Center for Coordination Science @ MIT


CCS Home  |  Working Papers   |  What's New?  |  People  |  Links  | Site Map  |  Comments

NOTE: If an electronic version of a paper is not available here, or you are unable to successfully download the paper that you want in the formats available, please contact MIT Document Services at (617) 253-5668, DOCS@MIT.EDU. Please do not contact the authors for copies of papers.


2001 Working Papers

CCS No. 216, Sloan No. 4159

Defining Specialization for Process Models

George M. Wyner and Jintae Lee

January 2001

Object-oriented analysis and design methods take full advantage of the object specialization hierarchy when it comes to modeling the objects in a system.  When modeling system behavior, however, system analysts continue to rely on traditional tools such as state diagrams and dataflow diagrams.  While such diagrams capture important aspects of the processes they model, they offer limited guidance as to the ways in which a process can be improved.  In this paper we extend the notion of specialization to process representations and identify a set of transformations which, when applied to a process description, always result in specialization.  We analyze specific examples in detail and demonstrate that such a use of specialization is not only theoretically possible, but shows promise as a method for categorizing and analyzing processes.  This paper makes two contributions toward answering this question: first, it articulates a formal definition of process specialization which is compatible with object specialization but allows us to reason specifically in terms of process representations.  Second, it develops the concept of the "specializing transformation" as a means for systematically generating and exploring process alternatives.  We illustrate these results by applying them to two commonly used representations:  the state diagram and the dataflow diagram. We identify a number of apparent inconsistencies between process specialization and the object specialization which is part of the object-oriented approach.  We demonstrate that these apparent inconsistencies are superficial and that the approach we take is compatible with the traditional notion of specialization.