Salurso, Maria Anna (1998) Extraction of objects from legacy systems: an example using cobol legacy systems. Masters thesis, Durham University.
In the last few years the interest in legacy information system has increased because of the escalating resources spent on their maintenance. On the other hand, the importance of extracting knowledge from business rules is becoming a crucial issue for modern business: sometime, because of inappropriate documentation, this knowledge is essentially only stored in the code. A way to improve their use and maintainability in the present environment is to migrate them into a new hardware / software platform reusing as much of their experience as possible during this process. This migration process promotes the population of a repository of reusable software components for their reuse in the development of a new system in that application domain or in the later maintenance processes. The actual trend in the migration of a legacy information system, is to exploit the potentialities of object oriented technology as a natural extension of earlier structured programming techniques. This is done by decomposing the program into several agent-like modules communicating via message passing, and providing to this system some object oriented key features. The key step is the "object isolation", i.e. the isolation of .groups of routines and related data items : to candidates in order to implement an abstraction in the application domain. The main idea of the object isolation method presented here is to extract information from the data flow, to cluster all the procedures on the base of their data accesses. It will examine "how" a procedure accesses the data in order to distinguish several types of accesses and to permit a better understanding of the functionality of the candidate objects. These candidate modules support the population of a repository of reusable software components that might be used as a basis of the process of evolution leading to a new object oriented system reusing the extracted objects.
|Item Type:||Thesis (Masters)|
|Award:||Master of Science|
|Copyright:||Copyright of this thesis is held by the author|
|Deposited On:||13 Sep 2012 15:51|