13
$\begingroup$

When a user asks their very first question on a Stack Exchange site, they are shown a modal window with some introductory advice. As explained in the March 2020 announcement of this feature, there is a standard Stack Exchange boilerplate wording for this advice, but the wording can be customized on a per-site basis.

On MathOverflow, we haven't yet gotten around to customizing this modal window, as Martin Sleziak gently reminded us in a meta post a few months ago. Martin's question is thorougly-researched and contains a wealth of information on the matter, including screenshots illustrating what we're talking about, and I refer there, as well as an accompanying discussion in the MathOverflow chatroom for further background. In the discussion on Martin's question, a consensus was quickly reached that that we certainly should customize this modal window. The intention of the current question is to close the loop and settle on the precise wording and formatting of the modal window. Once we have done so, the next step is simply for the mods to tag this question , sending it down the Stack Exchange pipeline.

In an answer to his question, Martin also provided a concrete proposal for what this wording and formatting should be, taking into account the wording of similar modal windows on other sites, the specifics of MathOverflow, and numerous suggestions from the comments. Martin's proposal reads as follows:

To improve the chances of your question getting an answer make sure that it:

Before you post, search the site to make sure your question hasn’t been answered

  • Summarize the problem
  • Provide details and any research
  • When appropriate, describe what you’ve tried

Note 1: As discussed in Martin's answer, the wording in the last 3 bullet points comes from SE and cannot be changed (although the third one can be omitted if needed).

Note 2: There is no hard limit to the length of the modal window, but by its nature it should of course not be too long. Certainly it should be no longer than the one on English Language and Usage, which looks like this, and probably closer to the length of the modal window for Math.SE, which looks like this. For comparison, the current MO boilerplate looks like this.

Question: What should be the wording of the modal window for first-time question askers? Should it be exactly as Martin Sleziak has proposed? Should it be a slight variant of his proposal? Should it be something else altogether?

Some possibilities of changes might include reordering some of the list items, tinkering with key words such as "graduate-level" (which was already changed from "research-level" so as not to clash with the later use of "research" in the non-customizable portion of the text), etc. Suggestions from the comments on Martin's question which didn't make it into his proposal due to controversy or to length limitations include some kind of admonishment against simultaneous cross-posting or some kind of link to Math.SE.

$\endgroup$
15
  • $\begingroup$ The answers I've added below are intended to get the ball rolling. Other suggestions are of course welcome! $\endgroup$ – Tim Campion Apr 23 at 23:50
  • 1
    $\begingroup$ A much more important question is: how do we get the new users to read the text in the modal window? $\endgroup$ – Dmitri Pavlov Apr 24 at 3:23
  • 1
    $\begingroup$ @DmitriPavlov We might consider propping their eyelids open as in A Clockwork Orange. Short of that, I'm not sure :) In all seriousness, your point is good. Fiddling around in private browsing mode, it seems that the modal window disappears after a single click anywhere on the screen, never to be seen again. Maybe we should petition SE to keep its text visible elsewhere on the screen for first-time askers or something... at any rate I'm not sure there are solutions which wouldn't require redoing the software. $\endgroup$ – Tim Campion Apr 24 at 3:28
  • 1
    $\begingroup$ A technique I have seen: no model window is presented, instead the main window is decorated with colored arrows pointing to various parts of the dialog, annotated with some hints. For instance, the input field for the title can say “Need an informative title. Do not use titles like “A question about measure theory”.” But this requires redoing the software. Other than that, I have a strong suspicion that most users now reflexively close modal dialogs. (I certainly do.) $\endgroup$ – Dmitri Pavlov Apr 24 at 3:34
  • $\begingroup$ Actually, SE already has some kind of algorithm which gives you feedback after you click "review my question". I just tried it out with a question titled "I need homework help!!!" and it told me that my question was ready to publish. Maybe we could ask for certain key words to be added -- if the question contains the word "homework" for example, we could have the algorithm recommend that the question might not be appropriate for the site. $\endgroup$ – Tim Campion Apr 24 at 3:34
  • 1
    $\begingroup$ You would think that by now there would be a neural net capable of automatically grading the MO-worthiness of a question. We certainly have a huge dataset. $\endgroup$ – Dmitri Pavlov Apr 24 at 3:35
  • 1
    $\begingroup$ Pie in the sky. $\endgroup$ – Tim Campion Apr 24 at 3:51
  • $\begingroup$ @DmitriPavlov I think that your suggestion sounds a bit similar to the Ask Question Wizard, which was tested on Stack Overflow and latter abandoned. See: What happened to the Ask Question Wizard / Guided Mode? and The Ask Question Wizard Is Now In Testing! $\endgroup$ – Martin Sleziak Apr 24 at 5:07
  • $\begingroup$ @MartinSleziak: Not really, in this approach several steps/forms are necessary, whereas in my suggestion there is only one step. $\endgroup$ – Dmitri Pavlov Apr 24 at 5:24
  • $\begingroup$ Another small detail I've noticed -- the hyperlinks on the English Language and Usage modal window behave inconsistently when clicked -- the one from the non-customizable section opens a new tab whereas the ones from the customizable section open in the same tab. I think it would be preferable to have all links open in a new tab. $\endgroup$ – Tim Campion Apr 25 at 0:54
  • $\begingroup$ And maybe the "helpful links" in the sidebar should be changed. Currently they link to the help page on how to ask which seems to be inferior to this FAQ meta question. $\endgroup$ – Tim Campion Apr 25 at 1:03
  • 2
    $\begingroup$ @DmitriPavlov: The best way to make more users read dialogues is to make the contents clear, concise, and useful. $\endgroup$ – Peter LeFanu Lumsdaine Apr 25 at 10:11
  • $\begingroup$ Are there any exemplary questions that could be linked in the modal? $\endgroup$ – Neal Apr 26 at 16:25
  • 1
    $\begingroup$ @DmitriPavlov A neural net capable of grading the MO-worthiness of the question based on the existing dataset is only useful insofar as the community's behavior over the last ten years reflects the ideal standard of MO-worthiness :) $\endgroup$ – Neal Apr 26 at 16:27
  • $\begingroup$ @Neal: Which it does, to a substantial degree, so such a neural net would be quite useful. $\endgroup$ – Dmitri Pavlov Apr 26 at 16:42
8
$\begingroup$

Vote up / down if you would / would not prefer to add to Martin Sleziak's proposal some language about cross-posting.

Voting for this change does not preclude voting for any other change as well; we may end up making more than one change in the end.

To improve the chances of your question getting an answer make sure that it:

If you have already posted your question on another site, please see our guidance on cross-posting.

Before you post, search the site to make sure your question hasn’t been answered

  • Summarize the problem
  • Provide details and any research
  • When appropriate, describe what you’ve tried
$\endgroup$
6
  • $\begingroup$ I would welcome any discussion about the wording of the cross-posting language, and the choice of meta link, chosen in this answer. In particular, it may be problematic that the meta question linked to is specifically about Math.SE cross-posts, whereas the language in the question refers to other sites in general. $\endgroup$ – Tim Campion Apr 23 at 18:20
  • 3
    $\begingroup$ If included, it seems worth creating a new meta question & answer which contains the same information but which is not specific to Math.SE. (But it could say "For example, if you post a question on Math Stack Exchange and want to post it again on Mathoverflow,.."). I don't think it's a big problem but making a new meta post and linking to it seems pretty easy to do. $\endgroup$ – Will Sawin Apr 23 at 21:39
  • $\begingroup$ @WillSawin Thanks for the suggestion. I have created just such a meta question, copied over Joonas' answer with some commentary, and updated the link on this draft to point there. Of course, that meta question is open for discussion as well. $\endgroup$ – Tim Campion Apr 23 at 23:20
  • 1
    $\begingroup$ As already discussed in another post, “graduate-level” is specific to North America. European-undergraduate-level questions are perfectly legitimate on MathOverflow. I suggest replacing “mathematics at graduate or higher level” with “current research in mathematics”, or “current research in mathematics, as found on arXiv.org” $\endgroup$ – Dmitri Pavlov Apr 24 at 17:18
  • $\begingroup$ Just a comment that despite advancing this suggestion and despite its apparent popularity, on reflection I do not personally support it. Cross-posting is not the most important concern for first-time askers -- somebody cross-posting usually has experience on math.SE at least and some idea that MO is different. Because the modal window is so easy to dismiss, we really need a laser-like focus on what's important in this window -- a single clear message which really jumps off the screen. I think the cross-posting point is a distraction. $\endgroup$ – Tim Campion Apr 26 at 14:20
  • $\begingroup$ After Dmitri Pavlov's comment which spawned this other proposed change most of the comments here were about that other change. Those comments have been moved to chat. $\endgroup$ – Tim Campion Apr 27 at 1:20
6
$\begingroup$

Vote up / down if you think Martin Sleziak's proposal should / should not be modified to use the phrase "research-level mathematics" rather than "mathematics at graduate level or higher".

(Note that this would actually bring the wording closer to earlier drafts of Martin's proposal; the term "graduate-level" had been introduced because the word "research" is also used in a different context in the non-customizable portion of the text. Some arguments against "graduate-level" center on the US-centric nature of the term.)

Voting for this change does not preclude voting for any other change as well; we may end up making more than one change in the end.

To improve the chances of your question getting an answer make sure that it:

Before you post, search the site to make sure your question hasn’t been answered

  • Summarize the problem
  • Provide details and any research
  • When appropriate, describe what you’ve tried
$\endgroup$
9
  • $\begingroup$ See Dmitri Pavlov's comment on another answer for some discussion of this proposal. $\endgroup$ – Tim Campion Apr 24 at 17:41
  • $\begingroup$ Could "Provide details and any research" be changed - for one, I'm not really sure exactly what it's supposed to mean - "Provide details and any relevant background", maybe? $\endgroup$ – Will Sawin Apr 24 at 21:15
  • 2
    $\begingroup$ @WillSawin I agree it would be nice to change that part a bit. Unfortunately, we've been told that those last 3 bullets are non-customizable. We could maybe make a stink about it; I'm personally reluctant to do so if it will delay things. $\endgroup$ – Tim Campion Apr 24 at 21:19
  • $\begingroup$ I have been using the phrase "original research in mathematics" when explaining to first-time askers that are confused about what MO is for. This way it makes it more clear that it's not just doing research in the sense of looking up stuff, or figuring out things on one's own. $\endgroup$ – David Roberts Apr 25 at 0:47
  • $\begingroup$ @theHigherGeometer Somehow it doesn't seem quite accurate to me to say the question must be "about original research in mathematics" -- would you propose something like "make sure that it is related to original research in mathematics" or "make sure that it is relevant to original research in mathematics" or something like that? $\endgroup$ – Tim Campion Apr 25 at 0:58
  • $\begingroup$ Or maybe "make sure that it is at the level of original research in mathematics"? $\endgroup$ – Tim Campion Apr 25 at 0:58
  • $\begingroup$ Yes, that last one is about the sort of thing I had in mind $\endgroup$ – David Roberts Apr 25 at 8:05
  • $\begingroup$ @theHigherGeometer I'm having a bit of trouble getting something like that to flow. If you'd like to add an answer suggesting a change along these lines, I'd encourage you to go for it. I would explain in the answer that the suggestion should be viewed as mutually exclusive from the answer I'm commenting on now, but not from any other answer. $\endgroup$ – Tim Campion Apr 25 at 19:01
  • 1
    $\begingroup$ Related to this part, I've also tried to air out some concerns I have about sufficiently emphasizing it in chat. If anybody has any thoughts on that, I'd be grateful to hear about it either here or over there in the chat $\endgroup$ – Tim Campion Apr 25 at 19:23
2
$\begingroup$

Okay, after several days with no further substantive discussion on this question, it appears that both changes which I suggested as answers are supported by the community. After tinkering with a couple other possibilities, I was not personally able to come up with anything making the words "research mathematics" pop out any further. So the plan will be to go with the following text. I will leave this here for a few more days for any further comments before accepting this answer and tagging ; if anybody has any other suggested changes, feel free to mention them or even go ahead and add another answer with the suggestion; if there's more feedback we can add more changes to the following text before accepting and tagging.

To improve the chances of your question getting an answer make sure that it:

If you have already posted your question on another site, please see our guidance on cross-posting.

Before you post, search the site to make sure your question hasn’t been answered

  • Summarize the problem
  • Provide details and any research
  • When appropriate, describe what you’ve tried
$\endgroup$
1
$\begingroup$

Vote up / down if you think Martin Sleziak's proposal should / should not be used as-is:

To improve the chances of your question getting an answer make sure that it:

Before you post, search the site to make sure your question hasn’t been answered

  • Summarize the problem
  • Provide details and any research
  • When appropriate, describe what you’ve tried
$\endgroup$

You must log in to answer this question.

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