Comments On The Two-Dimensional Landau-Ginzburg Approach To Link Homology