Not signed in (Sign In)

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

    As MO grows in popularity and size, people have been asking more and more technical questions. I've seen various meta threads discussing the danger of grad students using MO as a way to get others to do their thesis work for free. I'm wondering (as a grad student) how to prevent the opposite thing. I would hate to see my thesis question get posted on MO and have some tenured professor come by and solve it, or even give a clever idea and an outline for how the proof would go. This is in my mind because I've had a few scares recently where I thought a question was very related to one of my 2 theses (math/cs) till I started digging in. With just a little bit of bad luck those questions could have turned out to be my thesis and in that case I wouldn't really know what to do. If the thesis was almost done, then maybe I'd just post an answer. On the other hand, I might like more time to figure out the generalizations and corollaries myself, before I make my work public.

    I suppose the first thing would be to email the OP and reveal that this is related to my thesis. If he/she had also put a lot of time into the problem then perhaps we could collaborate a co-author a paper. If not, then perhaps I'd be able to convince him/her to leave it alone. But those reading the problem would not be able to follow this exchange and might come along and post an answer. Even if I convinced the OP to stop pursuing the question, I believe that if it has enough upvotes he/she would not be able to delete it (not 100% sure, though).

    To deal with this, I suppose I could leave a comment saying this is highly related to my thesis and please don't post an answer, but it's not clear to me that people would respect that or even read the comment. Do others have any ideas for how to deal with a situation like this? Perhaps the moderators would be willing to step in and close/delete the question if the OP agreed? Or try to come to some sort of courtesy agreement about not answering a question if some poor grad student leaves a comment begging people to turn a blind eye to the question?
    • CommentAuthorWill Jagy
    • CommentTimeDec 1st 2011
    Good question. I would contact the moderators immediately. You are right that a comment at a post is not automatically read. Meanwhile, documenting the fact that you are partway along a problem may turn out to be something you do not want to do in public...hmmm. I suppose what has changed is the speed, and public nature, with which problems and answers propagate now. I had a scare when my adviser asked an expert if my thesis problem was reasonable. The fellow said he thought so, went home, and solved it in one afternoon. But then he put his notes in a desk drawer and forgot about the whole thing, until I requested his notes many years later.
    • CommentAuthorgrp
    • CommentTimeDec 1st 2011
    I see your issue as a couple of issues: what your advisor thinks you need to do to have an acceptable thesis, and what you think you need to do to have an acceptable thesis. It may turn out that finding the clever idea and then busting your brain to make that idea work is exactly what your advisor expects you to do, even if the clever idea did not come to you originally.

    Think carefully and write down what you need to about the latter issue. Then try to imagine what your advisor would say about the first issue. Then work through a couple of likely scenarios (not ones you are afraid might happen) of some outside factor contributing to your thesis. Finally, TALK WITH YOUR ADVISOR about these issues and what they would recommend. If you are doing good work and producing something, you may get a thesis out of it, but only if you keep your advisor well within the loop. It may even involve less work than you might think.

    If you like this advice, I have further suggestions for how you can make the issue in your post less of a concern, e.g. asking how much collaborative material you can include in your work.

    Gerhard "Really, Talk To Your Advisor" Paseman, 2011.12.01
    • CommentAuthorDavid White
    • CommentTimeDec 1st 2011 edited
    Thanks for the replies so far. Just in case there is confusion, let me clarify that I am not currently in this position. I'm not seeking advice, more trying to brainstorm ideas for how the community should respond if this situation were to arise. On the other hand, any advice posted here will probably be of help to whoever the unlucky grad student is if this scenario comes to pass (which I think is likely, at some point). I asked both my advisors about this and both confirmed it was a really hard question and they don't know what they would do in that situation. I'll bet a lot of advisors feel this way, since MO is so new. I'd like for us to be prepared in case this situation occurs, so some poor grad student doesn't end up losing months or years of work so that an expert can get an extra 50 reputation points.

    Perhaps what this will come down to is trying to limit how technical we allow questions to get. I'm partial to the view that questions should not be the sort of thing theses (or even papers) are made out of, but right now plenty of very technical and involved questions end up on MO.

    I'd strongly encourage contacting the moderators quickly ( Personally, if I received a coherent explanation from a student that a question looked likely to cause significant difficulties to their thesis, I'd likely be happy to jump in and temporarily delete the question, and put the OP in touch with the student to see if something could be worked out. Maybe after the first time we'd decide that this was a bad way to handle the situation, but for now at least it seems a plausible enough solution.


    These things happen. When I was a grad student, one of the many projects I was working on was a version of something called "The Generalized Smale Conjecture". I had made some progress when some guy named Hyam Rubinstein announced he had solved the problem. So I met with him and saw that yes he was much further along on it than me, so I moved on to another problem.

    Somebody might solve your dissertation problem before you do. It's unfortunate if it happens, but it does happen sometimes. Ideally you'd have enough things going on for your dissertation that having one key problem knocked-out of the mix shouldn't matter. That's something you should talk with your supervisor about.


    Let me interject another point of view: If an expert solves your thesis problem in an afternoon, it is not a good problem for you. You should try to do something that no one else can do. If you cannot, think about changing directions. Keep in mind that it is much easier to make a living using mathematics than creating mathematics.

    So if your thesis problem is posed on MO and solved quickly by an expert who saw it, be happy that you did not waste time working on the problem.

    The "guy named Hyam Rubinstein," is a past president of the Australian Mathematical Society, recipient of the George Szekeres medal, etc., etc.

    This won't help anyone, but please indulge me while I tell a little story. I was nearing completion of my thesis when I saw a new book in the library with the same title as my thesis! For a few moments I feared all was lost, but fortunately when I calmed down enough to thumb through the book I saw there was very little relation between the two documents.
    • CommentAuthorRyan Budney
    • CommentTimeDec 1st 2011 edited

    :) Yeah, that's the Rubinstein I'm talking about. I later found out he's a lovely fellow, when I got over him crushing my then grad-student's heart (edit: that's me being melodramatic for fun). My point being, these things happen and they're natural. Rather than trying to protect yourself from them IMO it's best to just realize this is the world you operate in. Adjust to it now rather than have unreasonable expectations of owning part of the mathematical landscape.

    • CommentAuthorAngelo
    • CommentTimeDec 1st 2011
    My (rather limited) experience, and that of my students, is that when you ask a question that you have really thought about, in a subject on which that you, or your advisor, are an expert, no one will answer.

    There are a number of possible scenarios here (I am simplifying rather brutally).

    1) Somebody asks a question that is related to what you are thinking, but is not the central part. That's fine, you are free to use the answer, with proper attribution.

    2) Somebody asks a question that coincides with your thesis problem. This branches into two.

    a) Nobody answers. That's good; but on the other hand you know that someone else is working on your problem. You may want to contact the questioner to inform him or her that you are also working it. Or, depending on your character, you just work your ass off to beat the other person.

    b) There is no definite answer, but somebody suggests a plausible strategy to solve the problem. This is a combination of 1 and 2a; the alternatives above also apply to this case.

    3) An expert solves the problem. Then, as Bill says, it was not a very good thesis problem.

    In all these cases, seeing the post was a good thing. In case 1 it helped you. In case 2, it is better to discover that you have competition at an early stage, rather than when a paper gets posted. In case 3 you make an unpleasant discovery, but, again, it is better to know now before discovering that you have a thesis, but experts consider your result sort of trivial.

    I disagree with what I perceive to be the tone here which is that it's not really a problem, and I think that anecdotes are misleading: almost by definition, those who share anecdotes about this will be those for whom it (ultimately) made no difference. I think that Scott's response is, really, the only sensible course of action until we get more data.

    This is another argument for keeping tight rein on speculative questions: if it is perceived that MO is a place to ask broad-ranging questions then it is more likely that questions pertaining to an entire thesis will come up. But if MO is perceived as a place for asking small, focussed questions, then it is much less likely that it will occur.

    If I may say so, then I think that Bill's comment (endorsed by Angelo) is particularly simplistic. It's not just about who solves the problem, it's also about how much work it takes to get make that solution work. One could argue that very little in mathematics is truly new and that most of it is simply gathering together pieces already known and adding a little glue. There, the genius is in knowing which pieces to gather and that could take an expert an afternoon but be an entire thesis for a graduate.

    The true solution would be some way in which we, the mathematics community, adjusted our evaluation of PhD theses to take into account the existence of the internet.


    I don't think Bill's comment is simplistic at all. Moreover, I think Scott's proposal is really only one that should be acted on if there's reason to believe people are maliciously posting question's based on insider knowledge of another's current work (regardless of them being grad students). Say I had secret knowledge of one of Andrew's current projects and decided to sabotage it by asking questions on MO that would destroy any hope he had of claiming originality in his solutions. That's something I think we should not encourage. But if two people are working on the same problem (without each other's knowledge) and one posts a related question on MO when the other would not, IMO it's up for those two people to sort it out between themselves rather than getting a moderator involved.

    • CommentAuthorAngelo
    • CommentTimeDec 2nd 2011 edited
    To Andrew: I am not saying that it is not a problem; of course, having someone else working on your thesis topic is a major problem. What I am saying is that knowing this is better than not knowing it.

    Ryan, my gut feeling on this (and it is no more than that) is that there is a big difference between, say, you and I working on similar problems, than if one of us were still a PhD student. We attach such importance to the PhD thesis because it is ones "entry card" to the "Mathematicians Club" and this means that in the situation that you describe (in your last sentence) then the fact that one of them wants to get a thesis out of the problem does significantly change the situation.

    Basically, for a non-thesis problem then there's very little disadvantage in collaborating whereas for a thesis problem, part of the goal is for the student to show that they can do things by themselves so there there is a cost.


    I can see where you're coming from Andrew but still, when you're coming out of your dissertation it's a very bad idea to have all your publication hopes pegged to one idea/construction that could readily be done by someone on MO. If you're in that situation, it's unlikely you'll get anything published let alone get a first position out of your Ph.D. I would hope that anyone coming out of a Ph.D has several things going, or perhaps one very large thing going that people in the field would view as quite novel -- usually the latter is the former when one of the several things the person has going starts to "snowball" and fit-together nicely.

    There are still disadvantages in collaboration outside of your Ph.D. If you have no publications by yourself it could look like you have no ideas of your own and (say) the more senior person on your papers may appear as the "ideas" person and you're their technical typist. This kind of thing can come up when applying for grants, at tenure and promotion cases and such, even getting your first job. If someone applies for a job and all their publications are with their supervisor, that raises a red flag for me.

    • CommentAuthorStorkle
    • CommentTimeDec 2nd 2011

    "part of the goal is for the student to show that they can do things by themselves"

    Really? I thought the default assumption---in the vast majority of cases---was that the advisor gave a problem she already knew how to solve, and that the goal for the student is to show that they know how to write up the advisor's ideas.

    • CommentAuthorRyan Budney
    • CommentTimeDec 2nd 2011 edited

    Storkle, that's a bit on the skeptical side. There are many very "hands off" advisors that basically say to you when you arrive: "look, you're in a place where there's a lot of nice things going on, you get exposure to a lot of ideas. Pick one up and run with it, and get back to me when you're done. I'm available to chat about technical problems if you need it. " I think the number of advising styles is pretty open. There are also advisors that you can't tell too many of your ideas to because they'll quickly solve all the problems you found interesting - advisors that are competitive by nature and can't turn it off.


    Ryan: remember that not every Ph.D.-granting institution is Group I, or even Group II (or the non-US equivalent). The situation you describe is obviously desirable, but most students are not in this position when they graduate. Furthermore, many students will not stay in academia; for them the existence of the Ph.D. thesis can be much more important than the quality.


    Hi Tom, sure, understood. I suspect a safe strategy for advising someone at a non "Group I" or "Group II" institution would be to give them something that's a little more out of the way of current trends. That way if they prove something nice they'll still get noticed but if not, at least they've safely done enough to get their Ph.D without worrying too much about being scooped.


    > I thought the default assumption---in the vast majority of cases---was that the advisor gave a problem she already knew how to solve, and that the goal for the student is to show that they know how to write up the advisor's ideas. >

    I consider that behavior academic dishonesty on the part of the advisor. A dissertation is presented to the world as original research by the author. Assistance from the advisor is assumed, of course, but the thesis should not be research of the advisor that is written up by the student.


    I think this question is one we should discuss once it actually poses itself, rather than "brainstorming" about it now. It depends VERY MUCH on parameters that are not determined in this thread: whether the question posted on MO was an "innocent" one or a deliberate attempt at sabotaging the thesis; what kind of answers it already has received when somebody claims it is his thesis question (note: deleting substantial answers, even temporarily, has a high chance of pissing off answerers; we had this on the Art of Problem Solving forums); what work the grad student has already done on it (I am strictly opposed to "retaining seats" for a student who hasn't even starting work) and how credible his claims are; how much bad blood is already spilt before the moderators take notice of the problem; etc. These are lots of cases to consider if we want to do it in advance...


    In my opinion, if your advisor is doing his or her job, then this is unlikely to be a problem (but not impossible). Your advisor should be letting experts in your area know what problem you are working on and ask them not to work on it. It is possible that someone from another area becomes interested in the problem for some reason and works on it with different ideas, but this is unlikely, and even more unlikely if your advisor has properly warned you away from problems for which this is likely to happen.

    It is general common courtesy in the mathematical community not to work on a problem that a graduate student is working on. I would say this applies to MO as well.

    My suggestion for any graduate student who finds him or herself in this position is to get in touch with your advisor immediately and ask him or her to comment on MO, using his or her real name, that a student is working on the problem, and have him or her deal with the issues. This is because your advisor knows more about etiquette in the mathematical community than you do. I don't see how etiquette with MO is any different from etiquette before MO here.

    Because MO happens quickly, if your advisor is on vacation, it is acceptable to contact some other senior person you know (either another professor in your department, or, better yet, another person in your area whom you know) to do this commenting.

    Don't worry that 24 hours will make a difference. Unless your advisor has been remiss and let you work on a trivial problem, it will take more than a day to generate an answer that will have a real impact on your thesis.

    • CommentAuthorquid
    • CommentTimeDec 2nd 2011

    IMO if the student already has some (significant) progress, answering seems most natural to me. For example briefly saying 'I can prove this and that. I believe more could be true and am working on this extension and a write-up as my thesis.'

    With this there is even a (weak) documentation of priority and people are informed about the situation. Of course, one gave some information away. But I do not know how much this would hurt. The scenario that then others will start massively working on this with the information in hand seems rather unlikely to me.

    If there is anyway no progress I do not see the huge problem (it is perhaps some problem but not huge). Presumably (it is to be hoped) the expertise to be/already obtained to work on the one problem can be also be use on some other problem (if ever an answer shold appear) close by.

    However, I agree with Andrew that a preemptive measure to minimize the risk of such things is to keep MO close to its orginal purpose, and not turn it (also) into some giant online collaborative research activity, in contrast to what from other discussions it seems some people would consider desirable. From my perspective, one more reason not to go down this road.

    As implicitly alluded to by Darij, one annoying thing are public half-done (or quater or still less, or even only announced) projects, in particular when they turn essentially idle. (This applies to everybody not just students; indeed perhaps students, and people in closely related situations, should be the only ones with some 'right' to do this.)

    • CommentAuthorEmerton
    • CommentTimeDec 2nd 2011 edited

    This scenario has already occured:

    In that case the advisor responded in a sensible way.


    Yes, and this is one of the cases when no solution was posted before the response by the advisor, so a rather simple and benign case.

    quid: What do you mean by "[turning MO into] some giant online collaborative research activity"? In how far do you see a tendency towards that?

    • CommentAuthorquid
    • CommentTimeDec 2nd 2011

    @Darij:for example some days ago this was posted in a thread on extending the bounty period

    I think the idea of featured questions is great. Providing incentives to answer harder questions certainly improves the quality of the contributions, and may even lead to collaborations between users. [..] As a matter of fact, MO users do not invest in research. A bit embarrassing, isn't it? [...] But a week is too a short period to think to a hard problem

    Or, occassionaly people post problems they know very well are open and seemingly without direct motivation just to give them visibiliy I assume. (And here I do not mean some 'strangers' asking about the Goldach conjecture or there latest personal variantion thereof or something like that; but people knowing very well what they are doing.) If this were 'allowed' (as opposed to being essential against statements in the FAQs) I am pretty sure this would be a lot more frequent and some people would appreciate this.


    Oh, I see what you are referring to. It does not seem to me that many people would wildly give out bounties even if they become more flexible as suggested in that thread. In so far I don't expect much damage from that direction.

    • CommentAuthorquid
    • CommentTimeDec 2nd 2011

    Yes I agree that bounties would/should not be very effective (as documented in the other thread). The main point to me here is however that some seem to think to (really) work [everything above couple minutes of actual thinking counts as work; typing, looking things up and so on is something else] on an MO question and thus to ask MO question that require this is something desirable. Whereas to me the default assumption for an MO question should be: this should be known/obvious/easy to somebody with the right background. That sometimes this can then turn out to be different can of ocurse still happen; but at least the assumption should IMO be essentially this.


    Now I really understand your point. I can't say I agree with it (it reminds me of "immigrants are stealing our jobs", with "our" replaced by "academia's" and "immigrants" replaced by "online fores" - but there is still the same underlying idea that the right to work is a privilege). But in reality I must also say that for roughly 80% of my questions, before they got actually answered, I had no idea whether they would require work or be answerable in 1 minute by an expert. When you are asking "is this statement well-known/true/obvious", you usually don't know the answer, so you have no leverage over how much work will be done on that question; it depends too much on the answer. So the default assumption "this should be known/obvious/easy to somebody with the right background" is, I think, impossible to make for somebody asking a question.

    Anyway, if people think this is off-topic, I'll stop commenting on this point.

    • CommentAuthorWill Jagy
    • CommentTimeDec 3rd 2011
    This seems relevant. My adviser, Rick Schoen, asked people at the department if the thesis topic he had suggested was reasonable, he was not entirely certain. At a conference, he also asked our own Robert Bryant, who said it seemed good for a dissertation. He also went home and solved it that afternoon. But he placed his notes in a drawer and forgot about the whole thing. When I asked him later, in a bit of a panic, he said don't worry, it's your problem, finish it up and turn it in. It was different when I contacted Johannes Nitsche later, he said hurry up and publish before people scoop you, but even then his article did a certain base case and left the higher-dimensional cases for me.
    • CommentAuthorquid
    • CommentTimeDec 3rd 2011

    @Darij: Actually I am not at all sure we understand.

    In a narrow sense, we are talking about a specific online forum (MO), with certain technical merits and drawbacks. Its design is for question that get answered (definitely), typically in one try. [This is not always the reality, but the design-idea is this.] Not for some back and forth exchanges over extended periods of time that would seem more apt for collaborative work. So from the mere fact that I or somebody is 'against' something on MO, it does not follow the same would be true for other online fora (with different technical details).

    However, it is true that I am generally sceptical towards ideas of largish (online) collaboration. However, for sort of the opposite reason you seem to assume. Some people seem to think that such a developpment will result in a more open and/or egalitarian (math research) community. In my opinion if this would really catch on (beyond some experiments) the exact opposite would happen and the community would become more hierarchical; it would become like those sciences with lot of labor-sharing, large labs and so on [disclaimer: my knowledege of other, than math, scientific communities is only anecdotal].


    @quid: This is an interesting argument. If you are referring to sciences such as biology and medicine, I think that collaborative mathematics would still be of significant difference to these, because collaboration in mathematics is purely optional (we collaborate if we like it) and meritocratic (those who don't contribute anything of value are rarely mentioned among a paper's authors), whereas collaboration in medicine is enforced by practical necessity (not everybody has a lab at home...), the absence of intrinsic objectivity outside of mathematics and CS (the validity of a medicinic experiment rests on the academic status of the experimenter and his academic institution, the trustworthiness of the people involved and many other outside factors, as there is no such verification-falsification device as proofs in mathematics) and legislation (you wouldn't want people to know that you are cutting open monkeys' brains in your attic), and as far from meritocracy as science can get. I believe that obligatory collaboration of the latter kind results in (or at least cements) hierarchies, while that of the former kind is not particularly detrimental to the kind of anarchy that mathematics has always thrived in. I would change my opinion if I see signs of hierarchization (such as moderators on MO using their powers to push some questions of particular interest to them, the bounty system being overused by high-reputation members to make answering their questions more profitable, tactical banning etc.), but at the moment none of these is happening here on MO (the software is good at preventing these things, but I'm not sure they would come up in the first place).

    MO already has a good anti-hierarchical, or rather anti-feudal effect in my eyes: it makes people formulate (and even prove, on occasion) those "folk results" and "folk intuitions" that the experts on a given subject all know but never have really written up and published. This slightly levels the field to the advantage of newcomers who want to learn some modern mathematics (by "modern" I mean "not having a definitive and readable textbook written about it yet"; this does not mean anything like "50 years old or younger") without having to (physically) go to its headquarters and attend the right conferences. And potentially even find some good problems to think about without having to become a grad student of some of the relevant professors. When I read your post, quid, I thought this was exactly what you were trying to prevent. I am glad this was a misunderstanding.

    Wow, I had no idea this thread would see so much activity. What I was really seeking as an answer was given nicely by Scott Morrison and Andrew Stacey, i.e. the moderators would be willing to help with damage control should this situation ever occur. I also like Alexander Woo's idea of asking your advisor to leave a comment. I had no idea advisors get in touch with other experts about what their grad students are working on, mostly because I'm in a small department and no one ever talks to me about the kinds of things PhD advisors have to do. Anyway, I'm glad for this information and will keep it in mind if I ever end up being an advisor myself.

    I also really like Angelo's answer. That's definitely the sort of breakdown I should have done myself, and it does a lot to calm my nerves. One thing worth noting is that even if a PhD problem shouldn't be easy enough for an expert to solve in a few days (I'm sure some people out there think about MO problems for more than just one afternoon), many Masters theses are easy enough for this to occur. I'm mildly proud of the theorem which makes up most of my masters thesis, but I don't delude myself into thinking it's a groundbreaking result. It's sufficient for a masters in computer science and for a small paper, not for any sort of PhD. It also turns out to be a very natural question, so it's the sort of thing a person might think up and decide to post on MO simply out of curiosity. My fear of this happening before I had fully written the result up is what caused me to start this thread. Incidentally, this exact situation occurred ( for my first masters thesis problem, and my solution was to just drop that problem and move on. As quid said in his first comment, I still gained valuable knowledge from the project and I moved away from it before investing too much time. I would have felt differently if the question had been asked a month or two later and if I'd already had results.

    I was also planning to say something about the quid vs. darij debate being off topic, but as I thought about it, I realized this is not that far off from the sort of debate I envisioned when I started the thread. If MO questions are allowed to be very broad then that increases the odds of someone asking a thesis problem. If they are allowed to be very technical, then that increases the odds of some student asking lots of technical questions and building a thesis out of the answers without having done any real work. I guess I'd like to see the community reigning in questions which are too close to either extreme, but due to all the differences of opinion I'm not sure there's any hope for this except by the moderators in specific situations. Adding more to the FAQ won't help because it's already too big for most people to read. I agree with quid that there's some danger of MO turning into a "giant online collaborative research activity," and I'm becoming more convinced this would be bad for young mathematicians who are trying to make their way into the math community. Anyway, I think all the ideas above (for how to deal with this situation if it arises) were what I was really after, and I'm satisfied.
    • CommentAuthorAngelo
    • CommentTimeDec 3rd 2011
    Good luck, David. In any case, being scooped is something that can happen, to you or anybody else, with or without MO. It has happened to me, but, fortunately, not with my PhD thesis.

    Relax, and enjoy your work.
    • CommentAuthorStorkle
    • CommentTimeDec 4th 2011

    Dear Bill, Ryan

    I was being a bit hyperbolic. Let me clarify: My default assumption is that the advisor suggested a problem she already had a strategy for answering, and gave the student an idea of the strategy. My unscientific opinion is that this is what happens in nearly all cases, and that students who come up with their own problems together with strategies for solving them are exceedingly rare. The amount of intervention required on the part of the advisor obviously varies, and it is impossible in practice for anyone but the advisor or student (and maybe not even them) to really know how much of the thesis is the student's (though of course the student is expected to write all of it). Indeed, the thesis is really is a product of the interaction between advisor and student, and in most cases an attempt to assign (say) a percentage that expresses what portion of the work was done by the student is foolish, just as it would be in any other collaboration.


    If you regard this (typical) process as academic dishonesty on the part of the advisor, then we are guilty of academic dishonesty on a truly breathtaking scale! What do you propose to do about it?

    Best, Storkle


    @Storkle: There is a HUGE difference between your clarification and what you wrote earlier. Raising a problem and suggesting a strategy is not the same as knowing how to solve a problem.

    • CommentAuthorKaveh
    • CommentTimeDec 4th 2011 edited

    Two questions that one of my friends was given by his advisor to work on turned up as polymath like projects, so they had to change the question he was working on twice. It is not just on mathoverflow, it is happening all the time, every-time someone mentions an interesting problem and others start to think about it, Internet just makes that more transparent. On the positive side, when this happens you learn about others working on your problem faster.

    related: when should you say what you know

    This thread is relevant to me. Without giving anything away, my adviser has developed a new gadget in subject X which seems like it ought to have an analogue in subject Y, and more than just an analogy there should be a theorem relating the use in X to the use in Y. We don't know if anyone in subject Y has come up with the analogous gadget. There are two situations:

    1. If they have already discovered this gadget, and used it to do some interesting things, then this is probably not suitable for a thesis (it would most likely be a matter of slogging through the definitions, to come up with a precise comparison theorem).

    2. If the corresponding gadget has not been found in subject Y, then I would have to do a lot of work figuring out the right definitions, proving the comparison theorem, and most importantly applying the new tool to subject Y to see if I can get some novel results.

    The problem is how to find out which of the two situations I am in. No one in subject Y has heard about this at my home university, but the department is rather small. If I ask about it on MO, then I find out which situation I am in. If 1, then I avoided putting lots of work into something which shouldn't be a thesis anyway, which is good. If 2, then I am glad that my thesis topic is suitable, but I have also informed the world of an interesting idea before I have started working on it. So it seems difficult. I hadn't really thought about asking it on MO before this thread. Obviously I will ask my adviser for his opinion.

    What do you think?

    Dear Steven,

    Should you decide to work on this problem, you and your advisor need one of the people in subject Y in your home university to take responsibility for finding out if this gadget is known in Y and, should it become necessary, for letting others in Y know that a graduate student is working on the problem and gently asking them not to work on it.

    I hope that the mathematical community will develop standards of etiquette such that mathematicians can safely use the Internet to address these kinds of issues without direct personal contacts, but public discussions of research on the Internet is a relatively new development and not enough time has elapsed to develop such standards.

    • CommentAuthormarkvs
    • CommentTimeDec 4th 2011
    @Steven: You should not post your idea on MO or elsewhere. It is impossible to ask a mathematician not to work on your problem. It could be, for instance, that this mathematician started working on that problem before you, and your idea is the "missing link". The best (for you) he/she can do if he/she uses your idea and solves the problem is to give you credit or offer you to be a co-author. Both options are not very good for your thesis. @Alexander: The standards of etiquette in scientific community have been around for at least 500 years (remember Cardano, Tartaglia and Ferrari), and Internet will not change it significantly.