This blog will act as a journal through which to track progess on our term project in DCCT*4090 - Information Retrieval.

Friday, May 30, 2008

First Post

Well, it looks like this will be the diary for our group project in DCCT*4090 - Information Retrieval. The group members for this project are: Zeshaun Hasan, Matt Pintar, and Michael Kuredjian. Our aim is to produce a document search and result ranking algorithm with a web interface. We are very excited to be playing with new (to us) technologies, including:

  • Java J2EE Servlets
  • JSP
  • JBoss Application Server
  • AJAX
  • Subversion
  • Search and ranking algorithms
  • Regular Expressions
  • And many more!
 Presently, our project source is hosted in a subversion reposity which is made available to the world through the Web_DAV subversion module for Apache. The module is really handy, because it offers authentication and access control for multiple users and projects. Right now, the respository requires authentication for all users to checkout and commit code. However, we might change that in the future if anyone is interested in looking at our code.

Our primary reason for using subversion is that is doesn't require element types to be associated to checked in elements in the same way as Clearcase and CVS. We did not investigate Perforce. Furthermore, subversion is widely used and well respected as the successor to CVS. Finally, there are many useful clients for subversion, ranging from eclipse plugins to simple command line utilities.

The project is split into two top level folders and an ANT build.xml file; one folder holds the source tree, while the second, "Phase1.war", holds the Enterprise Web Archive application, which includes JSP files, CSS, images, javascript, and compiled Servlet classes. The ANT build script has two targets: compile and deploy, which copies the completed WAR file to the deploy folder of JBoss Application Server.

JBoss was chosen as the application container for our project because it is widely available and free. It is a happy medium between Tomcat's simplicity and Weblogic's complexity, offering JMS queues and topics, should they be required, yet will happily serve simple HTML and JSP files.

1 comment:

Anonymous said...

Looks good, nice idea!