Title:

Computer-Aided Refinement of Data Structures on Higher-Order Algebraic Specifications

Author & Affiliation:

Abstract. The paper studies the transformational refinement of data structures in the framework of higher-order algebraic specifications. We present novel procedures that mechanize the refinement of entire data structures within a single complex transformation step. The transformations validate a general refinement relation that captures different types of simulations. A uniform transformation rule describes algebraic implementations based on abstraction and representation functions. It is systematically specialized to transformations covering particular changes between data structures. All transformation procedures have been imple- mented in the LÄubeck Transformation System. The system uses analysis algorithms to establish the soundness conditions of the transformations by syntactic criteria. We report practical experiences when outlining the data structure refinement with the system. The paper summarizes results from the second author's PhD thesis .

Software quality management, which leads to improve the performance capability of project, has been studied on the many viewpoints. It is hard work to compare the each process level assessed with CMM, SPICE and Six-Sigma, which are the popular techniques for the industry of software and system management. This paper discusses the relationship between the three process assessments and the derivation of overall mapping table which relates the result levels assessed with SPICE, CMM and Six-Sigma. Additionally, we analyze the mapping-effectiveness, which is the most effectual way after mapping of the levels or before, using the Analytic Hierarchy Process method.

Deterministic end-to-end guanrantees for real-time applications in a DifferServ-MPLS domain

Measuring and Communicating Component Reliability

Much of the research on component-based software engineering assumes that each component has a single service. This simplifies analyses but it is a significant departure from actual components. This paper reports on an investigation of the feasibility of using design constructs as a means of treating several methods as a single unit. We introduce the CoRe Method for measuring and communicating component reliability based on grouping services according to the various roles a component can fulfill. Grouping the services of a component into a few sets satisfies the goal of simplicity while still providing the designer with a more realistic and useful model of component reliability. The CoRe method is described and results of a preliminary experiment in its use are reported.

Toward Component-based System: Using Static Metrics and Relationships in Object-oriented System

Object-oriented development had not provided extensive reuse and computing infrastructures are evolving from mainframe to distributed environments, where object technology has not led to massive development of distributed systems. However, component-based technology is considered to be more suited for distributed system development due to its granularity and reusability. In this paper, we present a process including the metrics and guidelines that can be applied to reengineer object-oriented systems into component-based systems. Components are created based upon generalization and composition relationships between classes. Resulting components are refined by using the metrics we propose.

Communication Style Driven Connector Configurations

Haeng-Kon Kim (Catholic University of Daegu), Hae-Sool Yang (HoSeo Univ.) and Roger Y Lee (Central Michigan University)

Abstract. Various advantages of XML such as flexibility and interoperability have given rise to a steady increase in the number of software documents using XML, and this growth has in turn necessitated new methods for systematically managing massive XML-based documents. Unlike general documents of planar structure based on lines, documents in XML internally constitute a tree structure. Therefore, traditional version control techniques that recognize documents of planar structure are not suitable in handling hierarchically structured documents. This paper proposes a new way of managing changes made in structured documents. While being a timestamp-based approach, the proposed method has the characteristics of maintaining version stamps for edges, rather than nodes in the tree structure, and only assigning version stamps when they are required.

Various types of COTS (Commercial-Off-The-Shelf) software are produced radically and its application areas are also being expanded. Because software quality evaluation depends on types of software and environments where software is operated, certification organizations need different certification programs that are foundations or basic frames for certifying software. We can certify new software products by making new certification programs or referring to previous ones typically. However, it is redundant to make new programs whenever certifying products and merely referring to previous ones may also cause problems because the previously certified software may not have the same quality characteristics as the new software has. For these reasons, we need to systematically derive certification programs for assessing new software, while considering types of software and software environments. Therefore, in this paper, we propose a meta model for systematically generating certification programs from previous ones. By using this model, certification programs are incrementally constructed based on hierarchical classification of software packages. This meta model is validated by generating certification programs with quality data on some certified software products.

Grochtmann and Grimmhave developed the classification-tree method (CTM) to facilitate software testers in generating test cases from functional specifications. While the method is very useful, it is hindered by the lack of a systematic tree construction algorithm. This problem has been alleviated by Chen et al.via their "integrated" classification-tree methodology (ICTM). In this paper, we describe and discuss a prototype system ADDICT that is built on ICTM.

Normalizing Class Hierarchies based on the Formal Concept Analysis

Class hierarchies often constitute the backbone of object oriented software systems. Building "good" class hierarchies is a very important and common task, but such hierarchies are not so easy to build and evolve. Therefore, their construction and evolution are very important issues in component-based and object-oriented software engineering. In this paper, we present a normalized form of class hierarchy based on the concept lattice of formal concept analysis. Our approach provides the theoretical bases for the creation and evolution of well defined object-oriented class hierarchy structures.

QoS-guaranteed DiffServ-aware-MPLS Traffic Engineering with Controlled Bandwidth Borrowing

In this paper we propose an integrated traffic engineering mechanism based on the DiffServ-aware-MPLS for Next Generation Internet that provides guaranteed bandwidth with controlled bandwidth borrowing among LSPs. In the proposed scheme, each DiffServ class-type is managed with per-class-type queuing controlled by committed data rate (CDR), peak data rate (PDR) and related burst sizes. A hierarchical DiffServ packet scheduler controls the aggregated packet flow of multiple DiffServ class types, where a priority-based scheduler is used for overall traffic, and a weight-based scheduler is used for the AF traffic flows. Another packet scheduler for MPLS LSPs is also organized with hierarchical configuration of weight-based scheduling and priority-based scheduling. We also propose a controlled bandwidth borrowing mechanism with extended MPLS signaling for maximized resource utilization. From the simulation results, we could see the usefulness of the proposed scheme to guarantee QoS provisioning and maximize resource utilization.

Architecture based Software Reengineering Approach for Transforming from Legacy System to Component Based System through Applying Design Patterns

Many reengineering approaches have focused on extracting an abstract representation through syntax analysis of legacy source codes. So, recovery of rationale behind the design decision, such as domain specific semantics and roles, has been ignored. In this paper, we suggest the architecture based reengineering approach using design patterns. A design pattern, as core element of software architecture, has integrated the concept of standardization about certain domain and expert experience into a set of related components that can perform specific functionality with better structure. We describe the reengineering process that defines a architecture of target system by refining architecture information of legacy system extracted through domain analysis, identifies the reengineering patterns that are applicable in that architecture, and completes a target system by mapping the identified reengineering pattern into target architecture. Also, we construct the Servlet2EJB prototyping system transforming servlet programs into EJB components as a case study for realizing our reengineering process.

Increasing the Efficiency of Cooperation among Agents by Sharing Actions

In this paper, we propose a new cooperative model for multi-agent systems which improves the performance of the systems by reducing redundant actions among the agents. Moreover, the results of disaster simulations confirm that our model is efficient for such complicated environments.

Applying Model Checking Techniques to Game Solving

In this paper, we automatically solve Push-Push game with model checking techniques which exhaustively explores all search space. Although model checking finds out the best solution for the game, it always suffers from the state explosion problem. To overcome this well-known problem, we use clever methods such as abstraction and pruning. In addition, we propose the relay model checking which decomposes the whole problem into a series of smaller one and then conquers one by one at a time. As a result, we solved all 50 levels of the game. Our solution shows better performance than human experts; that is, we solved the game with on average 22 fewer steps than human experts.

A Fuzzy Logic-based Location Management Method for Mobile Networks

State-of-the-art wireless communication networks allow dynamic relocation of mobile terminals. A location management mechanism is required to keep track of a mobile terminal for delivering incoming calls. In this paper, we propose a fuzzy logic-based location management method to reduce paging cost. In the proposed method, the location update uses an area-based method that uses direction-based method together with movement-based method, and the location search uses a fuzzy logic-based selective paging method based on the mobility information of mobile terminals. A partial candidate paging area is selected by fuzzy control rules, and then the fuzzy logic-based selective paging method pages only cells within the partial candidate paging area. The performance of the proposed fuzzy logic-based location management method is evaluated by an analytical model and by a simulation, and is compared with that of basic velocity paging (BVP) method.

A Next Generation Intelligent Mobile Commerce System

The proposed system has been designed and implemented to demonstrate its effectiveness through experiments.

Improving Software Engineering Practice with HCI Aspects

Reliability Assurance in Development Process for TOE on the Common Criteria