Any change to one component of a software system is likely to be felt
in other components, a phenomenon known as 'ripple effect', Impact ana
lysis is the process of identifying all those components of a system w
hich may be affected by ripple effect. This requires a knowledge of th
e dependencies between components. A software system consists not only
of the source code, but also of all life cycle work products includin
g requirements, design and test documents. Any of these components may
be the subject of a change request, and should therefore be included
in impact analysis. The ripple propagation graph (RPG) for impact anal
ysis, developed as part of the AMES (Application Management Environmen
ts and Support) project, models an application in such a way that depe
ndencies can be traced in order to identify all affected components. I
t is an object-relationship model consisting of the components of a sy
stem at some level of granularity, and the various relationships betwe
en them. This paper describes the AMES RPG, and how the results of the
application understanding (AU) toolset, developed (again as part of t
he AMES project) to support the process of gaining an understanding of
an application, can be used to provide some of the data required to p
opulate the RPG. (C) 1998 John Wiley & Sons, Ltd.