Abstract: In this post-genomics era, protein structural data are increasing so rapidly that we now have an unprecedented chance to study the structural, functional and evolutionary relationships among proteins. As long as we have suitable methods to efficiently analyze the tremendous amount of data, much new knowledge can be obtained to help us understand Nature and apply protein sciences. Extensive discussion of the applications is not the purpose of this talk; instead, by using two specific examples of structural encoding technique, i.e. Ramachandran Sequential Transformation (RST) and Angle-Distance image (A-D image) transformation, several issues will be delivered to the audience. What is the more suitable direction(s), e.g. accuracy, speed, broadness of applicability, or user-friendliness, for the development of new protein structural comparison methods? Are we now using the correct assumptions about the intactness and variability of protein structures for many basic researches? Is a novel phenomenon always rare? Experiences of the speaker are just for the reference of the audience, from whom the sharing and exchange of opinions will be appreciated. Some basic information: RST is an algorithm for encoding 3D protein structural data into 1D sequences. It has been applied to structural similarity searches and the study of circular permutation (CP). A-D image is designed to encode protein structures as 2D images and is now utilized to study the 3D domain swapping (DS) phenomenon. CP is an evolutionary event resulting in the fact that structural homologs may have different locations of termini. It has been applied in many protein engineering fields. DS is defined as two or more proteins exchanging part of their identical domain to form intertwined oligomers. Since it may be responsible for the formation of some deposition diseases like bovine spongiform encephalopathy, related researches may help to find new treatments. Besides, it shall be applicable to the design of auto-assembling biomaterials.