以實際模型基礎之物件導向需求工程及與物件分析/設計/程式整合及對軟體發展及維護衝擊之研究

THUIR > College of Engineering > Department of Computer Science and Information > National Research Council Research Report > Item 310901/10181

Please use this identifier to cite or link to this item: http://140.128.103.80:8080/handle/310901/10181

Title:	以實際模型基礎之物件導向需求工程及與物件分析/設計/程式整合及對軟體發展及維護衝擊之研究
Other Titles:	The Study of Empirical Model-Based Object-Oriented Requirement Engineering(More) and Its Integration with OOA/OOD/OOP and Its Impact to Software Development and Evolution
Authors:	朱正忠,薛念林 Chu, Cheng-Chung；Hsueh, Nien-Lin
Contributors:	行政院國家科學委員會東海大學資訊工程與科學系
Keywords:	需求工程；模型；物件導向分析；物件導向設計；物件導向程式設計；軟體發展及維護 requirement engineering；model；OOA；OOD；OOP；Software development andEvolution
Date:	2008
Issue Date:	2011-06-23T05:57:13Z (UTC)
Abstract:	在任何新的專案的生命週期裡，最重要的一個部份就是要定義出新系統的需求。需求階段的活動，其實是最需要專業的投入，但我們通常都以非專業的方式來處理需求階段相關之工作。在實例中，我們觀察到，與開發的案子有關的專案關係人(stakeholder) 通常不是沒有受過良好的訓練就是缺乏領域的相關知識和經驗。這樣的方式註定了失敗的高可能性，在需求工程(Requirement Engineering)的領域中，有許多的方法都被提出來。而其中物件導向需求工程(Object-Oriented Requirement Engineering)是將物件導向方法中好的特性用來幫助需求階段的活動，像是需求擷取，分析，協商，文件化，驗證和確認，管理等等。有些學者成功的應用了模糊理論的技術來處理不精確和不完整的需求。許多學者則著重在使用者案例和情節來將需求有系統的模型化(modeling)。大多數的學者則只著重在需求工程的某些片面方向(aspect)，而不是從整體來考量，像是整合 (Integration)的問題，例如怎麼讓需求文件與其他軟體工程生命週期階段的產物(Artifact) 作整合。因此，實際上大多數的需求文件仍然是以自然語言的文字格式(text form)寫成的，這樣子會造成了充滿模稜兩可(Ambiguity)、不一致(Inconsistency)、不精確(Imprecision) 及不完整(Incompleteness)的情形。除此之外，由於缺少精確的模型能力(modeling)，使得想要整合或追蹤其他階段的產物只能倚賴人工，而造成時間上的浪費以及容易發生錯誤。這個缺點也讓我們在導入CMMI 相關活動，例如人工維護相關產物(Artifact)（文件）的追蹤性(Traceability)，的時候會花費相當高的成本。為了能讓需求工程能夠有效的應用在實際的案例上，需求文件的模型化和與其他階段的整合需要作深入的探討與研究。在這個計畫中，我們提出一個以實際模型為基礎的物件導向需求工程(empirical Model-based Object-oriented Requirement Engineering(MORE))及與物件分析、設計、程式之整合。這個計畫的主要目標為： I. 研究在需求階段中，以模型為基礎的物件導向模型(MORE)的可行性 II. 從實務中來研究將」以模型為基礎的物件導向需求工程」(MORE)和」物件導向分析 (OOA)、設計(OOD)、程式(OOP)」的整合的可能性 III. 分析整合物件導向需求、分析、設計、程式對軟體的發展及維護所造成的影響 IV. 設計和發展支援 MORE 的工具在這個計畫中，我們將會研究相關的研究和實際上需求階段的議題，也會設計 MORE，並實作支援MORE 工具的雛型。我們也會將MORE 及其雛型工具應用在實際的系統上並證明我們方法的可行性。 Defining the requirements of a new system to be built is one of the most important parts of the life cycle of any project. Requirement phase activities, which indeed need the professional treatment the most, usually were treated in unprofessional way. They are then destined to the high possibility of the failure. From our observation, the stakeholders involved were either not well-trained or lack of domain knowledge or experiences. Many approaches have been proposed in the field called requirement engineering. Object-oriented requirement engineering which takes the good features of object-oriented paradigm have been proposed to assist the activities of requirement phase, such as requirement elicitation , analysis, negotiation, documentation, verification and validation, and management. Some researches have successfully applied Fuzzy technologies to deal with the properties of imprecision and incompleteness of requirements. Many researches focus on use case and scenario to systematically model the requirement artifacts. Most of researches focus on only some aspects of requirement engineering without considering the whole aspects that one needed in requirement phase from practice view point, such as integration issues like how requirement document can be integrated with other phases』 artifacts, e.g. artifacts in UML. Therefore, in practice, most of requirement documents are still written in text format in natural languages, which suffer the ambiguities, inconsistency, imprecision, and incompleteness. Besides, due to the lack of precise modeling capability, its integration and traceability to the artifacts of other phases may be able to only rely on manual efforts which are time consuming and error prone. This drawback has also created a high cost to the companies that implement CMMI related activities, for example manually maintaining the traceability of related artifacts (documents). In order to make requirement engineering practically applicable and useful to real practices, the modeling of requirement documents and its integration with the models in other phases need to be researched in depth. In this project, we propose empirical Model-based Object-oriented Requirement Engineering (MORE) and its integration with object-oriented analysis(OOA), design(OOD) and programming(OOP) from empirical view point. The main goals of this project are: I. Study the feasibility of model-based OO modeling in requirement phase II. Empirically study the possibility of the integration of MORE with OOA/OOD/OOP III. Analyze the impact of integrated OOR/OOA/OOD/OOP to software development and evolution IV. Design and develop supporting tools for MORE In this project, we will study the related research/practice issues of requirement phase, design MORE, and implement the prototype of MORE supporting tools. We will also apply MORE and prototype tools to real system application and demonstrate the feasibility of our approach.
Relation:	研究編號：NSC96-2628-E029-002-MY3 研究期間：2008-08~ 2009-07
Appears in Collections:	[Department of Computer Science and Information ] National Research Council Research Report

Files in This Item:

There are no files associated with this item.

Loading...