軟體在現今生活中已被應用於各個領域,成為不可或缺的一項重要元素,對人們的生活和工作都產生了深遠的影響。然而,生活中被人們廣泛應用的各類軟體卻難以避免地都存在著缺陷,也許是人們可接受範圍內的小缺陷,也許是人們可接受範圍外的大缺陷。缺陷將造成軟體使用者的不便,進而降低顧客忠誠度與產品銷售量等不良影響。為了減少缺陷所造成的損害,確保良好的軟體品質,在軟體的開發初期就需要執行軟體測試,並將軟體測試整合到軟體開發生命週期之中。由於軟體品質的提升需要耗費許多的人力與時間,現在已經有許多自動化的軟體測試工具被開發出來加速測試工作的進行。 使用測試軟體進行測試工作固然重要,但測試軟體挑選結果對測試工作的進行與效果影響甚鉅,並且挑選出適合待測軟體系統的測試軟體往往耗費許多人力與時間。因此本研究針對此議題提出一套以品質機能展開為基礎的測試軟體評選架構,並以德爾菲法為輔助工具,分析軟體系統的需求、測試軟體的技術,以及兩者之間的相關性,為待測軟體評選出最適合進行測試工作的測試軟體,以增進軟體測試工作的效率,進而提升軟體的品質及軟體開發的競爭力。評選架構更融入資料重複使用的作法,經由累積並參考先前的分析與測試執行結果,讓分析人員不需要對相同的資料一再地進行相同的分析,避免人力與時間耗費在追求已知的結果上,以提升評選的效率。
Nowadays, software has been applied and become indispensable in every area of people's daily life. However, software inevitably contains defects which may vary in scale of their influences. No matter what the effects are, they can reduce the customer loyalty and sales amount of products. In order to reduce the influences due to software defects and maintain good software quality, it is necessary to start performing software testing when software development life cycle begins. Since improving software quality is a labor-intensive and time-consuming task, many automated software testing tools are developed to accelerate the process of software testing. Although it is important using software testing tool to accelerate the testing process, how to evaluate and select the proper tools for the work is also a mandatory issue. A poorly chosen tool not only wastes much labor but also delays the whole schedule of the project. This research proposed a framework for software testing tools assessment based on Quality Function Deployment (QFD). Proper software testing tool can be selected objectively through the assessment of QFD. In the process of assessment, characteristics of the software being developed, functionalities of software testing tools, and the relationships between them are developed with the aid of Delphi Method. Having selected the proper software testing tool for the software project, the efficiency of software testing can be promoted and thus enhance the software quality. The proposed framework also takes advantage of reusing the data generated in previous software projects. By accumulating formerly analyzed data and the testing results, effects can be saved from pursuing known results.