Server

CS 460/594
Course Project
Fall Semester 1997

PURPOSE: The course project will allow students the opportunity to design and implement a client/server model of a data/information retrieval system. Students will have the freedom to design their own model using concepts and software described or referenced in lectures.

COMPETITION: Each student (or perhaps small group of students) will essentially design and implement a product which might be marketed by a competitive Internet Service Provider (or ISP). Each mock ISP in the class will be given the same data/information to process and index for query-matching. Students will be given the opportunity to evaluate other student ISP products at the end of the course.

GRADING: The number of points to be awarded for the course project will be determined by the instructor and a small (but distinguished) panel of experts in data and information management. Projects will be judged according to style, correctness, and efficiency.

ROLES: The role of Groups 1-5 and Group 6 will be different in terms of software development and packaging. The single-person group (6) will be responsible for the development of a key module needed to implement the vector-space IR model for this project. Group 6 will have to coordinate the development of their distinct modules in order to produce an archived library or software toolkit (written probably in C or C++) for conceptual IR modeling.

Groups 1-5 will use the library of routines produced by Group 6 to develop intelligent search engines for testing on a WWW-based document collection. These Groups (1-5) will develop customized IR client-server software environment (using Java, CGI/Perl, C, or C++) which is capable of conceptual retrieval, uses a simple user interface, allows relevance feedback, and returns a rank-ordered list of relevant documents and terms for natural language queries.

CORPUS: The text corpus for this project will be constructed from on-line (WWW) documents from the FedStats repository (supported by the Bureau of Labor Statistics), from the National Science Foundation's Division of Science Resource Studies, and the U.S. Census Bureau.

REPORTS: Groups 1-5 should assign one group member to submit responses to the questions contained in the report forms below. Please complete the forms with thorough explanations and submit the form by the specified date. Click here for a directed graph of evaluation assignments.

Report No. 1 Thursday, October 30
Report No. 2 Wednesday, December 17

PANEL: In-class presentations by each ISP will be judged by a distinguished panel of UTK experts in information technologies. This panel will judge the ISP's search engines according to level of effort, creativity, query-matching effectiveness, and in-class presentation. The panelists for this semester are:

Dr. Shirley Browne, Department of Computer Science

Mr. Scott Wells, Department of Computer Science

Prof. Mark M. Miller, School of Journalism

Groups:
Group # Students ISP Name
5th 1 Wesley Hooper, Robin Lee,
Qingping Deng, Elton Wells 		Information Pathways
4th 2 Michael McDaniel, Jennifer Jackson, Shilpa Singhal, Calvin Janes, Rachel Huff Safari Creations
2nd 3 David Zlotchenko, Elizabeth Dixon,Jamal Slappy, Bobby Theerathorn DigIn Products
1st 4 Kendra Jones, Luojian Chen,
Phyllis Sexton, Adam Partin 		Search Busters
3rd 5 Yen-Yit Chan, Fei See Tian,
Graeme Burke		 Tachyon
Group # Student Software Module
6 Kim Buckner SDD Algorithm
Ron Jurincie Document File
Safeer Ladha Ranking (Cosines)
Zach Walker Sparse Matrix
Brannon Moore Query Formulation

Update Last Update: December 18, 1997, Michael W. Berry