Not signed in (Sign In)

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

    • CommentAuthorEmil J
    • CommentTimeAug 13th 2012

    @François: In SE 2.0, these code blocks are exempt from interpretation by MathJax, so they appear to the user as the LaTeX source rather than a formatted equation.


    Thanks Emil, that clarifies the problem. I was finally able to log back into the sandbox. I think this is a correct interpretation of code blocks: MathJax has no business processing those. I think it would be best to fix this during migration rather than later. Thank you for pointing out that issue!

    I continue to feel that migration is a bad idea (not that I have a better idea). My primary concern remains that StackExchange employees will impose their worldview on us to the detriment of the site, in the interest of "global standards" (for example, creating global bans on certain types of signatures which drive away "Ask me about system design" Passman, or banning some other benign form of behaviour which some mathematicians happen to engage in). Can we have a symbiotic relationship with a company whose goals so widely diverge from our own? I continue to be pessimistic... but the alternative would be to break away somehow, which might be even less practical.
    • CommentAuthorbbischof
    • CommentTimeAug 13th 2012

    Along the lines of what Daniel is saying, I know on the SE sites ending questions with things like: "thank you", "thanks in advance", "I hope this makes sense", or beginning with "hello", "I have a question..." are banned.

    Maybe this isn't a big deal, but I have always felt the ban of these type of phrases completely pointless and irritating. The SE desire for streamlining feels nearly totalitarian in some ways.


    I suppose we need to require that a migration path to MO will be open only if there is a majority of consent between the moderators/Anton/the meta.MO community. This way we can at least diminish the amount of things trafficked into MO by non-moderator users.

    As for moderators on other sites, I'm sure they'll understand soon enough not to migrate things that are not fitting to MO.

    The only migration path that should be open from the beginning is MSE (I suspect that TeX.SE could be a useful option too, but with so little TeX traffic I'm sure the moderators here can manage manual migration).

    • CommentAuthorRalph
    • CommentTimeAug 14th 2012
    @Daniel Moskovich: I totally agree. I wonder if it hadn't been possible to find a university that is willing to host MO (similarly as Cornell University hosts the arxiv). Of course there would be drawbacks (mainly concerning the costs and programming the software) but in my opinion it would have been worth a try (actually I don't know if this has been tried, but it seems to me that the discussion Anton linked in the first post of this thread is primarily centered around pro and cons of migrating to Stack Exchange rather then searching for alternatives).

    The current proposal by Joel Spolsky as cited by Anton is completely insane. Has Elsevier's story taught us nothing? Do we want in a few years to find ourselves in the same situation as editorial boards of Elsevier's journals, who are unable to switch to a better publisher because they totally screwed up their legal homework when they originally set up their journals?

    The value of Joel Spolsky's promises is exactly zero. In fact, the text above might as well say “Stack Exchange Inc. is free to do to MathOverflow whatever it wants, without any consequences”.

    I find Anton's question “How can you justify such scepticism of their attitude (rather than anything concrete which can be put in writing) when there is such a strong track record (if there weren't, we wouldn't have anything to defend right now).” utterly naive. Did the strong track record of Springer and Elsevier (and their acquisitions like Academic Press) prevent them from charging ridiculously high prices for journals to which they add no value (and in fact, add negative value)? As a more recent example, did LiveJournal's strong track record in the first few years of its existence prevent it from screwing up its users (introducing ads everywhere etc.) a few years later? What makes you think that Stack Exchange Inc. will not be taken over by some evil Elsevier-type entity that will renege on all the agreements discussed above?

    If history teaches us anything, it's not to trust any promises made by for-profit entities (like Stack Exchange Inc.) unless they are made in the form of an actual legal contract reviewed by a lawyer.

    Again, math journals supply an example: the founders of the journal Compositio Mathematica did their legal homework right. The journal was owned from the beginning by Foundation Compositio Mathematica, not by a commercial publisher. When Elsevier became evil, the editors simply said goodbye and switched to publishing with LMS.

    In our situation, the appropriate way to go seems to setup some kind of “Foundation MathOverflow” (as proposed by part 9) which will not only own the domain name, but, contrary to the current terms in part 1, should also own MathOverflow's data and users (i.e., their private data like emails, OpenIDs, etc.). Then (and only then) it will be easy for us to transition to a different hosting.

    If this is not done, then the future evil entity that will own Stack Exchange Inc. will have an easy way to prevent a possible transition of MathOverflow to a different hosting system by refusing to honor the authorization clauses in part 8 or refusing to provide a public data dump. This has happened (in a different form) more than once in the past, and there is absolutely no reason to assume that Stack Exchange Inc. will follow a different path, despite its “strong track record”.

    Alternatively, we might as well switch to a free software platform like OSQA and severe all ties with SE.

    • CommentAuthorabergman
    • CommentTimeAug 14th 2012
    Having observed the entire theoreticalphysics process (debacle?) on StackOverflow, I think the culture of the management there is fundamentally broken. Their philosophy and other ideas about their site are incompatible with many of the features that have made MathOverflow a successful research community. They also were shockingly rigid during the theoreticalphysics process. I don't know if there is some technical reason why the current platform needs changing (why mess with success?), but I'll register my 'vote' against this.

    I still think that part of MO's terms should allow for us to keep this wonderful meta.

    • CommentAuthorEka-user
    • CommentTimeAug 14th 2012
    If and when MO joins SE, the ensuing culture clash, office politics and SE staffers vs MOflowers conflicts will be a deliciously entertaining drama to watch. All sides know this and are already preparing their part in the forthcoming game. The novelty of MO has worn off so the obvious next step is to have a big fight with the SE network. Is it a zero-sum game ? Who will win - the site or the network ? Individual cells can be sacrificed for the good of the organism.

    Dmitri, the problem with your all-corporations-are-evil-and-can't-be-trusted line is that Math Overflow is currently being run by Stack Exchange. 2.0 vs. 1.0 it's the same company, just a different interface. So if the switch to 2.0 is somehow wrong for corporations-can't-be-trusted reasons, then we're wrong for have started MO in the first place. Since MO is a pretty neat thing, the corporations-can't-be-trusted line seems to be the point in error here. There are no blanket statements of that form. I hope we can turn the conversation away from such conspiracy theory lines and towards more productive dialogue.


    Also note that a big part of the problem with journals is that they own the journal name and the back issues. The conditions we're being offered here are much better than what we have with journals.


    bbischof's comment is somewhat important. Note that SE does some of that stripping automatically (by script) across the whole network. I checked though, and they don't seem to have messed with Matt E's "dear x" comments. But I'm worried about things like one day arriving to find SE's deleted all of BCnrd's comments.

    Dmitri Pavlov writes "contrary to the current terms in part 1, should also own MathOverflow's data and users (i.e., their private data like emails, OpenIDs, etc.). Then (and only then) it will be easy for us to transition to a different hosting."

    I'd like to hear from Anton and Scott M to what extent this is true. Setting up "MO in exile" would be difficult in any event. Would the private data actually be the sticking point? I feel like it wouldn't be too hard to have everyone hit a banner the first time they come from the new site saying "Returning user? Please re-enter your contact information," with a link to a web form to do this. And we could seed the database with the logins from the 1.0 period, so some people could login as soon as the other problems of setting up the new site were solved.
    • CommentAuthorMariano
    • CommentTimeAug 14th 2012

    Are the details of the «theoreticalphysics debacle» written down somewhere?


    I've just been informed that the sandbox has been updated and the backticks hack has been stripped from all posts. As with any bulk edit of this kind, it's possible that different things were broken in the process, but these are hopefully far less frequent than the original problem.

    • CommentAuthorEmil J
    • CommentTimeAug 14th 2012

    Great, but it seems that backticks have been left in comments.


    Backticks were never necessary in comments.

    I have pulled all of the backticks from the posts in the sandbox - feel free to check it out and let me know if there are any other issues with the conversion with regards to the data.

    As a side note - we are working very hard to make this conversion as painless as possible and continue to appreciate all of the contributions of the MO community. We are solely interested in providing a place for a successful thriving community and will work around the clock if necessary to support MO. MOs success is a shining example of how a community can bond around a topic even through there may be disagreement among individual members. The primary goal for us is to help create a place to get expert answers to your research level math questions and create a valuable place on the internet that will help others learn. We will be happy to provide data dumps if you find the community isn't working using the free public park we have worked very hard to create.
    • CommentAuthorziggurism
    • CommentTimeAug 14th 2012

    I read the discussion on that Asaf posted and found myself sympathetic to both Henning's and Asaf's viewpoints regarding closing research level questions on math.stackexchange and migrating to mathoverflow. I really liked Charles Staats's comment that a software feature should be added that allows a post to appear in both places. That way the posts on mathoverflow can be a proper subset of the posts on, reflecting the fact that the topics that are considered on-topic on mathoverflow are a proper subset of Not sure how meta.mathoverflow community feels about the suggestion, but if you like it, perhaps it would make sense to make addition of such a feature a prerequisite for the move.

    • CommentAuthorabergman
    • CommentTimeAug 14th 2012
    A lot of the theoreticalphysics discussion seems to have disappeared when the site was taken down, and I was an observer rather than a participant in the management of the site. Nonetheless, the original discussion on formation is worth reading:

    and two pages about the aftermath:

    I would like to clarify my suggestion here: I by no means support, nor mean to suggest, any system in which "posts on mathoverflow can be a proper subset of posts on" The idea, as I understand it, is this: a number of people on math.SE object to the idea of questions being migrated to MO, even if they have gone unanswered for a significant time period and knowledgable members of the community believe they would be more likely to receive decent answers on MO. Math.SE people are worried that this might produce a drain on the supply of "interesting unanswered questions" and thereby be ultimately harmful to the math.SE community. A more subtle concern is that the sort of answers given on the two sites are different: a good math.SE answer (with lots of explanation of basic concepts) might come across as condescending on MO, while a good MO answer might be considered terse, incomplete, or incomprehensible on math.SE. At the same time, the fact remains that there are sometimes questions that are simply unlikely to receive any sort of satisfying answer unless and until they are brought to the attention of the MO community.

    My suggestion amounted to the following: I think it might at least make progress toward resolving the issues, if it were possible to "extend" a math.SE question onto MO rather than closing it on math.SE and migrating it. In my visualization, all the answers and edits would automatically be synchronized between the two websites, but--significantly--the voting would be entirely separate. Consequently, the answers would likely appear in different orders. In an "ideal" scenario, I would envision something like the following:

    1. A math.SE user posts a good question on math.SE that turns out to be difficult, and that might benefit by receiving more attention from the MO community.
    2. After a week on math.SE with no good answers, the question is "extended" to MO. Some MO users consider closing it, but forebear because they see how long it went unanswered on math.SE.
    3. Not to long after the question is "extended," MO users start writing answers. The culmination of this is an answer that beautifully and concisely lays out the answer to the question, in the process perhaps illustrating why the question is more difficult or subtle than it at first appeared. This answer receives many upvotes on MO and rises quickly to the top of the list. However, on math.SE, it languishes with few upvotes because relatively few people (including the OP) really understand it.
    4. Finally, a meta.SE who was unable to answer the question on his own but does understand the answer given on MO, writes an expanded, lower-level explanation of this answer as a separate answer. This answer languishes near the bottom on MO, where it is considered inferior to the original concise, brilliant answer. But on math.SE, the longer answer receives many upvotes, rises to the top, and is accepted by the OP. Finally, through a combination of the efforts of both communities, the OP has an answer to the question that he can understand.

    Of course, most questions would probably not work this smoothly, and the feature might prove impractical to implement (or the SE team might simply believe it is not a good idea). But I think it is at least something worth considering.


    In case that wasn't entirely obvious to everyone here, Charles Staats above is referring to the earlier post by ziggurism and these are ultimately continuing a lengthy debate on

    While this discussion is of interest to MO, it is tangential to the present thread. I recommend continuing this discussion on or on another thread here on, whichever is more appropriate.

    So far, Open Id login did not work for me at MSE, while it does on MO: The only way I was able to log in with my Open Id to MSE is to create a new account, different from my MO account (both link to the same Open Id).

    Misha, a handful of users (including myself) have been having similar issues. My understanding is that this is a temporary glitch. However, you can contact Geoff Dalgas (gdalgas at stackoverflow) so he can diagnose and fix your login issues.

    Dear Francois: Good to know that I am not alone! Of course, instead of MSE I meant
    • CommentAuthorWillieWong
    • CommentTimeAug 15th 2012

    Wow, I go away for two weeks and major changes are afoot. Cannot comment too much at the moment due to limited internet access (and lack of a well-formed coherent opinion on the matter). But can comment on a technical issue that has been raised with regards to question migration:

    • As mentioned before: ordinary users from other StackExchange sites cannot vote to migrate questions that are not on a migration path made available to them. It should not be impossible to request SE Inc to not open a general migration path to MO on any other StackExchange sites.
    • Moderators on other StackExchanges sites have the power to migrate to any other StackExchange site they see fit. For borderline cases, or cases where the Moderators are not normally familiar with the culture/scope of a possible target site, we are encouraged to post a query prior to migration. This is done via the Moderators Only chat channel on chat.stackexchange, where one is given the ability to "ping" all the moderators of a particular site. I can attest that most moderators are reasonable people and have put this feature to good use (at least when migrating to Math.StackExchange is concerned).
    • As part of the Moderator Tools on SE2 is a function to monitor question migrations to and from the website. If we were to really want to be isolationists, it is not too hard to task the moderators to check on the question migrations log and close any questions migrated to MO from other sites. (I am not advocating this, just stating that this is a possible option.) The would-be migrating question would then be bounced back to its originating site and marked plainly as "off topic" there.

    There are some other concerns that I may have with regards to moving MO to SE2, but I don't think we need to be worried too much about influxes of other questions due to migrations from other SE sites. (Caveat: I have not read nor participated in the Meta.MSE discussion that Asaf opened; see my first two sentences at the top of this comment.) If anything, I think the migration will actually help clear out clutter by making it easier to redirect questions to MSE.


    @Ryan: Please read again my comment. It's not “just a different interface” as you claim. Currently SE does not own the MathOverflow data and users. This will change if the new terms are accepted, and it will make the potential future transition to a different hosting system much more difficult.

    As for “I hope we can turn the conversation away from such conspiracy theory lines and towards more productive dialogue.”, do you think that the current situation with Elsevier's journals is a conspiracy theory? Or do you claim that my proposal to retain ownership of users and data is somehow unproductive?

    I just noticed this interesting comment

    "Assuming the [private] dumps have the same format as the normal data dumps, they’ll contain a md5 hash of the email address(unfortunately not validated). This could be used to write a “reclaim account” feature on a third party site."

    This seems like a way to make the "break glass" option more functional, while respecting SE's privacy concerns. (By the way, can I say that it makes me happy to see a social networking company that does notice the privacy issue here, even though it causes a problem for us?)
    @ David:

    "Further, there is also a chosen-prefix collision attack that can produce a collision for two chosen arbitrarily different inputs within hours, using off-the-shelf computing hardware"

    • CommentAuthormarkvs
    • CommentTimeAug 15th 2012
    Is it correct that SE is less "community driven" and more "moderator driven" than MO? For example, can people with 10K+ reps vote to delete questions? Also am I right that a user cannot make his/her question or answer into CW without a moderator approval? If there are such differences, then I hope we'll know it before the move.


    Users with 10k reputations can vote to delete closed questions two days after closing, or so. Users with 20k reputation can vote to delete downvoted answers, as well immediately vote to delete after a question is closed.

    It is true, however, that a moderator is required to ask a CW question.

    • CommentAuthormarkvs
    • CommentTimeAug 15th 2012
    About moderators and CW is may be not a bad idea especially if the request can come not only from the OP. Sometimes it is hard to make an OP turn a question into a CW even if it is obviously CW-like, and if moderators will be more actively involved, it might be good.

    One can flag for a moderator attention and request the question be made CW, or one can leave a comment and hope a moderator comes by and acts on that.

    I'm not sure whether or not it's preferable or not. Maybe it will ease up things in that aspect, but I also think that it's quite annoying personally.


    Using the CW hammer is already a regular task for moderators on MO. When we hit a question with the CW hammer, all the answers become CW as well. This is not the case when a user makes their question CW, only subsequent answers become CW in that case. We often go back and hit CW questions with the hammer.

    Also note that voting to delete answers is not currently possible on MO, only the original poster and moderators can do that. Same thing with immediately voting to delete closed questions.

    • CommentAuthormarkvs
    • CommentTimeAug 15th 2012
    Thanks for clarification. I hope it (and other similar changes) will be collected somewhere and publicized, possibly in a special section of FAQ.

    I only have a minute. Let me comment on David's thoughts about the feasibility of migration to another platform using only the public dumps. Assuming "reclaim account" is solved (either by authenticating with SE to reclaim an account or by some other method), there is one other important technical problem that needs to be addressed. The public dump does not contain the owners of any votes. This means that we would either have to allow double voting on old posts, or disable voting on old posts. Neither of these is very appealing. I've asked the SE team if they could include the vote owners in a one-time custom dump without violating their privacy policy. I'll keep everybody here posted on what happens with this.

    Community Wiki is one of the places where there's a cultural difference between and MO. It's hard for me to make the difference precise, because (even though I spend lots of time at it's not clear to me what exactly the policy is. I guess what it comes down to is the community feels people should get credit (that is, points) for contributing a useful answer to, say, a big-list question, so they oppose making such questions CW.
    • CommentAuthormarkvs
    • CommentTimeAug 15th 2012
    @Gerry: And they are going to impose their point of view on us? Are they going to interfere if our moderators turn a question into a CW? I do not think it is very good for MO if somebody who is not a research mathematician makes decisions about questions posted on MO.
    Most of the time the decision will most certainly be silly. Is it possible to make sure that they do not interfere?

    Mark, although your point of view is just as valuable as Gerry's, it is wrong to attribute this to "them" (whoever "them" may be). This debate has been happening here for a long time, without outside influence. Even the moderating team is split on that issue, though it's never been a source of internal strife. CW processIng has been running well for several years and it should continue that way. If and when it stops running well, we will address that issue at that time. There is no need to create issues where there are none.

    • CommentAuthormarkvs
    • CommentTimeAug 15th 2012 edited
    @Francois: I was talking about new rules, at SE where we are supposed to move, not the current situation at MO (the whole thread is about the move, isn't it?). The meaning of "them" is well explained in the previous comments. I did not express any opinion, I asked a question: will anybody from outside MO interfere in the MO business (like making questions CW, but not only). It would be nice to get an answer to my question.

    Mark, it seems like I might have misread your first comment. To answer your question, no I don't think anything will change drastically. Some users might have to take the habit of raising a flag in addition to making a comment but a lot of people are already doing that. I don't think the adaptation will be significant to anyone concerned.

    Other than that, SE people have not expressed any desire getting involved in everyday MO business. They have not expressed disdain, but the total lack of interest is significant enough for me at this point. Right now, everybody involved is focused on making the transition happen as smoothly as possible.

    @markvs, I take it your "they" refers to the community, since that's the only group I referred to in the post of mine to which you responded. I can't imagine any circumstances in which the community would have the ability to impose on MO their point of view on CW or anything else.

    But perhaps your "they" refers to the people who run the SE network. I think they will find the MO interpretation of CW to be quite peculiar, and not what they have in mind; I doubt they would interfere with what we do, although I don't feel that I am sufficiently well-informed to claim any authority on the question. It may be something that our folks want to discuss with their folks as part of the negotiations. Or it may be better to let sleeping dogs lie and not bring up the question and hope no one at SE notices our approach to CW. I don't know. I saw CW raised earlier in this discussion, and was only trying to draw attention to the differences in understanding and usage of CW in different places.

    Something else I thought might be an issue, and I had to turn off all my ad-blocking and do-not-track plugins to see them, is the cross-platform ads (they show up on the rhs just over the 'N people chatting' link). I think having ads for MO plastered across the SE network may go against the isolationist preference of some (including me). They may be context sensitive (not showing up on, say, travel.stackexchange) but I could image they are not (I got one for serverfault on Likewise, we will also get ads on MO2.0. I also saw ads for PlanetMath, OEIS and GeoGebra, which is ok, but I do not know how ads are determined/distributed.

    EDIT: ah, here is how:


    I think that the SE overlords have accepted the fact that m.SE (and MO) use CW differently. They still look at us like we're weird, but we are mathematicians... so it makes sense regardless.

    The only problem which may arise is that new users might take the "SE direction" a bit too much, and not only behave in contrast to site norms -- but will also quote the SE blog as the authority which allows them to act like that. It happened on m.SE not too long ago...


    @Anton: Would it be possible to clarify the reason why Stack Exchange demands that the ownership of MO's users and data be transferred to them?

    If the current model (in which SE does not own users or data) worked just fine for the last three years, why can't it work in the new version (which is ”just” a different interface, as Ryan Budney claims)?

    I cannot see any possible positive uses for such ownership, but I do see a lot of potential for abuse (e.g., Stack Exchange is bought by some big company like Microsoft or Facebook, which then decide to extract more money from Stack Exchange website, like I explained in one of my previous comments).

    The only possible reason that comes to my mind is that SE does not want to grant exceptions to individual websites. But then again MathOverflow receives very special treatment from SE, and not owning data and users seems like a relatively small part of this special treatment.

    The fact that the Stack Exchange representative chose not to comment on any real issues like ownership of users/data, but instead goes on whining for a whole paragraph on how good Stack Exchange is (as if somehow free software like OSQA, whose authors do not demand ownership of users or data is much worse than SE) also seems disturbing, not to mention his claims of SE being a “free” “public” park (both of which are false).


    @Dmitri: Yes, I'll try to clarify. There are two inseparable changes that come with a move to SE 2.0: (1) updating to the most recent version of the software, and (2) joining the SE network.

    As far as I know, absolutely everybody is in favor of (1). I suppose an alternative like OSQA could work, but would be a step down from doing nothing and staying on SE 1.0. Have a look at these two posts on meta.OSQA, taking particular note of the bit that says "[OSQA is] only an entry-level solution. We offer superior commercial solutions to paying customers" (those commercial solutions are closed source). Functionality aside, though Stack Exchange is a commercial enterprise and OSQA is open source (driven by the commercial enterprise DZone), I feel SE has been more transparent about its intentions and has a more idealistic spirit.

    The source of our concerns is (2). Specifically, users of the SE network can easily create users on sites within the network, and their new profile includes private details. SE's privacy policy dictates that they cannot share these private details with third parties. The yet-to-be-created managing board of MO which might decide to move us to another platform in the future is such a third party. Does this clarify?

    For further clarification, I feel I must take issue with your point of view. You say "Stack Exchange demands that the ownership of MO's users and data be transferred to them." This seems to have the express goal of generating fear in the reader. What exactly is being transferred? Is there something currently owned by you or me which would be owned by SE after migration? You make it sound like using an SE site is a form of slavery, but SE does not own its users, and it does not own the content on any of its sites. It is appropriate to keep private user details private, and user-generated content is owned by the people who produce it and licensed to SE under the CC BY-SA license. Please look at section 3 (Subscriber Content) of the SE legal statement. I hope this makes it very clear why the comparison to Elsevier is severely flawed. If I have misunderstood the comparison, please explain how moving to another platform would become significantly more difficult after migrating to SE 2.0.

    Note that by migrating, we are in no way obligated to delete dumps of the MO database. As David said, this information could be used as a seed in the event that we eventually move to another platform. I don't think it's so onerous of SE to say that private user details will not be available en masse once we migrate to SE 2.0. Though it is an inconvenience, it's the right thing to do from a user privacy point of view.

    I am bothered by your criticism of Geoff. I hope my clarification above is sufficient to explain why he did not speak to the question of "owning users and data" (nevermind that he has been given the task of caring to our technical needs, not our political ones). I do not understand your objection to his phrase "free public park". What is not "free" about it? Is it that some SE sites may eventually have ads on them, generating a profit? Is there some unreasonable way that SE restricts the behavior of its users? Please give an example of a correct use of the word "free" which clarifies why Geoff's use of the word is "false". What is not "public" about it? Is it because there is private information in the database which is not accessible to us? It seems this objection applies equally well to argue that no government in the world is "public". (Note: if you understood that copyright for user-generated content on SE sites was transferred to SE, then your objection makes sense to me.)