Location dependent message behaviour
Original Publication Date: 2001-Sep-16
Included in the Prior Art Database: 2003-Jun-20
Disclosed is a mechanism by which distributed objects may exhibit markedly different behaviours in different processing environments/locations. The different processing environments may provide or imply different constraints (processing capacity, storage capacity, security). Objects can be distributed across communications channels. The messages are represented at either end as object code. The class of the object is transmitted as part of the distribution protocol so that the correct class (hence behaviour) can be reconstituted at the far end. This invention is the use of different implementations of the same (named) class at each end of the communicatons channel. The behaviour approriate to the sender is implemented in the class on the sending side, the behaviour apropriate to the receiver is implemented in the class on the receiving side. The 'transmutation' of the class happens as a result of the the fact that only the data and class name is serialised, transmitted, and associated with different processing code. For example: On server side, an object could stored in a database (eg IBM's DB2*) and it would be very useful for the message to implment some behaviour to facilitate this storage. On the client side, for reasons of security or resources, this code wuld be undesirable. The server-side class could implement the neccesary code without polluting the client side class. IBM and DB2 are registered trademarks of International Business Machines Corporation.