The extant "answer" to this question: http://mathoverflow.net/questions/49244/give-an-example-of-monoid-with-property-m2-m3 seems to have troll-like characteristics. I got as far as writing a comment in reply (along the lines of mistaking cause with effect) and then decided not to.

I've flagged the comment as spam, I urge others to do so.

(Note that the question itself is highly likely to be closed, but that is independent of troll-like behaviour)

Yes, after posting this comment I noticed that that user had a non-trivial positive reputation so I went on a little trip and found those answers. I can't make head nor tail of them, though, so don't know how to classify them.

(The original answer that provoked this has now been deleted.)

I recognize the phrase "this is somewhere in Laumon Moret Bailly, I don't have it here" because Ben Webster recently noticed and commented "WTF?" on an answer (I can't find it at the moment, perhaps it was deleted as spam) that was essentially ripping off this answer, in a similarly nonsensical way. There is also this answer, another strange copy of the same text, but under the guise of a different user; however that user appears to have given two substantive answers as well.

EDIT: I believe the answer Ben commented on was on this question, though it was deleted one way or another (could a 10k user or moderator confirm?)

Zev, you're right. Here's the text:

Yes. The criterion for a vector bundle on a connected paracompact space to be invertible is that it should have unramified diagonal (this is somewhere in Laumon Moret Bailly, I don't have it here). If the vector bundle has a total Stiefel-Whitney class invertible in its cohomology ring, the diagonal is a monomorphism, and a monomorphism is certainly unramified.

In the comments, someone remarks that this is plagiarised. The original is here and the text is:

Yes. The criterion for an Artin stack to be Deligne-Mumford is that it should have unramified diagonal (this is somewhere in Laumon Moret Bailly, I don't have it here). If the stack is fibered in sets, the diagonal is a monomorphism, and a monomorphism is certainly unramified.

For comparison, there's the "answer" to Harry's question:

Yes. The criterion for f∧g to be in An is that it should have unramified diagonal (this is somewhere in Laumon Moret Bailly, I don't have it here). If An is a class of monomorphisms anodyne w.r.t. a separated segment on such a category of presheaves, the diagonal is a monomorphism, and a monomorphism is certainly unramified.

Two further occurrences of this remark:

1. http://mathoverflow.net/questions/42109/semicosimplicial-totalization
2. http://mathoverflow.net/questions/43537/property-t-and-subgroups-of-finite-index

(Apologies for cross-posting with Kevin - I got distracted by my hunt for further occurrences)

Interestingly, the accounts that left those remarks have other activity which does not appear to be trollish.

The best explanation I can think of is that it's some student with some mathematical knowledge, who every now and then gets intrigued by people using so many words that he has never heard before. So he starts doubting whether these people are actually saying anything that makes sense to anyone but them and decides to conduct an experiment. He produces some Gibberish that is not immediately recognisable as such by people who are not familiar with the subject matter. Ironically, some of that Gibberish actually has - at the point of writing this - a positive vote count (whoever votes up answers that they don't understand?), so if my explanation is correct, then this student must be very content with the outcome of the experiment.

Edit: Something I in turn find curious is that now, that these examples have surfaced, nobody seems to be flagging them as spam. Anyway, I am going to.

I've flagged them as spam. However, given that this person has been posting these answers under multiple accounts, there exists the possibility that they voted themselves up, to lure more people into upvoting than otherwise would (I'm thinking along the lines of a panhandler putting some of their own coins in their cup).

@Spiro: now that's a thought. MO uses OpenID, which is, IIRC, susceptible to the FireSheep exploit of a month ago. So it is not completely inconceivable that someone was able to sniff log-in credentials over wireless and posting as random users. But I do find it slightly unlikely.

I vote up answers I don't understand when I have some other reason to believe that the answer is substantially correct (e.g. the answerer has given other correct answers in the past, the answer already has a high vote count) and when I think that the answer deserves more votes (e.g. clearly a lot of work has been put in but it is hard to check). Maybe this isn't a good idea.

Quoting Todd

whoever votes up answers that they don't understand?

I'm convinced that the answer is "lots of people".

Amen. And I speak as someone whose fairly banal observation about expressing the trace of a matrix as a weighted average of its numerical range continues to get occasional upvotes ;-)

@Peter: Harald's description is the one I understand to be the case. The authentication process is not the problem. The problem is that then you are granted a cookie stating that you've logged in, and shouldn't be checked again for your identity.

I think Harald, at least, is talking about Firesheep. http://codebutler.github.com/firesheep/

optima says:

Yes. The criterion for a symmetric monoidal functor I to be equivalent to the datum of such a fully dualized object is that it should have unramified diagonal (this is somewhere in Laumon Moret Bailly, I don't have it here). If I is from an n-cobordism to the n-category of n-families over a fixed symmetric n-category, the diagonal is a monomorphism, and a monomorphism is certainly unramified.

Has the mystery been solved yet? What's up with these weird posts?

Has the mystery been solved yet? What's up with these weird posts?

It's just some troll mimicking one of Angelo's answers. Is there something more mysterious about it?

The current user optima (who I assume is the same as the former user?) requests (in the comments to this answer) that this answer be deleted.

I think it would be great if a moderator could delete both answers, as one is being heavily downvoted due to apparent trolling and the other only exists because the first lowered the user's reputation below the commenting threshold. Thanks.