Why Do Projects?

1
Why Do Projects?
My Thoughts on Things

• Omer Rana
• omer_at_cs.cf.ac.uk

2
What is required and what should I do?

• A project should push you as an individual
  slightly beyond your existing skills -- IN
  UNIVERSITY right?
• A project should enable you to explore something
  you like related to your course - INDULGE
  YOURSELF
• A project should be doable in the allocated time
  -- DO NOT BE OVER AMBITIOUS
• A project should look nice on your CV/Resume when
  applying for jobs -- GOTTA WORK

3
Why are you here?

• You are at University for a variety of reasons
• Finished Alevels, Finished first degree, Didnt
  want to work yet, Unemployed, Want to change
  career, Want something different, Too Lazy
• Aim Change your way of thinking -- your way of
  understanding the world around you
• More important -- than giving you training in a
  given area

4
Lets be idealistic for a while?

• I am here to learn to REASON BETTER
• I am here to CHALLENGE MYSELF of what I can do
• I am here to QUESTION NOTIONS THAT I HOLD
• I am here to see how much I can learn in a
  LIMITED TIME
• I am here to PLAN my time well, write good
  REPORTS, learn to APPLY what I have learnt before

5
A Good Project is
Well Planned (Scheduled), Well Reasoned, Well
Written, Shows what you have learned, Shows
Critical Assessment and Contains Good and Well
Documented Code
6
My Projects

• In three main areas Distributed Computing,
  Parallel Processing, Neural Networks/Data Mining
• Most are open ended -- will be customised to your
  skills and interests in the first visit
• Require weekly contact with schedule for next
  week -- it is your responsibility to attend the
  meetings at the allocated time
• Aim to ensure that you do the best project
  possible based on your own ability

7
Projects OverviewNeural Networks and Data Mining

• Data Mining Suite (DMS) -- neural networks,
  ID3/C4.5, regression analysis, Bayesian analysis
• Write DMS in Java or C -- preferably Java
• Develop infrastructure in Java -- with pluggable
  Java components
• Can work on Writing code for Data Analysis,
  Visualisation or Connecting to Databases with
  Data sets

8
Projects OverviewAgent based Computing

• Use AGENTBUILDER to develop an E-Commerce
  application -- for music, travel agency etc
• Use of mobile agents using AGLETS for network
  management, automatic software configuration etc
• Investigating Agent security -- what is
  important, implemention to demonstrate concepts
• Study of scalability in Agent systems -- use of
  Petri nets or other modelling approaches

9
Projects OverviewCORBA and Distributed Objects

• Use of JDK1.2 ORB or Visigenics VISIBROKER for
  wrapping C code into Java/CORBA objects
• Analysis of horizontal CORBA services Event
  service
• Using Java for linking distributed objects --
  write an ID3/C4.5 application in Java
• Developing a mediator using a Java/CORBA object

10
Projects OverviewExpert Systems using JESS

• Write rules for solving problems -- and develop a
  front end in Java
• Domains Scientific Computing, Medicine, Network
  Management etc.

11
Projects OverviewRobotics

• Use of the Mindstorm Kit from LEGO
• Develop robots that can be programmed from LEGO
  Blocks
• Jointly with Dr David Marshall and Mr Mike Daley

12
Projects OverviewParallel Computing -- Center
for Computational Science

• Learn parallel Java/C programming
• Developing financial applications on a 30
  processor Sun Machine using MPI
• Steep learning curve in the beginning -- but a
  very challenging project
• Jointly with Dr Roger Philp (Engineering and
  Physics)

13
Projects OverviewWeb Page for Parallel and
Scientific Computation

• Enhance Web site for Parallel and Scientific
  Computation
• Use of new Web technologies Java, Servlets,
  CORBA, XML and a number of other technologies
• Ability to generate new pictures and show some
  creative flair
• I will be very selective on who I pick for this
  project! -- this is in no way an EASY project

14
What Next?

• If you like a project, come see me sometime today
  for the fine print
• If you would like more time to think, see me
  sometime next week -- send me an email to arrange
  a meeting

15
Good luck
Look Mom! - I got it!!!

• Its your project -- aim to get the best degree
  possible
• Dont let your supervisor decide -- push your
  interests and agenda!
• Do what you like -- not what is currently HOT in
  industry or what your supervisor thinks is
• Read around -- lots of trade press, Web and see
  what people think!
• Available at http//www.cs.cf.ac.uk/User/O.F.Rana
  /Project/