0

I occasionally get into the situation that I think an answer has issues, and I'd like to propose an improvement. Sometimes the improvement is trivial and obvious - and a plain edit is fine - but sometimes it's a little trickier, and I'd like to original poster to shine his light on the issue; and integrate the suggestion as they see fit.

Right now, that somewhat collaborative review process is only really possible if you have low rep: then you can propose an edit, and it's implicitly just a suggestion to the original author (or other reviewers). The review process is valuable and leads to better answers, and I'd like to see it supported even if you're a frequent poster who may have the privilege of rewriting an answer.

6
  • what stops you from posting your own answer?
    – user221081
    Nov 8 '13 at 11:06
  • @mehow: adding an answer adds mostly repetetive content, and it still doesn't get you the advantage of interaction with the original author. Also, I think it's impolite to copy someone's answer wholesale just to make small, controversial changes. I do repost an answer if it's really different, but that's not the norm. Nov 8 '13 at 12:28
  • @psubsee2003 can you elaborate? Why would they be rejected as an invalid edit when they are in fact valid? Conversely, when they are "generally regarded" as invalid, then rejection is exactly what I want: but that does not happen now because the edit is silently applied due to the rep "priviledge". Nov 8 '13 at 12:33
  • @EamonNerbonne maybe you can elaborate on what kinds of changes you want to be able to propose. The way I read it, I am interpreting "non-trivial" changes to mean that you think there might be an improvement to the code or changes that would change the context of the answer. If that is not accurate, then please clarify. Nov 8 '13 at 13:09
  • Note that a very small percentage of suggested edits are voted on by the post's author. The vast majority are reviewed by other reviewers first.
    – Servy
    Nov 8 '13 at 15:02
  • @psubsee2003 I mean changes that for instance clarify and shorten code, but may require (say) a slightly newer version of the library used. Or perhaps it'll run a little slower. In any case: the change warrants some consideration. For text, a change might be to use a different example that's more directly appropriate to the question, but touches many sentences in the answer. Basically, changes that might be quite significant, but aren't intended to change the core message. However, that's just intent: it might not be achieved due to the complexity and nuance of the change. Nov 10 '13 at 9:42
4

A completely peer reviewed editing system doesn't exist in the way you want, and suggested edits are not the way to add the functionality. You are looking for a way to propose a modification to an existing answer and have the the author or several people who are experts in the topic (or at least well versed) look at it and say "Yes, that is an improvement".

It sounds great, but I think you are grossly overvaluing the quality of the review system. In most cases, the reviewer is not going to be well-versed enough in the subject matter to make a judgement (and in many cases, they will have no clue). Likewise, suggested edits that change the answer are often discouraged (partially for this very reason).

So in the end you are going to have 1 of 3 things happen

  1. The edit will be approved by people who have no clue, and as such the peer review you were hoping for did not happen.
  2. The reviewers are going to see an edit that changes the answer or adds new information, and will reject it (usually as an "Invalid Edit")
  3. You get extremely lucky and happen to get 3 reviewers who know the subject matter or the original author and say "Yes that is better" and approve the edit.

Bullet 3 will be rare since there suggested edit queue moves so fast (at least on Stack Overflow). Edits are reviewed on a first come first serve, so there it would be extremely rare to get the OP or 3 such reviewers who happen to be in the queue at the same time as you propose your edit and they all happen to be shown your edit. So in the end, you are either going to get your edit approved by 3 people who didn't really evaluate the quality your changes like you want or it is going to get rejected.

The appropriate way to handle this is to comment on the post and let the author of the answer know what you think might be a better way. And then let the author make the judgement himself and whether to edit it in or not. Or they might make a comment back to say it is good, and let you edit it in.

The other option is to write your own answer. Obviously this depends on how different your "non-trivial" change is and how it would change the answer. If your code is indeed better, then maybe it deserves to be its own answer.

5
  • Fair enough. But none of these choices are very good. Nov 10 '13 at 9:22
  • Comments work very poorly for review because you can't easily identify the bit of text you mean, but need to describe it. That's OK for simple edits, but if you need to change two bits of text in a correlated fashion it already gets tricky. Similarly if you need any formatting to clarify things; especially code+indentation, then you simply can't express that at all practically. Comments are ok for discussion, but they don't work for review very well. Nov 10 '13 at 9:24
  • Writing your own answer is usually a bad option for two reasons: Some questions already have more than enough answers; an extra answer just isn't going to get any attention. With attention disparities, the stackexchange voting model breaks down - more attention means more votes, means higher ranking, means more attention. Secondly, I get the impression that copying people's answers and making slight but positive changes is frowned upon in practice. You can't attribute your upvotes; so it feels unfair to claim someones work like that unless you really have changed quite a bit. Nov 10 '13 at 9:33
  • @EamonNerbonne I understand adding an answer isn't usually ideal for the reasons you specified but it is an option in some cases. And comments aren't perfect, but are really the best option since the "suggested edit" system is not going to work. Nov 10 '13 at 12:18
  • How about suggested edits only the post author(s) see? Nov 10 '13 at 12:24

You must log in to answer this question.

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