Collaborative editing systems support a group of collaborators to edit or view the same graphics at the same time from geographically dispersed sites connected by communication networks. The research on this is of great challenge. The multi-versioning technique in collaborative graphics editing system is discussed. This paper puts emphasis on object identification and its strategies, some existing problems. A general object identification solution for these problems is proposed, which solves their object identification and system maintenance between operations with causality relationships, and with identical relationships, that the existing strategies cannot resolve. On the other hand, the size of object identification will become larger and larger with the progress of collaborative editing works and the generation of conflicting operations, thus affect the efficiency of the system. Therefore, some rules suitable to the compression of the object identification is analyzed, an object compression method is presented and its validity is also analyzed. Analysis of some examples and experimental results show that all strategies and algorithms presented in this paper improve the efficiency of the system.