Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 157 Stack Exchange communities.

What is meta?
Here's how it works:
  1. Any Stack Exchange user can ask a question
  2. The community provides support, votes on ideas, and reports bugs
  3. Your voice helps shape the way Stack Exchange operates

Some suggested edits like this one attempt to fix non-trivial mistakes in posts, including the addition of a sentence or paragraph explaining why. Since this does change the meaning of an answer in a severe way, we are supposed to reject such edits as too radical (we are, aren't we? but then again how does this make internet a better place?). Indeed I think this edit should have been a comment instead so the OP could either clarify a misunderstanding or learn from their mistake. As long as low-rep comments are not allowed, such an edit is unfortunately the only way (apart from a should-be-a-comment-answer) to communicate their spotting, but apart from that:

In order to make clear that such a modification should not just be an edit, should the system

  • reject too long edits, since a long edit is likely to severely change the meaning, or
  • at least give a warning that he suggested edit might be better suited as a comment, would you like to post one instead?
share|improve this question
I do understand there may be reasons to object this, but I tagged this discussion specifically to understand them, so please don't just downvote to disagree but leave an answer or comment stating e.g. when a long edit is good. – Tobias Kienzler Jul 11 '13 at 8:00
Taking php code out of quote and formatting it into code marks whole code as edit on the version comparison box. Would these edits be rejected as to long too? they require adding 4 spaces per line, more inside lists. With 20 lines of code in list (like 1. do something 2. use this code) it's easily 160 added characters. – Mołot Jul 11 '13 at 8:09
@Mołot Ah, good point. Though Jeff claimed whitespace changes don't count. Anyway, your point speaks at least for only a message instead of plain refusal. And maybe the length of the longest connected edit should be considered instead of total edit length. – Tobias Kienzler Jul 11 '13 at 8:22
You can't [successfully] dictate the way voting works with a tag or a comment. Voting on meta traditionally indicates agreement or disagreement with ideas and the direction is almost always determined by the title of the post. If you ask "Should there be X?" votes will indicate whether people think X should be. Simple as that. If you want a different pattern you need to frame your questions differently. If you want to learn why people don't think there should be X, ask "What is the reasoning behind not having X?". Then make sure the question is framed the same way in the body. – Caleb Jul 11 '13 at 8:34
FYI, the edit you linked to is wrong and the original answer was correct. This is one of the reasons these edits should not be allowed: because the reviewers are not qualified to judge the technical correctness of the edit. They can only judge whether it improved the formatting/spelling/etc. – interjay Jul 11 '13 at 8:42
@Caleb I'm aware of that and appreciate the downvotes, but I'd like to get to know some situations where a long edit makes sense, such as the codification Mołot mentioned – Tobias Kienzler Jul 11 '13 at 8:43
@interjay Indeed, that's why I rejected it - though I assumed there was some correctness in it, it's still too radical and should be a comment instead. – Tobias Kienzler Jul 11 '13 at 8:45
hm, is there a badge for -10 votes? – Tobias Kienzler Sep 18 '14 at 15:45
up vote 0 down vote accepted

I don't think that the proposed measures (automatically preventing large diffs/suggesting to use comments) are a good idea.

In the particular case that you linked, the edit should have been a new answer instead - and apparently this is what the user did: propose a new, corrected answer. The new answer has significant differences (which were rejected as edit) to an existing answer, so this makes a perfectly valid new post. (If there are still non-trivial parts of the original post left, you'd need to attribute them to the original author though, as required by the CC license.)

So instead, the system could propose to create a new answer in case the edit exceeds a certain threshold. But IMHO these cases are just too rare to justify any change to the current review process.

share|improve this answer
Your proposed solution would only work for a limited subset of question types where the answers are code and even the slightest change often constitutes a different answer. For most sites on the network, making a few modifications or expansions on an existing answer and re-posting it would just be plagiarism. The edit system exists for a reason and its use should be encouraged where possible. No automatic system is going to be smart enough to detect the minority of cases where it would be better to re-post. – Caleb Jul 11 '13 at 8:51

While the specific example you link to might have been a bad edit, the fact that it was long was not what made it bad. In a majority of cases, length is a spurious signal not directly related to either the quality or correctness of an edit. If anything my experience is that longer edits tend to be better except in a handful of cases where they are vandalism.

In fact many of the best edits I've seen across the network are long. These often

  • rewrite entire paragraphs for clarity, often because of poor English.
  • incorporate new material the OP said in comments but neglected to integrate into their posts.
  • incorporate expanded versions of quotes or references.

The review system exists for a reason. At some point a judgement call needs to be made that an algorithm is unable to make because there are no direct signals. If you try to lock down the whole system on a strict set of rules you will A) end up restricting all the best possible activities B) force people into using kludges to "get around the rules".

Lastly you really don't want to do anything that encourages comments over edits. If anything it should be the other way around. The length of comments is restricted for a reason.

share|improve this answer
Ah, including comments is a very good point, thanks! Maybe the suggested edit view should show comments in order to make judging that easier – Tobias Kienzler Jul 11 '13 at 8:46
Also removing entire paragraphs (of nonsense / irrelevant information). – hayd Jul 11 '13 at 9:06

You must log in to answer this question.

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