We have devised novel methods to evaluate the structural similarity of proteins and we compare them. In each method, a hash vector is associated with each fixed length fragment of three dimensional protein structure. Then, we analyze the similarity between fragments by evaluating the difference between true hash vectors. The novel aspect of the methods is that the following property is proved theoretically: the root mean square deviation between two fragments is small, so the distance between the hash vectors associated with the fragments is small. The methods were compared with the previous methods using PDB data, and were shown to be much faster. One of the new hashing methods is already included in PROTEIX, a database management system for protein structures. The features of PROTEIX are described.