RATIONALE

"Brown, DSPL/AIR-CYL"

D. C. Brown & B. Chandrasekaran, Design Problem Solving: Knowledge
Structures and Control Strategies. Research Notes in Artificial
Intelligence Series, Pitman Publishing, Ltd., London, England.  


DSPL is a language for writing down the knowledge needed to do routine
design by "hierarchically guided refinement".  AIR-CYL is a mechanical
design system built using DSPL.  Decomposition and refinement is done
in this work by selecting and executing plans at each level on the way
down.  

DSPL makes a very strong committment to separately describing each
type of knowledge, such as what's needed to select plans, to identify
known types of constraint failures, to do redesign, or to make a
decision about a single parameter.  The types of knowledge were
detected by studying a human designer.  

The issues that this work raises are concerned with routineness, types
of knowledge, whether the task is parametric or configuration design,
and the role of plans in designing.