應用主題地圖建立物件導向程式庫樣式

Abstract

隨著電腦技術進步，資訊化、數位化的速度也越來越快，越來越多人與企業相當依賴資訊系統。所以資訊系統的發展也越來越重要，如何能夠更快速、更正確的發展出對人與企業有用的系統呢？建立一套軟體的程式庫是相當重要的。但是如何建立、使用電腦軟體的程式庫不是一件簡單的事，程式庫給軟體工程師使用時，要讓他知道正確方式來使用，且使用這些元件時還要注意哪些細節，尤其建立一套應用程式時，通常是許多的程式庫交叉使用，如何釐清彼此關係而不至於誤用，且將複雜度降到最低，這是在發展應用程式時很重要的議題。

資料探勘的技術主要是想從大量的資料中，探勘出有用的資訊，將大量的資料中找出有價值和可利用的部份，應用在我們平常的活動或決策中，能夠更快速且正確地完成每件工作。

主題地圖可以提供、展現主題知識的方法，是一種使用數位化的方式來表現多個主題的資訊與關連，每個主題是一個題材名稱並關聯一些資訊與概念，有了主題地圖，人們可以更視覺化的來快速學習一些事情。

本研究是利用資料探勘的方式來挖掘程式庫與程式的關係，主要是從應用程式翠取分析出程式庫與程式之間的關係，以提供未來程式重整時的依據，並建立主題地圖供程式人員快速學習，來減少學習曲線。這個方式的優點是可以節省程式人員的時間，也減少公司所付出的成本。

在本研究的架構下，完成了使用自動化的機制，讓程式樣式能夠維持最新的版本。可輸入多個專案程式原始碼，得到這些程式碼的關係。我們可以觀察哪些程式影響哪些程式，哪些程式的品質對於專案的品質有較大的影響，這些明確的結果可以提供專案軟體工程師與程式庫軟體工程師體認當程式發生問題時該問題的嚴重程度，知道工作的優先順序與重要性。

Following a rapid pace of Information Technology, the process of translating data into digital form and the utilization of computer systems have stronger momentum. A growing number of workers and companies closely depend on information systems, and naturally increase the importance of them. Although a software library can significantly contribute towards an efficient and effective development of software, to establish and apply a library is never a simple task. When a software engineer uses a library, he has to know proper procedures and details of each component. Besides, a library can frequently overlap various libraries in any software project. Obviously, it is a very important topic to clarify relations of libraries and avoid misapplication of them, as well as minimize their complexity.

Data Mining is a technique which is trying to dig out worthwhile information from massive data. If we can seek valuable and useful parts from abundant facts, and adjust our action or strategy based on them, then we can fulfill every task properly and precisely.

Topic Map is not only a method to provide and display thematic knowledge, but also a method to present information and relationship of multi-themes. Each theme is a name of a subject and a collection of information and concepts. With the aid of Topic Map, people can learn faster in a more visual way.

This study is to investigate the relationship between libraries and programs by the methodology of Data Mining. We extract the associations and analyze them so as to form a guideline for any future refactoring. In addition, Topic Map can help software engineers to shorten the learning curve and save both programmers' time and companies' costs.

Within our framework, we use automatic mechanisms to maintain up-to-date versions of software. After inputting source codes, we obtain relationship of programs in a project. Therefore, we can observe how programs influence each other, and predict key programs when quality of the project is concerned. Thus project engineers and library engineers can acknowledge the possible damage when a specific program breaks down, and then they can know the severity and prioritize their tasks accordingly.