14
$\begingroup$

We get a lot of questions on this site whose titles are not written as questions. Recent examples include

  • I don't know how to start with this exercise
  • Finding tension in strings
  • List of international, English language physics grad programs
  • Schrodinger Equation - History

and so on. In and of itself, this is fine; after all, some of these are perfectly good titles. But some of them are not. And I've been getting the sense that on this site, the titles written as questions often tend to be better than the ones that are not.

Improving the quality of question titles is a good initiative that can boost the overall quality of our site. And if title quality really is correlated to asking a question, one way we can do that is by encouraging posters, especially new posters, to pose their titles as questions. I think it's good for two reasons:

  1. Titles in question form tend to be more focused on what the person actually wants to know. You get a specific sense of what the question is about by looking at a question title, rather than just a vague sense of what topic is involved.
  2. Perhaps more importantly, we get a fair number of homework questions here, and in several of those cases the poster hasn't put in even the bare minimum of effort to make the question conceptual, as required by our homework policy. Almost invariably, the titles of these questions are simple statements, like the first couple examples in the list above.1 I believe that in the process of figuring out how to phrase their title as a question, the people who post these sorts of questions will put in some thought that will help them write a better question.

I would like to see us get into the habit of phrasing the titles to our own posts as questions, when we can do so without sacrificing quality. Similarly, let's edit the titles of questions by new posters to also be questions, again when we can do so without sacrificing quality. If we establish a consistent pattern of titling posts with questions, it will encourage new posters to follow that trend when posting their own questions, which hopefully leads to better title quality overall.

Caveats:

  • I'm not saying "all titles should be questions." As I mentioned above, there are perfectly valid question titles which are not questions.
  • I'm also not saying that we should edit titles to be questions just for the sake of making them questions. After all, editing "I don't know how to do this problem" into "How do I do this problem?" does nothing to improve the quality of the title, and remember, title quality is what it's all about. But starting with "I don't know how to do this problem," given the choice between "Period of a pendulum" and "What's the formula for period of a pendulum?", I'd choose the latter. It shows exactly what the poster is really asking.

Do you agree? I'm not only posting this to encourage people to adopt this guideline but also looking for feedback.

And yes, I do realize the irony of titling a question about titling questions with questions with a statement :-P Just go with it.

1There is a separate issue as to whether those questions demonstrate the bare minimum of effort required to meet our policy. The issue of whether those questions should be closed is a separate one that I would be happy to discuss if anyone has input.

$\endgroup$
1

3 Answers 3

Reset to default
6
$\begingroup$

There are really two types of question titles:

  1. Good titles that describe the content of the question in an accurate way.

  2. Everything else.

You want #1, and definitely not #2.

There are many ways to get there, but you don't want to start arbitrarily forcing everyone to prefix their questions with "How do I.." because that would actually harm the titles. See the excellent guidelines here:

https://meta.stackexchange.com/a/10648/1

TL;DR -- you want good, descriptive titles, a question title in the formal tone of a question can be every bit as bad as one that isn't.

$\endgroup$
7
  • 1
    $\begingroup$ Sure, if the choice is between title "X" and "How do I X?" then of course just "X" is preferable. This is not an initiative to require all our titles to be questions. But overall, in my experience, titles phrased as questions strongly tend to be better, especially when the alternative is something like "I don't know how to start with this exercise." $\endgroup$
    – David Z Mod
    Commented Jan 5, 2012 at 20:58
  • $\begingroup$ So you prefer "How do I start with this exercise?" the point is that good titles have defining qualities that transcend rote formula, and that is what you want to be teaching -- per the meta post I linked above $\endgroup$
    – Jeff Atwood
    Commented Jan 6, 2012 at 3:54
  • 1
    $\begingroup$ No, of course "How do I start with this exercise?" is no better than "I don't know how to start with this exercise." But for whatever reason, we see the latter kind of title more often than the former. Like I said in the question, experience suggests that titles which are questions tend to be more descriptive, even if I couldn't tell you why that's the case. So we're doing this as a step toward improving the overall question title quality, not just to have questions as our titles. $\endgroup$
    – David Z Mod
    Commented Jan 6, 2012 at 4:38
  • 1
    $\begingroup$ Can't say I agree with that; every word in the title is rather important, there are only so many words allowed in a title (both in Google and on Stack Exchange) and people read from left to right. Better to start with a useful action word than a bunch of boilerplate text like "How do I..". The problem with this "teach to ask in the form of a formal question" methodology is that it masks the underlying issues we want to truly teach, namely, writing brief, accurate titles that are reflective of the underlying post. $\endgroup$
    – Jeff Atwood
    Commented Jan 6, 2012 at 7:15
  • $\begingroup$ Well, OK, if you insist, I'll drop it. Though I thought I was starting to see an improvement even after just the couple days it's been. $\endgroup$
    – David Z Mod
    Commented Jan 6, 2012 at 7:32
  • 2
    $\begingroup$ I'm not insisting (and there is no need to accept my answer), I just want to make sure you're focusing on the core issue, teaching people the many factors that go into great titles, first and foremost. I worry that people read this as "all I need to do is state it in terms of a question and my title is perfect!" when of course, that's not true. Quite the opposite. I support whatever title forms work and make the site better (and variety is better than any formula), my concern is that it's easy to learn the "wrong" thing if you zone in on a title format rather than "what makes a title great?" $\endgroup$
    – Jeff Atwood
    Commented Jan 6, 2012 at 8:50
  • $\begingroup$ Yeah, I think an edit to the question to clarify that is in order. (FWIW I think/hope I'm justified in trusting that the people who read this question will have the sense to actually read it, and not just take the title and run with it) $\endgroup$
    – David Z Mod
    Commented Jan 6, 2012 at 22:27
3
$\begingroup$

1) Very good initiative! First of all, titles should be informative about the physics question. When a user reads the frontpage of Physics.SE, he (or she) should at least have a general idea what a question is about before clicking on it. Usually it is good to pose the title as a question, but ultimately it is more important that it is informative about the physics topic.

2) Moreover, avoid abbreviations in the title. If an abbreviation is only explained in the question formulation, one still has to click it to find out.

The following is not directly related to the opening post, but rather a couple of scattered thoughts on how to write questions (which someone in the future may write up as, say, an extended version of the faq on how to ask questions ).

3) In the question formulation, if you are reading a book, and there is something in that book you don't understand and you want to ask about, please give the reference! In fact, in different areas of physics, notions can mean completely different things! By providing a reference, you ensure that the answers will be much more focused and useful to you.

4) When linking to the arXiv, link to the abstract page rather than the pdf file.

5) When linking preferably use perma-links if possible to prevent link rot.

$\endgroup$
1
  • $\begingroup$ re 5) the title of the question is actually ignored in links, it has just been considered more useful to provide links with more meaningful titles than just a post number: meta.physics.stackexchange.com/q/1000 But in a more general sense this is important, link rot to external sites, including abstract pages, is a serious problem $\endgroup$
    – Tobias Kienzler
    Commented Jan 4, 2012 at 11:42
2
$\begingroup$

I agree on that the title should be a question. Maybe one could program the requirement of a question mark for acceptance of the title?

As for closed questions. I would be happy if there were a qualifier next to the word "closed" in the main list. For example "closed/duplicate" in the main list encourages people to open the question and find the link to the duplicate. "closed, not a question", or "closed, FAQ", will allow to skip in good conscience. "closed/migrated" would encourage opening the question to see where, etc.

$\endgroup$
2
  • 2
    $\begingroup$ I believe requiring a question mark in the title by technological means has been brought up on meta.SO and probably declined. If it were implemented people would probably just put question marks after whatever nondescriptive statements they were going to write anyway. The thing about more verbose close reasons might be one to bring up on MSO as a feature request. $\endgroup$
    – David Z Mod
    Commented Jan 2, 2012 at 21:15
  • $\begingroup$ counterexample to the question mark filter: "Can you help me?" But I think there is some kind of "poor question title" filter already in existence (though not at meta.SO it seems) that tries to detect vagueness and adding the question mark criterion at least would not hurt. $\endgroup$
    – Tobias Kienzler
    Commented Jan 4, 2012 at 11:51

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .