Kinematic design is an important aspect of the design of mechanical sy
stems with moving parts. The performance of such systems may stiffer f
rom the presence of parts that bind during their motion causing jams,
or that execute undesired motions. Kinematic analysis needs to be cond
ucted and the design of the system modified Analysis-redesign-reanalys
is cycles are common in reaching a satisfactory design. If the system
is large and the incidence of redesign frequent, then it becomes imper
ative to have fast analysis methods. This work addresses this problem
by providing computational techniques for doing incremental kinematic
analysis. The mechanism is represented as a graph of objects with conn
ection constraints among them. A recursive procedure traverses the gra
ph to establish instantaneous kinematic properties. A store-and-reuse
strategy is employed to increase the efficiency of re-analysis runs. T
his involves storing intermediate results of computations for reuse in
case of incremental redesign or iterative analysis. The store-and reu
se strategy is also applicable when retrieving elements of previous de
signs, and their associated kinematic analyses, and adapting them to n
ew design problems. A comparative, quantitative study of the normal, '
'from-scratch'' method and the incremental method is provided to prove
the greater efficiency of the latter.