24

Recently, I received this answer and a comment asking me to run a particular command to get the precise Last Modified timestamp of a file:

wmic datafile where name="c:\\foo.txt" get lastmodified

When attempting to run this on the example ("e:\\test1\testIMG.JPG"), I got a message which just said "Node - " followed by the computer name and then "

ERROR:
Description = Invalid query

The solution was to use double-slashes for every slash, not just the one after the drive letter (which is sometimes different). I tried specifying this in an edit and even explained the change in the edit comment, but that effort was rejected as "This edit does not make the post even a little bit easier to read, easier to find, more accurate or more accessible. Changes are either completely superfluous or actively harm readability."

The edit changed that line from producing an error to producing an answer, which in my opinion makes it more accurate and accessible rather than harming things.

Why do these kinds of things get rejected?

21

Imagine how bad Wikipedia would be if original article authors always rejected or rolled back helpful, correct edits made by contributors. It's amusing but sad to think of an equivalent to SE answers on Wikipedia:

https://en.wikipedia.org/wiki/Cheese/User1'sPointOfViewOnCheese
https://en.wikipedia.org/wiki/Cheese/User2'sPointOfViewOnCheese
...
https://en.wikipedia.org/wiki/Cheese/User100000'sPointOfViewOnCheese

One of the most frustrating things about browsing SE sites for answers, as a user, is when there are six different answers, each with a half-truth, various errors, etc. and the most upvoted/accepted answer is wrong or obsolete. People are so afraid of editing someone else's answer, as if it were their "intellectual property", and accidentally "stepping on their toes", that they avoid it and rather post their own answer to add to the clutter instead of helping one answer be the best answer.

Answers on SE are Creative Commons licensed for a very good reason, folks.

And if you think that people conscientiously go back to edit their answers to make the most upvoted / accepted answer better so that users don't have to scroll down pages and pages of answers to get to the best one, that's an extremely optimistic point of view. Maybe you think that you always do that (maybe you always, in fact, do) -- but I can't tell you how many times I've found wrong, incomplete or less-than-ideal answers with a considerable number of upvotes and the accepted answer. In a few instances, the objectively best answer gets voted to the top, but that doesn't always happen, either.

I'm a big proponent of "just edit it" (and, necessarily, "just approve it" if you're a reviewer), working under the assumption that the edit improves the answer without introducing opinion. BTW, I think it's perfectly fine if an answer gives an opinion, and then an edit is made that provides objective evidence supporting the OP's opinion. It's only bad form if an edit to an answer changes the meaning of the OP's opinion, or reverses it entirely.

But a lot of people seem to be under the impression that, just because answers have the OP's gravatar/profile link next to them, they are some kind of an indelible "forum post" that no one but the OP should be able to edit, even helpfully. Both editors and reviewers often seem to have this opinion, and thus, the edits never even get submitted in the first place, or if they do, they get rejected, like yours.

This is a real shame, because it makes SE sites much less accessible and increases the frustration and difficulty of finding good answers. It's only somewhat better than a phpBB forum, because the voting mechanic does tend to (often, but not always) push the most useful answer to the top. But it could be really great if a well-written answer with a typo or an invalid command gets corrected, instead of making someone else "compete" with the original answerer for votes, which might take years of Google searchers coming and upvoting it until it gets more votes than the incumbent - especially if software patches from a vendor have semi-recently broken the original answer's code but a simple edit could fix it.

Unfortunately, as long as SE gives answerers the option (in terms of policy permitting this attitude, not the literal option on UI) to reject an edit on the grounds of "No! It's MINE! I can be wrong if I want to be!", the site will be hamstrung in its usefulness, requiring answer-seekers to run a linear-time algorithm each time they find a potentially applicable question:

for each answer in answers_sorted_by_upvotes_descending
  try answer
  if answer doesn't brick machine and doesn't solve problem then
    next answer
  else
    break
end
  • 3
    Just to be clear, in this particular case it was not the answerer who rejected the edit (comments suggest the answerer would have approved it, given the chance) but other high-rep users who may have assumed the answerer prefers the ownership perspective you described. – WBT Feb 11 '17 at 21:59
  • 2
    Thanks for that clarification @WBT, I don't disagree. My answer was more general and not actually focused on the specifics of your question, which might've been a mistake on my part (your question, indeed, is about this specific scenario, not the general case). But I hope I also succeeded in sharing with you my view of how I think some subset of the users (who both review suggested edits, and post answers) view the editing "etiquette" on SE. I personally find this sort of position to be a detriment to the site, but it's tolerated or even encouraged by influential folk, so it continues. Meh – allquixotic Feb 11 '17 at 22:02
  • 2
    The question was meant to be at least somewhat more general than one specific edit, so that answers can be applied as guidance for the future. I generally agree with your sentiments here, but just wanted to clarify. – WBT Feb 11 '17 at 22:05
  • @allquixotic Out of curiosity what do you think of the highest voted answer to the Meta question What are the guidelines for reviewing? instructing reviewers to reject "Edits that attempt to add clarification to an answer, like “this doesn’t work in Windows 8”...."? – Twisty Impersonator Feb 11 '17 at 22:10
  • HAHAHAHA, this answer goes against the "establishment" on MSO. I wonder what was the turning point. – Braiam Feb 19 '17 at 13:44
  • 1
    @Twisty that was written exclusively with the SO mentality. SE in general welcomes edits that improve posts as you can see in the help center: If you see something that needs improvement, click edit! I changed the post (HA!) to make sure that it stays close to the help center rather than a specific site philosophy. – Braiam Feb 19 '17 at 14:14
  • @Braiam I agree with the edits you made to the MSE post. – Twisty Impersonator Feb 19 '17 at 21:47
  • 1
    It's incredibly annoying when I try to edit a post because there's a comment saying the current answer has an error in the code, but the edit is then rejected with the message "This should have been left as a comment", EVEN THOUGH THERE ALREADY WAS A COMMENT THERE! :( – numbermaniac Feb 23 '17 at 7:37
7

The edit suggestion changes c:\\foo.txt to c:\\test1\\foo.txt to demonstrate that directories need to have their backslashes escaped as well. The reviewers likely thought that it was already clear that all backslashes, regardless of what they're delimiting, have to be doubled up to be escaped. In that light, the edit looked like it was just using a different example path, which doesn't add anything. Admittedly, if you haven't used programs or languages that require this kind of escaping, it might not jump out at you. (But now you know!)

General advice follows.

Strictly speaking, edit suggestions shouldn't attempt to change the technical details, since reviewers aren't expected to judge technical accuracy. It's debatable whether providing a potentially-more-useful example is a technical change, but some would say that altering code intended to be typed/run changes the meaning of the answer. To play it safe, leave a comment to share your newly discovered clarification.

Post owners receive notifications about edit suggestions, and if they get to it before others review the suggestion, their vote is binding.

  • 4
    "Strictly speaking, edit suggestions shouldn't attempt to change the technical details, since reviewers aren't expected to judge technical accuracy." Therefore, if there's a technical bug in the answer and I've gone through the effort to figure out how to fix it, I know now I should not be suggesting that edit to fix it. It will be up to later readers to figure that change out for themselves - and if I'm feeling like longer explanations, maybe they'll be able to find it buried in comments. Got it. (Not convinced this is the best answer, but thanks for answering.) – WBT Feb 11 '17 at 22:03
  • 1
    @WBT As unfortunate as that situation is, yes; that is my understanding. On the bright side, once you have >2K reputation, you can fix it yourself because your edits won't have to be reviewed. – Ben N Feb 11 '17 at 22:04
  • 4
    Interesting discussion here, and even more interesting that you seem to feel this situation is "unfortunate" but "is the way it is", @BenN. If a policy is actively detrimental to the site, and you feel that way, why not also offer your opinion stating that fact? Perhaps if enough people feel the policy of "don't judge technical accuracy when editing" is problematic, and argue convincingly for the opposite, we might at least change some users' behavior, if not the official word from on-high as to best practice when reviewing edits. – allquixotic Feb 11 '17 at 22:06
  • 4
    I'm not sure we'd (ever) be able to succeed at changing this culture SE-wide, mind you, as there's a lot of momentum around the status quo on stackoverflow in particular; but the SU community is small enough and agile enough that we could enact meaningful change that helps the site if we had the willpower to push for it. :) – allquixotic Feb 11 '17 at 22:07
  • 2
    For what it's worth, I am familiar with the fact that many technical settings require escaping, but enclosure in quotes is a common way for such escaping to happen (and is how spaces are escaped here). As double slashes are needed after the protocol component of a URL, sometimes double backslashes are needed at the beginning of a file path following a drive letter or share. Also, I did include a direct link to the edit review in the question. – WBT Feb 11 '17 at 22:39
  • @WBT Oops, I didn't notice that the "edit" text was a link, sorry. I've removed the duplicitous part of my answer. – Ben N Feb 12 '17 at 2:31
  • 2
    Could this be made into a rejection reason so that it's easier for people to understand this nonintuitive policy? The canned "This edit does not make the post even a little bit easier to read, easier to find, more accurate or more accessible. Changes are either completely superfluous or actively harm readability" does not seem to be a good fit for helping people learn about this rule. – WBT Feb 12 '17 at 4:00
  • 1
    @WBT The usual reason for rejecting technical changes is something along the lines of "deviates from the original intent of the post." In this particular instance, though, the reviewers probably objected to what they saw as the triviality of the change, so the used reason makes some sense. – Ben N Feb 12 '17 at 4:46
  • 3
    I haven't gotten into the review process yet, but since my rep is getting close I did look, once, at what's involved in the process. I seem to remember a mention of the idea that if you're not sure whether to accept or reject a review, just pass on it. Seems to me that if an edit involves a technical detail, and I don't feel qualified to judge it on its merits, I should pass, and allow someone more capable to handle it. If the editor also included comments on the edit page that clarifies why the technical edit is needed, it should help reviewers judge the accuracy if they know the subject. – Gypsy Spellweaver Feb 14 '17 at 8:27
  • @allquixotic "Perhaps if enough people feel the policy of "don't judge technical accuracy when editing" is problematic" actually, this culture is only prevalent on SO. Most sites don't agree with it (your answer here, on UL, Shog himself (I'm sure you can find examples elsewhere). But one of the policies of SE is that if someone can improve something, there's a button to do just so: edit. – Braiam Feb 19 '17 at 13:54
  • @GypsySpellweaver man, you will be a hell of a reviewer. That's actually what we look for. – Braiam Feb 19 '17 at 13:55
  • 2
    I agree. If something says C:\\file.txt then people should generally assume that subdirectories are double-slashed. Such an edit may be a bit more clear for some people, but makes the Edit history more cluttered. The answer, as provided, wasn't wrong in any way; a comment added below the answer would accomplish the same positive thing. – TOOGAM Feb 21 '17 at 23:50
4

If it was one of my answers I'd accept it. Otherwise, it depends, but I'd not auto accept it and I'd more likely than not pass on or reject the edit.

Part of the problem is the lack of context. In a comment you could go "This answer is fine but X does not work when Y happens, so you can workaround it".

Without this context, even if I tested it, I'd probably test the original version and go "Hey, this works, WHAT IS HE ON ABOUT", and be inclined to reject. Yeah, the edit reason says so but it still looks almost pointlessly trivial

I'm a big fan of comments here. Its relatively unambiguous, is there whether or not its accepted, and can be confirmed by other comments from people who have an interest in the answer. The OP can also roll in your comments and flag your old comments for obsolence.

Ideally once a comment's made I'd try to edit anyway if I were a new user pointing at the comment. That way the information is not lost.

  • Yeah, it could be a sane approach to test both the before and after edit, and if I don't have time, just skip. – Braiam Feb 19 '17 at 13:56
0

The purpose of the site is to provide answers to questions, obviously. If so, as a user, then I need to rely on both the answerer's expertise, (regardless of the vote count), and anyone that has knowledge about the technical aspects of the answer that could have edited it. If I am not aware of this policy and I am relying on the technical validity of the answer, then dire results will occasionally occur.

Since correcting technical details is not acceptable policy, we should definitely make it very clear that, "if someone's answer is judged to be technically wrong as posted, it is not allowed to be corrected by editing", (if it changes any technical details, especially code or command lines).

OK, put me on record as saying this policy drastically reduces the value of this site, IMO. If not made clear, you are misleading users. How? Common sense would tell users to expect that, if editing was allowed, any wrong or incomplete answer, answers that need caveats, or it becomes wrong at a later time, (for instance, when Win-10 breaks it), it would likely be corrected by anyone knowledgeable.

Frankly, I'm very discouraged about using, contributing, or recommending this site with this policy without clearly disclosing it. If recommending it, I will definitely write about and let users understand this policy, so they won't rely on any answers being vetted by experts on that subject.

There IS a possible solution / workaround, if this following policy is allowed to be added:
If you disagree with the answer, instead of editing over the answer, or relying on a comment to correct it being noticed by users, add your corrections at the bottom of the answer, in a separate section for of a dissenting opinion. Hopefully your dissent/correction will include supporting evidence and or links. I'd also suggest a convention of putting asterisks on the OP's incorrect part of the answer. This way, if there was a technical debatable point, I would notice it, and would see all the evidence inside one answer.

Please feel free to correct and edit this answer, so it will help others. If I'm way off on this or misunderstood, then I will delete this answer so it doesn't further confuse users.

  • I would add to this that when adding a supplement that technically conflicts with the original answer, include an attribution to make clear that those are not the words of the original author. For example, start the addendum with "Edit by DaaBoss" and make the name a link to the user profile. – fixer1234 Feb 24 '17 at 1:23
  • Notice: is part of SE core philosophy to improve every post by editing. It's explicitly encouraged in the help center about editing. So, there's no need for "workarounds" of policy. The policy is clear in intent and wording: you feel it should improve it, nobody can tell you not to edit. – Braiam Feb 25 '17 at 15:38
  • Notice: this specific question represented unusual circumstances. The OP originally authored the question with the title "Why was this edit suggestion rejected?" We had a rather lengthy discussion explaining why, and the OP was nice enough to concede that they didn't fully understand the process before asking their question. Unfortunately, the moderators deleted that entire conversation. With good intentions, the OP changed the topic subject to read "Should users suggest technical edits?", when that really wasn't appropriate... it should have been an entirely separate question. – Run5k Feb 25 '17 at 19:12

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .