @user36790.

Rolling back is an option that corrects things, IF the original author reads the site often. We have many good answers by very good physicists who have stopped looking in everyday or even month and some not at all. A distortion of their answers might go unnoticed and get approval by ignoramuses and destroy the integrity of the site. Moderators cannot be reading everything as postings go through first page very fast.

• Possible duplicate of Why do we allow anonymous users to suggest edits? – ACuriousMind Nov 3 '15 at 16:28
• As that other question is a discussion and this is a feature request, I'm inclined to think they're not quite duplicates. – user10851 Nov 3 '15 at 16:37
• Let me reiterate that if anyone sees an edit being approved which really should not be, please bring it to the mods' attention via a flag. (I'm already looking into this one.) – David Z Nov 3 '15 at 16:39
• Effectively, it seems like a duplicate since a feature request usually belongs on the main meta.SE site rather than here. – Qmechanic Nov 3 '15 at 17:00
• @Qmechanic I do not contribute there, and the tag is here too. I suppose the moderators can transfer the feature request. In any case the answers in the duplicate do not address the same problem. I am not only asking but also suggesting a change. If when an anonymous edits only the author is alerted this would solve the problem. – anna v Nov 3 '15 at 17:10
• If when an anonymous edits only the author is alerted this would solve the problem sounds like an interesting fix, I'd be curious to see how the SE team thinks of it. – Kyle Kanos Nov 3 '15 at 17:25
• Also, it seems the proposed dupe is asking if anonymous edits are useful whereas this one is asking for a new way to deal with those edits. – Kyle Kanos Nov 3 '15 at 17:27
• I've seen this problems quite often. Don't know who accepted the edit; probably he was bit reluctant or probably he mistakenly pressed the accept button. But even if the edit has been accepted, you can rollback to your previous unedited version. Anonymous users rarely & barely suggest a good edit, after all. – user36790 Nov 4 '15 at 3:31
• "Maybe anonymous users should not be allowed edits", or maybe we could do something to improve the reviewing standards! That can be corollary, if I'm right? – 299792458 Nov 4 '15 at 12:55
• I don't think users should be allowed to edit somebody else's answer. I wouldn't dream of doing so. – John Duffield Nov 8 '15 at 14:40

Single cases and anectodal evidence are tricky bases to work with here. We notice the ugly suggested edits, but who's to say that there isn't a large body of reasonable suggested edits by drive-by anonymous users.

To sort this out, I've written two SEDE queries:

The obvious observation is that anonymous edits do have a much higher rejection rate than edit suggestions from registered users - 56% approved for anonymous suggestions vs. 87% for registered users, over the latest 500 edits of each.

They are also much less common: as Frequency of anonymous suggested edits shows, about one every 10 suggested edits in the entire site history is an anonymous suggestion.

I feel that overall this is an acceptable state of affairs, and that the ~45% of bad apples in the anonymous-suggestions apple cart is well handled by the combination of a review queue and a notification for the OP. The thing to do is for people to dig into the actual edit suggestions by anonymous users and form an opinion on whether they're generally salvageable or mostly just rubbish that occasionally passes the bar, and whether there are hidden gems in there that make the whole pile worth it. (I'll leave that bit to others - writing SQL is more fun than looking at a hundred edit reviews.)

I will also note that in this particular case the problem is not the fact that the suggested edit was bad, but that the review was bad: this was one bad edit that might have got through the review queue when it shouldn't. (It's speculation whether someone else would've let it through had anna not intervened. I tend to think it wouldn't, with other reviewers stepping in and the review queue working like it should, but it's a moot question now.) This particular bad edit happened to come from an anonymous user, but I think it's a red herring unless one can rule out the existence of similar bad-edits-that-slipped-the-net from registered users. Unfortunately those are really hard to find, and getting reliable statistics will be even harder. However, I don't see why the review queue should function less well for anonymous suggestions than for registered-user ones - if anything, the anonymous source ought to make reviewers more wary of the edit.

Let's make this decision based on the actual overall quality of the edit submissions we get, rather than anecdotal impressions from a few ugly ones.

• You have not addressed my latest edit: what if some anonymous"crackpot" subtly changes posts of former users who will never look at the alert, but are very good posts from good physicists enriching this site and will become useless as a physics data base. – anna v Nov 6 '15 at 15:33
• @annav I see that as covered by the second-to-last paragraph. I think the problem is bad reviews and not anonymous edits (unless one can argue convincingly that anonymous edits are more likely to get bad reviews, but I just don't see why this would be the case). If we have a bad-reviews problem, then that's what we need to address (and there's a huge toolbox, developed for StackOverflow, to do just that). – Emilio Pisanty Nov 6 '15 at 15:42
• On the other hand, I think that the first response should not be to completely ban anonymous edits, but to increase the number of required reviews to three or four approvals. This is sustainable at the current conditions (that queue is never in stress, and anonymous edits are ~10% of the total), it can probably be justified well (given the much lower accept rate on anonymous edits, plus your valid concerns on valuable users no longer visiting the site), it exponentially decreases the chance of bad edits getting approved, and it can probably be implemented easily. Why not ask for that? – Emilio Pisanty Nov 6 '15 at 15:47
• well , maybe for anonymous edits only. I think it works fine with registered users. On the other hand , why would it not be as easy to implement that for anonymous users, the author should do the accept. In this way, if an old good answer is changed, nothing will happen to the answer. If it were a bad answer needing a correction a registered user would probably have intervened. – anna v Nov 6 '15 at 16:56
• Again, I still don't see why anonymous edits would be more susceptible to a bad review, and you haven't addressed that. Please step back and think in terms of the broader picture: you got a bad anonymous edits, but you're not necessarily seeing the bad-edit-plus-bad-review cases from registered users if they are there. The problem isn't bad edits, it's bad reviews. – Emilio Pisanty Nov 6 '15 at 16:59
• From a practical standpoint, this sort of feature request is much more likely to get traction with the SE team if it's simply altering a few lines of code (demanding three reviews instead of two if the edit is anonymous) which works within the current review mechanisms and doesn't break functionality (the possibility for any visitor to improve this answer), against something that does break current functionality. I think the team will be quite likely to roll with it if we ask for it - but you're not (yet) actually asking for that. – Emilio Pisanty Nov 6 '15 at 17:07
• The probability that somebody with his/her own theory will try to push it on a site that is becoming a data base ( often in google searches a physics.SE answer comes up) is much higher in anonymity. No repercursions. Whereas a registered user has given some details and can be tagged. That is why I think the anonymity is insidious. – anna v Nov 6 '15 at 19:02
• Again, I still don't see why anonymous edits would be more susceptible to a bad review, and you haven't addressed that. We have mechanisms that work well for this, and it's at most a matter of fine-tuning them. – Emilio Pisanty Nov 6 '15 at 21:23
• The review is a gate. Let us suppose that they are equally susceptible to pass the gate. I believe that the percentage of bad edits that will pass review will be much higher in anonymous edits, because of the 56% to 87% that you found. . Anyway, I am just raising a question. Maybe nothing will ever happen. It is all a matter of probabilities. One can wait until it happens and is caught to do something about it. – anna v Nov 7 '15 at 4:04