Skip to main content
16 events
when toggle format what by license comment
Feb 28, 2012 at 14:43 comment added Thomas Richard This paper of D. Ramos arxiv.org/abs/1109.5554 gives a way to smooth out the cone points without messing with gluing. It constructs a smooth solution to Ricci flow for surfaces with conical singularities.
Jan 29, 2012 at 3:16 history edited Vitali Kapovitch CC BY-SA 3.0
deleted 67 characters in body
Jan 29, 2012 at 3:14 comment added Vitali Kapovitch @Anton No I don't have a short proof. I admit it was wrong to call what I describe an "idea of the proof" as the hard part of the proof is missing but I haven't looked at Alexandrov's book in a long time and I could only remember the general structure of the argument when I wrote my answer.
Jan 29, 2012 at 2:07 comment added Anton Petrunin BTW, Once you are trying to make the "idea of the proof" work, you get into tedious details. $$ $$ That means: If you have a complete clean and short proof then I am very interested.
Jan 28, 2012 at 23:05 comment added Vitali Kapovitch @Igor: Alexandrov doesn't mention smoothing but that is a minor point. He proves that $X$ can be approximated by polyhedral surfaces with the same lower curvature bound. By construction the approximating surfaces are smooth and have constant curvature outside of the cone points. Cone points are isolated and it's trivial to smooth such metrics with the same lower curvature bounds the way you smooth a flat cone by a surface of revolution. The fact that he considers lower curvature bound 0 is a minor point too as all arguments are local. He discusses the general case in chapter 12.
Jan 28, 2012 at 19:43 comment added Igor Belegradek @Vitali, I see nothing about smoothing in Lemmas 1-3. Also he only considers the case of positive curvature, but this must be a minor point. Smoothing at vertices is not a minor point. It may be fixed by the realization theorem in Chapter XII, section 2 which says every point in a space of curvature $\ge k$ has a neigborhood isometric to a convex surface is the 3-dimensional model space of constant curvature $k$. Of course, any convex surfaces can be approximated by smooth ones, giving the smoothing, and hopefully these smoothings can be glued.
Jan 28, 2012 at 18:44 comment added Vitali Kapovitch @Igor sorry,it's in Chapter 7. In case the chapter number differs in your edition the chapter is called "Existence of a closed convex surface with a given metric".
Jan 28, 2012 at 18:41 history edited Vitali Kapovitch CC BY-SA 3.0
added 11 characters in body
Jan 28, 2012 at 17:54 comment added Igor Belegradek @Vitali, on your edit: which chapter is this in? I have the new edition of Alexandrov's "Intrinsic Geometry of Convex Surfaces" published in his collected works and almost every chapter has section 6.
Jan 28, 2012 at 15:53 history edited Vitali Kapovitch CC BY-SA 3.0
corrected references
Jan 28, 2012 at 14:15 comment added Igor Belegradek After looking at Alexandrov-Zalgaller's argument, I must admit I would not be comfortable quoting it as an answer to the OP's question. Given that the question is of considerable interest it is unfortunate that there seem to be no proof in the literature.
Jan 28, 2012 at 13:32 comment added Thomas Richard Thanks, I didn't realized that Alexandrov already proved the result about the topology of 2-dimensional Alexandrov surfaces. Maybe my mistake come from "Geometry IV" of Reshetnyak, which defines surfaces of bounded integral curvature assuming at the beginning that the underlying space is a topological surface. By the way, Alexandrov-Zalgaller is on my desk these days, I'll have a look at it. Thanks again.
Jan 28, 2012 at 13:27 vote accept Thomas Richard
Jan 28, 2012 at 0:22 comment added Vitali Kapovitch @Igor sorry, I don't have the book and I haven't looked at in in quite a while but my recollection is that I read it there.
Jan 28, 2012 at 0:06 comment added Igor Belegradek Where exactly is this argument in Alexandrov-Zalgaller's book?
Jan 27, 2012 at 23:31 history answered Vitali Kapovitch CC BY-SA 3.0