How do we deal with outdated answers?

Question

I don't know if this is a solved problem already, but particularly when questions arise about "new space" that are non-historical in nature, the answer might change over the years. Obviously, I think it is too much work to actively maintain every single question and keep it updated a-la-wikipedia, however I think there should be a mechanism to enable this, or at least, change the "accepted answer".

The post that brought this to mind is this one: How long does the refurbishment process of a Falcon 9 take?

The question, asked about four years ago, was promptly and correctly answered, and the answer was accepted by the original poster. Since then, however, the answer has significantly changed and the random internet users that will stumble across this Q/A likely won't read past the first answer with the massive green check mark next to it.

So, what should the procedure be in this case? I can think of several options, but I'm not sure which one is the "accepted" one or even if there's already a procedure in place for this:

1. Do nothing, let the newer, more correct answer languish underneath the old answer
2. Edit the old accepted answer with an addendum containing the updated info or pointing towards the newer and correct answer
3. Somehow change which answer was accepted. If the user who originally asked is still active, this might be possible, but in many cases questions are asked by dead accounts.
4. Something else?

Answer 1

An answer being accepted doesn't mean it's correct in the first place, so I'd immediately eliminate thoughts of changing which answer was accepted. Where the check mark goes is solely up to the judgment of the asker, and the askers don't always have good judgment. Coming to the Q and A later as a learner and only reading the accepted answer is a major failure on the part of the learner. If there's many answers to a question, there's a reason for that, and it's a good idea to look at the community's feedback on both the question and the answers to understand what's correct and what isn't.

I think that updating the content of a different user's answer is generally wrong (and likewise not encouraged by the reputation mechanics of the network). Updating an answer with a note about its limitations is probably better. I prefer comments for this, but comments may be less-ephemeral on Space SE than elsewhere.

Within the mechanics of the site, I think the best things to do are:

• upvote the more recent / updated answer if you think it's correct and/or good. The new answer in the question you link to is a quote from an investment newsletter that goes on to cite a reddit thread and an NTRS paper, so I think at minimum that answer should be updated to use a better source if those upstream sources support the claims, but maybe I'm a pedant.

• leave a comment on the accepted answer directing people to the answer that you think is now correct, if there isn't already one

• leave a comment on the question to see if the asker is open to changing the accepted answer, if they're still an active account. If they're not, oh well, hopefully the future learner sees the comments and the more-recent answer.

Here's an example of uhoh adding a comment on my answer to uhoh's own supplementary answer that had more-recent info.

It isn't clear to me how the rest of the network prefers to deal with outdated answers; the relevant Meta SE QA kicks off in 2009 and has a LOT of history and no accepted answer. The Outdated Answers Project seems related but I guess that's really for when none of the existing answers apply anymore.

Answer 2

While this question focuses on the approach of keeping answers from becoming outdated over time by using various maintenance techniques, we should also consider that the phrasing of the question may directly impact how much "answer maintenance" will be needed.

We can create more maintainable questions by providing revision suggestions in the comments or by directly editing a question. A more maintainable question will reduce the likelihood that its answers do not age well.

For example, instead of "How long does the refurbishment process of a Falcon 9 take?", the question could ask something like, "How is time that it takes to refurbish and re-flight-qualify a Falcon 9 changing year-to-year?"

An answer to this question is more likely to include dates and perhaps a chart. The reader may even be able extrapolate a bit if they see a trend. Most importantly, the reader is less likely to be misled by an out-of-date and unmaintained answer.

Answer 3

Questions about progress don't make for good answers.

There are discussion fora where such things would be better asked.