DL Project Suggestion: ETD Federated Search

• Number of people: 2-10 (1 or 2 groups)
• Goal: Replace the federated search system of the Electronic Thesis and Dissertation (ETD) collection of the Networked Digital Library of Theses and Dissertations (NDLTD)
• Contact information: The Virginia Tech contacts are Robert France (france@vt.edu) and Marcos Andre Goncalves (mgoncalv@vt.edu), Professor Mann-Ho Lee, and the instructor. There are technology partners in several places including Cornell University and University of Oldenburg (Germany).
• Required background: To focus on the federated search and combination of results, students should have some background or interest in Java and information retrieval, etc. To focus on the cross language issues, students should have some interest or experience with other character sets and languages.
• Description: In 1997-1998 a simple system was devised by James Powell and others for searching a number of university sites that have theses and dissertations. It handles multiple sites, and carries out dictionary-based translation to German for searching a site in Austria. This system was documented in an article in D-Lib Magazine. It can be seen from the online WWW site. There are two separate efforts needed:
  1. Most urgently, we need to replace the current approach with one using the MARIAN system. This work is the focus of efforts by Mr. Goncalves, with help from Robert France. Interested students would join that team and so be able to collaborate with 2 PhD students.
  2. We now have partners all around the world interested in working with us to extend this to handle other character sets, fonts, and to have even more effective searching. We wish to improve the presentation of the results as well.

     Regarding other languages, some initial work was begun through an independent study in Spring 2000 by Libo Xin, to see how Chinese could be managed using MARIAN. A good project might extend his work to handle languages like Chinese and Korean (with help from Professor Mann-Ho Le). Then the full solution would be carefully documented so that it will be easy for other universities to adapt.

• Students involved: The following students are working on this project:
  • Avnish Kumar Chhabra, avnish@vt.edu