Using the structure of certain Springer fibers and their components, I'll describe a geometric construction of an algebra which is painfully close to being isomorphic to the arc algebra defined by Khovanov, but in fact, isn't. I'll then hopefully explain why this is actually a good thing.