Not signed in (Sign In)

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


    Question #132632.

    Remark: The typical pattern for using CW mode in this way is:

    • Have a simple question, e.g. a statement of the original bound, a link to the original paper, and a statement of goal.
    • One accepted answer with the main info: the latest known improvement and a timeline of earlier improvements.
    • Possibly additional answers with auxiliary information.

    In my opinion, cramming everything into the question text is a rather confusing use of this site.


    Definitely for a blog rather than MO.

    Or, as mentioned by Nurdin, for a polymath project (which I actually think is a good idea).
    • CommentAuthorHJRW
    • CommentTimeJun 3rd 2013

    Since there haven't been any comments in defence, and since even the principal participants seem lukewarm about using MO in this way, I just cast the final vote to close.


    I would have been curious to see if it worked out. The polymath projects have had uneven success, and I think part of that is finding the right format. It's not entirely clear to me that MathOverflow is actually worse for this than a series of threaded blog posts. Perhaps this would only make matters worse in the eyes of some, but such a MathOverflow hosted polymath project would naturally result in multiple questions, corresponding to new, smaller problems that arise along the way --- thus it wouldn't necessarily be a long jumble of disconnected answers on a single post.

    I'm not proposing reopening, but perhaps this could be a useful place to think about future polymath projects.

    @Bill, why do you think it's so clear this sort of thing is definitely for a blog?

    • CommentAuthorquid
    • CommentTimeJun 3rd 2013

    @Scott Morrison: there are two issues I see with this general idea.

    1. Where does one draw the line? That is, if sometimes/for some problems it is alright they are posed on MO, then there would need to be some criteria to decide which problems can be posed on MO. Ever so often somebody already poses a problem, this is typically closed. Now, I agree that this current one is for various reasons a relatively good choice, but still we will face the question sooner or later (rather sooner than later, IMO)

    2. Due to the way how MO works it is very sensitive to 'noise.' And, if there is too much on one subject at the same time and/or in rapid succession this causes in my observation (and also personal experience) quite quickly a certain annoyance. To formulate it a bit stronger than I mean it: it feels somewhat impolite to effectively force the entire MO-community to follow the progress some make on some problem in real-time. (The one at hand is rather close to my interests so, in this case I personally would be rather on the good side, but still I am against this on general grounds.)


    David Roberts suggested here a more focused version of the question, namely "Can the techniques of Morrison-Tao be used to get the prime gap bound below 10 million?", which I believe would have been better than the more open-ended current version. Anyway, the current bound is at most 4,982,086.


    It's pretty clear to me that MO is not the ideal place for the discussion and development of strategies in improving the bound. However, I see no reason that MO couldn't act as a "central repository" for the improvements. But it has to be done in a useful way; see the first post for a reasonable organizational plan. Auxiliary answers could be used to point to different places where strategies are discussed but I don't think anything other than highlights should make it to MO.

    Scott has a comment on the main site with a link to polymath. It would be helpful if people could vote that up so it becomes visible.

    Terry Tao made the following comment on the Polymath Blog:

    Seriously, though, it may perhaps make some sense to define the scope of the project a bit; I don’t think the project could try to cover every single possible way to develop Zhang’s breakthrough, and there should still be plenty of room for more traditional research in this area.

    This suggests that MO could still be used as a way of keeping track of progress similar to what I suggested above. I would be willing to reopen the question if someone volunteered to rewrite it in a useful form. Alternately, start a new question and I can reclose this one as duplicate to redirect people.

    (See this meta.SO question for an example of using a simple community wiki question with a community edited answer summarizing all the relevant info. There are lots of other examples if you look.)

    • CommentAuthorVit Tucek
    • CommentTimeJun 4th 2013
    I am happy to rewrite the question.
    • CommentAuthorquid
    • CommentTimeJun 7th 2013

    It is not quite clear to me how this "question" should operate now. Specifially, I am asking since there is now a paper on the arXiv by Pintz getting 768534, so about two orders of magnitude over Zhang's original and almost one over what is listed as "current record." The paper is already included in the polymath-timeline and polymath seems to be past that point already and even it seems was independently before.

    Should this paper and some text on it be added now?

    If so as, as an independent answer or via an edit. In my opinion, since this question now exists, it would make sense to me to add it, likely as an independent answer; on the grounds that there is a substantive improvement in the form of a complete write-up. However, since I am somewhat dubious regarding this and also did not want to 'steal' this answer, and finally would anyway prefer if somebody else would do so, I ask here.


    Community wiki questions like this are intended to be community maintained. Anyone with 100 points is strongly encouraged to update it if they have additional information to contribute. Your questions, quid, indicate some deeper problems with this idea. I can think of two potential explanations:

    • The community does not understand the idea and/or is not receptive to collaborative editing.
    • The question is not sufficiently well organized. Ideally, everyone should know immediately where and how to add tidbits of relevant information.

    There is some evidence for both; the second is easier to correct than the first.

    Another possibility is that this question (partly) duplicates other resources such as the polymath wiki. That one is easy to fix: replace the duplicate content with a link to the more appropriate source.

    • CommentAuthorquid
    • CommentTimeJun 7th 2013 edited

    @François G. Dorais: Thank you for the reply. Good, if nobody else does it within some time or there is a different developpment in the discussion, and I find time to write some meaningful accompanying text, I will add this as a new answer.

    Another possibility is that this question (partly) duplicates other resources such as the polymath wiki.

    Well, yes. This is why (among other reasons) I was rather against this "question" in the first place. But now that we have it, it seems odd to me if it were not maintained properly (as this would reflect rather poorly on the site, in my opinion) or if one wishes to see it like this the most relevant (or only relevant if one wishes to be critical) information is a link in a comment to the answer.

    That one is easy to fix: replace the duplicate content with a link to the more appropriate source.

    The link is there but this seems a bit too little content in my opinion, besides it not being very visible. In addition to being against the general philosophy that the content of the site should be at least also somewhat meaningful without the content of the linked resource. (At the risk of drifting off-topic: On the network they seem to be rather strict, even in mechanical ways. Better we prepare and learn collectively for this now, than being then faced with many complaints, in either direction. To avoid a misconception, I am looking forward to this aspect of the change as I quite dislike 'link only' information in particular for relevant information, though I admit sometimes I do this myself.)

    As said if noone beats me to it, I will do this later. Thanks again for the feedback on the matter.

    Added: upon reading a bit on the wiki and a blog not sure I will actually do this as there is some discussion I do not fully oversee. So, there is some risk of creating some confusion, which I'd prefer to avoid.