Editing highly upvoted answer to redirect attention to different answer with less votes?

Question

Update 07/21/2015 - First, to @SaultDon, I beg your pardon. I jumped to the conclusion that the Open FileGDB driver was also installed as part of the OGR_FileGDB library as opposed to being part of QGIS core. This conclusion led me to think that the Open FileGDB driver would not work unless you followed the process I laid out, thus making your statement of it being a shorter process incorrect. I was wrong.

To All, the primary differences in our answers are that one driver is read-only, the other read/write. This makes neither one more correct than the other, and in a way, both are equally necessary depending on your workflow.

I stand by my statement that the edit was arbitrary, but by the moderator not the asker, as the mod accepted and refined the edit to direct people to a different answer. I think if you have a better answer, add it as a new one. If you think another answer is better, vote it up, especially if it is new and more updated. If someone wants to shift acceptance to a newer answer, that is their prerogative. If the newer answer proves to be better, people will vote it up. That is what the voting process is for.
As others have said in their answers and comments, I don't think there many times when it makes sense to edit an answer, unless fixing links and the like. Even in cases of rashly incorrect answers, I'm torn, but tend to think that removing them if bad enough is acceptable, but short of that, adding a comment to the answer and there pointing to a newer/correct/updated answer.

---

This question is in reference to an Answer I put up a long time ago regarding Support for a File Geodatabase in QGIS.

My answer was edited to include the following statement:

Update: This question was asked long time ago and there is now a direct way to open File GDB in QGIS. Please see answer from SaultDon

I in no way believe my answer is the best answer in any situation, or the only answer. My answer has receive 50 upvotes and was accepted. The answer linked to is a much shortened version of my answer, which includes a few of the same steps mine does.
The steps included in my answer, while dated, are still, so far as I can tell, the only valid means of enabling both Read and Write support for the File GDB in QGIS.

My questions is in regards to how this matches with the editing guidelines as shown in the Help Section:

When should I edit posts?

Any time you see a post that needs improvement and are inclined to suggest an edit, you are welcome to do so. The original author of a question or answer may always edit their own post, regardless of reputation level.

Edits are expected to be substantial and to leave the post better than you found it. Common reasons for edits include:

To fix grammar and spelling mistakes
To clarify the meaning of the post (without changing that meaning)
To include additional information only found in comments, so all of the information relevant to the post is contained in one place
To correct minor mistakes or add updates as the post ages
To add related resources or hyperlinks

I could completely understand if the other answer was more accurate than mine to provide a reference to it. My answer is not wrong though, and adding this reference to the other answer seems to be saying that it should be disregarded. Isn't the upvote process supposed to be the means of indicating the relative strength or quality of an answer? There are only 3 answers, and the answer being referred to has 20 votes, so it isn't like someone might just miss it if they are looking through the list.

This edit just seemed arbitrary as opposed to really improving the answer provided.

Is a question asker allowed/encouraged to Choose a Correct answer based on their opinion and then make another answer point to it?

Answer 1

I'll say a couple of things about this specific example first. I note that, as already pointed out, the edit was made by the asker and that raises the question why they did not change the accepted answer if they wanted one to appear above another. The accepted answer always appears first, regardless of vote count, unless it was posted by the asker in which case vote count dictates order. And of course equal vote counts are random on page load.

I can't really speak to the degree to which answers are correct/incorrect/out of date/whatever. I would normally see that kind of edit as a comment on the answer rather than an edit, but there are already a lot of comments on it - and comments don't bump the question or call attention to themselves like edits or new posts do, so while they can be upvoted as well, they have to be noticed first. This case probably could have been handled better, but I don't think it was entirely out of line either (whuber's comments pretty much cover this).

---

Now, the problem(s) in general with 'outdated' answers and suggestions for addressing them.

First, votes and acceptance can be unreliable when evaluating answers. There's no way around this. There are old questions and answers with a tremendous number of votes that are either flat-out wrong or do have better solutions because of software changes. As discussed at Old questions with correct (but now dated) answers and a few others here, 1) hardly anyone goes back and starts doing mass revotes because of a new answer on a question, 2) even if they did it wouldn't be fair to downvote what was a correct and good answer when it was given. So even if they only upvote the new one, the old one still will have a tremendous vote lead over the old one - likely forever.

Changing acceptance can help there, but again most people don't do it. It would not really be appropriate, as you say, to pick an answer as 'best' and make all the others point to it. Unless of course the answer is now flat-out wrong (not the case here, just so we're clear). In that case I don't see any problem with an edit like this. I see it happen quite a bit with QGIS questions, particularly since we have so many people involved in that project on the site. At the same time I've pinged a number of old answers in need of update, which I feel is the best first step anyway.

In my opinion, the best process to handle an outdated answer is:

1. Leave a comment on the answer pointing out what you think the problem is and could improve it. Give the original author a chance to revise their answer based on current info.
2. If no response, choose your next course of action:
   • Post your own updated answer or upvote one already there. You might leave a comment on either the question or old answer that directs to yours/the updated one. You'll have to weigh how many comments are already present and how likely your information is to be noted when considering if and where to do this.
   • In an extreme case, where comments will be buried, an answer is completely incorrect now, or similar, and you really feel strongly something must be done, you can edit the post with an 'update' block - what has changed, why it matters, and yes, even direction to another solution if present. When doing this it's important to ensure you don't change the existing answer too much - an update block would be preferable to a wholesale revision of someone else's answer.

How and what you make that edit is up to you, but generally the less intrusive an edit the better. Fixing or eliminating deprecated links, correcting code formatting, things like that are fine. Improving code, even by just a few lines, has always warranted a new answer to me and I reject most such proposed edits. Replacing one method with another is way out of bounds, that's a new answer. Adding alternatives to deprecated links if unrepairable would be ok. Maybe it just needs to point out the answer is outdated and not point to a specific alternative.

You are correct in that a lot of (most?) cases answers are left to stand on their own, outdated or not, and new answers with better solutions are simply found lower in the stack - it's up to the person reading the question to find that out and vote for whichever answers they found helpful. Personally, I stick to leaving a comment as opposed to editing an answer in all but the most extreme cases - can't even remember if/when I've done an update block that wasn't on my own answer, but I might have. As I mentioned I most often see it happen with QGIS questions and if not by the OP, then by another person involved in the project.

Answer 2

I made a mistake when reviewing that Q&A by enhancing the edit instead of rolling it back and suggesting that the asker unaccept the first (~50 votes) answer in order to accept the second (~20 votes) answer. That would have brought the second answer to the top.

I thought the intent of the asker/editor was clear and free of any malice. They were trying to highlight that the second answer appeared more likely (to them) to be the one which would provide visitors of today with the quickest solution.

---

Related reading on Meta SE: How do we encourage edits to obsolete/out of date answers?

Answer 3

Regarding point four from the 'when should I edit posts' quote above: In what way, then, would it be OK to 'add updates as the post ages'? It goes without saying that technology is prone to deprecation, so something should have been done. The options for those who feel the need to edit:

• edit the original post to include the more modern answer
• add a comment to direct people to the modern answer

I think that the edit in question was a bit shortsighted and could have been more detailed. However, there was already a post which detailed the solution so why re-invent the wheel?

I've seen numerous SO questions with hundreds of votes against the correct and modern answer, while the actual accepted and deprecated answer only had a handful. It's also quite common to see the other way around. Here's one I found this morning: https://superuser.com/questions/519628/clear-html5-local-storage-on-a-specific-page

Leaving a comment to say either 'This is out of date, see new answer below' on the question or 'this should be the accepted answer' appears to be the acceptable trend in the larger community.

Perhaps, then, the edit would have been better as a comment and let the up votes speak for themselves over time? Unfortunately, the GIS community is relatively small and we don't get the throughput necessary to up-vote modern answers as they surface.

To clarify and summarize, I don't think the wider community looks highly upon editing an answer to refer to another answer - even if the latter is the more relevant solution as the post 'ages'.