How much improving of an answer should be allowed?

Question

I've seen a issue about reviews developing, and have finally found a good example.

On Stack Overflow we preach that posts are community-owned (to a point) and encourage collaborative editing of content. I have seen numerous times in the suggested edits review queue where I felt the editor edited too much extra into the answer, and rejected them. I just saw an instance here where an answer was updated with some more current information (to cover Iphone 5), and thought it was appropriate and approved it, but it was rejected by three others.

It seems like we need to be careful not to lose our proclivity for updating information.

I don't have any solution yet, just wanted to bring it up for discussion.

Answer 1

Problems

• There are many reviewers who don't know the proper way to review (at least, the effect of low quality reviewers is so grim that review audit has since been implemented and the number of reviews per suggested edit has to be increased).

  What is the proper way to review is also not plainly in sight. I am also not sure if there is any meta post describing in details the steps in a review process. Nevertheless, if you need to search for it, then there is even lower chance people are going to read it.

• Reviewing for grammatical mistakes and formatting is easy. The knowledge necessary to review is low. We only need to have good English background and a bit of knowledge about the language syntax, plus Markdown.

  When I said it is "easy", it is easy compared to major changes (discussed below), but in no way I imply it is trivial work. There was one case where a spam link is slipped into the suggested edit, which requires people to pay attention to the Markdown diff.

• Reviewing for major changes in the code and content is hard, since this requires as much research as the editor, or you must have the knowledge beforehand. It may also require intimate knowledge about the language or technology. Such major changes are sometimes result of an update to the technology, or to fix problems that was not commonly aware of in the past.

  The above are the 2 cases where major changes should be encouraged. In other cases, an answer might be applicable. However, between editing a highly-voted answer and write a new answer, the former will expose the new content more than the latter.

• Currently, the review system trains people (like how you would train an agent in AI) how not to approve of garbage. But it does not train people to find gem.

  There are also ways to detect audit. There is a hole that allows userscript to detect with 100% accuracy and 0 false positive, at least 2 weeks ago - not sure whether the code has been changed. Disregarding the problem with the implementation for a moment, the "training system" may also train people to detect audit and avoid them.

• The site is collaboratively edited. However, in reality, there exists a strong sense of ownership, since the post is tied to the reputation.

  The common suggestion is to leave a comment to the author to improve the post. It is a fair suggestion, and good etiquette. This assumes that the author is around, and is updated with the technology. From my experience, the assumption above is OK in most cases. However, in case it does not hold, important information is buried inside comment - a secondary citizen.

Answer 2

Improving is always welcome, but I think the issue in this particular question you mentioned might be the 'risk' of the edit being wrong.

Because the editor is not the original author, if the added information is not valid then the whole answer could turn invalid too. This could have been the reason behind the rejects.

I would probably suggest the new information in a comment, and let the OP of the answer decide whether it's better to add it or not.

Answer 3

My thoughts — not so much on what should be allowed or disallowed, but what should be practiced.

So, yes, if you see some peripheral or complementary information missing from an answer which you think would make it better, then you either add a comment or you make a suggested edit.

If, however, the answer lacks crucial information due to neglect or ignorance, then I would recommend not editing the answer. Rather, I would recommend posting a new answer which features the necessary information. The concern here is not a petty greed for rep, of course, but that if a hypothetical user does not agree with the necessity of the hypothetical information, then it does not seem proper for their account to credited, whether positively or negatively, for such information.

Of course, we should be flexible, and judge each situation upon its own special conditions. Rules are only as good as the one making them.