Just out of curiosity I anonymously clicked on the "improve this question" link and see this message:

We welcome all constructive edits, but please make them substantial. Avoid trivial, tiny one-letter edits unless absolutely necessary.

In Wikipedia it makes sense to change the meaning of the article if this helps Wikipedia. StackOverflow is different. The emerging consensus seems to be: You only "fix" a question: typos, broken links, formatting, etc. You don't "improve" a post. If you want to change the meaning, write then a comment or post an answer of your own. Am I right or did I misunderstand something here?

I see many well-intentioned edits being rejected as invalid edits. I think this is because beginners don't realize the difference and are misled by the word "improve" and not guided thereafter. I propose a different wording. Perhaps "fix" instead of "improve", and perhaps an additional sentence

Avoid changing the meaning of the question. Please write an answer of your own or post a comment.

at the top of the edit window?

This applies to "improve this answer" as well.


3 Answers 3


I had somebody edit one of my answers with some reasonable fixes to add clarity and give some more detail about what I was recommending. He also added an entirely new recommendation. I rolled it back.

He insists that what he did was right because of the creative commons license which means I don't own the answer. I insist he was wrong - at least in spirit - because he substantively changed the nature of the answer. I do not think a wholly new recommendation should be piggy-backed on my rep and my work. As it happened, I would not have recommended the specific technology he added - it is restricted in scope and not portable - but I think he would have been wrong in any case.

So I think I agree with you that there is confusion and not just among newbies.


He was right to add mention that file ownership is not changed but wrong, in my opinion, to add in the RichACLS recommendation.

Edit: To be clear, my instinctive reaction is that it is in some way misleading to add new recommendations to an answer or to change the content substantively. The reader is denied the chance to assess the answer properly; the thought process, opinions and reputation of the original author are significant pieces of information in making that judgement.

It should be clear that if somebody adds a spurious recommendation to an answer provided by somebody who has consistently shown good judgement, that the added recommendation gains undue respectability. It also falsely gains the votes accrued thus far by the original answer. I do think that other substantive changes are equally misleading.

If you have your own recommendation. add it yourself. Let your own rep speak for it.

  • 1
    That's your opinion. But I do feel that your opinion coincides with the emerging consensus about what are acceptable "improvements" to a post.
    – nalply
    Commented Oct 19, 2012 at 10:47

I've made a fair number of edits that fix factual errors in posts. Generally, they've been well-received, and I generally accept edits to my posts that do the same...

However, there are a couple of things I see, particularly in suggested edits, that tend to lead to rejection or worse:

  1. Failing to explain the changes. If you're fixing typos, that's usually self-explanatory. Not so much if you're re-writing parts of the post to mean something completely different. If your revision comment boils down to, "It was wrong and I fixed it", then you stand a pretty good chance of rejection - so explain how and why it was wrong, and how your change corrects it! If your edit includes external references that back up the change, so much the better. Also, leave a comment detailing the problem and summarizing the solution you've provided in the edit - not only is this polite, it provides a fall-back if your edit is rejected anyway.

  2. Fixing something that wasn't broken. Just because the answer didn't work for you doesn't necessarily mean it was incorrect - you may have had a different problem! Adding solutions to related issues to an existing answer can work, but unless you have a solid enough understanding of both problems to relate them in some way, you're probably better off writing your own answer... Or even your own question.

Remember, when you edit someone else's post, their name remains below it. Don't treat it like a casual wiki grabbag of tips and tricks, strive to make that original author look good by leaving a clear, comprehensive text with a well-documented history in your wake. If you can't pull that off, write your own.

  • Thanks. Would you update the wording for the "improve this question" link, or is it good as is?
    – nalply
    Commented Oct 25, 2012 at 12:31
  • I think the wording is fine as it is; there's a whole culture surrounding editing, not to mention an endless variety of personal opinions and preferences among individual authors - trying to capture all that in a small bit of text is futile. In general, substantive, constructive edits should do well - hence the current instructions.
    – Shog9
    Commented Oct 31, 2012 at 15:35

I see we have a different definition of the word improve, so let's get that straight out of the away:

Wiktionary definition:

  1. (transitive) to make (something) better; to increase the value or productivity (of something)

    Painting the woodwork will improve this house.

    Buying more servers would improve performance.

  2. (intransitive) to become better

    I have improved since taking the tablets.

    The error messages have improved.

My definition:

It sucks less when I'm done.

The definition on Stack Exchange (pulled out of thin air):

To make a post better readable, fix typos, repair broken links, adding or removing formattting, removing taglines/salutations, adding or removing tags and (for a very limited scope) add additional information.

Additionally, to express it in my favorite syntax: improve != change meaning

Any further questions?

  • 3
    I downvoted the answer because I don't want to discuss the definition of "improve". The subject is: Beginners seem to be misled. Is this true? What can we do?
    – nalply
    Commented Oct 19, 2012 at 9:56
  • @nalply: And I wanted to imply that the word "improve" should not be confusing in the first place. Commented Oct 19, 2012 at 10:29
  • Sorry for the downvote. But I just do not want to discuss the word "improve".
    – nalply
    Commented Oct 19, 2012 at 10:45
  • 1
    @nalply You proposed changing it. Its meaning is therefore relevant. Commented Mar 13, 2013 at 9:56

You must log in to answer this question.

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