(no subject)

Anybody knows how to define spin structure on a triangulated Riemann surface combinatorially? (The 3d analog would also be interesting). Should it be something like picking a 1-simplex in every 2-simplex in some consistent way?