# Possible improvements of the “How to answer” page

I was looking over the How to answer help page, and upon reading it, I and a few others agreed that it was not as helpful as it could be. And so we came up with a list of things that might improve that page Here's a list of what I think maybe should go on the help page:

2. We recommend you use English, as it is the language most users can read. If they can't read it, they can't like it.

3. Refrain from making the fonts hard on the eyes. All caps, bold font, colored text, etc. should be avoided except for emphasis

4. Understand the question. The deeper you can understand what is being asked, the deeper you can provide an answer, and that is almost always a good thing. If you don't understand something about the question, ask. A clearer understanding of the question saves misguided answers from happening.

5. Stay on topic and think your answer all the way through before posting. Hints, for example, that are only dead-end paths will not make your answer useful.

6. Be courteous. Treat everyone on an equal playing field, regardless of whether or not you may think they should already understand something.

7. Watch the tags. Questions with tags you are familiar with are more likely to be answer-able by you, and it also helps cut down 'question searching time'.

8. Learn from the comments and read other answers. It is usually the case that you are not the only one answering, and by learning from others, you can return the knowledge to produce better answers in the future.

9. Look it up. Many problems have good related content from the internet that could improve your answer.

10. We recommend pictures sometimes. While not necessary, it is like icing on a cake, and often times a picture is what was truly needed to answer the question.

11. If you think something out of immediate context is needed in your answer, reference it. This helps avoid readers' confusion and clarifies and distinguishes between things that may be similar.

12. KIS. Keep it simple. Unnecessarily complicated answers are less likely to be understood or fundamentally undermine efforts to help the OP.

13. Use MathJax. Answers that don't use MathJax or use it incorrectly are subject not only to downvotes, but often times an unhappy crowd.

While not all of it should go on the page, I think there is a lot of useful content here. Feel free to edit this and add on.

• I would like to post an answer going through these point by point, suggesting rewordings and some removals (and possibly some additions), but it would be easier if I could click the "edit" button and copy the raw text. :/ – Wildcard Jan 30 '17 at 22:33
• Overall I think the "How to answer" page could indeed be improved, but that these points are far, far too much detail and also mix the unimportant with the important. It's better to stress only the important on the page itself, and then perhaps link to this Meta post for more detail (for those who can read large volumes of text easily). – Wildcard Jan 30 '17 at 22:34
• @Wildcard That sounds like a great idea! And you could copy/paste most of this and use however you want. Feel free. – Simply Beautiful Art Jan 31 '17 at 0:54
• After a short search, the only site I have found with a customised /help/how-to-answer page is Code Review. Even Stack Overflow uses the default. (I haven't checked all sites, but there's only a handful I would suspect to have had this page customised.) (As an aside, I wonder if the effort made to have this page customised would have any discernible effect on the quality of answers posted to the site.) – user642796 Jan 31 '17 at 18:59
• @arjafi♦ likely so, but the MathJax page has had an amazing impact. The only thing I can truly imagine from this would be good, and so I hope. – Simply Beautiful Art Jan 31 '17 at 19:06
• I can understand @arjafi's parenthetical comment (wondering if a revised/customized page to replace the standard "How to answer" guide.) Also it my educated guess that the majority of answerers here have not yet read the "standard" guide to answering questions. – Namaste Jan 31 '17 at 20:38
• That said, I do think that the discussion about answering standards/expectations here on meta is worthwhile. What has been repeatedly addressed here (almost to the exclusion of addressing quality of answers) is the quality (or lack thereof) of questions on MSE. So I do think that the propositions posted here can, at the very least, hold answerers, and not just askers accountable to maintaining some level of standards, particularly when there is significant community consensus to emphasize what already are standards, and perhaps to even raise the bar. – Namaste Jan 31 '17 at 20:41
• @amWhy exactly. If there was one thing I would insist needs to be stressed heavily is proper english and a somewhat standard way of always including math into the sentences of the post. It might sound strict, but it's how formal math is written and it isn't really that hard to do. A little extra work on the part of users will go a long ways in making posts easier to understand and more worthwhile. – user64742 Feb 1 '17 at 4:58
• I feel like the guide should contain needed information but be as concise as possible. Some users, seeing a huge flood of text, will just not read the page – suomynonA Feb 2 '17 at 3:31
• I object to "bad questions should be avoided from answering unless you can provide an answer that outweighs the question." For one thing, what does it mean to "outweigh the question"? This is vague enough to make it almost meaningless. More importantly, inexperienced students need help, and they might not know how to ask good questions. Those who feel inclined to help them should be encouraged to help them. We don't have to upvote weak questions, but if we can provide a helpful answer we might as well do that. In general we should be open to inexperienced and weak students. – littleO Feb 2 '17 at 22:27
• @littleO But a major problem: Does answering bad questions promote bad questions? Overall, I think it may be better to avoid "bad questions", unless you can make it good. I agree it is vague, but I have little more to offer. – Simply Beautiful Art Feb 4 '17 at 13:52
• @littleO, I think points 1 and 2 in my set of guidelines cover this well. Also note that one "off-topic closing" reason is: This question is missing context or other details: Please improve the question by providing additional context, which ideally includes your thoughts on the problem and any attempts you have made to solve it. This information helps others identify where you have difficulties and helps them write answers appropriate to your experience level. Oddly, this is not mentioned at all in the "off-topic" help page. – Wildcard Feb 5 '17 at 7:55
• "KIS. Keep it simple." Is that not contradictive to itself? Why add the abbreviation "KIS"? – Therkel Feb 6 '17 at 11:23
• @Therkel As far as I know, KIS is a well-known acronym. – Simply Beautiful Art May 7 '17 at 23:35

I went through your thirteen points and consolidated them to the following points, which I've arranged in a logical sequence:

1. Understand the question before you answer it. (This implies incoherent questions should not be answered.)

5. Use good taste in formatting. MathJax should be used for formatting equations and notation. Boldface, italics and the like should be used (sparingly) to make your answer easier to read, but not to shout. Paragraph breaks are often helpful.

6. Be courteous. Do not disparage someone for lack of knowledge, even basic knowledge. Help them to learn. (Lack of willingness to learn is different from lack of knowledge, but be courteous even then.)

7. Provide links where appropriate or helpful. The internet is a wide resource. Prefer links that you think are likely to stay valid for a long time.

8. Diagrams can be extremely helpful.

9. Simplicity is simpler than complexity.

(I plan to add relevant links later on, such as "MathJax" and "off-topic." I don't plan to change the text.)

As much as we Meta-users all (obviously) enjoy writing and reading, guidelines should not assume users who like to read. They should be short, succinct, to the point. Not elaborative.

Celtschk has further consolidated these points.

Let me do another consolidation step based on Wildcard's:

1. Understand the question before answering. If you misunderstand the question, your answer is unlikely to be helpful. Also take the tags into account.

4. Make your answers understandable. If the answer cannot be understood, it won't help anyone. Don't make things more complex that necessary. Where appropriate, include diagrams. Provide helpful links for further reading, but don't replace essential information by a link.

5. Make your answer readable. Try to use correct English. Use MathJax for formulas. Use emphasis and boldface where it improves readability, and avoid it otherwise. Avoid overly long sentences and paragraphs.

6. Be courteous. Answers are there to help, not to humiliate or to show off. If the question author lacks knowledge, don't criticize him for it. Provide that knowledge instead.

Edit: Improved points 1 to 3 based on the suggestions in Wildcard's comments.

• I like this consolidation a lot. Mine are more "zen-like"; the original list (in the question post) are extremely verbose; this is a nice middle ground. My biggest suggestion is that you choose another action word besides "make"; it seems overused. – Wildcard Feb 7 '17 at 6:53
• "Deserves an answer" gets into thorny emotional ground. "Avoid answering bad questions" is clear but still does not preempt individual judgment. – Wildcard Feb 7 '17 at 6:56
• Point 5 is really a subset of point 4, but there are a lot of sub-points in both.... – Wildcard Feb 7 '17 at 6:59
• "1. Understand the question before answering. 2. Don't reward off-topic questions with answers. 3. Review existing answers so you aren't redundant." 4, 5 and 6 in your post are perfect; I can't suggest any improvements. :) – Wildcard Feb 7 '17 at 7:01
• @Wildcard: Thanks for the suggestions; I've worked them in. – celtschk Feb 7 '17 at 7:50
• @Wildcard: I just noticed that I somehow had overlooked your previous comments ;-) I made some improvements based on those, too. Also, this resulted on the bold points always being at the beginning, which is also a positive effect. – celtschk Feb 7 '17 at 8:20
• There is one thing that is not very clear to me. What is the way to know when a question is a duplicate?, for sure there are some "basic questions", specially in arithmetic, that one knows that are probably duplicates, but in general I think there are no rules about how to identify duplicates. – Xam Feb 8 '17 at 19:21

Here's my opinion of the guide:

I think this all depends on how the question is bad. A question can lack context but that doesn't make it inherently bad. Furthermore, a question with horrible grammar than be edited and fixed still deserves an answer. I would recommend something along the lines of "questions that are too trivial or not mathematical". After all, many questions can be 'bad' but still fixable via edits. We need to encourage a bit of participation in that aspect and not punish people solely for some people understanding versus other people not.

1. We recommend you use English, as it is the language most users can read. If they can't read it, they can't like it.

Additionally, it should say "Also, while this is an informal website site one should use the same sort of grammar and writing as they would in a school setting or formal mathematical proof."

1. Refrain from making the fonts hard on the eyes. All caps, bold font, colored text, etc. should be avoided except for emphasis

I personally think that those things should be avoided period except to make comments in a post for direct revision (like right now). Honestly, I would make it clearer that they are more for use in chat/comments than in answers and questions.

1. Understand the question. The deeper you can understand what is being asked, the deeper you can provide an answer, and that is almost always a good thing. If you don't understand something about the question, ask. A clearer understanding of the question saves misguided answers from happening.

That is good, but I think a "the deeper you can provide an answer" sounds weird. Perhaps just say "The deeper you can understand what is being asked, the more accurate and understandable of an answer you can provide, and that is always a good thing.". I emphasize understandable as context/previous knowledge is fundamental in a proof.

1. Stay on topic and think your answer all the way through before posting. Hints, for example, that are only dead-end paths will not make your answer useful.

I think a second sentence saying something about normal posts is good too. We don't want people thinking this only applied to hints.

1. Be courteous. Treat everyone on an equal playing field, regardless of whether or not you may think they should already understand something.

This is generally false. People know more than others. I think that while being open to what they know and not judging is a great thing, making your answer use only simple terms so as to assume they don't know advanced things is a bit absurd. I think this should refer more to the concept of previous knowledge. Here's my version.

Proving or answering any question in mathematics always depends on the intended audiences level of previous knowledge. Always keep an open mind to what the knowledge is and be respectful when an asker or future poster needs clarification on knowledge they did not accept a true or know was true.

1. Watch the tags. Questions with tags you are familiar with are more likely to be answer-able by you, and it also helps cut down 'question searching time'.

This seems like a weird thing in a guide for answer-writing. I think it should be here but yet it feel irrelevant. I'll leave this open for debate. I cannot decide one way or the other.

1. Learn from the comments and read other answers. It is usually the case that you are not the only one answering, and by learning from others, you can return the knowledge to produce better answers in the future.

Good advice in general. I agree with it. However, I would like to stress that one can never assume their answer is the only answer. If someone answers using an unfamiliar technique or different reasoning, one shouldn't assume that it is immediately "not an answer".

1. Look it up. Many problems have good related content from the internet that could improve your answer.

I would warn against links to wikipedia and/or wolfram alpha solely to replace the body of a question. Links should be thought of best as little tags one can click on to see the definition of an unfamiliar word like in a digital textbook.

1. We recommend pictures sometimes. While not necessary, it is like icing on a cake, and often times a picture is what was truly needed to answer the question.

Once again, warn against overuse and stress diagrams not images. Images can be construed as scanning in a photo of written work. That is bad.

1. If you think something out of immediate context is needed in your answer, reference it. This helps avoid readers' confusion and clarifies and distinguishes between things that may be similar.

Refer back to my previous knowledge comment. Make sure to state explicitly that the person isn't using the knowledge of the post but invoking their own previous knowledge to solve the problem or answer the question. This is mote for proof-like or numerical questions. Literal question about facts can certainly use citations though I personally find it strange.

1. KIS. Keep it simple. Unnecessarily complicated answers are less likely to be understood or fundamentally undermine efforts to help the OP.

Agree/disagree. Answers that are more complicated but make the entire problem clearer for the user is far better. Think of it this way. Someone might solve some integral by direct algebra. However, if one makes a diagram and explains why that is in fact the integral and why it makes sense... it's more complicated but yet it is far more worthwhile. Complexity is not about word count. It's about the believably of a method. Yes, math is concrete but if one wishes to know something they wish to know more than just "the steps to solve it". They usually want to know why it is that way even if they don't explicitly say so.

1. Use MathJax. Answers that don't use MathJax or use it incorrectly are subject not only to downvotes, but often times an unhappy crowd.

Three words: vague, vague, and vague. Please state how mathjax needs to be used "correctly". Here's my version.

Use MathJax whenever you write anything referring to a mathematical object in your post, such as numbers, geometric entities, variables, formulae, etc. Math should also follow English standards. This means that mathematical statements should be written within a complete sentence and as nouns and also sentences should not begin with mathematical symbols as they are not truly words. Answers that don't use MathJax or use it in ugly formats are subject not only to downvotes, but often times an unhappy crowd.

• I mean; the OP actually suggested that people don't use bold and italic font except for emphasis. For the sake of readability, perhaps consider outlining the original text using a block quote, and keeping your input as the regular font? – user400188 Jan 30 '17 at 1:02
• For the record though; I agree with all of your comments. Of course; we need to keep the "How to Answer Page" short enough to be readable. So somehow we will need to incorporate all these ideas into just a few concise tips. – user400188 Jan 30 '17 at 1:08
• @user400188 of course. I'm mostly nitpicking. Concise is definitely good. I just want to avoid making things vague or confusing or leave it open to an interpretation where the post ends up bad. For instance "diagram" instead of "photo/image" since a well made diagram is more appropriate then a photo of scratch work. – user64742 Jan 30 '17 at 2:30
• @user400188 for that, see point 12. – Simply Beautiful Art Jan 30 '17 at 12:29
• @SimplyBeautifulArt no... he means make the actual page itself short and simple. He's not mentioning a tip, he's mentioning how the new list itself should be written. – user64742 Jan 31 '17 at 3:31
• I agree point 7 is irrelevant to "how to answer." It's about how to use the site in general—how to find what to answer. – Wildcard Jan 31 '17 at 4:39

Interesting debate! I will just analyze points that somewhat trouble me.

I mildy disagree with the bold part. Many questions just have a good spirit and a bad formatting, for instance. But also: there is no absolute authority drawing the line between good and bad questions. Many users here are Calc-I or Calc-II students asking questions that are just a bit less than completely trivial. Should we ignore them just because we are "so superior"? I don't think so.

1. Refrain from making the fonts hard on the eyes. All caps, bold font, colored text, etc. should be avoided except for emphasis.

I don't know. I find the blockquote feature quite useful for separating intermediate lemmas in a proof, for instance. And I do not find distasteful to stress some numeric result with colors or a small box. Overuse is annoying, of course, but are we sure this recommendation deserves $\#$3?

1. Watch the tags. Questions with tags you are familiar with are more likely to be answer-able by you, and it also helps cut down 'question searching time'.

We are quite stating the obvious. I think it would be more important to recommend once you are confident with the tag management system, watch the tags. An efficient tag system allows many users to save a lot of time. Anyway, this has little to do with writing answers or questions.

1. Look it up. Many problems have good related content from the internet that could improve your answer.

I completely agree but also see a danger. I would bet on some users being ready to give the following interpretation: we are happy to see you have purchased a copy of Mathematica $29$ and you may answer a lot of questions in no time at all, without even understanding what they are asking for. That has to be avoided at all costs.

1. We recommend pictures sometimes. While not necessary, it is like icing on a cake, and often times a picture is what was truly needed to answer the question.

Diagrams is far better. You may also add a picture of yourself, but only if you are very good-looking. Just joking now :)

1. KIS. Keep it simple. Unnecessarily complicated answers are less likely to be understood or fundamentally undermine efforts to help the OP.

I am a huge fan of overkills and appreciate very much the artistic part of math (like the OP probably does too, given his nickname). In my humble opinion, elegance and usefulness are more a matter of efficiency rather than simplicity. Creative, non-conventional but efficient answers should not be discouraged: they are an essential resource, giving readers a chance to learn new and interesting things. Erdos loved to say my mind is open, and we should not do the opposite here - answers are not for the OP only.

I don't think that is a useful recommendation. Questions are the main entry point on this site. Questions are tagged so that they can be found easier, question titles are displayed on the main page of math.stackexchange, the title and the first lines of its text are displayed on the all questions and the question is the first thing you see if you press on a title.

Bad questions should be improved before they are answered. The person that poses a question has the main responsibility for the quality of a question but the other users should support him, especially if it is a user that has little experience with this site. If answers are posted before they are improved and it actually solves the problem of the person that posted the question, often he has not further interest to improve or clarify his question. What should the community do now? Should other user now overwork the question that it fits to the highest voted or accepted answer? That sounds ridiculous.

So I can't support this recommendation 1 because this sabotages the question improvement process by posting an early answer.

• I can't agree. There are lots of questions that are simply to vague to answer, for example. – Simply Beautiful Art Feb 7 '17 at 11:59
• The reason for posting an answer to a bad question is that someone might learn something from reading it. – Gerry Myerson Feb 7 '17 at 12:07