RG: A Case-Study For Aspect-Oriented Programming
RG is an image processing system that allows sophisticated image
processing operations to be defined by composing primitive image
processing filters. An implementation of RG using OOP is easy to do
and quite manageable, but its performance is severely lacking. The OOP
approach turns out not to be well-suited to addressing the performance
problems because the performance issues we want to handle do not
respect object or method boundaries and cannot be effectively
addressed in a paradigm tied to those boundaries. Aspect-oriented
programming is an approach designed to handle complexities arising
from such cross-cutting issues. This paper presents a case-study of
how this limitation of OOP was overcome using aspect-oriented
programming techniques, such that the performance problems were
adequately addressed without compromising the original OOP
architecture of the system.
© Copyright 1997 Xerox Corporation. All rights reserved.
Download paper.