Program

09:00©¤09:10
Welcome and Workshop Information

09:10©¤10:10
Keynote Speech:  Towards Domain-specific Software Component Models
©¤ Dr. Kung-Kiu Lau from University of Manchester, UK

Abstract:
A software component model defines components and their composition mechanisms, and thus provides the basis for a component-based approach to software development. For domain-specific applications, we believe it is best to use a domain-specific component model. Morever, we believe such models can and should be derived from domain models. In this talk, we will outline our ideas along these lines.

10:10©¤10:40

Paper 1

Using Responsibilities for Early Identification of Frozen and Hot Spots Reused in Frameworks Modeling

Author: MATOS Simone N.1+, FERNANDES Clovis T. 2
1. UTFPR and ITA (Department) of Computer Science, University, Ponta Grossa 84016210, Brazil
2. ITA (Department) of Computer Science, University, São Jos¨¦ dos Campos 12228-900, Brazil

Abstract

Most approaches for the development of domain frameworks usually determine frozen and hot spots from class diagrams of the domain concrete examples. As a result, a delay at the understanding of which spots should be kept stable or flexible occurs during the framework modeling. This work presents a responsibility-driven method aiming to identify frozen and hot spots in the analysis phase. Its purpose is to allow the designer to have an earlier domain understanding, as well as give the opportunity of reusing as many software artifacts as possible from the domain concrete examples, thus increasing the consistence at the framework modeling level. The proposed method has a supporting tool and has been illustrated with the elaboration of a framework for Car Racing Games domain.


10
:40©¤11:00
Break

11:00©¤11:30

Paper 2

Software Reuse to Support Earth Science

Author: MARSHALL James J. 1+, DOWNS Robert R. 2, SAMADI Shahin 1, GERARD Neil S. 1, WOLFE Robert E. 3
1. Innovim / NASA Goddard Space Flight Center, Mailstop 614.9, Greenbelt, MD, 20771, USA
2. Columbia University, Center for International Earth Science Information Network (CIESIN), 202 Geosciences, Lamont-Doherty Earth Obs., Palisades, NY, 10964, USA
3. NASA Goddard Space Flight Center, Mailstop 614.5, Greenbelt, MD, 20771, USA

Abstract

Earth science relies on software to prepare, distribute, access, manipulate, and analyze data. Scientific replication often requires reuse of the original software. Similarly, efficiently distributing, accessing, or analyzing scientific data often requires software reuse. Recognizing the value of software reuse, the Earth science software development community has encouraged and fostered reuse within the community. The NASA Earth Science Data Systems (ESDS) Software Reuse Working Group (WG) was established to support software reuse among members of the community. The WG developed the software reuse portal to inform the community about reuse issues and available resources and educate the community on improving software reuse practices. Requirements for a Reuse Enablement System (RES) have been developed and are being used to implement a prototype system to provide access to software for potential reuse. Reuse Readiness Levels are being developed to document criteria and improve capabilities for evaluating and selecting software for potential reuse.


11:30©¤12:00

Paper 3

Feature-Oriented Software Product Line Design and Implementation Based on Adaptive Component Model

Author: Yiming Yang, Xin Peng, Wenyun Zhao
Computer Science and Engineering Department, Fudan University, Shanghai 200433, China

Abstract

In current feature-oriented methods for SPL development, requirement-level variability analysis, representation and application-oriented customization have been well understood and supported. However, it is still difficult to map customization and tailoring on requirement level (feature model) to implementation level (architecture and components). In this paper, we propose an SPL development method based on the feature-oriented adaptive component model proposed in our previous work. The adaptive component model introduces feature-based domain model as the semantic basis of component ports (including internal and external ports). On the other hand, the adaptive component model has a micro control structure within the component, which enables the adaptation of the component behavior, including inter-component interactions, interaction sequence and style. In order to implement the feature-oriented customization on the program level, an in-component control center is separated to enforce the customized behavioral protocol and port semantics for each component according to the mapping specification. This separation of component coordination and computation facilitates the application-oriented customization on component behaviors. Then, in application development, requirement-level feature customization can be mapped to architecture- and component-level adaptations on architectural structure and component behaviors.


12:00©¤12:30

Paper 4

Multi-View based Customization of Feature Models

Author: Haiyan Zhao, Wei Zhang, Hong Mei
Key Laboratory of High Confidence Software Technology, Ministry of Education of China
School of Electronics Engineering and Computer Science, Peking University

Abstract

Feature models have been widely used to capture reusable requirements in a specific software domain. Through customizing a feature model, customers can specify their own products according to different application context. However, the customization of feature models usually involves many stakeholders and each of them may be only interested with parts of the feature model. Different stakeholders¡¯ concerns are often intermixed together and scattered in the entire feature model, which results in the difficulty in customization for stakeholders. To address this problem, we propose a multi-view based approach to facilitating the customization of feature models. The underlying idea is to cluster information scattered in a feature model into separate views according to different stakeholders¡¯ viewpoints, so that stakeholders can customize the feature model on their own views. We demonstrate the feasibility of our approach through an case from ATM domain.