On the local structure of stacks - MathOverflow

On the local structure of stacks

2013-02-03T20:09:18Z

1) Is it true that any Deligne-Mumford stack is locally a quotient stack $[X/G]$ with a finite group $G$?

2) Is it true that any Deligne-Mumford stack can be globally presented as a quotient stack $[X/G]$ with a non necessarily finite group $G$? For example, Geigle and Lenzing give such a presentation for stacky projective lines here.

3) What about Artin stacks?

Answer by Jacob Bell for On the local structure of stacks

2013-02-03T20:37:14Z

the one I know without having to look in the literature is 1)

Lemma 2.2.3 of http://arxiv.org/pdf/math/9908167v2.pdf

I think 2) is true as well (maybe you need to add the adjective tameness appropriately?) and for 3) there should be a result of Kresch saying that your stack can be stratified by quotient stacks. But I'd have to look this stuff up.

EDIT:

for 1) I should say etale topology.

for 2) I was thinking about this result (Theorem 4.4 and Proposition 5.1) by Kresch http://www.math.uzh.ch/fileadmin/user/kresch/publikation/geodm.pdf

for 3), the result I was misremembering was Proposition 3.5 of http://arxiv.org/pdf/1002.4372.pdf, and the first paragraph of the proof. (it's for stacks with affine stabilisers)

SECOND EDIT:

There is a paper by Edidin-Hassett-Kresch-Vistoli where the investigate when an Artin stack is a quotient stack. It turns out that this is closely related to the pushforward of the structure sheaf of a smooth atlas to admit a surjection from a vector bundle. Here is the review by Vezzosi. http://www.ams.org/mathscinet/search/publdoc.html?pg1=IID&s1=611835&vfpref=html&r=34&mx-pid=1844577

Answer by David Carchedi for On the local structure of stacks

2013-02-03T22:08:36Z

I am not an expert in the algebraic category, however, I know that 2) is an open problem in the differentiable category; it is not known if every smooth orbifold is a global quotient stack. It is true in the differentiable category when you ask for the orbifold to be effective ("reduced" in algebrogeometric lingo). I have good (topos-theoretic) reasons to suspect that if $\mathscr{X}$ is a Deligne-Mumford stack such that for each scheme $T,$ the subgroupoid of $\mathscr{X}\left(T\right)$ on those maps $T \to \mathscr{X}$ which are etale is equivalent to a set, then $\mathscr{X}$ is a global quotient. (This condition is basically means that the isotropy groups act faithfully). I would be very surprised if 3) were true.