Not signed in (Sign In)

Not signed in

Want to take part in these discussions? Sign in if you have an account, or apply for one below

  • Sign in using OpenID

Discussion Tag Cloud

Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.

Welcome to Math2.0
If you want to take part in these discussions either sign in now (if you have an account), apply for one now (if you don't).
    • CommentRowNumber1.
    • CommentAuthorMark C. Wilson
    • CommentTimeOct 30th 2012

    This question is of course not all about me, but I will keep it personal to make it clearer - I am interested in the general problem. Background: I have been working with a senior coauthor for (on and off) more than a decade on a topic (asymptotics of multivariate generating functions). There is one Math/CS community A that uses generating functions a lot, and we have presented work at their conference. In addition we have a survey article in SIAM Review. By 2008 it should have been completely clear that our way of solving problems of type X was superior to the older methods. I find it very hard to see how anyone from community A solving a problem of type X would not be aware of our work, and even if they had forgotten, the standard Google, etc, keyword search we should all be doing would have found us.

    In the last few weeks while researching for our forthcoming monograph on the topic, I have come across two papers in 2012 by members of community A (some of whom I recall being in the audience at our talk in 2007)) which use the old methods and don’t even mention our work at all.

    I have sent a friendly email to one group and will likely do it for the other. Beyond this, I could write a short paper (maybe one for each group) showing how somewhat better results can be obtained with much less effort. The latter will take some work, and the papers won’t be published anywhere great. I am concerned that even if I try to publish the reply, it might not be accepted (no longer “new”).

    I am far from confident that this won’t happen again (hopefully not with the same researchers). And I am really quite annoyed. Proper recognition of other work is important. And it is even more important not to solve not-super-exciting problems using the wrong method, thereby giving others the wrong impression that this is state of the art.

    By the way, the referees presumably didn’t pick this up (Discrete Mathematics, Bulletin Australian Math Soc). Another reason to be a bit worried about the state of refereeing.

    How would you react in such a situation?

    • CommentRowNumber2.
    • CommentAuthorzskoda
    • CommentTimeOct 30th 2012
    • (edited Oct 30th 2012)

    This is not easy. Community A understand their own methods better, and might consider the methods which you perceive as better, more complicated and cumbersome from their point of view. Ignorance is hafd to fight unless it is one of the following clear cases: They ought to cite if either

    • somebody had the same result before them, or if
    • they used a method which other devised.

    Did the community A use any of your methods ? Apparently not. They use obsolete traditional methods of their school (in your own description).

    Did you obtain the same results before ? I don’t know, but if you did not, then they have the precedence and have the right to publish even if you now state that you can get the results in a simpler or better way. The fact that there are people somewhere there who work at the higher level is not a license to ban other people who obtained or announced the results first to show and obtain the credit for what they have.

    Of course, the editor can say that the community A proliferates examples of a method which is considered obsolete by experts, not interesting and so on, but this is often the matter of the opinion, and depends on the background of the editor.

    • CommentRowNumber3.
    • CommentAuthorMark C. Wilson
    • CommentTimeOct 30th 2012

    Thanks, this helps somewhat to sort out the issues. What we did is to prove a general formula that applies to many special cases including those in the 2 papers mentioned, and explained its use in detail in more than one paper. One problem in combinatorics is that there are many alternative descriptions of counting problems. However since both papers used generating function methods, the authors should not have been prevented from seeing the similarity. In paper 1 the result is not something we can claim to have done already, and so I guess I will try to write a paper redoing it and extending it using our methods. In paper 2 the main result not only is a special case of our main results, it is even given as a running example in more than one of our papers. From the point of view of professionalism, both cases annoy me, but I guess the second one is worse, since the results are even less novel. In that case I will definitely point it out in Math Reviews (I found out about it through being assigned it for review), and write to the authors, but beyond that I am not sure.

    • CommentRowNumber4.
    • CommentAuthorHenry Cohn
    • CommentTimeOct 31st 2012

    In paper 1 the result is not something we can claim to have done already, and so I guess I will try to write a paper redoing it and extending it using our methods.

    That sounds like a good idea. One issue in general may be a lack of comfort with several complex variables: aside from analysts and algebraic geometers, many mathematicians are only vaguely aware of how contour integrals and residues generalize beyond one complex variable. It’s easy to imagine that your techniques are more complicated or less beautiful than they really are, in which case it could seem easier for people to stick with what they already know. (Assuming they’re aware of your work it would still seem natural to mention it, but it can feel awkward to say “you can probably also do it using technique X, but we don’t really understand it well enough to say for sure.”) Writing a paper showing how to apply it to this case may help speed up the transition towards your approach.

    In paper 2 the main result not only is a special case of our main results, it is even given as a running example in more than one of our papers.

    This is more troubling.