Reverse Engineering Research Should Target Cooperative Information System Requirements

Abstract

One of the premises of this panel is that the Information Systems (ISs) of tomorrow will be component-based and distributed. Until recently, ISs were designed to support single functions (e.g., purchase orders or payroll). Brodie argues convincingly that future ISs are to support business processes that involve many cooperating functions [Brod98]. To build Cooperative Information Systems (CISs), existing ISs will have to be wrapped and adapted to support cooperation and interoperability. SoRware engineers expect to leverage distributed object technology and other middleware as well as reverse engineering and reengineering technology to achieve this migration to CISs effectively. However, many obstacles and challenges lie ahead.Firstly, there is no clear vision (let alone a standard) on what infrastructure to build future CISs. There is intense competition among software companies such as OMG (CORBA), IBM (NCF), Microsoft (Active), and Sun (JavaBeans) to roll out the dominant infrastructure for network-centric applications. Consequently, we seem to be years away Tom effective and stable infrastructure technology for CISs and yet we are trying to build CISs today. Instead of concentrating on the interoperability among information systems, we are forced to build infrastructure interoperability bridges.Secondly, the reverse engineering community is concentrating on the automatic extraction of myriad software artifact horn source code instead of concentrating on the harder problem of identifying business rules. Thirdly, the software reengineering camps are busy migrating imperative code to object-oriented platforms (e.g., C++ or Java) instead of migrating to cooperative agents.