How much does it matter that an answer actually answers the question?

Question

Oftentimes I come across an answer that reads something like

On a related note, this other thing happened in the same decade on a completely different computer than the one you're asking about.

or

Yes and here's something else you might not know about assembly.

So I'm definitely not writing about things which are obviously malicious or spam. Just an enthusiast got enthusiastic, just like we all should do.

Sometimes I comment and I put something like

Yes, but how does it answer the question?

and then I flag it as "not an answer". Some of my "not an answer" flags were marked as helpful, some weren't, and some of the answers are still there.

So I wonder if we need to have a think about how this community needs to treat these answers. If we're a question-and-answer site, and need to vote these down more aggressively or remove them, as would happen on our sister site stackoverflow, or if we're relaxed and more interested in these interesting tidbits.

Answer 1

As we all know, the purpose of the site is to provide answers to questions. There is not always one single direct answer to a question and so we have the mechanism of voting for the community to select the best.

Sometimes, on this site and on other SE sites, posts provide a tangential answer that is intended to aid the questioner's understanding of their problem. These do often attract NAA flags.

My approach is to try to follow the SE guidelines;
If the answer has no visible relevance to the question - it is NAA and deleted.

If the answer does directly address the question but is just plain wrong, then it is an answer and an NAA flag will be declined.

If the answer is tangential but does appear to be a valid attempt to provide relevant information then it is an answer. It is then left to the community to downvote if so inclined. Any flag will be marked helpful as it is worth making a check on such answers. Comments on such answers asking how does this answer the question? are useful as it can prompt the poster to give further explanation and hint to other readers to read the answer with care.
An approach I have sometimes taken is to post a comment and wait a few days for the OP to edit their post. If no edit is forthcoming - then delete.

We do tend to get most such tangential answers on questions that have made the HNQ list. The answers are from users new to this site drawn here by that honeypot. However, that doesn't affect the way they should be treated.

The upshot of that is keep on flagging.
Posts that are interesting (to some) but irrelevant are not answers.
Posts that provide supporting evidence are useful and may stay.
Posts that directly address the question are best.

The system isn't foolproof and I would always try to err on the side of keeping poor posts rather than deleting good ones. Some dodgy posts will always get through - we are only human.

Answer 2

It matters quite a bit. This is a question and answer site. There are questions, and there are answers. Answers should be answers to the question, and not a different question; if you want to answer a different question, either find that question or ask it (if you can't find it). There shall be no "replies" or "I had this question too, but I wanted to know about the TI-81 instead of the ZX81" posted as answers. If you want to discuss something, take it to chat – don't go back and forth in comments unless you're civilly debating the proper tags to use for the question or asking for clarification of this term and that concept, and if you're going on about tags for too long it's probably something you should bring here.

Oh, and No Fun. Fun Bad. The reason that we shouldn't have fun on this site is because fun causes people to enjoy themselves, and that distracts them from the menial task of scribing their knowledge into emotionless, hard-to-read walls of text with no paragraph breaks before becoming bored and abandoning their futile quest to convert their passion into fun-free words so that others may read but not enjoy what they have written, knowing full well that nobody will read this far into such a boring paragraph so I can write what I want. As such, I propose that we eradicate fun from this site...

Actually, scratch that.* The rules are there for a reason: to guide and allow you, the community, to create a high-quality database – an archive – of information about Retrocomputing. About history, and preservation, and getting this thing to work with that thing. The rules wouldn't be serving their purpose if they got in the way of that. So treat them as guidelines. That's not to say that you should go against them – questions must be questions and answers must answer them – but feel free to include relevant information that helps the reader to get a deeper understanding of what you're writing. Let's say you've got a question about [INSERT QUESTION TOPIC HERE]. Which is better?

[RUIN EXAMPLE HERE]

or

[INSERT EXAMPLE HERE]

We need to be concentrating on the latter, not the former. And if that means that you have to go off on a tangent about [INSERT DETAIL FROM EXAMPLE] here, then so be it, if it helps to answer the question and increase the net knowledge of this part of the internet. That doesn't mean that this would be a good answer to a question about [INSERT QUESTION TOPIC HERE]:

[INSERT OFF-TOPIC ANSWER HERE]

If you want to write that, ask a question about [INSERT OFF-TOPIC ANSWER'S TOPIC] so that people searching for that can actually find it.

---

*: This applies to the paragraph about fun.

Answer 3

Oftentimes I come across an answer that reads something like

On a related note, [...], or _Yes and here's something else [...]

Just an enthusiast got enthusiastic, just like we all should do.

I assume you are talking about 'Answers' that only contain such sidenotes/remarks but don't even try to tackle the question asked?

Sometimes I comment and I put something like

Yes, but how does it answer the question?

and then I flag it as "not an answer".

Exacly like I do as well. This stuff might be realy interesting, but unless the author doesn't try to answer, at least in part, it's not what should be there. The distinguishing feature of C is being a Q&A site focused on concrete and related answers, not user stories or beautiful memories (*1)

Some of my "not an answer" flags were marked as helpful, some weren't, and some of the answers are still there.

Usually it's a community decision, and some may see this point less strict as you and me. Some actually like such stories - especially when about some of the favourites. Being a community, we may have to live with it - still, I will flag it.

So I wonder if we need to have a think about how this community needs to treat these answers. If we're a question-and-answer site, and need to vote these down more aggressively or remove them, as would happen on our sister site stackoverflow, or if we're relaxed and more interested in these interesting tidbits.

I'm well in favour for a rather strict handling of the topic and answer handling as you might know. And to be honest, me flagging such 'answers' down won't change even if there is a more relaxed aproach adopted, as I feel them to be worth a downvote.

---

Now, when the answer in question also includes a real andswer, just to be followed by lenthy stories, we need to judge it more diferentiated.

For one, if the story is just additional and there is a also a related answer, it comes down to the quality of the answer. I'd be rather relaxed if this is the case - even more so if the answer realy tackles the question before going astray.

Second, there may still be the case that such a story is related to the question. Maybe jsut anecdotal. In this cas I may restain frommarking them down, but this is realy a case by case decision.

---

*1 - I admit, I'm guilty of this - but I try to put my nostalgia always second to answering the question.