Not signed in (Sign In)

Vanilla 1.1.9 is a product of Lussumo. More Information: Documentation, Community Support.

    I think that we need to work out what to do with 'soft questions'. In my opinion, these shouldn't be allowed. From a personal point of view in terms of usability of the site, they clutter up my screen for no positive benefit, merely blocking other legimate questions from my view (and presumably consume our computing resources at the same time). I see mathoverflow as a place where we actually discuss and do mathematics, while if people want to post blog entries encouraging the audience to share their favourite maths jokes, then these sorts of things can go on someones personal blog.

    I both agree and disagree, I think. I think there are too many of them now, so I tend to agree. On the other hand, this is a natural place for people to ask such questions, since this is in other ways a natural magnet for mathematicians. Now there are some things you can do: You can go to your preferences and declare soft-question as a tag to ignore. You can even decide you don't want to see questions with ignored tags altogether. Problem solved?

    I think it's likely that “soft questions” will pop up no matter what we do. At least by having this tag, we are making it easier for people not to be bothered by them. But maybe we should have degrees of softness, and questions that are beyond soft, to the point of being downright squishy, ought to be closed. I don't know. I haven't been here long enough to have a strong opinion, but I do skip many soft questions myself.


    I was going to post something similar. By my counts we have about 20 about-math/general-math questions, including about teaching, wick rotation, plus a couple more very elementary questions — not far from half out of ~50 active questions on the home page.

    I don't have any specific suggestion, but one could think about:

    • if the soft questions are really interesting to research mathematicians (some of them probably are)
    • if there should be a different venue, essentially a discussion forum, for them
    • is there a drop in hard questions because of weekend?

    Again, I don't have any particular opinion at what's best at this point, but I want to point out that adhering to a strict policy might be not as easy as it sounds — the way SO/SE is designed was actually to include as many questions as possible of all types (Joel Spolsky said that Stack Overflow is an encyclopedia).


    I added [soft-question] to my ignored tags a few weeks ago, so they show up as greyed out for me, but even then, there have been too many of them for my tastes lately. I think MO is the right place for a lot of these questions, but I'd like to have a policy roughly along the lines of "there can be at most X 'live' soft questions at a time." It's too easy for people to get sucked into talking about philosophy (because it's easy) rather than get sucked into thinking about really interesting problems (which is the main purpose of MO). One thing I was hoping would happen with soft questions is that there wouldn't be very many distinct questions one could ask, so at some point, just about any new soft question gets closed as a duplicate. I'm not sure if this will happen, but it would be nice.

    I'll give this problem some thought. One idea that comes to mind right now is that when the density of soft questions is high, we could just close new soft questions immediately, reopening them only if there is enough demand (in the form of votes and comments).


    I think this is a tough nut. I agree with you that on some esthetic level, the number of soft questions has gotten a little out of hand, but the flip side is that I think my soft question was quite relevant and useful, and probably most of the other people who asked soft questions thought that too. I certainly would be willing to support some kind of policy for temporary closure of soft questions, but it would need to be thought out.

      CommentAuthorJon Awbrey
    • CommentTimeNov 14th 2009 edited

    I think of this as a problem in system architecture. MOFlow is using a classic, almost nostalgic blackboard architecture that traces its heritage back to the early days of AI and cognitive science, where it was taken as a model of attentional processes, and which it still models for people who pay a lot of attention to the main board.

    Different users apparently have different ideas about the purpose of the system and the current set-up of the system forces them to compete for a resource that appears to be bottlenecked by design.

    I think it might be a good idea to spend some of our spare time "thinking outside the bottleneck" and critically examining some of these built in design assumptions to see if they really achieve our espoused and long run goals.


    I think your soft question on two body job searches is appropriate for mathoverflow, since it is a question that is professionally relevant to mathematicians. As a question where you are asking for expert opinion, it belongs here (and more under career than soft-question). I see the problem with the majority of soft questions as they are explicitly violating "Math Overflow is not a discussion forum" (eg

    This brings up the problem with ignoring a tag, that there are useful questions that will be missed under this policy.

    • CommentAuthorSteven Sam
    • CommentTimeNov 15th 2009
    In response to useful soft-questions that may be missed to due to ignoring tags, maybe the right thing to do is to create a new tag like "not math related" and retag things like "Your favorite math joke" from soft-question to that. Then people wouldn't feel like they're missing anything by ignoring such a tag, and people who want to discuss off-topic kind of things can do so.
    • CommentAuthorESQG
    • CommentTimeNov 15th 2009
    Since I hypothesize that people who like the soft-questions are less likely to come to the meta page to discuss them, I would like to point out that a few of the soft-questions, including the one that pointed me here, are not only active but receive high voting scores. They must be of interest to mathematicians! It would be simple enough to split soft-questions between career-motivated and "not-math-related" tags as Steven Sam suggests. Does MO have enough server space to shunt the "not-math-related" questions to the side, so that they can be searched for optionally rather than always show up unless ignored?

    Looking through the list of soft-questions, on the other hand, I agree that many of the lower-voted ones are not ones I would really like to keep on the website. How about closing/deleting such a question if it has not received at least n votes within a week (where my estimate of n is 10)?

    Elizabeth S. Q. Goodman

    I'm in favour of having fewer soft questions. Here's my proposal:

    • if you have 3000 rep (not many people yet, but many more soon), vote to close any soft question that you wouldn't vote up.
    • if you have less rep, but don't like a particular soft question or are sick of their frequency, flag it for moderator attention. I'm very willing to close borderline questions if I see someone has taken the time to flag it.

    I suppose I should weigh-in on this as I asked three soft questions in rapid succession, at least one of which sparked another by someone else. In total, I've asked 11 questions, 5 of which are marked "soft-question" (one other is marked "idle-speculation"!). Three of my soft questions have gained at least 10 votes, in total they account for 45 votes and 15 favourites, compared to 22 and 5 for the rest of my questions. The corresponding number of "views" exceeds my ability for mental arithmetic this early in the morning.

    Let me try to explain why I asked them. There are generally a couple of reasons:

    1. The first two questions, which incidentally would get chopped by ESQG's limit, were because I am very interested in how best to get mathematics on to the web (more generally, in how to make the best use of technology for mathematicians). So I want to know about how mathematicians use the internet, and what works well for them. Whilst these aren't specifically mathematical questions, I hope to be able to use the answers to make mathematics easier to do over the internet. Being able to ask such questions of a broad range of mathematicians (who wouldn't think of going to a specific form/blog/wiki on the subject) is extremely useful and, due to its popularity, MO is a great place to ask them. I'm curious, actually, if those complaining about "soft questions" would include these two in their list.

    2. The other three questions were more "It's nearly the weekend" type questions. Of course, being nearly the weekend here in Norway doesn't mean that it's nearly the weekend in the US! Two of the questions, "Dimension Leaps" and "Theorems for nothing ..." were aimed at helping people get a feel for a subject. I'm one of those people that knows a lot about a little and nothing about a lot. So finding examples and curiosities in other areas of mathematics helps me get a handle on what's going on in them. I can't ask a more specific question because I don't have the knowledge to ask it! Plus I don't really want to know more than a vague "what's it all about".

    3. One aspect that I think hasn't been addressed is that of community. I think that MO needs a community feel to realise its full potential. Without the feeling that you are getting to know people, I think it will turn into a homework site. One of the huge benefits of the internet is that mathematicians in different parts of the world can work together and not be isolated in little departments with no-one else interested in their work. But mathematicians are people (more or less) and every now and then need to be reminded of that.

    4. To some extent, all of my questions have been "soft questions". Well, not quite all. Two questions are directly related to research. And guess what! One is bottom of my list on all metrics, the other is next to it in terms of views and answers. So for MO to be useful to me, at the moment it's got to allow some spectrum of "soft question".

    Having said that, of course I agree that there's a limit. Given my last point above, it would certainly be useful to me if those who are complaining about the number of soft questions could look at the questions that I've asked and tell me which ones they would rather I hadn't asked. Then next time I think of one, I'll know who I'm likely to annoy by asking it!

    • CommentAuthorJason Dyer
    • CommentTimeNov 18th 2009 edited
    I'm for letting the votes decide. This is supposed to be a community site, and if we're closing questions with 10+ votes and thousands of views just because someone doesn't like the overall topic there is a mismatch with what the community wants.

    The "ignore" tag is supposed to handle this situation. If greying isn't enough for you, the settings can be tweaked so that the ignored tags are invisible.

    (This brings up it would be a nice feature to have two levels of ignore -- a "skim" level where it's not in your field, but you aren't actively offended, and an "ignore" for the soft-questions tag and anything of its ilk.)

    I am sympethetic that the soft questions seem to be popping up too often, but I'm of the opinion there are too few "hard" questions rather than too many "soft" questions, so the solution would be to ask more "hard" questions.

    (ADD: Something else that indicates to me that following the voting is the best way to go is to note that nearly all the homework problems have been slammed in their reputation. So it's certainly possible for the community to deem a particular type of question unworthy.)

    Actually, Anton and I were discussing this Monday, and pretty much decided that "10+ votes and thousands of views" wasn't going to stop us closing stuff! I think it's still early enough that we're willing to use our dictatorial powers to try to steer the community. On the other hand I'm certainly open to reopening questions if there are complaints in the comments or over here on meta. The basic problem is just that soft-questions naturally get more positive votes, because a larger fraction of the users understand and care about any given non-mathematical question. As such, comparing the high positive votes on these questions with voting on technical questions is comparing apples to oranges.

    Perhaps the soft questions should be seen as a loss-leader. Yes, they not as desirable as hard questions. But they keep people coming back to the site.

    Speaking for myself: there are relatively few "hard" questions here which I have more than a vague understanding of, and even fewer that I can answer. Certainly, there aren't enough "hard" questions to keep me coming back every day.

    I'm sympathetic to closing the soft questions after a while, when it's clear nothing new can be added to them. But it's depressing to see them get cut off right away.
    • CommentAuthorJason Dyer
    • CommentTimeNov 18th 2009
    I wouldn't compare votes on community wiki questions versus normal technical questions, no, but I'm not sure how that's relevant to the fact that 10 or more people liked a question enough to click the "this question is useful and clear" button.

    All the comparison means to me is some excellent technical posts have far too low a rating.
      CommentAuthorJon Awbrey
    • CommentTimeNov 18th 2009

    After a period of due rumination, I have come to suspect that these so-called "soft" questions may be needed to preserve the cultural embeddings and practical integuments that the more acrobatic exercises of mathematical training need in order to stay fresh and vital.

    Of course, acrobatics is all a matter of balance …


    The solutions I envision are more about encouraging people to upvote and reply to "hard questions" than about closing questions (which will also waste our time discussing which questions should be closed and which not).

    For example, I went out and retagged several questions from October with no accepted answers. This brought them to the top of the list. If people do similar thing once an hour, the proportion of hard questions on the home page will be at its right level.

    1. I have absolutely no issues with my questions getting closed. It's your baby, you decide what happens. I'm a little annoyed at the "argumentative" tag for the blackboard question as I tried really hard to phrase it so that it wouldn't be argumentative and I felt that the answers had been in that spirit. However, I'll get over it!

    2. I would like some rules, or at least good guidelines. I asked for some guidance at the end of my previous post here but didn't get any feedback. That's not a complaint, but an observation. Being British, my next resort is to case law. At this point in the game then I'm still trying to get a feel for what a good question on MO is. My research-level questions don't get any responses, my questions inspired by teaching get some but these all come under the heading of "idle speculation", my other questions get the most and I'm (usually) very interested in the answers (and, being a professional mathematician, that at least means that they are "interesting to a mathematician"!). As I try out different types of question, naturally you guys are looking at them and deciding whether you want those types of question or not. That's fair enough.

    3. I second both Ilya's and Charles' points. I'm with Charles in that there aren't enough "hard" questions to keep me coming back to the site. Filling in gaps in other people's references is going to get boring after a while, and working out whether or not a particular operator is skew-adjoing (or whatever Theo's question was about) when I'm never going to see that operator again, is quite frankly going to lose it's shine fairly quickly. And to Ilya's point, this is one of the basics of parenting: encouraging good behaviour is far more productive than discouraging bad behaviour. I'm not saying it's an "either-or", of course.

    4. I'm coming to MO to meet other mathematicians, not to get particular questions answered. I'm trying to make contacts, learn techniques, pick up the odd tip here and there, but most of all to interact. I know that MO isn't meant to be a discussion forum, and that's fine by me, but given that mathematics is what I do, rather than a means to an end (as most programming is), then I want to be able to have fun with it as I do it, and to find out how other people do it to. As a simple example, I couldn't care less about the Fell-Whatsit problem; but weighing in has meant that I now know of a functional analysist who I might be able to plague with some questions. Some of my other questions mean that when Hanche gets back from sabbatical then we'll have a few conversations about functional analysis (I hope). I'm not really all that bothered about doing away with Riesz rep. Other questions have put me back in touch with some people I knew (at least, I think they have! I suppose there could be some duplicate names out there).

    Please don't misunderstand me. I am not complaining. As I said, it's your baby. I really like the general idea, which is why I'm hanging around to see if it develops into something truly useful. If I can help with that then I will, but as I'm not one of the "in-crowd" then inevitably there will be clashes between what you want and what I think you want! That's fine by me; I hope it's fine by you.


    Hi Andrew,

    I would certainly encourage you to keep asking questions exactly as you see fit, irrespective of local patterns in which ones get closed, at least until we have some sort of coherent policy! I like your questions, even if I'm the most enthusiastic closer of them. :-)

    For now at least I'm happy to close questions over-enthusiastically, in order to test the bounds of what the community actually wants. As we've recently seen with a question of yours that I closed and then reopened, sometimes I do actually learn that the community disagrees with me!

    More generally, I'm very glad to have you at MO, my inclination to close some of your questions notwithstanding. :-)



    responding more specifically to your irritation at the "argumentative" tag on your blackboards question: the things that worried me about that question were

    • I'm interested in opinions both as speakers and listeners. Your answers are allowed to be different from the different perspectives!
    • I'm possibly completely out of date on this - maybe everyone has now fully Embraced Their Inner Beamer.

    which seem to explicitly ask for opinionated answers. Anton has specifically argued against such questions, and the "argumentative" tag seemed most appropriate to indicate that this was the complaint. I think that any question which has to say: "In a - probably useless - attempt to forestall a flame war, let me try to narrow down the question a little." should just have been much narrower to begin with. I'm hoping that the consensus on questions at MO will end up disallowing even the possibility of a flame war -- those are for blogs.

    I also think that much more narrowly focused questions are better because they play to the software's advantages: specific answers can be sorted, but a big open-ended question like the blackboard one encourages big sprawling answers, which are harder to compare.


    @Scottv2: Can you link to where Anton has specifically argued against opinion-based questions? I'm guessing that there will be other useful bits of information about how he sees MO which would be useful to know (if it's in the FAQ then I apologise as I presumably read it at one stage, though I may not have read the most recent version). But I think that this is an impossible dream. Mathematics is what I do. It's not the means to an end, it is the end. So of course I have an opinion on it, the right way to do it, the right way to present it. I have an opinion on the right way to think about smooth spaces. I have an opinion on the right way to think about metric spaces (should I delete that question?). Tom Leinster's question on distributions could be viewed as asking for opinions - my opinion on that is that everyone else who answered has the wrong view of distributions, and given that my answer is top of the list on that one then I'm not alone in that view.

    I put the "flame war" (how many mathematicians know what one is?) and the "out of date" bits in just because I'm a human being (or at least a reasonable approximation of one) and if I didn't write things like that then I'd go mad. Sometimes it gets me into trouble (the n-crew are starting to get used to me), but I'm used to that and am always happy to remove the humour or apologise.

    "consensus"??? Betweeen mathematicians? Who are you kidding?

    I agree that there are is a certain type of question that plays to the software's strength, and I'm trying to figure out what that is. However, just because I have a question that doesn't quite fit doesn't mean that I'm not going to ask it! You've got a seriously large group of mathematicians here and that's what I'm going to make use of. My philosophy on software is that it is meant to help me do what I want to, not to force me to do what it wants me to do. I suspect that I'm not alone. If there is a difference between me and others, then it's that I will take the time to find out what the software was meant to do first, and then I'll adapt it to do what I want (which is why the n-forum is so much better than this place!). I suspect that others will just throw the software in /dev/null when they find that it doesn't do what they want.

    In reply to your first of the two posts, that's absolutely fine. Now that we've got that clear then I think that we can work together on this. I'm not saying that I should put up questions specifically for you to close, of course. One thing you should do is make it clear to everyone else that you are still in the "working out what fits" stage. If you look at the comments to the "Walking in the rain" question, you'll see that not explaining this can annoy people.

    (Incidentally, I don't know if you can see this as an admin but that "Walking in the rain" question is the only thing that I've written on MO so far that has received any down votes! (In case you can't see that, if I'm interpreting things correctly then the score is 12-8 in its favour. Of course, I agree that you shouldn't be ruled by votes.)

    Finally, "those are for blogs". Sorry, I completely disagree. I think blogs are almost a complete waste of time. If you'd said "forums" then I'd've agreed, but complained that there isn't one where I could ask that question and get answers from mathematicians.

    Okay, really finally this time (gosh, I do go on a bit, don't I!). Here's a bit of free advice. There are some people who are going to use MO almost no matter what. You probably have a better idea of who these people are than I do. These are absolutely the wrong people to ask for opinions on what makes it work. It's the next level of people that you need to look at: the ones that thing "This looks interesting, I'll give it a go" but aren't really engaged with the whole idea yet. On the basis of Charles Rezk's comment, I'd put him in that category. These are the ones that you need to figure out how to keep. Sorry if that sounds patronising, it isn't meant to be.


    Just one minor thing: when you do reopen a question, it looks a little odd at the moment as there are usually a load of comments saying "Why has this been closed?" when it isn't closed! And the record of why you closed it seems to disappear. Either leave all the comments (including your reason for closing), and add one saying that it's been reopened, or delete all the comments (though I can see problems with deciding which to delete so I'd go for the former).


    I'm staying out of the main discussion but DON'T DELETE those comments. It is a rule of the internet that people get very annoyed when their comments are deleted, and it should only happen with good reason. If you care, you could add a comment saying "reopened". I think it is fine the way things are.


    @David: my point was that that isn't the way things are. When Scott's reopened a question, he's removed at least his comments on why it was closed. This can make some of the other comments seem odd. It also just occurred to me that as we are all learning what we want from this site, knowing that a question was borderline would be useful information so if a question was closed and reopened it's good to know that.

      CommentAuthorJon Awbrey
    • CommentTimeNov 19th 2009 edited

    I think it's important to recognize that the system property we're discussing here is Attention Management. It appears that many attendees walk in the door thinking they've come to a "Birds of a Feather" session, only to discover that the confab in progress is far more omnibus than they expected.

    There are many more options for handling that than simply hiring bigger bouncers.


    @Andrew: I'm kind of surprised at you. You've always struck me as a "right tool for the job" guy, but now I feel like you're saying "never mind the intention or functionality of the site, there's no way I'm going to pass up this large and audience."

    MO should primarily be a site for asking and answering mathematics questions. Socializing and having a good time are (intended) side effects. Opinion questions are fine so longs as the primary goal is to develop as a professional mathematician, rather than to just have a nice chat. The objection "there aren't any other sites that draw the right audience for my question" is like complaining in response to getting a speeding ticket, "but there aren't any nearby roads with a higher speed limit." If there aren't enough hard questions to keep you coming back every day, don't visit every day; visit every few days. I have no interest in making MO as big or as popular as possible (though I admit it's easy to get caught up in those metrics). It was big enough to be actually useful to me a month ago. I'd rather lose traffic by banishing unprofessional questions than have people say a year down the line, "MO was pretty good at first, but then it degenerated."

    Of course, I want to hear ideas about how MO should evolve. But keep in mind that the goal is to do this one thing really well, and if it happens to fill some other niche in the process, that's great. If not, I refuse to stretch the software or the principles of the site to fill that other niche just because it needs filling. Let's try to solve one problem at a time.

      CommentAuthorJon Awbrey
    • CommentTimeNov 19th 2009

    @Anton : I see everyone here as dedicated to "the primary goal … to develop as a professional mathematician", but there are many different ways that the present or future "groupware" — collaborative software environment — can facilitate that aim. If people with a modicum of experience in mathematics do not participate in the development of that groupware, then it's almost inevitable that tools more Procrustean than Promethean will limit their ability to do the right job.


    @Andrew and David above, re: deleting comments after reopening.

    Mea culpa, I realise this was a bad decision, motivated by a misplaced desire for cleanliness. I won't delete any comments in future without permission (excepting of course anything offensive, wildly off-topic, etc. etc.)


    @Andrew: re "@Scottv2: Can you link to where Anton has specifically argued against opinion-based questions? "

    Sorry, Anton's office is just around the corner from mine, and a lot of offline discussion happens...

    • CommentAuthorgilkalai
    • CommentTimeNov 20th 2009
    Perhaps the most desirable questions are described in the faq

    "Math Overflow visitors should know how to learn new things and do mathematics on their own, but we all get stuck sometimes, and this is where MO saves the day. When you're stuck, you can come to Math Overflow and say "I'm trying to do X. How can I do that? Does this work? Does anybody have a reference?" The idea being that for an expert, it should take very little effort to understand your confusion and set you on the right path. Or maybe a non-expert has come across the same sticking point and can explain how she resolved it."

    I suppose that only a small percent of the questions (maybe 5%) really represent a place where a mathematician is stucked and needs help; maybe we should have some tag for "guys, I really need the answer". It is especially unreasonable that people have dozens of problems which represent places where their research projects are stucked.

    All the rest of the questions, soft or not, are of a different nature. Pleople are trying to ask good questions in order to widen their knowledge and knowledge of others. I see no difference in the soft questions; some of them are better than others but overall they are quite interesting. Especially at this early stage where the topics of averall questions are rather concentrated, the soft questions can appeal to wider audiences.

    (Regardind discussion: It is correct that the software is not a convenient platform for discussion. This already gives negative incentives for discussions. It is a logical flaw to conclude from this that additional measures should be taken to supress discussion.)

    @Scott: I suspected as much. Just remember that the rest of us aren't privy to those discussions. If you're going to change the rules, can you announce them here beforehand (and preferably with a little lead time before you actually do so)? I'm not complaining about you changing the rules - I'd be more worried if you didn't.

    @Anton: I guess I'm a little more complicated than that. I do try really hard to use the right tool for the right job, but when there isn't a "right tool" available then I'll take the nearest one and try to make it the right one. I have an area on my website called "How Did I Do That?" and the tagline is:

    Collected here are a load of hints, hacks, and howtos that I have used at some point to get my computers to behave the way I want them to (as opposed to me behaving the way they want me to).

    So if there is a "right tool" then yes, I'll use it. But if there isn't then I'm not going to let that limit what I do. I'll go out and write the darn program myself. If you look at the "How Did I Do That" then you'll get a sense of just how often I've had to do that.

    However, in this case I think that you are missing my mark. I'm not trying to subvert the software. I'm trying to subvert you! I'm trying, albeit indirectly, to persuade you (and Scott and the other admins) of something. You (plural) are the "not quite right tool" that I'm trying to mould into the right one!

    More precisely, I think that MO is a great idea. I think you all are doing a great job. But I think you're missing something that will make it fantastic.

    I think that there's a heck of a lot that mathematicians can learn from software programmers, in particular making the most of the available technology. I even went so far as to setting up a blog/forum/wiki to discuss and develop that idea. But to make the most of that idea then it's important to get the analogy right. It is not "mathematicians = software programmers" but "mathematicians = library writers". We're the ones that write the libraries so that the software guys don't have to worry about implementing an algorithm to parse HTML, they just "use HTML::Parse;" (or whatever it is). We're the ones that write the regexp libraries so that when you do s/(^|[^A-Za-z])\*\*($|[^\*])/$1<strong>$2/g then it just works.

    Stackoverflow is, I suspect, primarily populated by program writers not library writers. So I caution against sticking too closely to their model because that isn't us! Perl Monks is probably a lot closer, and if you look at their site then you'll see places dedicated to perl obfuscation and perl poetry. I'm not advocating a code shift, absolutely not! But I am saying that you should look at other sites that do a similar thing and take what you like from them and put it in your model.

    Actually, Jeff Attwood makes this point in an article on his blog. When he (rather smugly!) comments on the fact that there's nothing open-source that's quite like stackoverflow, he actually makes my point: it's not the software that made SO what it is, it's the community that uses it. Software should follow, not lead. I suspect that all of what I'm advocating (not totally sure what this is!) could be done within the SE framework, but if not, so what? As Jeff said, it'd be the work of a weekend to write the software (indeed, I'm considering doing it because I think that this would work great for a class but there's no way I'm paying for something that less than 100 people will use for less than half a year). It's the community that makes it work.

    Which brings me back full circle. Concentrate on the task, and you'll find the right tool. Concentrate on the tool, and you'll limit what you can do.

    Incidentally, as you don't know me, you won't know that I tend only to write such reams of waffle when I actually care about a project (or whatever) succeeding, and when I feel that there's someone listening (not necessarily agreeing!). Otherwise, I just walk off and say nothing. So take this as a compliment!


    Incidentally, some top guys on Stackoverflow are core developers of C# and Python (both are quite advanced languages used by developers in mission-critical projects).

    Those people are really similar in their position to research mathematicians — they sit down and think, write articles and give talks, and then other people use their work. Another example of concentration of brainpower on development happens on python-dev list.

    The SO founder, by the way, worked at MS too and he seems to be the one behind Excel macroses :)

    Speaking about HTML parsing, there's a hilarious answer with 1856 upvotes (at this moment).


    @gilkalai: that part of the FAQ is not meant to refer exclusively to research. Sometimes I think about math that's not my research and I get stuck, and sometimes I think about how to be a good mathematician and I get stuck on something. So long as I can state what it is that I'm stuck on, and I have reason to believe that other mathematicians can offer a solution (rather than an involved discussion), it's a fine question for MO. I'll think about how to change the wording to reflect this better.

    (Regardind discussion: It is correct that the software is not a convenient platform for discussion. This already gives negative incentives for discussions. It is a logical flaw to conclude from this that additional measures should be taken to supress discussion.)

    I disagree. When somebody finds a neat new tool, they're likely to try to use it for everything, even if it's terrible for some purposes. They might not even realize there's a better tool for the job unless somebody tells them so. If we allow discussions on MO, I think it will just set the precedent that it's okay to try to have discussions. Maybe you're right that eventually everybody would figure out that it's a bad idea, but until they do, I'll find visiting the site painful, and I imagine others would too.

    @Andrew: I think the two of us agree on the nature of software, but disagree on the best way to improve the lives of mathematicians. If the goal is to bring mathematicians together and have them learn from each other, it is best done with "hard" problems. I don't think, no matter what software it ran on, would take off, and if we converted MO into mathchat, I think people would stop visiting. I know that math problems bring mathematicians together in a very powerful way. I think allowing "discussiony" and "squishy" questions on MO will decrease the strength and utility of that connection, not increase it.

    I also know that we all like to have discussions and talk about squishy stuff sometimes. I realize it sounds kind of paradoxical to say that adding those things to MO would make it less useful, but I think it's true. A lot of us also like to teach calculus to undergraduates sometimes, but it's clear that catering to that desire would be a disastrous move.

    • CommentAuthorgilkalai
    • CommentTimeNov 20th 2009
    Dear Anton, I think some amount of discussion is a good idea on MO and there is some type of welcome discussion that does take place. I think the endeavor is exciting and successful (and you have a large share in its sucess), and we can let it evolve with maximum tolerance to what the participants choose to do and the various ways they want to widen their knowledge and understanding of mathematics. Some questions are more technical than others, some question are softer than others and for some questions some form of discussion is anavoidable, and it is all part of what mathematical understanding is. Of course, we also have different tastes and visions but it will not be wise if the tastes of some will be strongly enforced. --G.

    @gilkalai: I think we mostly agree. Some discussion will happen on any given question, and that's good. But I think it's important to discourage questions whose intention is "pure discussion" at MO, or to encourage them to happen elsewhere. I hope you agree that the moderation to date has been fairly permissive. I try to only actively moderate against behaviors which I think will be harmful to the rest of the endeavor.


    @Ilya: I know. That's why I picked that example. It's well worth reading.

    @Anton: I completely agree about ''. I have no interest in such a place. I want to do mathematics, I want to learn mathematics, and I want to present my mathematics, and I want to know how to do that better. Sometimes, though, I also need to remember that it's possible to have fun with mathematics. I'd like to be able to do all of these in the same place.

    I don't want MO to turn into a "soft question" discussion forum. But I think Charles got it right: it's a loss-leader. It's what keeps me coming back to the site each day. Otherwise I would get overwhelmed by all the algebraic geometry that goes on. It's what means that I'm there to spot the questions that I can help out on.

    My blackboard and beamer questions weren't intended to spark a discussion. I wanted to know what others thought, that's true, but also I wanted others to see what people who have actually thought about these issues think. I want everyone to read Tim Gowers' answer and come to the same conclusion that Mark Meekes noticed: it's worth actually thinking about how you give a seminar. Once someone does that, I don't care how they do it. They can carve it on a hillside for all I care: their talks will be that much better because they've thought about it and that will, I hope, make it less painful to listen to them! (See, ultimately I'm selfish!)

    My worry about MO is that there aren't enough "hard" questions out there. And that a "hard" question of the sort you are thinking of isn't something that I just want someone to tell me the answer to! I want to solve it for myself, not be told the answer, because I'm far more interested in learning how to solve it than what the actual solution is. That's the real difference between programmers and library writers. Programmers just want to get the job done, library writers want to do the right job.

    I don't see any real difference in what we want. I don't want "soft questions" cluttering up the main area of MO. I'd just like it if there was a side room where I could go and relax a bit every now and then. It'd be nice not to be asked to leave just because I took off my tie!


    Okay, okay, it looks like I was getting too hypothetical. I think the beamer and chalkboard questions fit well within my standards of the purpose being to develop as a mathematician. I have no problems with you taking off your tie; I even take mine off sometimes. Just keep your shirt on, and don't try to argue that you need to take off your pants because you "need to use them to demonstrate what a Frobenius algebra is."

    Even though I think soft questions are acceptable, I do want it to be clear that if the density of soft questions gets too high, we're going to start closing them to prevent MO from getting mushy. I think even good soft questions will drive mathematicians away if there are too many of them.


    Anton, I agree completely.

    The number of places in the UK that insist on ties is decreasing rapidly, but I think that almost all still insist on pants being worn.


    In Norway, back in the seventies where many restaurants still required ties, a well known comedian showed up at a restaurant dressed in nothing but a (huge) tie. He had made sure the press was present, of course. Unfortunately, I don't remember what happened next. But it probably contributed to the tie requirement being phased out sooner.

    • CommentAuthorgilkalai
    • CommentTimeNov 22nd 2009
    Perhaps, the best way to influence people is to ask them. For example, when people click on the box "ask a question" there could be some requests like:

    We have at present too many soft questions and other very general questions. Consider avoiding or delaying such a question and also even in these categories try to restrict yourself to questions with strong mathematical content.

    At present, the algebraic geometry tag has more than double the number of questions as any other tags. If you plan asking a question in algebraic geometry make sure it was not asked before. Consider postponding questions in directions which are overrepresented in math overflow.

    (@Anton, of course, I agree...)
    • CommentAuthorHarry Gindi
    • CommentTimeNov 23rd 2009 edited
    I think soft questions should always be community wiki and displayed somewhere other than the main question page.

    And I've spoken to a few people in person who were turned away from MO because of the number of soft-questions.
    • CommentAuthorgilkalai
    • CommentTimeNov 23rd 2009
    (Just to make my position clear, I do not support any policy meant to "balance" the mathoverflow topics, and neither do I support a policy, at this stage, to limit soft questions. However, whatever policy you would like to have the best strategy, in my opinion, is to ask people to follow it and to be tolerant to those who don't.)

    @Gil: I second your parenthetical clarification.

    I like @fpqc's idea of having different "streams". It's a bit like the Perl Monks' "Perl Poetry" and "Obfuscation" areas. I wouldn't want MO to be too segregated though. Can tags be used to display "soft-question" and "all-but-soft-question"? In particular, is there a negation?

    I guess that if I choose to ignore "soft-question" then in the normal run of things then I wouldn't see those posts (is that right? Or are they still there but just in a different colour?) but if I then clicked on "soft-question" then I would get them, so that would be functionally the same as the separation proposal.

    The point is that new visitors shouldn't be greeted by, "Good math jokes", or "Who's the best mathematician in history of all time". This type of stuff tends to scare away the people who we want coming here. Well, if not scare away, at least lead them to believe that the level of discourse here is beneath them. As I noted above, I have spoken to several people in person who were turned off exactly by those types of nonsense topics.

    @Ilya: excellent. That should be publicised. I wonder if there's a way to exclude certain tags from the main list so that people aren't put off by "soft questions".

    However, I find that I am still not sure what a "soft question" is! And nor are others, to cite David Corfield's recent question as an example. As I'm a repeat offender on this matter, I genuinely would like to know which of my questions are objectionable (by which I allow that it's not the question itself that is objectionable but the quantity of questions that are like it). I shan't take it personally, I promise! Otherwise, I'll have to think up lots more borderline questions to test the matter.


    @Andrew, the author of the tag meant it to apply to all questions that are not "hard", which is, roughly, "the ones without formuli", or, more precisely, "the ones without technical details".

    I'd like to leave completely aside the hard problem of what type of questions are appropriate on Math Overflow, and point out ((that this thread is mostly discussing the deficiencies of home page feed, a technical, rather then a policy, question)) ---- corrected below ----


    It didn't start that way, though. Perhaps it's morphed into that - it that everyone else's perception? That the issue is more about keeping these off the front page than about whether or not they should be there?


    My bad, a more correct statement would be it turns out that this thread is mostly discussing the policy question even though most (though not all) arguments voiced against soft questions are that they clutter the home page; the last problem is largely technical.

    Also, when I say technical question, it still doesn't mean we'll be able to solve it — the way home page works probably can only be changed by the Stack Exchange team.


    To harp back to an earlier comment by Ilya in response to my "mathematicians are not programmers" comment, I just looked at Jon Skeet's profile on stackoverflow. That makes my point rather well. If you look at the ratio of the number of his questions to the number of his answers, and also look carefully at the type of question he asks, you'll see that he doesn't really use stackoverflow to help him with his day job. I'd guess that for him, stackoverflow is a bit like how we might regard Dr Math (or whatever that homework forum is): somewhere we go to feel like we're "doing our bit" for the community but not our main place for asking questions that we want the answers to in our research. (That's not to say that helping out at Dr Math can't be useful, but it's not the main reason for going there.)