The point of reading the papers is:
The coursework for the class consists of a number of readings (24 at least, see the schedule for the list). Each student is responsible for 20 1/2 page "Observations" on the readings. Since we'll cover more than 20 papers, you'll get to skip some. What's an "observation"? It's anything to do with the paper that's NOT a summary or a list of conclusions. For example, related personal ancedotes, a weakness in the paper, a contrast with other papers, and ideas for future work, are all valid ovbservations. In short, you can say anything interesting related to the paper that's not in the summary or conclusions.
Observations are due the Friday BEFORE class! It's pretty useless when everyone turns in their obervations 5 minutes before class starts. I'll post the observations up the Monday before class. See the submission form for how to turn in observations; you'll need to make up a URL containing each observation.
The goal of the project is to give experience testing out your ideas. You should work in groups of 2-3. Project ideas should be of high enough quality that at they could lead to publication in a systems conference. Performance analysis, porting applications, and design and evaluation studies are all starting points. In addition, I'll post a list of project ideas in the next few weeks. See the schedule for when the project ideas as well as the project checkpoint. A project write-up (in HTML or PDF) will be due at the end of course. The write up should be an 8-15 page paper describing your work.