When does it make sense to post an additional answer?

Question

Suppose you see a question, which has an answer of which you think it solves the problem. Assume further that you'd now an alternative way of answering this question. Meta-Question:

Under which conditions should one post an additional answer?

Possible attitudes, which are not mutually exclusive, include (but are not limited to):

• Post your additional solution. It cannot hurt to add some information.
• Post your additional solution only if it has an advantage compared to the already existing answers.
• If the problem is solved, save your energy to solve unsolved problems, regardless of potential weaknesses of the existing answer(s).
• If none of the previous answers has been accepted, this is a sign that the OP is not happy with those, and it is thus perfectly fine to add an alternative even if it has some weak points.
• Only post an additional answer if none of the previous answers has been accepted. If an answer has been accepted, the OP seems to be fine with it.

What is the aim of this post (and what is it not)? Since I have seen this discussion here and there, I'd like create a more official anchor in which different opinions can be presented and discussed. The aim of this post is definitely not to coin some "master" opinion which others are supposed to share and/or adopt. Nor is the aim to restrict anyone in their freedom to write answers. Rather, the aim is really to collect different opinions such that everybody who wants to can share their thoughts on this without entering a comment orgy under a regular question or answer on the main site.

Answer 1

If you can solve the question in an alternative way, then post an answer. Getting to the same destination in different ways often helps in diversifying the knowledge. One solution might be restricted based on package compatibility, or something else.

If you have an answer that is similar but you feel it's an improvement, I'd post a comment with a suggestion in favour of my own code or an improvement. If the number of suggestions are numerous, I might post my code in a link to Pastebin, or just post my answer.

I've seen the latter case happen often when a user wants to answer the question first, focussing solely on providing the means to answer it, without any other consideration. For example, other formatting considerations or alignments or code improvements.

Answer 2

Members of TeX.SE adding answers (I don't mean questions by new users that are accidentally posted as answers) to questions indicate that they appreciate the question and want to contribute solutions.

Those solutions can consist of a new way no one has thought of yet, i.e. based on a variant of existing solutions or of a different format, i.e. using ConTeXt instead of pdflatex, perhaps.

Sometimes a question provokes a new package that provides new possibilities to solve the question not yet found in already existing answers.

In this case it is quite useful to provide a new answer that is different, perhaps even more elegant.

And the package author can shamelessly advertise/promote his/her package ;-)

There is no doubt that questions might lead to further developments in existing packages/classes or even new packages, but the 'other way round' other (new) packages might provide facilities that can solve problems in the relevant question such that a new answer is totally justified.

So whenever a new feature of an existing package or a completely new package might provide another (I don't say better) way of solving the question, it is legitimate to add another answer.

Now, on the other hand, 'low reputation' user may be repelled of adding new answers to questions that already have answers that those my have been provided by high reputation users and being upvoted beyond sky level -- new answers are placed at the bottom of the page (most users sort by votes ;-)) and as such, the answer is to be found somewhere between Earth's crust and Earth's core, but definitely not at sky level ;-) 'No' one sees the new answer, 'no' one uses the new answer later on...

This does not mean that no new answers should be added -- but providing them can frustrate low rep users.

Answer 3

Of the five possible reasons that you suggest, I would ignore the last two. Basically as a long term information resource whether the person asking the question used a particular solution is not that relevant. Of course it's initially relevant, as a site you want to make sure someone's immediate question is answered, but that shouldn't stop you adding more information later, as another answer.

In practice the third one has some merit. That is, often I would not look at, or only briefly scan, questions with existing answers.

Of the first one I don't post second answers normally, in particular not if I think it would "steal a tick" from a newish user (stealing a tick from egreg is of course a plus)

For the second, I would rephrase and say that an alternative answer doesn't necessarily need to be "better" than the existing one, but should demonstrate some tex feature not demonstrated. So perhaps its a picture mode answer to add to existing tikz ones, or an expl3 one, or one using some primitive tex manipulation rather than using an existing package, or whatever.

Answer 4

I am writing this hoping that this is the first answer in 2019 (UTC). ;-) Here are some views which a very similar to those described in the answers by Werner, Christian Hupfer (which unfortunately got deleted, and which I'd love to see undeleted) and David Carlisle, but not identical.

1. I would like to argue that adding an answer makes only sense if it brings in a new advantageous feature compared to existing answers. The advantage does not have to be huge, and might be just the use of a different package such that users with compatibility issues have an alternative, as explained by Werner. However, I would refrain from adding an answer without additional feature or advantage.
2. I also would like to argue that it is valuable to explain what distinguishes the new answer from earlier post. After all, if you write a scientific paper, you first comment on earlier works and then explain what your analysis adds. In this context, it might be worthwhile to avoid strong language. Instead of "You do not need a sledge hammer like ..." you may just say "This can be achieved with arguably more basic tools ...", which contains the same information but cannot be misinterpreted as an aggression.
3. If there is already a comment going in the same direction as the answer I am considering to post, why not acknowledging this comment? Acknowledging this comment does not mean that I learned how to answer the question from this comment, it is just fair play IMHO.
4. Likewise, when elaborating on the weak points of previous answers, why not also honorably mention the strong points? Sometimes, when reading the second or third answer, from what is written there one can get the impression that the earlier answer was a complete failure, which it is not in most cases.
5. One should also take into account the history of an answer IMHO. It sometimes takes several iterations to make the OP clarify their question. Of course, after the hard (and sometimes, yes, somewhat frustrating) work by the first answerer has been done, and the question has become clear, it is rather easy to add a quick answer, which may be more elegant than the one by the first answerer who was initially being lead in the wrong direction. I think that in such cases one should ask oneself to which extent posting an additional answer is "cherry picking", and consider letting it be.
6. And then there is of course the sociological aspect mentioned by David Carlisle. If someone answered a question in the sense that the problem got solved, and a second answer pops up, this may easily make the first answerer feel that their answer is considered inferior. This may have a particularly strong impression on new contributors. Therefore, once again, I would like to encourage users to avoid strong wordings and to politely explain what is new or better or different in the later answer. Even if the previous answer is by a more seasoned user with sufficient self-esteem, the posts will still get read by newcomers alike, and strong or inappropriate wording may discourage them from contributing. Always try to remember the days when you were new to this site, and how you felt then.