The course project is a short paper on a topic related to relational database management systems.  In addition to writing a paper, you will be expected to read and review some of the papers written by your classmates.   This is an individual project.  You many not work with a partner.

Step 1: Determine your topic

1. Index Selection:  (Lehui Nie, Fei Yuan, John Paul Steinberger, Hichem Fadali, Xuhui Li)

Indexes are an important physical design tool in relational database management systems.   Normally, the number of indexes that can be created is limited either by storage capacity or by the cost of maintaining the indexes.    Given these constraints, which indexes should be created?
• Sheldon J. Finkelstein, Mario Schkolnick, Paolo Tiberio: Physical Database Design for Relational Databases. TODS 13(1): 91-128 (1988)
• Gary Valentin, Michael Zuliani, Daniel C. Zilio, Guy M. Lohman, Alan Skelley: DB2 Advisor: An Optimizer Smart Enough to Recommend Its Own Indexes. ICDE 2000: 101-110
• Surajit Chaudhuri, Vivek R. Narasayya: AutoAdmin 'What-if' Index Analysis Utility. SIGMOD Conference 1998: 367-378
2. Approximate Query Answering: (Xiangyu Hu, Ray Sweidan, Yang Sun, Wentao Luo, Stuart Pollock)

For some applications, approximate answers to relational queries may be sufficient, particularly if approximate answers can be produced more quickly than exact answers, or if the approximate answers can be incrementally refined to produce answers that are more exact.   What might it mean for a query result to be approximate, and how can approximate answers be produced quickly?
• Swarup Acharya, Phillip B. Gibbons, Viswanath Poosala, Sridhar Ramaswamy: The Aqua Approximate Query Answering System. SIGMOD Conference 1999: 574-576
• Robert L. Read, Donald S. Fussell, Abraham Silberschatz: A Multi-Resolution Relational Data Model. VLDB 1992: 139-150

• Michael Ley's databases and logic programming bibliography server  contains many links to on-line papers.  It is searchable.
• The ACM Digital Library   is also searchable.  Access is by subscription only.  However, UW maintains a campus-wide subscription to the ACM Digital Library, so you should be able to search it and retrieve from for free if you are coming from any machine on the UW campus network.
• The IEEE Computer Society Digital Library should also be available for free if you are coming from a machine on the UW campus.
• The UW Library has copies of many database-related publications.
• Database conferences:
• The ACM SIGMOD International Conference on Management of Data (usually just called the SIGMOD conference).
• The International Conference on Very Large Data Bases (called VLDB)
• The Symposium on Principles of Database Systems
• The International Conference on Data Engineering

Step 2: Write an Initial Draft of your Paper

Your paper should describe the problem identified by your assigned topic, and survey proposed solution(s) to that problem.   Your target audience is other CS648 students.   Assume that your audience understands relational database management systems as covered in class.    Your task is to convey to your reader an understanding your assigned topic you are writing about.    You are limited to six pages.  The literature addresses these topics in greater breadth and depth than you will be able to achieve in that amount of space.    You must balance the breadth and depth of your presentation so that you  cover as much material as possible, but not so much that your presentation cannot be understood.

Your paper must be no more than six pages long, including bibliography and appendices.  Use a font size of 10 points or larger and make sure that it is formatted so that it is easy to read.
Your paper must be original work.  The use, without attribution, of material (text, figures, examples) from your sources is plagiarism, which is an academic offense.  You may make limited use of material (text, figures, examples) from your sources, but any such material must be properly identified, and its source must be cited.  Your paper must include a proper bibliography.

You should submit the following by the initial paper submission deadline:

• three identical hard copies of your paper, each including a title but not your name
• one cover sheet, indicating your name, your student ID, your topic, and the title of your paper.

Step 3: Read Others' Papers and Write Reviews

You will be given two or three of your classmates' papers to review. For each paper, you are expected to prepare a brief, constructive review.    The purpose of these reviews is to provide feedback to the papers' authors so that they can improve their papers.    Each review should be no more than one page long.  Other than this, there is no required format for reviews.   In your reviews, try to identify a paper's strengths as well as its weaknesses.   Be as specific as possible.  Be constructive: provide suggestions as to how the paper's weaknesses might be improved.

Reviewing will be double blind.  That is, reviewers will not be told the names of paper authors, and authors will not be told the names of their reviewers.  Each of the papers you are given will be marked with a number.  At the top of each review, you should clearly indicate the number and the title of the paper that is being reviewed.  Do not identify yourself in your review.

You should submit the following by the review submission deadline:

• two identical hard copies of each review
• one cover sheet, indicating your name and your student ID

Step 4: Revise your Initial Draft

You will be provided with two or three anonymous reviews of the initial draft of your paper.  Based on these reviews, you may revise your initial draft.    Like the initial draft, the final version is limited to six pages.   Your final draft should not be anonymous; please clearly indicate your name and your student ID number on the paper.

Place one hard copy of the final version of your paper in the CS448 assignment boxes on the 3rd floor of MC by the final paper submission deadline.

Evaluation:

Your project mark will be based the reviews that you write, as well as the final version of your paper.

• Paper (75%):
• Presentation:  Is the presentation well organized and clear?  Does the paper make use of diagrams, concrete examples, or other devices to improve its clarity?   Are there grammar and spelling errors?
• Coverage:  Are the depth and breadth of coverage reasonable, given the limited length of the paper?
• Originality:  Does the paper include some substantial original elements, such as examples, classifications, or a novel organization of the material?    Or is the paper a simple, shallow summarization of something from the literature?
• Reviews (25%):
• Do the reviews identify specific strengths and weaknesses?  Are the comments justified?   Are the reviews constructive - do they include specific suggestions for improvement?

Maintained by K. Salem