Say, over the time every piece of software evolves, new packages appear. And what if better than the accepted answer appears? I think, especially for questions that are about doing a task (making it a lot easier for example) it's important.
We've had a similar discussion some time ago:
In a nutshell, besides the person who posted the question, nobody can change the accepted answer, even if it is blatantly wrong. What you can do instead:
- Comment on the answer explaining why it is wrong, obsolete or Not The Right Thing To Do (tm). If the author of the question agrees, then he may be able to change the accepted answer. Most often, this doesn't happen if the user is not very active on TeX-SE.
- Answer the question yourself, and contrast with the accepted answer. This will bump the question back to the top of the question list, and you'll get visibility and upvotes, if your answer is good. On top of that, there's a Necromancer badge to sweeten the deal if the question is old and you manage to get a few upvotes.
The TeX-SE community is usually quick to spot problematic (i.e. wrong) answers, but unfortunately some older questions haven't been updated. So, if you see a question that can be improved (and is not community wiki), by all means, post your answer and comment.
One thing to remember here is that the acceptance of an answer does not mean "This is the right answer.". Acceptance of an answer means "This is the answer that helped me most at the time that I asked the question.". With that interpretation, it's clearer that another answer can be the right answer.
Of course, it isn't obvious to the casual visitor that that is the right interpretation which is why the community needs to make it clear along the lines that Martin Tapankov has said. In extreme cases (extremely rare) the accepted answer can be edited to say: "Warning: this answer is now out of date. See below for a more recent one.".