Cylindrical contact homology as a well-defined homology?