57
Essences of Interaction
Epistemology of Systems Engineering Integration
The epistemology of systems engineering integration intends to provide a
basis for the measurement of a system as a system (through its properties,
traits, and attributes) (described by David Cropley as “characteristics, features, or properties” (Cropley 1998)). But that is rarely the case. Founded on
the somewhat roguish testing regime followed for building products or services, we test what is convenient and expedient, not what is necessary and
sufficient. In short, because it can be tested as part of development, we conspire to test and then presume that the test in some fashion reflects on what
is both beneficial and pertinent for integration. This portrayal of testing
should not be taken as an incisive rail against current practices in systems
engineering. Indeed, it is not. The very nature of systems engineering is iterative; each pass at the work designed to improve on the previous incarnation,
while systems engineering integration is recursive by design. It is expected,
planned, and appropriate to test. However, testing to suggest improvements
in work is not adequate to provide sufficient confidence in an artifact that it
is ready for integration. Were the artifacts designed according to an idealized set of perfect requirements, and specified without fault, the integration
work would then proceed with difficulties mired in different interpretations
of what was done, wrapped in a medley of social behaviors. Epistemology
provides the basis for certainty (Ferris 1997). But within epistemological
thought, the reasons for believing that measurements are accurate and precise are mirrored by the reasons that data are marked by inaccuracies and
imprecision. These issues of error are further complicated by trust or lack
thereof in the data (or in the means of measuring or recording the data).
Metrics
The concern for integration is to develop a set of metrics for determining how
well integration is proceeding. Those metrics could be in terms of the amount
of time taken to achieve some specified level of integration; the amount of
time that is remaining to achieve some specified level of integration; the trust
(or lack of trust) that verification or validation will be satisfactory; the level
(percentage or degree) of completeness (or lack thereof); the utility that can be
ascribed to the integration that is accomplished (or remaining to be accomplished); and the utilization of resources to accomplish the level or the time to
complete integration. In all, numerous metrics can be devised to determine
how successful the integration efforts are compared to a standard or a measure in absentia. The concern for integration is to know what the leading
(or lagging) indicators are to better gauge how well the work is progressing