CS227 - Topics in Object-Oriented Databases:
Final Project Page
Overview
The final project for this course is intended to have you focus
on one aspect of query optimization. We are open as to the form
your final project takes (research paper, literature review,
implementation etc.) and the nature of the effort
(individual, group). Naturally, we expect the
project scope to reflect the number of people who are
contributing to it.
By Wednesday Mar. 20, you should submit a project proposal
which describes your intentions for the final project.
(These are not carved in stone).
Below we propose some ideas for a final project,
as well as what we'd expect both by way of the project and the project
proposal.
Research Paper
If you believe that you could contribute something original to
the field of query optimization, we encourage you to write up
your ideas in a research paper. Of course, new ideas are not
enough. You need to stress why your approach is better and
different than those out there, either through convincing argument,
analysis or empirical study. Any aspect of query optimization
is a valid topic for a research paper.
For the project proposal submission, you should indicate
the area in which you have a research idea that you'd like
to flesh out, and sketch the idea.
Literature Review
There are a number of topics related to query optimization that
we haven't had the time to discuss in this course. For any of these,
you could choose to write a "literature review" that presents
a survey of the important work that has been done in the area.
We would expect a "Literature Review" to be roughly 25 pages,
and to present no fewer than 8 papers not part of the reading list.
More than summarizing the papers, your review should summarize
the area - categorizing the various issues and approaches
relevant to the subfield. Among the topics that could
be considered for "Literature Reviews" are:
been
- Rule-Based Query Optimization
- Semantic Query Optimization
- Parallel Query Optimization
- Distributed Database Query Optimization
- Query Language Expressibility and Complexity
- Data Mining
- Constraint-Based Querying
- Query Optimization in Commercial Databases
- ...
You could also choose to "zero in" on a topic that
was covered in a cursory fashion in the course.
For example, a review could provide a more thorough analysis
of the various approaches to processing joins or of
the various structures available that could be used as indices.
For the project proposal submission, you should choose one
of the above topics (or another), as well as a list of
readings (at least 8) that you'd like to draw from in preparing
the review. Note that these papers should not be taken from
the reading list for the course.
Implementation
If you decide that you'd like to work on a KOLA or COKO-related
implementation, we've a number of project ideas. Feel free to stop
by to discuss them. Of course, other implementations are possible -
for example, implementations (for the purpose of empirical testing)
of any of the ideas presented in a paper discussed in the course could
make for an interesting project.
Naturally, for the project proposal submission, you should describe
the implementation that you intend to build.
Mitch Cherniack