文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > Automatic Generation of Rule-based Software Configuration Management Systems BRIEF DESCRIPT
侵权投诉

Automatic Generation of Rule-based Software Configuration Management Systems BRIEF DESCRIPT

We propose a model-driven methodology and toolset for automatic SCM system repository creation and feature composition using code generation and rule engine technologies.

Automatic Generation of Rule-based Software

Configuration Management Systems

Guozheng Ge, E. James Whitehead, Jr. (Advisor)

Dept. of Computer Science University of California, Santa Cruz

{guozheng,ejw}@cs.ucsc.edu

BRIEF DESCRIPTION

We propose a model-driven methodology and toolset for automatic SCM system repository creation and feature composition using code generation and rule engine technologies.

1. INTRODUCTION

1.1 Research Opportunity

Software configuration Management (SCM) systems [4-7] provide control over the evolution of complex software systems. Most existing SCM systems are built as monolithic systems using proprietary data models, unique system architectures and non-reusable SCM components. This makes SCM system customization and evolution both a difficult task and an expensive investment. Most customers lack standard SCM domain knowledge and modeling tools to identify features and compare capabilities of various systems on a common ground. Few rapid development solutions are available to automate and customize the SCM features and integrate them with customers’ existing development process. To make things more complicated, once a SCM system gets deployed, it is hard and expensive to change feature configuration and redefine operation semantics when customer requirements change. So we believe it is an important research problem to provide automated SCM system development solutions where customers design data models and semantic models to drive a code generator to automatically produce a SCM system that implements these model specifications. When customers change model specifications to reflect new requirements, an updated SCM system can be quickly generated with ease.

1.2 Approach Overview

Whitehead, et al. has proposed a uniform data modeling framework called Containment Modeling Framework (CMF) [1-3]. It has been successfully applied to define and compare repository models for more than twenty Content Management and SCM systems with a standard set of modeling primitives. This opens the possibility to automatically generate SCM repositories based on their CMF descriptions. For defining SCM features on top of the repository, we have designed a Semantic Modeling Framework (SMF) to specify SCM features and operations. With the semantic model specifications, we can also automatically generate feature implementations. Based on CMF and SMF, a prototype SCM system generator called Bamboo has been developed and it can generate a replica of Revision Control System (RCS) now. Other advanced features, such as advanced version control, workspace management, distributed collaboration, process support,

accounting and auditing etc., will be identified and integrated into SMF after a detailed domain analysis process like FODA [8]. Rule Engine technologies [9] provide a viable solution to separate SCM application logic from other parts of a SCM system. Various SCM features and operations are encoded as rules using special rule languages. The rules are centrally stored and managed in a rule repository. Other parts of the system send SCM service requests to the rule engine, the engine queries the rule repository and execute the matching rules. Centralized management of application logic reduces the maintenance effort and cost; it also makes dynamic change and customization of SCM features possible.

Combining the two technologies together, we propose an automatic generator for rule-based SCM systems, in which we generate the repository layer based on containment models and create the SCM application logic layer based on semantic models. The generator translates the models into executable Java code and business rules. Java code implements repository initialization, repository API, SCM utilities services, glue code for business rules, and a demo command line client. The business rules are executed in the rule engine to provide SCM features and operations that can be triggered by input events or data. When customers change SCM requirements, they redesign the repository and semantic models (with assistance from a visual modeling tool that we create) and an updated SCM system can be quickly generated according to the new model specifications.

1.3 Contributions

First, we carry out a comprehensive domain analysis to identify SCM features and operations by analyzing use cases and representative SCM systems. Through the domain analysis process, we also set up domain models and reference architecture for new SCM system development.

Second, we propose a repository modeling framework and a semantic modeling framework as the open standard for SCM system design and modeling. Using these two frameworks, we can formally compare repository structures and feature structures for various SCM systems.

Third, we develop an automatic SCM system generator based on the domain analysis and two frameworks. The generator makes fine-grained feature customization and requirement change easy and quick: developers design the models and SCM systems can be automatically generated.

Fourth, we apply rule engine technologies to compose SCM features. SCM features and operations are encoded as rules using special rule languages. This is a unique and novel way to provide

第1页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email