How many papers with awful notation are actually the reverse engineering of some (barely) working code, cobbled together from random libraries and coefficients?