So, in my first paper based on a blog post (well, loosely adapted), Catharina Stroppel and I have finally posted our paper on 2-block Springer fibers. It was longer in coming than I had expected, but I’m certainly glad that it’s up there now.
What this paper does is describe the structure of 2-block Springer fibers (stuff which is basically already known, but not very well collected), and then use this to describe a convolution algebra structure on the cohomology of intersections of components, which is self-consciously modelled on the arc algebra of Khovanov. However, the first big complication is that it seems to not actually be isomorphic. There’s an obvious vector space isomorphism, but one just can’t get the multiplication right. This is one of these points where having a coauthor helps. After, you’re unlikely to both go crazy and miss an obvious isomorphism.
The other thing we do is consider the connection of this convolution algebra to coherent sheaves on the resolution of the corresponding Slodowy slice (and thus to Cautis & Kamnitzer’s geometric knot homology). Unfortunately, this didn’t come outas cleanly as we hoped. We do have a vector space isomorphism of some interesting algebras, which we don’t know can be made into a ring isomorphism.
That is, the Ext-algebra of certain coherent sheaves associated to crossing-less matchings by C&K’s picture is isomorphic as a vector space to our convolution algebra (and thus also to Khovanov’s algebra) but we aren’t sure which multiplication is right. Hopefully one of you guys can figure it out.