173
Foundations in Systems Integration
integration is complete. Even if all the issues, rework, and problems were both
tractable and recognized so they could be included in the early integration
planning, not all of the daily changes can be accomplished within the period
allocated for the specific integration tasks. The allocation of project resources
and the requisite skills of the individuals working the tasks constrain the
completion of work that is anticipated but as yet unplanned. Further, all interfaces and all data exchanges for all objects need to be defined for the entire
system before the assumption of piecewise continuous, successful integration
of parts is an effective strategy, deterministically. The issues with piecewise
continuous in conjunction with integration are twofold: first, the individual
tasks associated with integration of an object-to-object strategy are nondeterministic, and second, the number and duration of iterations to complete the
integration is also nondeterministic. Piecewise continuous integration (objectto-object) is inefficient, portending unexpected delays in integration and perhaps unplanned, additional expenditures.
A more effective strategy for integration is to represent the totality of the
system’s objects, identifying the expected (1) system-level functionalities,
performances, losses to achieve those performances, and boundary and
boundary conditions; (2) physical entities and their mechanisms, EMMI,
boundaries, and boundary conditions; and (3) the expected behaviors from
users of the new system as well as their behaviors due to their anticipation of
tasks for the new system, their boundaries, and boundary conditions. In
essence, it is a simulation model of what the system will do and how the
system will operate when completed. A simulation model (Hoover and Perry
1989) uses the variables that comprise relations between the system functions in logic that addresses the impacts of context and environment through
system behaviors. System properties and attributes are discussed in terms of
objects and EMMI. The importance of using a simulation model to facilitate
planning for integration is to predict how each object will interoperate with
the system (as a whole). Joining each object through EMMI with the systemlevel perspective reveals the service each object provides to support system
functionalities. The paradigm of integration is not achieved by an object-toobject but rather an object-to-system model. Change the model of the system
and simultaneously change the actions of all objects. To integrate is to unite
an object with the system model—the result of revealing, identifying, specifying, describing, and detailing the functions enabled by the system’s interface with that of the object. Therefore, integration is not object to object, not
interface to interface, and not data for data.
Unlike systems engineering that is intensely iterative, integration of
human-built systems is system focused to achieve end-to-end performances. Integration is neither systems engineering nor profoundly repetitious. Repetition in integration is expensive, time consuming, and tactically
inefficient. Moreover, iterative integration is strategically ineffective.
Human-built systems integration achieves an architecture that provides
services to objects with EMMI enabling the mechanisms of objects.
Foundations in Systems Integration
integration is complete. Even if all the issues, rework, and problems were both
tractable and recognized so they could be included in the early integration
planning, not all of the daily changes can be accomplished within the period
allocated for the specific integration tasks. The allocation of project resources
and the requisite skills of the individuals working the tasks constrain the
completion of work that is anticipated but as yet unplanned. Further, all interfaces and all data exchanges for all objects need to be defined for the entire
system before the assumption of piecewise continuous, successful integration
of parts is an effective strategy, deterministically. The issues with piecewise
continuous in conjunction with integration are twofold: first, the individual
tasks associated with integration of an object-to-object strategy are nondeterministic, and second, the number and duration of iterations to complete the
integration is also nondeterministic. Piecewise continuous integration (objectto-object) is inefficient, portending unexpected delays in integration and perhaps unplanned, additional expenditures.
A more effective strategy for integration is to represent the totality of the
system’s objects, identifying the expected (1) system-level functionalities,
performances, losses to achieve those performances, and boundary and
boundary conditions; (2) physical entities and their mechanisms, EMMI,
boundaries, and boundary conditions; and (3) the expected behaviors from
users of the new system as well as their behaviors due to their anticipation of
tasks for the new system, their boundaries, and boundary conditions. In
essence, it is a simulation model of what the system will do and how the
system will operate when completed. A simulation model (Hoover and Perry
1989) uses the variables that comprise relations between the system functions in logic that addresses the impacts of context and environment through
system behaviors. System properties and attributes are discussed in terms of
objects and EMMI. The importance of using a simulation model to facilitate
planning for integration is to predict how each object will interoperate with
the system (as a whole). Joining each object through EMMI with the systemlevel perspective reveals the service each object provides to support system
functionalities. The paradigm of integration is not achieved by an object-toobject but rather an object-to-system model. Change the model of the system
and simultaneously change the actions of all objects. To integrate is to unite
an object with the system model—the result of revealing, identifying, specifying, describing, and detailing the functions enabled by the system’s interface with that of the object. Therefore, integration is not object to object, not
interface to interface, and not data for data.
Unlike systems engineering that is intensely iterative, integration of
human-built systems is system focused to achieve end-to-end performances. Integration is neither systems engineering nor profoundly repetitious. Repetition in integration is expensive, time consuming, and tactically
inefficient. Moreover, iterative integration is strategically ineffective.
Human-built systems integration achieves an architecture that provides
services to objects with EMMI enabling the mechanisms of objects.