Dmitri’s point in #11 (repeated in #13) isn’t even to do with homotopy, it’s a rather basic observation about sites and sheaves.

Dmitri brought up a possible subtlety in the definition of Stein manifolds only as an attempt to find a technical loophole clause that would explain why Lárusson doesn’t consider what seems to be the immediate conclusion.

]]>Oh, sorry, I was kinda ignoring the homotopy sheaf material :-)

]]>if the dimensions of the pieces are unbounded, then the disjoint union isn’t Stein

That’s not the example I had in mind, though.

Disjoint unions of points (finite or countable) are 0-dimensional Stein manifolds.

Finitely supported complex-valued functions on such manifolds form a presheaf that satisfies finite homotopy descent, but does not satisfy descent with respect to covers of arbitrary cardinality. This would seem to constitute a simple counterexample to the question stated in Section 5 of his paper, but probably I am just misreading something.

]]>I should add that one doesn’t need all the dimensions of the disjoint pieces to be equal, but you *do* need a bound on the dimensions. I spoke with Finnur today and he seemed to indicate that if the dimensions of the pieces are unbounded, then the disjoint union isn’t Stein. Maybe this is it.

Re #12: Basically, I am wondering why the example in the last two paragraphs of #11 is not a (trivial) counterexample to what Lárusson is suggesting in his paper https://arxiv.org/abs/math/0101103v3, at the beginning of Section 5, where he asks “It is natural to ask whether a finite homotopy sheaf on S satisfies descent.”

]]>I don’t see why not. Use a characterisation of n-dimensional Stein manifolds that involves proper embeddings in C^m for some m. In fact there is a uniform bound on the needed m in terms of n, so for countable disjoint unions there is a common m. Then properly embed embed the countable copies of C^m in C^m+1 as parallel affine hypersurfaces.

Or else I can ask Finnur today, as his office is next to mine, if that’s not convincing enough. Why do you ask?

]]>Just a sanity check: are disjoint unions of Stein manifolds again Stein manifolds? Including countably infinite disjoint unions? The answer seems to be yes, simply by staring at the definition, but perhaps I am missing something?

The reason I am asking about this is that Lárusson in his paper https://arxiv.org/abs/math/0101103v3, at the beginning of Section 5 asks “It is natural to ask whether a finite homotopy sheaf on S satisfies descent.”

Here S is the site of Stein manifolds and holomorphic maps, and a finite homotopy sheaf is a presheaf that satisfies the homotopy descent condition with respect to finite covers.

It would seem to me that the answer to Lárusson’s question should be negative as stated, e.g., because we can take the presheaf P that assigns to a Stein manifold M the abelian group of holomorphic functions on M that vanish on all but finitely many connected components of M.

Then P satisfies descent with respect to finite covers, i.e., is a finite homotopy sheaf in Lárusson’s terminology. However, it does not satisfy descent with respect to (say) countable disjoint covers.

]]>adjusted the wording of the example of open Riemann surfaces (here) and added pointer to the classical reference:

- Heinrich Behnke, Karl Stein,
*Entwicklung analytischer Funktionen auf Riemannschen Flächen*, Mathematische Annalen volume 120, pages 430–461 (1947) (doi:10.1007/BF01447838)

adjusted the wording of the example of open Riemann surfaces (here) and added pointer to the classical reference:

- Heinrich Behnke, Karl Stein,
*Entwicklung analytischer Funktionen auf Riemannschen Flächen*, Mathematische Annalen volume 120, pages 430–461 (1947) (doi:10.1007/BF01447838)

For what it’s worth, I have found and uploaded an old copy of the file (here)

]]>Zachary Maddock is on Linkedin so you might be able to contact him and put a copy of the document on the Lab if it seems worth it.

]]>somebody just alerted me:

this page here has been and still is referring to

- Zachary Maddock,
*Dolbeault cohomology*(pdf)

for proof of some of its statements (existence of “good” Stein covers). However, the link to that pdf

```
http://www.math.columbia.edu/~maddockz/notes/dolbeault.pdf
```

no longer works, and Google seems to see no other trace of it either.

(?)

]]>I am not saying that (about the analytification), but precisely what you say above: affine are cohomologically trivial in the sense as proved in chapter 3 of Hartshorne’s book and this is a usually given statement when algebraic geometers look at analytic spaces. Many other deep properties are also analogous.

]]>Or close at least. I have added something here.

]]>Affine schemes of algebraic geometry.

]]>Added the statement of Cartan’s theorem B and added to the Idea-section a remark that therefore Stein manifolds play the role in complex geometry of Cartesian spaces in smooth manifold theory, for purposes of abelian (Cech)-sheaf cohomology.

]]>have added a tad more content to *Stein manifold* and cross-linked a bit more