11

I have just attempted to make my first edit of another user's post - correcting a reference to a single-letter class name, which otherwise could throw a novice (see paercebal's answer in In C++, what is a virtual base class?).

I was then confronted with the "6 letter limit" for edits because my reputation is below 2K.

Trying to understand the reason for the 6-character limit, I have found and scavenged the following threads: Legitimate edits of less than 6 characters, Change this behavior to allow for spelling corrections and the like: "Edits must be at least 6 characters", How to overcome "Edits must be at least 6 characters"?, Allow under 6 character changes if they are inside code tags, and Suggestion: Allow 1-char edits, if they are within the code block.

These threads are filled with people who disagree with the rule and provide tips and tricks for workarounds. One of these "tips and tricks" is the "accepted answer" for the third link I just gave, which is a question about how to overcome the 6-character limit. Here is the relevant quote from the answer:

look for other changes that can be made

It should go without saying that this is not an answer to my question.

Of the links above, there is only one reference to a justification (that I can find). It says (see comment beneath second question linked above):

The limitations imposed on low-rep editors are largely there to discourage incomplete edits

Is this rationale based on experience - that (without the requirement) users below 2K would otherwise tend to overload the system with edits below 6 characters that are trivial or incomplete, such as the fixing of typos, but otherwise tend to make good edits (when over 6 characters), such that simply removing the ability for users under 2K to edit would detract from the community?

Is it really true that for users under 2K, it is known that edits under 6 characters tend to be incomplete, but that edits over 6 characters tend to be complete?

If so, I understand the requirement. But in this case I would like to recommend that StackExchange add an "orange box" at the top of the "edit" page, appearing only when the editor's reputation is below 2K and only when the number of characters in the edit is less than 6, that says something like:

Edits must be a minimum length of 6 characters for users with a reputation below 2K. This is to discourage incomplete or trivial edits. If you have a legitimate edit of less than 6 characters and your reputation is below 2K, please add a comment, which will be received by the post's author.

On the other hand, if the "incomplete" rationale is not justified, can StackExchange please either remove the 6-character limitation, or simply remove the ability of users under 2K to make edits?

Quite obviously, for highly popular code-oriented sites such as StackOverflow, code edits can be critical, and the length of the edit has no bearing on its legitimacy.

I suspect that the reason this issue has not been attended to is that most people who develop the StackExchange sites, and other users who make frequent edits, have reputations higher than 2,000. In my opinion, to the extent that this is true, this is not a justification to maintain an unnecessary and/or unhelpful requirement for users whose reputation is below 2K (that is, if I'm correct that the requirement is unnecessary and/or unhelpful).

Can something please be done about this? I have taken the time to write this post after running into this issue in order to help the StackExchange sites, so I hope someone else can follow through and recommend that something be changed.

10
  • 3
    Why the downvote? Commented Nov 12, 2012 at 17:08
  • 3
    Downvotes on Meta (per the FAQ) are often used to indicate disagreement. It might very well be that someone disagrees with your feature request. The downvote does not necessarily indicate "bad question" in this case.
    – Bart
    Commented Nov 12, 2012 at 18:25
  • 3
    @BenVoigt I have carefully looked at the two links you provide, and they do not answer my question. In particular, the second states the same answer I've highlighted: Was the post otherwise so perfect that absolutely nothing in it could be improved?. The answer also contains a comment claiming that users under 2K "make trivial bit-twiddling changes". I requested evidence for the latter, and the former rationale is not something that applies any more to those under 2K than those over. This is a serious question I am asking. Commented Nov 12, 2012 at 19:26
  • 1
    @Dan: I am not saying that the question has been answered. I am saying it is not a new question. The other user asked the same question, and is unfortunately still waiting for a good answer. Asking the question again won't somehow produce an answer that the first question didn't get. Instead, raise awareness of the question already out there.
    – Ben Voigt
    Commented Nov 12, 2012 at 19:32
  • 2
    @BenVoigt - I thought carefully before I posted my question. I noticed that the other questions did not provide alternatives. In my question I made a point to provide alternatives, and to ask about these alternatives. I did so in my title, not just the body. Does this not count as a new question? Commented Nov 12, 2012 at 19:41
  • @Dan: I see you added those two links to your question now. Unlike the ones you first found, they certainly did offer alternatives. Your whole question pivots on "code edits can be critical, and the length of the edit has no bearing on its legitimacy". This is exactly the same point made by those two earlier questions.
    – Ben Voigt
    Commented Nov 12, 2012 at 20:16
  • 2
    @BenVoigt - No. My question provides two alternatives. (1) Disable editing entirely for users under 2K (this is in the title of my post), and (2) provide better, dynamic text feedback in an orange box when the editor window is active (not in my title, but only for lack of space). Both of your links, and the links I originally found, offered alternatives in the answers and in the comments - but not in the questions. Because the topic of alternatives is important - as you seem to agree - I created a new question specifically re. alternatives. Commented Nov 12, 2012 at 20:22
  • 1
    @BenVoigt (Also - I considered saying "I added your links to my question" in my comment, but it seemed irrelevant so I did not. But since you raise the point, I will point out that I looked very carefully at the links you provided, came to the conclusion that they did not discuss my question directly, and therefore added them to my list of links that "are filled with people who disagree with the rule and provide tips and tricks for workarounds", which is not what my question is about.) Commented Nov 12, 2012 at 20:30
  • Dan, you're unfortunately wasting your time. This is one of Jeff's personal idiosyncrasies, and [nothing whatsoever ](meta.stackexchange.com/questions/2645/…) will convince him to change his mind. Commented Mar 21, 2013 at 8:07

7 Answers 7

9

My suggestion would be that the character count shouldn't be checked when any change is made inside a code snippet (either inline or block)... even a single character there can produce dramatically different results, in nearly any language.

Since <2k edits are going into a queue for review anyway, I think that's the only change needed.

In the meantime, you definitely should leave a comment on any post that contains an error if you can't fix it yourself.

Of course, this has already been proposed:


Please note that I'm not expressing an opinion for or against allowing low rep users to edit code... just that the number of characters changed is a useless metric of whether a change is trivial, when the change occurs in code.

4

Indeed. The arbitrary limits and hard coded regulations on edits have a very annoying effect on editors wishing to help other people.

Just let people edit and let the community moderate itself like it always has.

2

As you have probably seen, the 6-character limit is completely asininearbitrary, so you should feel morally in the clear to abuse it by making pointless edits just to exceed the character limit.

One popular solution is to link to the Wikipedia page for one of the concepts in the text.

0

Everyone has their excuses, but In my humble opinion, Trivial edits(formatting, spelling, grammar, etc.) are more appropriate than non-trivial edits.

Otherwise you risk changing the meaning of the post to something that the author didn't intend.

4
  • I would think that this has nothing to do with the length of the edit - if anything, it's more likely that an edit under 6 characters will meet your criteria and should be allowed, with edits over 6 characters disallowed. But those are just my reflections regarding your answer. Do you believe the 6-character limit is relevant to the issue you've raised? Commented Nov 12, 2012 at 16:13
  • @DanNissenbaum You are correct in your assumption that i disagree with the base philosophy behind the 6-character minimum limit, but you would be incorrect in your assumption that that needs to translate to an X-character maximum limit Commented Nov 12, 2012 at 16:16
  • Thanks - I have a slightly off-topic comment: Can something be done about this? Commented Nov 12, 2012 at 16:18
  • I consider the neg reps on this comment a badge of honor Commented Nov 12, 2012 at 19:13
-1

The 6-character minimum is arbitrary, just like the 30-character minimum for new answers or the 15-character minimum for comments.

They're based on the belief that, generally-speaking, doing something worthwhile requires a minimal amount of effort. It's pretty easy to find examples that violate this rule, and just as easy to find examples that support it. All things being equal, it rarely hurts to write more...

It's somewhat harder to hit the 2K reputation mark that makes minimum edit lengths irrelevant. But worth it.

1
  • 3
    I don't understand how, generally speaking, it helps to write more when code errors are being fixed. Commented Nov 17, 2012 at 1:49
-2

I would say the system worked as it should. Code Edits are generally frowned upon for anything more than cleaning up the formatting. If you see a problem with the code, you should leave a comment so that the poster can verify and fix, or explain why it is not a problem.

EDIT: Incorrect code: Edit, comment, or provide a correct answer? I would like to point out it is a gray area and that is why I said generally.

The problem is that when you do not have enough rep to do it yourself, and it goes through the suggested edit review queue. I can tell you personally, when I see a code edit to an answer, and the answer was posted by a high rep user I tend to decline it. Rationale being that the high rep users are generally active on the site, and if the editor left a comment pointing out the problem, they would fix it or explain why their answer is correct.

2
  • Can you please provide a reference or justification for this? It adds value to the community to fix a major code error on a software development forum. Making such corrections can sometimes save people a great deal of time and headache, and can help prevent the spread of misinformation. Please explain why this is not a good idea. Commented Nov 12, 2012 at 19:39
  • Re. your edit: this reasoning applies equally as well to code edits >6 characters as it does to code edits <6 characters. My question is specifically in regards to the 6 character limit. As noted, one of my suggestions is simply to disallow code edits for users with <2K. Once code editing privilege is granted, I would think it means that the user is deemed to have a high enough reputation that they can be trusted. Again, why should the 6-character limit have anything to do with the trust level - one way or the other - regarding code edits? Your answer does not discuss this. Commented Nov 12, 2012 at 20:08
-3

When a user with less than 2k reputation suggests an edit, the idea is that they should not just be fixing one problem with the post; they should fix all of the problems with that post. This limit is targeted towards what are clearly small fixes for single, narrowly scoped problems, and those are not encouraged by the system. When someone is fixing all of the problems with a post it should involve editing more than 6 characters.

The reason for this is that the suggested edits need to be reviewed, and those reviews consume time and effort of multiple (2-5+) higher reputation users. Filling the queue with lots of small edits ends up consuming much more of their time, which could be spent on more valuable edits (or other types of reviews).

Other costs with small edits are that, from the time the edit is suggested until the review is complete the post is locked. If someone else would like to make a substantive edit but there is a very minor pending edit the substantive edit can't be made. Additionally there is a limit of 200 pending suggested edits for the entire site at any one time. If the system is flooded with lots of very minor edits the queue can end up sitting at capacity and thus preventing anyone from suggesting edits to any posts. This is less of a problem now, with some of the changes to the review system (instead of constantly sitting near capacity during peek hours; it's usually sitting pretty low) but it's still a concern.

As for what you should do; it's fairly simple:

  1. If you notice a small problem with a post that doesn't involve editing more than 6 characters, first be sure that there are no other problems with the post. If you can find any other problems you should fix them. Don't just change things that aren't wrong though; your edits should be correct and improvements.
  2. If you still haven't edited 6 characters you can comment on that post. Describe the small fix that needs to be made. Either the OP, or another 2k+ user who comes along can make that edit. If the change isn't important enough that you feel it's worthy of a comment, then clearly it's fine to leave it. You can delete your comment if/when the change is made.
  3. You can wait until you have 2k+ reputation to make such changes. Note that even if you have 2k reputation you should still strive to fix all problems with a post when you edit it, but this isn't enforced nearly as much as with <2k users.
49
  • Any non-whitespace (and in Python, whitespace too) change to code is a substantive edit.
    – Ben Voigt
    Commented Nov 12, 2012 at 18:32
  • @BenVoigt Changing one or two characters may have a significant impact, but it can also be quite a bit more substantive to fix all of the problems with a post. In any case, I have described what the appropriate action to take is if a single character code change is the only problem and it's a major issue.
    – Servy
    Commented Nov 12, 2012 at 18:34
  • I am not referring to minor edits, as you indicate, just because they are small in length. This is not an answer to my question. You imply that edits of less than 6 characters are unsubstantive (If someone else would like to make a substantive edit but there is a very minor pending edit ...). I made clear in my question that I am referring to substantive edits of less than 6 characters. A blatant code error of 1 character is substantive, even though it is only 1 character. Can you please explain why substantive edits of >6 chars are more worthy of others' time than those <6? Commented Nov 12, 2012 at 19:33
  • Also, your reasoning that users who make a critical edit should be obligated to "fix all of the problems with that post" makes no sense to me. Why should users be obligated to change all problems with a post, not just critical ones? Commented Nov 12, 2012 at 19:34
  • @DanNissenbaum The entirety of my post is explaining, specifically, why the general guidelines is that anytime anyone edits a post they should fix all of the problems. Asking that makes it sound like you didn't get past the first paragraph of my answer.
    – Servy
    Commented Nov 12, 2012 at 19:42
  • @DanNissenbaum As for definitions of "minor" or "substantive". A minor edit, from the point of view of this post, is one that edits only a small number of the problems in that post; a substantive one edits all, or a significant percentage, of the problems.
    – Servy
    Commented Nov 12, 2012 at 19:44
  • I do not believe you can possibly be correct. Any reasonable system would take into consideration the importance of the changes, not just the number of changes. In other words, an extremely important change should obviously weigh higher than an extremely trivial one. For this reason, I request that you provide justification for your assessment that "the general guidelines is that anytime anyone edits a post they should fix all of the problems" with your corresponding implication that the importance of the issues has no bearing. Commented Nov 12, 2012 at 19:49
  • @DanNissenbaum I am not the one who put the system into place. I am merely telling you what the current policy is, and how the situation ought to be handled. If you don't want to fix all of the problems with a post then simply don't edit it; comment with the problem that you see. While it's a side topic that's not really specific to this question, note that you should be very careful when it comes to "fixing" errors in code. Changing the content of a reply isn't appropriate in an edit; only the OP should make such a change. If the meaning of the post is changed it's an "invalid edit".
    – Servy
    Commented Nov 12, 2012 at 19:53
  • Please justify your assessment of the policy. I don't believe it. (And I don't believe it for the reasons in my previous comment - I strongly suspect that the importance of the post is a factor in both the current policy, and the current guidelines.) Commented Nov 12, 2012 at 19:55
  • Regarding your statement that "you should be very careful when it comes to "fixing" errors in code" - are you suspecting that I am not very careful? I note that there are posts with legitimate code errors, and where correcting those errors provides a service. In the link that triggered my question (see top of question), the error was obvious - an incorrect variable was referred to in later text (an easy mistake to make) - and it could save someone a big headache to just fix it. Do you agree that it is sometimes a good idea to fix code mistakes, such as in this case? Commented Nov 12, 2012 at 20:03
  • @DanNissenbaum One of the rejection reasons in the review queue is "too minor" for which the response text is: "This edit is too minor; suggested edits should be substantive improvements addressing multiple issues in the post.". I can't seem to find the FAQ entry on it at the moment, but what you're proposing has been brought up a number of times (you've linked to several yourself) and you said that you've commonly seen this answer given. Asking again won't change the answer as the policy hasn't changed.
    – Servy
    Commented Nov 12, 2012 at 20:28
  • @DanNissenbaum As for code snippets, if it's clear that the OP meant to type something but got it wrong, fixing it is okay, but if the OP meant to do something and you think that it's wrong then changing it in an edit would be inappropriate. In this case, it does appear that the OP didn't enter what they meant, so editing it would be fine.
    – Servy
    Commented Nov 12, 2012 at 20:30
  • Do you honestly believe that if an edit is extremely important, but only addresses one issue, that the edit should be rejected on the basis that only one issue is addressed? It seems to me that you are being nit-picky. I am addressing the guidelines and the established culture; again, please provide evidence that the culture and guidelines support outright rejection of important edits on the basis that only one issue is addressed (rather than multiple issues). You mention that you cannot find the FAQ's. Are you suggesting that you believe the FAQ's will make such a statement? Commented Nov 12, 2012 at 20:36
  • @DanNissenbaum Yes, I am saying that if you fix just one small issue in a post in a suggested edit and leave many other glaring errors unaddressed the edit should be rejected given the "too minor" reason, which has the text that I described. That is the established policy, given that the rejection reason exists; but I can only find it reference from the point of view of the reviewer, not directed at the person suggesting the edit.
    – Servy
    Commented Nov 12, 2012 at 20:40
  • 1
    @DanNissenbaum I can tell you that On multiple occasions before I reached 2k, I would find questions where code blocks weren't formatted as code blocks. When the asker posts that it looks horrible. Sometimes multiple lines of code merge into the same line. indenting blocks of code is just adding more spaces, but the 6-character limit doesn't count white-space, so if you want to just indent a block of code, you'll find yourself having to invent other reasons why the post needs to be edited further Commented Nov 13, 2012 at 0:32

You must log in to answer this question.

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