This is a 3-year standard award. The need to merge different versions of an object to a common state arises in collaborative computing for several reasons, including optimistic concurrency control, non-WYSIWIS coupling, absence of access control, latecomers, and mobile computing. Current systems either support restricted forms of merging, or require programmers to completely implement merge procedures. The hypothesis of this research is that it is possible to automatically generate a large variety of merge procedures from high level specifications of these procedures. The project will investigate several issues that must be resolved to establish this hypothesis, including simultaneous merging of an arbitrary number of objects, merging of user-defined objects, realtime merging of user-interface objects, merging in a mobile environment, abstractions for specifying merge procedures, and the relationship between merging and other collaboration functions such as concurrency control. It is uses an experimental method wherein the new concepts are implemented and used in validation experiments. If successful, it will allow merging to become a common operation in a collaborative environment without requiring significant application-specific programming.
