OBJECT ACQUAINTANCE SELECTION AND BINDING

Large object-oriented systems have, at least, four characteristics tha t complicate object communication, i.e., the system is distributed and contains large numbers. e.g., thousands, of objects, objects need to be reallo cated at run-time and objects can be replaced by other objec ts in order to adapt to the dynamic changes in the system. Traditional object communication is based on sending a message to a receiver obje ct known to the sender of the message. At linking or instantiation tim e, an object establishes its acquaintances through name/class based bi nding and uses these objects through its lifetime. if this is too rigi d, the software engineer has to implement the binding of objects manua lly using pointers. In our experiments we found the traditional acquai ntance communication semantics too limited and we identified several p roblems, related to the reusability of objects and selection mechanism s, understandability, and expressiveness. We recognised that it is imp ortant to separate a class or object's requirements on its acquaintanc es from the way an object selects and binds its acquaintances in actua l systems. Based on this, we studied the necessary expressiveness for acquaintance handling and identified four relevant aspects: type and d uration of binding, conditions for binding, number of selected objects , and selection region for binding. To implement these aspects, we def ined ''acquaintance layers'' as part of the ''layered object model.'' Acquaintance layers uniformly extend the traditional object-oriented a cquaintance handling semantics and allow for the first-class represent ation of acquaintance selection and binding, thereby increasing tracea bility and reusability. (C) 1998 John Wiley & Sons, Inc.