Project Part # 1

Web Database Management Systems
Term Project
Part #1, due March 25, 1999 by 5 p.m.

In this project, you are required to work in groups of two or three. This project is a completely conceptual exercise. You will be asked to define a problem and propose a solution. You will be asked in part 2 to describe your approach and define why it is going to be a good solution. You will not be required to solve the problem in detail, but your work in the project can lead to a research paper. In part 1, you are required to describe in detail a research problem in web database management systems (WDBMS). Your problem has to be:

Related to web. This means that the problem is not a regular database problem, it depends on the new challenges brought by the introduction of the Internet or an Intranet to the data processing methods.
Scientific. This means that the solution to the problem is not a mere programming task, but it requires the design of a conceptual framework.
Interesting, useful to somebody and needs a solution in the near future.
Unsolved. This means that it is not fully explored in the research community yet, no known comprehensive solution exists.

Almost all the papers we are reading in the course this semester are related to the query language aspects of WDBMS. You are free to choose any subject you like. You can propose a problem that is either

a complete framework for providing a type of service, or
an extension to the existing database technology for a new functionality.

SAMPLE PROBLEM
As an example, here is a sample research problem that you can work on. We know that databases are really good for storing "consistent" data. But the data coming out of databases is dull, it is not edited. most of the languages we have seen in this class will generate a web page from tuples, one way or another. However, we would like to allow people to generate their own documents, put pretty pictures in them, maintain them, make them look as they wish. In doing so, the complete control of the document is passed to the user. As an example, suppose each page contains a text field that mentions when that page was last modified. Hence, if the user changes a single line in this file, this field has to be updated. There might also be associations between pages, if a new page is added to a directory, then its index and title has to be listed in a list in another page. The problem is then letting people create their pages and also create such associations in a reasonable way. How can ordinary users specify such things, what is the meaning of specific types of rules and how do you define a consistent database of web pages with respect to these rules? How can these rules be maintained without forcing users to do a lot of shuffling between pages? Hence, from a framework point of view, you can define a model for implementing such a system. You can discuss how a database may be used as part of such a system, how users can edit files and how they can be notified of possible changes to the document database after they create a new document or change an existing document. From an extension point of view, you can specify the syntax and semantics for rules that can be defined in such a system, when these rule become true, when a rule is allowed to trigger a change to a document, and when a rule is allowed to trigger another rule.

DELIVERABLES
You are required to present a project report of at least two at most four single spaced pages (with reasonable margins) by the due date written above. Your report should have the following sections:

Problem description should define exactly what problem you are trying to solve. Try to be succinct and to the point.
Justification should describe why this is an important work. In particular, you have to describe how a solution to this problem may become part of product, what type of users will use it and what types of gains they will have.
Research areas section should describe which fields in computer science, mathematics, etc. contain tools and theories that might be useful to your problem domain and why. For example, the web based query languages were based on SQL-like declarative query languages from database systems and also search algorithms from graph theory. A work that finds web usage patterns of different users may use many different data mining techniques, it may also correlate the content of web page with the amount of time spent at that page using database type query capabilities.

Note that you are not proposing a solution in this part. You should define the problem clearly and think about other areas of research that might provide useful methods. In the second part of the project, you will be asked to read several papers that constitute related work and discuss the details of your solution.