A Personal Information Agent on The Object Web

• Main Authors: Kuo-Tung Yu, 余國棟
• Other Authors: Hsing-Tai Chung
• Format: Others
• Language: zh-TW
• Published: 2000
• Online Access: http://ndltd.ncl.edu.tw/handle/05778623856909162253

碩士 === 淡江大學 === 資訊工程學系 === 88 === The thesis of this project is to describe a personal information agent (PIA) system, to save the users from time-consuming and tedious search activities for information over the Internet, as well as to provide a better service in terms of information retrieval. PIA builds a working environment for the user based on the object-oriented paradigm, which encapsulates program modules of traditional programming. The fundamental functionalities are as follows. (1) The PIA is operated by a user-friendly interface - Web browser. (2) A task editor is provided for the user to schedule a set of tasks, which can be executed automatically by PIA in the time periods specified. The execution status and results can be reported to the user when requested. (3) PIA provides an easy-to-use, effective interface for information retrieval activities such as Web search, newsgroup search, E-mail notification, and files search and transmission. (4) PIA can be accessed and controlled by a standard Web browser on a remote host. In order to provide a better and more professional information retrieval service, and to adapt to the future needs of Object Web, PIA further supports an object plug-in platform for the following needs. First, it allows different service providers to design their own proprietary interfaces, which can be embedded into the platform. After a user specifies his/her user preferences, PIA downloads information accordingly from the service providers so that the users always receive the newest information in time. This avoids the inconveniences of the time and labor a user has to put on searching for information frequently over the Internet. Second, it allows distributed computing of Object Web under the ongoing CORBA. Third, it allows the development of middlewares that can access information from different Web sites or CORBA objects based on the plug-in specification. This provides a solution to the information filtering and fuse that is difficult to achieve under current Web architecture. In addition, the retrieval process is totally automatic and transparent that requires no operation overhead for the user.