Abstract: In an object oriented computer system, a framework mechanism defines an infrastructure for allowing a user to flexibly define keys and their corresponding attributes. A special type of key called a specification key allows combinations of attributes to be specified, as well as set and range capabilities. An attribute retrieval policy is defined by a programmer to specify an algorithm or criteria that is used to calculating the desired attribute value. An attribute key controller contains key/value pairings for the attribute, contains the attribute retrieval policy, provides a maintenance interface for changing the attribute values; and provides a client interface for retrieving the attribute according to the attribute retrieval policy. The framework thus allows a user great flexibility in defining a suitable key/attribute retrieval mechanism by defining a suitable policy, defining one or more specification keys, and defining an associated controller.

Abstract: In an object oriented computer system, a framework mechanism defines a mechanism that separates the ability to complete a process from the results of that process. A task is defined in the framework in terms of a sequence of processes. The framework includes a coupling mechanism for flexibly coupling any of these processes together in any suitable order to define the run-time behavior of the framework without changing the core functions of the framework. The coupling mechanism includes the definition of a processable interface and a processing interface. A particular process generally takes detail information as input and produces detail information that is used in a subsequent process. The detail information input into a process is said to be “processable”, and therefore implements the “processable” interface. The detail information output from a process is said to be “processing” detail that is output to the next process, and therefore implements the processing interface.

Abstract: In an object oriented computer system, an object oriented framework defines an interface mechanism that provides a common interface to a non-object oriented persistent datastore. This common interface can be customized by various types within the framework by effectively hiding attributes on the common interface that are not needed by a particular type. Attributes defined on the common interface may be accessed at any level in the framework through a process of delegation from one level to the next. In addition, one or more of these attributes on the common interface may be defined at various levels, which causes those attributes to be accessed at their respective levels instead of delegating to different levels.