Should we strive for perfecting answers? Or should we avoid enhancing answers through edits?

Question

After an answer is posted, I often find myself reading it over and finding ways to improve it, and I usually do so.

Some times, a comment points to an aspect of the question that I've missed so I go on and improve the answer.

In other occasions, new answers are posted not aiming to be a comprehensive answer but just adding one or few points. So I feel tempted to improve my answer considering what I missed (as I did in this question but I don't know if I should have done so).

Other times, weeks or months after I post an answer, I learn something new that I find worth adding, so I go on and improve the answer.

All this situations arise just out of striving for perfecting the answers. Which, sounds good, but might be messy as one answer might show upvotes/downvotes that were in fact given to a older version. Or in other cases might look like I'm stealing knowledge of others (coming from comments or partial answers) and presenting them as my own.

I didn't find a straight answer to this issues on the SE Meta, so I decided to ask what's the Earth Sciences community stand on these issues.

Should we always strive for perfecting answers? Or should we keep the core of the answers unchanged and post a new answer if relevant improvements are needed?

Answer 1

My take on this is that we should we always strive for perfecting the answers. Otherwise, multiple incremental answers would make the answers more difficult to follow and confusing (and the edit history provide that approach).

And such improved answers will provide a much better service to pople visiting the question later.

I think making "summary answers" is good in some cases. And including stuff from comments, partial answers and even full answers is good too, but we should always acknowledge who brought the point first (something that I did not do originally on the linked case above)

Answer 2

I figured I would take this opportunity to respond with a similar take to Camilo's:

If you answer a question, you should always feel free to edit it to include improvements or new information, techniques, data, etc. Like Camilo's answer mentions, posts have a revision history which—barring extreme circumstances where moderators get involved—cannot be redacted by anyone to hide revisions, so there's no worry with people getting confused about what changed or when a change was introduced. (The exception of course is the grace period given to posts for the first few minutes, where multiple edits can be made without leaving a trail.)

However, if you see an answer from someone else, you should not edit it to add additional information or new techniques - that's what the comment feature is for - constructive criticism, clarification requests, etc. Or, if you have enough new information or a substantive different technique to share, you should write your own answer. The exception to this is Community Wiki posts, where the original answerer marked the "community wiki" checkbox to indicate the post is a 'community-managed' effort.

Answer 3

I would say in my case incrementally updating answers are the way to go for answers that relate to ongoing meteorological events. In this case data that is available is mostly GFS data (not always reliable or a pointer to what actually happened). The "real" data comes in the form of reanalysis two or three days later and the high quality reanalysis comes in months later. So incremental updates make for the best approach.