5 Design Methodologies
A review of the existing design
(and design management) process with its roots in analysis and reductionism,
is followed by a revision involving interaction and purposive behavior.
General Solutions: Design in the Mass/Energy Domain
Demand-activated production is a radical shift from classical engineering
design, which was the major paradigm for the first part of the 20th century.
Handling of Uncertainty: Respond to problem with
need to know - pre-analysis and reductionism
prediction - all cases handled in advance
control - rules for handling all cases
Intractable Complexity (story of trivial and non-trivial machines): it is
impossible to handle all cases in advance
Worst Case Design: implementation will require maximum safety for worst
case scenarios
Inefficiencies in implementation (and also in design): resources required
for pre-construction analysis and construction expend large percentages
of total effort for least-likely cases
Inflexibilities: final implementation cannot handle variation outside of
design specification (shallowness); un-foreseen situations not handled well
(brittleness)
Effect of Classical Design
on Present-day Production
The primary uncertainty in the production chain (what will sell and when)
is handled classically by maintaining inventory. This has a series of implications:
Large amounts of cash are tied up and non-liquid.
The investment that inventory represents is made before the moment of proximity,
is inflexible, and adds considerable inertia to the system, which is therefore
less able to respond to changes.
Because product is buffered at all points in the production chain, inefficiencies
in production are simply concealed and hence cannot be identified and removed.
The efficiencies of classical "mass-production" are to be found
in economies of scale and through-put; therefore the concern is for maintaining
production schedules, and product quality becomes a concern when it is too
late, i.e., after the production run has been made (at which point the inventory
has been invested in useless results and the reasons for poor quality are
probably lost).
Customer becomes too much a part of the QA process, which leads to customer
dissatisfaction as well as added expense for the producer in repairing the
problems.
Specific Solutions: Design
in the Information Domain
Demand-activated production is based on the broader approach of cybernetic
engineering design, which arose in the second half of the 20th Century.
For each point there are implications for new production methods (in contrast
to classical design approach of General Solutions):
Handling of Uncertainty: emphasis on dynamic interaction using
error detection - variation from desired goal
error correction - active response to variation
regulation - monitoring for overall stability
Handling Complexity: responding dynamically with system's malleable structure
handling complexity with complexity
relinquishing need to understand non-trivial machines
eliminate hesitation in awaiting thorough understanding of citizen; instead,
respond and interact
Design for coupling to environment: close connection to environment assures
consistent data about current state for interpretation against goals
Concomitant efficiencies in design and implementation: effort expended on
flexibility and dynamic response
Robustness and fault tolerance: system naturally more redundant, and is
made to have a response to un-foreseen situations
The value-added is design. The implications for social productivity,
GNP, and resources required to handle complexity are vast.
