Princeton University Computer Science Dept.

Computer Science 425 Database & Information Management Systems Andrea LaPaugh

Fall 2006

Project Overview

Examples

Here are some suggestions. You may choose one or use them as guides to specifying your own project.  Note that many of these suggestions talk in terms of one of database systems or information retrieval (search) systems.  In reality, database and information retrieval are combined in many tools and applications;  your project certainly can combine them.

1. In earlier offerings of COS425, the same project was done by all students. It was the implementation and evaluation of the main query optimization algorithm presented in Ramakrishnan and Gerke Database Management Systems (3^rd edition). A detailed specification of that project can be found in the COS 425 spring 2001 project description.
2. Implement an application that requires database support. Implement the user interface, the application interface to the database and the database. This application needs to have some complexity in functionality, constraint maintenance, reliability or user interface.  The user interface may be minimalist if the focus of the project is elsewhere.   The application should be something in which you are interested and for which you can obtain or generate a reasonable set of interactions and data for testing.  The application need not be "serious":  one previous student implemented a database and rule system for the game Warhammer.    The database may be relational or in XML.   Depending on the system being used and the student's background, learning the use and configuration of an API for a database system may be considered a substantial project goal.  (See Chapters 6 and 7 of Ramakrishnan and Gehrke, Database Management Systems (3^rd edition), for a discussion of application development for SQL. See Chapter 10 of Silberschatz, Korth and Sudarshan Database System Concepts (5^th edition) for a brief discussion of application development in XML.  Problem sets 3 and 4 give you pointers to SQL and XML servers.) 
   
3. Database techniques and systems exist for special kinds of data. For example, Chapter 28 of our text presents techniques for spatial (geometric) data. Your project may focus on techniques for a special kind of data. Possibilities include:
• Select a special kind of data and write a review paper describing the important techniques and issues, with references and critical analysis.
• Implement and evaluate an important method for specialized data, e.g. multidimensional indexing (see Section 28.3 of our text).
4. There are also customized information retrieval techniques for special kinds of data.
   
• Select an example of a special kind of data (e.g. music, images, video) and write a review paper describing the methods that have been developed for information retrieval beyond the use of keyword search of text labels.  What are the important techniques and issues for information retrieval of this information medium?  Provide up-to-date references and critical analysis.
• Explore the use of  information retrieval software posted on the Web for some specific type of non-text data;  do a set of well-designed experiments or implement a small application that uses it.   A home-grown example of such software is Marsyas  for music (begun as a Ph. D. thesis in CS at Princeton).
5. Do an empirical study of two or three search services that use different search engines, i.e. the crawling, indexing and search software are different.  To the best of my knowledge Google, Yahoo, Ask  and MSN Search are all different, but you should check that nothing has changed in this fast-changing area (e.g. until 2004, Yahoo used Google for its Web search).   To do your study, read about well-established studies like those conducted by TREC.  Be forewarned that such studies involve testing the search engines on identical queries and analyzing the results, which can be very time consuming. 
   
6. Explore methods for enhancing and changing  queries that are specified by search terms -- i.e. the kind of queries we use every day.  Such enhancements and changes attempt to deal with ambiguity, synonyms and other natural-language issues.  Read about state-of-the-art  methods for changing queries.  Do some analysis of methods -- for example,  do a comparative analysis from the literature or conduct experiments to evaluate methods.  You may design and analyze a method of your own design, but you must have knowledge of prior methods and results.

Requirements

1. By 5pm on Wed. Nov 8, 2006 send email to Professor LaPaugh containing a one-paragraph description of your project. Each individual must email Professor LaPaugh to confirm partnerships.
2. During the week of Nov. 27, 2006 meet with Professor LaPaugh for 15-20 minutes to discuss project progress and issues.
3. Submit by 5pm Tues. Jan 16, 2007 (Dean's date) a report that describes your project. This must include the goals of the project, your methodology and the results   If it is an application implementation, you need to describe the application, your design requirements, the major implementation decisions, and your assessment of the result. If it is an experimental algorithm study, you need to describe what was implemented, the major implementation decisions, how you designed the experiments, and the experimental results.  If it is a theoretical study, you need to describe the problem, review what was known about the problem before your analysis, and give the details and the results of your theoretical analysis. If it is a literature-based project, you need to describe the major issues under study, summarize the major techniques and experiments presented in the literature and critically analyze the results; you must have a bibliography that includes recent research.  For any project that involves programming, all source code you write should be in an appendix or made accessible on the Web.
4. After the project report is submitted and before 5pm Mon. Jan. 22, 2007 each individual or pair must meet with Professor LaPaugh for a project demonstration (where applicable) and discussion.