270
Engineering Systems Integration
together, but rather is a separate evaluation after the completion of the work.
Integrating objects merely brings their boundaries into some sort(s) of conjunctive relation(s) through the exchange of EMMI. However, the mere conjunctive relation(s) of individual parts, units, components, subassemblies,
assemblies, or subsystems does not satisfy the requirements for a system
unless all the various kinds of objects are operative. The goal of a project
may be to build a system, but a system is not a system until it satisfies the
requirements for a system. Integration is the means for building a system, as
it is the means to put anything together. The process of integration is necessary to build a system, but it is not sufficient.
It is a learned response, whether by a previous stressful experience
(Kinnaird 2003), by warnings, or by procedures that systems engineers have
turned into the habit of testing what is determined that needs to be tested.
Sometimes items are included in that determination that has no more than a
reflexive stimulus (Gabora 2001) to evoke testing. The impetus for testing an
item should be based on something other than fear or instinct. As the practice of systems engineering has evolved and taken its lessons from engineering, the basis for a test is problematic. Certainly, we are unwilling to test
every part (e.g., every resistor, every wire, every washer, and every transistor
under every condition that is likely to occur in conjunction with a finished
product or service). Instead, systems engineering and projects have built-in
formalisms to provide a level of assurance that certain standards (ISO 9000
series) for the improvement of organizational performance are followed for
production of individual parts.*
Testing is a process to determine the difference(s) between an object’s properties, traits, and attributes under certain conditions in a given set of circumstances with that of a representation (or test model) of what is desired. The
representation includes the test setup; the test procedures; the test plan; the test
personnel; the test objectives; the data analysis (and tools); the theory in which
the measurements are planned, executed, and interpreted; and the biases (of all
parties). It is tacitly assumed that all factors not included in the representation
(or test model) are factually extraneous (and therefore not significant either to
a specific test or to a concatenation or totality of tests). Great care must be exercised in planning and executing the testing of an object. Testing is a means of
comparison: comparing an object to the test model. Tests do not prove anything; they only show a correspondence to an expected result (Aerts 1983).
What to test is extremely important for any project. Testing impacts on the
schedule, use of resources, budget, and final performance(s) of the product or
service. The high costs and impacts of remediating defects discovered after
the product or service is in operation cause great consternation when deciding what to test (Boumen et al. 2006). From an integration perspective, the
* International Organization for Standardization, About ISO introduction http://www.iso.ch/
iso/en/aboutiso/introduction/index.html, website modified February 16, 2004, accessed
April 18, 2004.