Volume2, Number3, 2001

 

 

Title:

A Formal Characterization of Electronic Commerce Transactions

Author:

Sylvanus A. Ehikioya

Abstract:

E-commerce provides a vital function in a business's overall information technology architecture and provides functionality that transcends departments, locations, and international boundaries. The development process of such e-commerce systems should be sound because the consequences of any error may severely impact the survival of the company. In addition, fixing an error after a system's deployment is expensive and the error may manifest in many different parts of the system, than if the error was detected and corrected early in the development process. Thus, appropriate rigorous methods for modelling both the static and dynamic aspects of e-commerce systems are necessary. This paper examines e-commerce and some of its requirements and how the development of applications in this domain can gain from using formal methods approach. We formally specify an abstract model of e-commerce and then show how that specification can be used to create a real-life e-commerce application. The specification, done in a mathematical modeling language called Z and combined with verbose informal descriptions, significantly explains some critical parts of the electronic commerce problem domain. Using the formally specified model and modern programming languages and techniques, a working system can be created quickly and efficiently.

 

 

Title:

A Study on Component Reuse System Using Reengineering

Author:

Haeng-Kon Kim, Roger Y. Lee, Tadanori Mizuno, Atsushi Togashi

Abstract:

Software reuse techniques make reapplication of various well-organized information and knowledge to system development so that improve productivity and make it easy to maintain software. This paper describes the design and implementation of C3RUS (C and C++ Component ReUse System) which can extract reuse components using re-engineering, and store, retrieve, and merge them written with C and C++ programming language. The construction of reuse components using re-engineering has advantage in software quality assurance because they are reliable components already tested in real environments.

 

 

Title:

A Task Analysis and Design Framework for Management Systems and Decision Support Systems

Author:

Michael Herczeg

Abstract:

This contribution describes a framework with organizational and operating concepts for the management, i.e. supervision and control of complex dynamic systems (processes), like networks, factories, power plants or transport systems. This kind of system operation has also been called supervisory control [10], [15]. These applications are characterized by a complex allocation of tasks between several users (operators) and machine agents (system functions).

For systems to be designed, the concepts of the framework have to be transformed into applica­tion-specific terms according to the new management application. This step will create an application-ori­ented framework to design new management systems in a way that they are able fulfil the operational and human requirements to meet the business goals. For existing systems to be analysed, the entities have to be transformed into the typical application terms in respect to these management applica­tions. This will provide an application-oriented framework for an analysis to evaluate the system, identify problems, improve and optimize the system.

     This framework has been developed mainly in the area of telecommunication network management and has been used to design telecommunication management systems for digital broadband networks. The concepts are expected to applicable in other application areas as well.

 

 

Title:

Periodic Scheduling of Distributed Tasks with Consistency and Hard Real-Time Constraints

Author:

Soumaya Kamoun, Pascale Minet

Abstract:

A key issue in the design of real-time distributed systems with persistent objects is to overcome both consistency constraints and hard real-time constraints. The problem becomes harder when distributed tasks have a graph structure. In this paper, we first show that basic results in uniprocessor scheduling are no more valid in this context. We then propose an algorithmic solution with its feasibility conditions. Feasibility conditions enable to determine whether a set of distributed tasks is feasible (i.e. all tasks meet their termination deadlines). We show how to establish necessary and sufficient feasibility conditions. Their computation is done by a tool, called feasibility oracle, we have designed. The oracle declares whether a given task set is feasible and if yes, provides the worst case response times. In the reported experiments, we show how to choose the best values of the scheduling parameters.