|
Page 3 of 3
"Ten Basic Design Principles: and some
Implications"
|
8.
The ‘Design does not
impact Function’ principle:
Design, by definition only impacts performance and cost attributes. It has
no impact on the fundamental function of a system. Nor does Function impact
design. But design must fit function!
Implications;
-
Function requirements must not contain ‘design’ (they too often do !)
-
Design must be appropriate to the function of a system; meaning we cannot
design in ignorance of the functions, to reach some abstract performance
levels
- To classify design as
a function requirement is bad practice, since it robs us of freedom to
address the real performance requirements, and it probably obscures the
real performance requirements themselves. Real performance requirements
often don’t get defined because the design, called falsely a function, is
all there is to represent them
|
|
9. The
'Architecture' Principle:
Architecture is the highest level of design
for a given system. It is the framework and control over all other
more-specialized design in the system. But Architecture follows all
principles of design.
Implications:
- architecture must have more authority and
power than the narrower disciplines
- architecture needs to be settled before
serious engineering design can proceed
- architecture is the same discipline as
design, but controls a higher level of the system, and consequently has
different abstractions of information to handle
|
|
10. The
'Design Evolution' Principle:
Design, at all levels needs to evolve in
small steps of confrontation with reality, analysis and re-design. It cannot
adapt to the inevitable continuous stream of information from different
sources, different timings, experience feedback, problem solving insight,
economic change, political change, and technical change in any other
reasonable way.
Design cannot be simply and correctly
completed at once, any more than a child can be mature, or a new business
perfectly designed.
Implications: the design process must be
taught and practiced as an iterative process
- the design process must know how to
collect data about design performance, and how to adjust design itself to
better meet the real needs of stakeholders
- the design specifications must not be
prematurely frozen
- but design specifications must neither be
changed without clear logical and profitable reason
|
Conclusions:
- designers need formal training and
leadership in these principles
- design as practiced today is too often
failing to systematically address the multiple stakeholder needs
- these principles apply more, the larger
and more critical the system at stake
- we can ignore these principles if the
risks we thus incur are more tolerable than the cost of such systematic
engineering
Right now the failure rate of projects is so
uncomfortably high that we need to look at the option of investing more in
the systems engineering intellectual processes.
_____________________________
© Tom@Gilb.com
2005
|
Back: to
System Development Back:
to Page 2 of
3
|