Research On Domain Assets Management And Adaptability Evolution

With the Internet technique development and prevalence, the numbers of the usable web services online is increasing in a high speed. And our society today requires higher level on the form as well as development process of software than ever before, e.g., seamless integration of various businesses, agile development of application system, automatically creation of application software oriented specific domain etc. To battle for these requirements, one kind of running on Internet and service-oriented software as a new form is recognized and applied recently by the more people.Networked software is one of such software, which is of service-oriented and networked, and embodies the principle of "Software as a Service(SaaS)". The development lifecycle of networked software is a kind of process which is user-centric, domain-oriented, and flexible. During the lifecycle of networked software development, the reusable and common domain assets play the important role for its agility and satisfying the client goals of personalized and diverse. The domain modeling of networked software is a process of creating the reusable and common assets of application domain.The final goal of networked software is to provide service which is individually, diverse, and highly quality of experience. for its important foundation, it is by no means trivial that how to manage large numbers of domain asserts of multi-level, various granularity, and accomplish evolution of domain asserts accommodated to requirements change. Efficient management of domain asserts is favorable to control the scale of asserts and promote reuse of asserts, while dynamic evolution of domain asserts can guarantee the usability, advance the reuse scale of asserts, and be convenient for creation of target application system.As a further study of the meta-modeling framework O-RGPS, this thesis focuses on the organization, management, and adaption evolution of domain assets. Therefore, the scientific problem studied in the thesis is, based on the meta-modeling framework O-RGPS, how to efficiently manage the domain assets and accomplish the evolution adapt to requirements change by taking advantage of domain problem ontology. Considering the scientific problem, the main contributions of thesis are summarized as follows:(1) Oriented to DPO and RGPS meta-modeling framework, a management framework for organizing and management domain assets is proposed. The core of this framework is domain problem ontology (DPO), and the function of DPO is relation to creation methods of its. In this framework, domain goal models are clustered through semantic similarly calculating, at the other hand, domain role models, process models, service models are clustered also based on the correlation among them. And then domain problem ontology is created by use of basic information abstracted from these models. At last, management for domain asserts is accomplished through the domain problem ontology organizing to these asserts. Their function is including some aspect, such as eliminating the superfluous models from domain asserts bases, measuring the completeness of domain asserts bases, and providing the function of semantic query for domain models by establishing semantic relation among these models.(2) Domain assets weighting algorithm based on dynamical requirements priority is proposed, which make the domain assets weighting is related to its frequency be used. In other words, users’interesting and concern is captured according to their immediate requirements, meanwhile, requirements priorities is calculated dynamically based on the features of networked software such as type of requirements, role of requirements proposed, frequency of requirements proposed and time of requirements proposed, and then domain assets weighting is figured out. By this means, the correlation is established between domain asserts weighting and requirements of target application system.(3) Based on ripple-effect analysis of requirements change, a method of domain assets evolution is proposed. In this method, once evolution is triggering automatically, evolution object is selected autonomously, and then the influence of evolution is analyzed qualitatively and measured quantitatively. By this way, a evolution path of minimum expense is chosen from several alternative evolution paths in order to reduce the influence scope. So the cost and complexity of asserts evolution is decreased and the evolution is completed quickly and effectually.At last, In the O-RGPS domain modeling tools which is designed and implemented, the management and evolution for domain asserts is realized by use of the related theories and methods brought forward in this article.