在英文字彙學習發展上，字彙數位學習系統的運用層出不窮，如套裝軟體、網路學習等，其主要的架構模式多數為主從式架構（Client/Server）的發展。然而隨著應用平台的發展，主從式架構也大致可分為視窗應用程式（Windows Application）與網頁應用程式（Web Application）兩種型式。前者屬於傳統主從式架構，其硬體設備的優劣將直接影響整體應用程式的運行，稱之為胖客戶端（Fat Client）；後者則為了降低硬體設備和提升跨平台性的支援，使用者僅需要透過網路的方式便能進行活動，稱之為瘦客戶端（Thin Client）。這兩種方式也根據不同的需求廣泛應用。 針對近年來字彙數位學習發展上諸多需求的考量，包括：軟硬體設備、網路能力、管理策略、RIA概念興起等，在上述主從式架構的發展中，有一新一代技術與架構的整合，稱之為富客戶端架構（Rich Client）。其優點在於結合胖客戶端開發獨立運作之應用程式與瘦客戶端易於佈署與管理的優勢，使主從間資源利用上達到平衡。因此，本論文之研究將以富客戶端為基礎，建置出富客戶端架構的數位學習系統。其中客戶端主要著重於視窗應用程式的多元性，強調有效運用客戶端的資源與擁有離線獨立運作的能力，並且能方便學習者於不同平台與作業系統上安裝和自動更新；同時伺服器端則以服務導向架構（SOA）為主，發展個人化學習資源的同步技術，並且以降低處理學習資源時所造成的伺服器負載（Overloading）與延遲（Latency）問題為訴求。 本論文透過實作英語字彙數位學習系統－MyVLS，並為了降低客戶端計算要求與最佳化伺服器的運算，本研究提出了：（1）客戶端應用程式支援連線與離線的學習、（2）提供個人化學習資源，與處理主從間學習教材與歷程紀錄的上傳下載和同步機制、（3）自動化產生個人化數位教材，並提出預先計算（Pre-Calculating）用以最佳化伺服器運算與提升系統效能。最後，經過本研究之實證評估得知，個人化學習資源的上傳下載相當合宜，其預先計算的使用也確實有效提升整體系統的運作。 In the development of English vocabulary learning, vocabulary learning system is used worldwide. However, the architecture design of client/server which used in e-learning system is changed by the development of applications. One is windows application which is regarded as traditional client/server architecture and the quality of hardware devices will influence the operation of applications directly. We call it Fat Client. Another is web application which for reduce requirements of hardware devices and to enhance supports of operating systems. User can only operate through internet. We call it Thin Client. This two type are used depend on demands. In addition, some considerations of vocabulary e-learning system like: hardware devices, quality of network, strategies of management, and RIA conception etc. which develop in a new generation we call it Rich Client. It combines Fat Client and Thin Client which can develop independent applications and have the advantage of deploy easily. This makes the use of resources on the balance between client and server. Therefore, think highly of the ability in online/offline modes and easy to install or update for client application are the most important things. Moreover, to solve the problems of overloading and demand latency results from produce individual digital resources, the server have to optimize systematic calculating. The thesis implements a vocabulary learning system: MyVLS, and to bring up: (1) Client application supports learning in online/offline modes. (2) To deal with synchronization of digital resources and portfolio between client and server. (3) The thesis emphasizes the use of Pre-Calculating which can optimize systematic calculating and enhance system quality. Finally, by experimental evaluation, we can find the design of MyVLS is truly works.
