# What do we mean by welcoming when we're a site aimed at researchers?

I was just browsing the main page this morning, and I decided to look at some of the posts by new people. One of the questions had garnered three votes to close and a comment from a regular user to the following effect:

'Welcome! This website is for research level mathematics, and what you've posted is an exercise. I suggest you post it on M.SE'

I sort of stared at it for a moment, and I thought about the definition of welcoming. Now certainly the message did indeed begin with a hearty welcome. It then immediately turned the user away (inherently not welcoming).

But here's the problem! The comment was right (ignoring the initial comment of 'welcome')!

So there have been complaints that MO is not welcoming recently (I have grown to despise the word at this point, but okay). But the question is this: Can we simultaneously be welcoming and also turn away such a large number of questions? Isn't this simply a contradiction in terms?

I think that the correct answer is really that we ought to be welcoming to people who actually belong here (and that we cannot be welcoming to people who don't belong here). Am I misunderstanding the idea here? Has welcoming become some kind of catch-all buzzword for being 'nice' or 'polite'?

Note: This is not a rhetorical question. I am sincerely confused.

• We have all grown to despise the word "recently" recently. Ugh. What a year! – Asaf Karagila Jan 11 at 9:37
• Read the initial “welcome” as “welcome to the Stackexchange network”. The commenter did not turn the user away, but to a more suitable place on SE. I don’t see anything wrong with it. – Emil Jeřábek Jan 11 at 11:09
• Well, that's "You are welcome. Your question is not really welcome because...". – YCor Jan 11 at 17:54
• Anyway I see little point in these posts that are visibly exercises copied without efforts by undergrad students and this is not the issue. There are also first posts that don't sound like exercises and quite likely are written by people related to research in some way, but are poorly received (because poorly written, showing little effort, because the question is not judged as research-level, because the post is not really a question, is speculative, open-ended etc.), which results in down/closing votes, or possibly discouraging comments. – YCor Jan 11 at 17:55
• @EmilJeřábek Aren't these functionally equivalent for a new user? – Harry Gindi Jan 11 at 20:07
• No, not at all. It is a constructive suggestion what to do with the question. – Emil Jeřábek Jan 11 at 21:44
• Welcome! Your application to our PhD program might be more suitable for another university. – Monroe Eskew Jan 11 at 23:13
• People who ask questions here that are not on-topic might be asking questions that are on-topic in a few years. If they feel MO is generally welcoming, we have a higher chance of seeing them when they are ready, compared to if we give them a curt brush-off. If someone asking an off-topic question feels helped by MO, even if to know where to ask that's more appropriate, then that is still a win for the site. They will tell their friends they met friendly people here, rather than complain about how rude we all are, amplifying our direct actions either way. – David Roberts Jan 12 at 2:20
• I've had people explicitly thank me when I've showed them the MO door, so it is possible. Your experience may be different, of course, and I leave you to reflect on why. – David Roberts Jan 12 at 3:04
• Also, how you "show someone the door" doesn't only affect them, it affects everyone who is reading the site and wondering how they will be treated. – David E Speyer Jan 12 at 3:44
• @FrancescoPolizzi There is already one thread which has some comment templates - although for another situation - questions more appropriate for MSE. There could be similar discussions for other suitable comments - or MO could to the same thing as they did on Mathematics, where various other sites, where there is a single thread where the comment templates are posted in answers: cstheory, TeX - Latex, Mathematics. – Martin Sleziak Jan 12 at 9:30
• Rudeness is culturally dependent, so it may not be visible to everyone. "With the greatest respect" can be an insult. – David Roberts Jan 13 at 0:38
• @DavidRoberts: sorry, I do not really get your point. I suspect that >90% of the users here had no idea that "With the greatest respect" can be an insult. Of course, if someone tells me "Please do not use this expression with me, in my country it is offensive" I will not use it anymore with them. What is your proposal on this respect? Write a list of all apparently innocent expressions in common English that might be offensive somewhere in the world and forbid their usage here? – Francesco Polizzi Jan 14 at 5:43
• @FrancescoPolizzi: I found a lot on Google. Here's the BBC: bbc.co.uk/news/world-us-canada-46846467 As a Brit, I would tend to agree with the thrust of these light-hearted articles. I think, when online, I have learnt to read things in "international English", as it were. Though I still find reading "the so-called ..." in a Maths paper hard: I always hear this as sarcasm, which I think is a uniquely British-English meaning. – Matthew Daws Jan 14 at 9:49

Here are some examples of people we would like the site to be welcoming to:

1. Prominent mathematicians who are less "online" than the typical MO user.

Currently MO has a pretty bad reputation with both such groups. In neither case is this primarily about the relationship with math.SE Certainly the reputation of the site among graduate students in my department is that MO is too intimidating for them to use.

(It's not entirely clear to me what we can do about this reputation or exactly where it's coming from, but that's an entirely separate question.)

• Can the graduate students ease into MO by answering questions anonymously? – Matt F. Jan 15 at 19:29
• I agree with this, but I think those in group #1 might be hard to recruit to the site, while those in group #2 should be gettable. – Sam Hopkins Jan 17 at 17:25
• @MattF. This may depend on the field, but in some areas, I have the feeling it's really hard to answer a question (before someone else does it). Not that it's a bad thing, but I doubt it helps attracting graduate students. – ARG Jan 19 at 10:01
• @SamHopkins: As regular users, yes, but the reputation of the site with mathematicians who don’t use it regularly does still matter. – Noah Snyder Jan 19 at 13:58
• @NoahSnyder I agree that the reputation of the site with mathematicians who don't use it is important, but I think we should exercise extreme caution with regard to certain mathematicians who use twitter to (disingenuously) kick up a fuss about everything. We shouldn't put ourselves at the mercy of a vocal minority who are, frankly, not making criticism in good faith. I have in mind a particular group of people who, to me, come off as a bunch of online bullies. I think criticisms from these people ought to be disregarded wholesale. – Harry Gindi 2 days ago
• It is true that on the social media we often witness the tyranny of the vocal minority. One should be careful and check that the data are not biased. For instance: how many mathematician after all have a positive or neutral opinion about MO, but stay silent? – Francesco Polizzi yesterday
• @FrancescoPolizzi I think there is some truth to the last sentence, but then one has to weigh up whether the positive/neutral opinion about MO is "the good outweighs the bad, but we would like it to be better in certain directions" or "I like it as it is, and would be uncomfortable with changes". (Taking the arithmetic mean rather than a Boolean/Hamming distance) – Yemon Choi 22 hours ago
• @YemonChoi: I agree. I in fact suspect that many mathematicians are just happy to find the answers they are looking for by browsing MO, but are not interested to become an active member. On the other hand, I also suspect that many of those who complain do this on the base of a single experience, and never tried to be seriously involved with the site and its mechanism. It is a bit like politics: there are few activists, while the majority of people just wait and see what happens. And many of the people complaining about politics have no idea what active politics really looks like. – Francesco Polizzi 12 hours ago

Martin's comment and his recent thread on the welcome window hit the nail on the head. The correct way to be welcoming towards this kind of asker is having good "how to ask" instructions that essentially show the asker that this isn't the right place before they post. Mathoverflow doesn't have those instructions, and this is something that we need to change.

One way to reconcile the contradiction is to incentivize our more experienced users to clean up borderline questions. Then instead of sending the message, "Only people who speak our secret language are allowed in these parts," we can send the message, "Hey there! If you don't mind, I'm going to reword your request a little bit to help make sure your question gets attention from our users."

I had a recent success story in this direction with this question. The question had 3 or 4 votes to close when I found it, but I thought the core of the question was interesting and appropriate, so I left a comment and wrote an answer illustrating my point. Someone else then rewrote the question a bit, and now it has 44 upvotes and no votes to close. The original poster hasn't done much on the site since (in hindsight, even my comment defending the question could have been more welcoming) but I think it was a more positive experience than if the question had simply been closed.

I think we should do this more. When high rep users come across a problematic question, it's quick and easy to vote to close it - this has the upside of keeping MO clean, but maybe it has the unintended downside of making MO more hostile. If we can encourage high rep users to try to improve a question before closing it, either through the rep system or some other means, maybe we can get a virtuous cycle where more people are asking questions and then more people join to answer them.

• BTW you have linked to your answer. Maybe you wanted to link to the question - which is On which regions can Green's theorem not be applied? (Or maybe you wanted to include both links somewhere in your answer - but link to the question seems to be a natural guess in the place where you write "the question".) – Martin Sleziak Jan 18 at 6:01
• @MartinSleziak Oops, good catch - it's been fixed. It was easier to search for my answer than the question, and I forgot to change the link. – Paul Siegel Jan 18 at 15:46
• Also sometimes if there's an obvious oversight, it can preferable to edit than commenting. The problem with commenting is that it can generate an irrelevant discussion (such as: "but this counterexample is not interesting, of course I'm assuming this", etc). Of course the editor's role is to judge correctly that this is the right correction. For thorough changes, it's let obvious: in the case raised by Paul, Tim Chow's changed were done after the question was closed. But the problem is that it's not always that clear what is meant behind a poorly phrased question. – YCor Jan 18 at 18:43

I may add my 2¢ : I have the feeling that the single most unwelcoming aspect of MO is the down vote. There are several reasons for that: unlike a close vote, it is visible for all and therefore tends to have the effect of a "slap in the face". When new users respond aggresively it is typically after their first down vote. Moreover, while a close vote may eventually lead to closing of the question, this then comes with suggestions to improve the question, so it has a more constructive feel. Down votes also tend to arrive very soon after the question is posted, giving the impression that users did not spend any time thinking about the question.

I ask myself if down votes serve a purpose, and I am not so sure they do. Yes, they advertize on the front page that a question is not worth looking into, but seeing a question with 0 votes that has been around for some time conveys the same information. If a question is spam, then flagging it as such is quite effective at triggering a deletion, down votes do not seem needed for that purpose. Down votes do serve a ranking purpose for community wiki, big list, types of questions, but those are few and I guess that counting up votes would work just as well.

I don't think we can change the way Stack Exchange works, I presume the option to down vote is here to stay, but we could decide as a community that we will not down vote off-topic questions, just close them. All for the goal of making MO a more welcoming place.

• Re: I ask myself if down votes serve a purpose. One thing which is influenced by the negative scores is auto-delete (roomba). This topic was discussed a bit in some threads on this meta: Decluttering MathOverflow and Should we try to re-start manual deletions or is the situation fine anyway? – Martin Sleziak Jan 17 at 12:47
• Now I noticed that you have also mentioned the relation between negative score (downvotes) and roomba in an older answer here on meta: Remove down-voting or show down-votes in the separate counter. – Martin Sleziak Jan 17 at 13:50
• what about downvoting answers? – Francesco Polizzi Jan 17 at 14:58
• downvoting answers: I don't think that is a main factor in the welcoming/unwelcoming issue, since new users are typically on the asking side; and yes, down votes have a purpose in housekeeping, but as I follow this discussion on Meta I am inclined to think that their negative impact on new users outweighs the housekeeping benefit. – Carlo Beenakker Jan 17 at 15:08
• I'll just mention the fact that a negative balance of $\le -3$ allows to user with enough rep ($\ge 10K$?) to vote deletion as soon as the question is closed while in general it requires a few days after closure. This is particularly useful for questions that are highly off-topic such as high school exercises. – YCor Jan 18 at 0:13
• For context, let me copy what appear when the mouse is on the downvote button: "This question does not show any research effort; it is unclear or not useful." For an answer: "This answer is not useful." – YCor Jan 18 at 0:18
• @MartinSleziak have you got any data confirming Carlo's claim that new users are "typically" on the asking side? Esp. counting questions that aren't closed. Are new users starting with an answers really that negligible? I actually asked my first question in MO 3 years after giving my first answer (Sep 2014 vs April 2011) and had more than 100 answers then. Of course this is extreme, but I wanted some familiarity with the site before asking questions. – YCor Jan 18 at 0:26
• @YCor I have posted something in the MathOverflow chatroom - although I am not sure how close that is the data you're looking for. – Martin Sleziak Jan 18 at 5:39
• I agree that 0 upvotes with many views is as good an indicator as anyone needs that a question might not be the most interesting thing in the world, and with the close vote/flag system in place I also agree we don't really need them for very off topic questions. I know that all of stack exchange isn't going to change, but I also know that MO has a special contract the details of which I routinely forget; we might be able to remove the downvote option on questions if the community decides it wants to, I'd be interested to hear from someone well-informed on this topic. – Alec Rhea Jan 19 at 9:15
• @AlecRhea That kind of flexibility was lost when we moved to MO 2.0. What remains of our special deal with SE is that we can take it or leave it, and if we leave it, they'll give us our data. Also they don't run ads here. – Harry Gindi Jan 19 at 9:19
• @HarryGindi Ah, cest la vie, thanks for letting me know. – Alec Rhea Jan 19 at 9:36

Short answer: To be welcoming probably means to

1. Be patient and polite when explaining the scope of the website to those who aren't very familiar with the scope of MO.

2. Give newer members an opportunity to get involved with voting to close questions. Having only 6-7 names appearing in every closed question can be intimidating.

3. Encourage professional behavior in order not to turn away those that are familiar with the scope of MO.

Whether Mathoverflow is perceived as a welcoming community depends on the observer. If it is someone who is not very familiar with the website, then I think that being welcoming boils down to politeness. This is not just about redirecting questions that may not be research level, but also explaining various other procedures such as other members of the community editing one's posts, changing the tags, closing as duplicate etc.

There are aspects of the website where the community's consensus is constantly evolving. When it comes to deciding what constitutes a "research level" topic, some users have a higher bar than others. As a corollary, the experience of a new contributor will be shaped differently depending on whom they come into contact with. For example, I have noticed that user abx has been vocal about casting votes to close on algebraic geometry questions that others might find ok or borderline, you may even have experienced this in a recent question of yours. I don't think that there is anything wrong with this (in fact I believe that many topics on MO could benefit from "raising the bar" a little bit) but perhaps it would be helpful if a new user could tell the difference between the opinion of a handful of users and the community as a whole.

One way of achieving this is to discourage users from identifying themselves with all of MO. In the context of borderline questions, using phrases like "I think that your question is covered in a standard textbook like X and might have a better chance at attracting more elaborate answers at website Y" is probably better than an absolute "your question is not acceptable/welcome/appropriate on MO". They essentially achieve the same goal, but the perception can be different. I guess the desired perception should be more along the lines of "community moderated" and less "policed by a small set of users". The first scenario is one where every member can feel like they have a voice whereas the second is one where some members are left to feel sort of helpless.

I think that the case of borderline questions is truly the place where you can clearly see how welcoming we choose to be as a community. Perhaps if we were able to treat them with more dialogue we might start seeing a change in perception. Here is what I think is a great example, where instead of getting closed, a question received constructive criticism as well as an edit that improved it.

Now, for the case of someone who is familiar with the scope of MO, there are other things that matter more. The impression I get form personal conversations is that some of the factors which may paint an unwelcoming picture of MO include: 1) Non-mathematical debates that turn offensive or rude. 2) The lack of a unanimous agreement against hate speech/symbols (without bias on which country or culture they come from). 3) Users regularly posting controversial statements and then deleting them for the sake of creating confusion etc. Running into such threads on MO or meta can give the impression that this is not a professional environment. This can turn potential users away because they do not feel like the level of decorum is high enough to protect them from personal attacks, prejudice, bullying etc.

• I do rather like the first 80% of this answer. But I don't know that there is even a solution to the last paragraph apart from something that amounts to a political purge of the userbase here. The complaint about a lack of unanimous agreement, for example, is about the inner thoughts and feelings of members here. It's kind of totalitarian. – Harry Gindi Jan 12 at 2:14
• @HarryGindi My impression was that some of these examples arose as a result of us being too confident that we could handle talking about any topic. The solution is not to find out who is right, so we can talk more about the topic, but rather find a meeting point where we can continue discussion while maintaining civility, and then direct everything else to blogs and social media. – Gjergji Zaimi Jan 12 at 2:15
• If there is a debate about MO policy that ends up intersecting with some hot-button issue (I'm thinking about the confederate flag case), I can't imagine a better place than meta.MO for people to have a discussion about how it should be dealt with. But anyway, thank you. At the very least, the final paragraph is informative to me (whether or not I think it's a reasonable complaint, it is at least something concrete). – Harry Gindi Jan 12 at 2:20
• @HarryGindi I want to add that I'm afraid I disagree with your interpretation of totalitarian. I was born and raised in a country that used to be totalitarian and I think I would recognize it. Putting ground rules in your workplace (eg. Don't graffiti the walls when you come to my office hours) is a healthy thing to do. – Gjergji Zaimi Jan 12 at 2:24
• I think the hallmark of totalitarianism is not just trying to control people's behavior but also their thoughts. It's in that sense that I used the term. – Harry Gindi Jan 12 at 2:27
• I agree with 1. and 3. I am critical about 2., because voting to close a question is somehow delicate and requires IMO some experience and involvement with the site, so I would not grant this privilege to absolute newcomers to MO. "Having only 6-7 names appearing in every closed question can be intimidating.": this might be true, but in MO there are much more than 6-7 people with the privilege of closing questions. If there are only 6-7 people willing to use it, I do not see how extending the privilege could really help. – Francesco Polizzi Jan 12 at 6:38
• @FrancescoPolizzi I didn't mean to extend privileges to newcomers. Perhaps I should have qualified it as "new 3k+" users or some such thing. As for your second point, I was imagining that if some of us who have been around longer took a step back when it came to closing votes then others might step up to the task and it might 1) bring a fresh perspective 2) look better on the outside. (I am referring mostly to the situation of a borderline question. When it comes to something that is obviously inappropriate I don't think it matters much who casts the votes.) – Gjergji Zaimi Jan 12 at 6:53

One might compare the overall population of MO and of sci.math.research, active in the 200$$\ast$$'s. On sci.math.research, there were no rep, no down/up votes. Posts were filtered by moderators. My first-glance impression was that these were similar populations, although MO's is broader. It would be interesting to have data about this, and feedback by moderators of sci.math.research (which are listed here, at least at some point).

• This seems right on the money! – Harry Gindi 2 days ago

I am not sure what specific suggestions I can make for how to make MathOverflow more welcoming, but I do want to say that the MathOverflow site—while tremendously interesting—does seem to be fairly unfriendly to new users, even ones who might be worthwhile contributors to the community. Although I first started posting at MathOverflow with a full understanding of the level and goals of the site, I nonetheless felt that the responses I received were rather shallow and dismissive.

Let me give a few words about my own background. I am a high reputation user on a number of StackExchange sites, so I know all the mechanics of how the stacks function. I am also a peripheral member of the research mathematics community. I have a Ph.D. in mathematics, although my regular job is as a professor of physics, and most—but not all—of my research is in theoretical physics. Several years ago, I made a couple of posts on MathOverflow, and I definitely did not feel like I got a warm reception. The responses I got could best be described as "chilly and cliquish." I felt like the way I framed my posts (based on my way of thinking as an applied mathematician) was not especially appreciated or even subtly sneered at. Besides that, there just seemed to be essentially no interest in the kinds of approach I was interesting in taking.

I considered and ultimately decided to suspend my account, and although I continue to read this stack, I did not post here again for many years. I only very recently reactivated my MathOverflow account, after I was encouraged to do so by some other colleagues. The response to the couple of questions I have asked since then has not been rude, but it has still been somewhat disheartening. I have endeavored to formulate my questions in a very pure mathematics language, although there have been more applied motivations behind them. The one substantive response I received was shallow and rather missed the point of the question, by trying to abstract it even beyond the level of artificial mathematical abstraction that I had already inserted into my query, in the hopes of attracting some interest from the very pure crowd who appear to predominate on this site. I am not going to be delete my account again, but I do not hold out much hope of getting useful answers to the kinds of applied math questions that I might really want to ask.

• Somewhat similar observation was made also in some older posts: applied math being underrepresented, biased a more towards pure math: What areas/aspects of mathematics are underrepresented on MO? and arXiv vs MathOverflow - popularity of disciplines. – Martin Sleziak Jan 15 at 5:59
• Where is the question you are talking about in the last part of your answer? Did you delete it? – Francesco Polizzi Jan 15 at 6:12
• @FrancescoPolizzi my guess is this one: mathoverflow.net/questions/379474/… (the other question asked by Buzz and still visible has no answer, currently) – David Roberts Jan 15 at 7:18
• @DavidRoberts: If it is this one, it has two upvotes, an answer with three upvotes and there are also a constructive comment. I honestly fail to see how this reaction can be considered "disheartening", unless the OP's definition of "disheartening" is "failing to provide the answer I was hoping for". – Francesco Polizzi Jan 15 at 7:21
• Also, Buzz didn't accept the answer or follow up to clarify why it didn't answer his question. I think as with most things, the more you put into MO, the more you'll get out of it. – Harry Gindi Jan 15 at 18:01
• @Buzz, Thank you for sharing your experience. I think it is valuable to enrich the current discussion with concrete examples, like yours. Obviously, your perception of the feedback from MO (say on the regularization question) would be different from that of an experience and somewhat successful user (like those commenting here, myself included). Perhaps it is worth nailing down that difference. Any open ended question may attract answers and comments that don't go in the direction intended by the OP. But the same feedback may be perceived differently. ... – Igor Khavkine Jan 16 at 23:02
• For instance, how did your initial perception (from that question) weigh these two extremes? (-) Your question is not "interesting," here's an answer to an "interesting" question you should have asked. (+) I don't know the answer to your question, but it's interesting enough for me give some feedback; here's my best guess at some information that might be relevant. What could have made your initial perception more positive? My own weighing, as an experienced MO user, would be 10%(-) 90%(+) or better. Others should feel free to share their perception, to get at least some anecdotal data. – Igor Khavkine Jan 16 at 23:10
• I would add (by repeated personal experience) that obtaining a good answer is sometimes the result of a dialectic process. You obtain a first answer, and maybe it is not what you were really looking for; then you make a comment, the answerer replies and there is the possibility that, after a few steps, you obtain what you need. In the case under consideration there is an answer with 5 upvotes, that the OP considers nevertheless "shallow and missing the point of the question" (without providing details). Perhaps making a comment under the answer could have led to a more satisfactory outcome. – Francesco Polizzi Jan 17 at 8:31