Welcome to COS597E!

Announcements

Course expectations have been changed in order to reduce the overall class workload. Please reiew the "class information" for updates.

Office Hours will be Wednesdays 4:30-6pm in 194 Nassau street, 2nd floor. Enter the center door (between both banks), and take the elevator or stairs up one floor.

The schedule for the semester

Class Information

In this class, we explore the way that two different research communities approach solving problems in the domain of systems. Each week, we will read two 'major' papers and one 'vision' paper (from two different communities), and explore the differences between their approaches---not just in terms of what one paper did and what one didn't, but in terms of community norms such as the standards of evaluation, or the aspects of the research that each paper presented as most important.

The readings: what's due each class

Every class, we will read one major paper and will be ready to discuss it. For Tuesday class, we will also read one minor paper. In order to keep us all honest, we're associating homework with this reading. Due at the beginning of each class, each student must submit:

Submit homework by sending me an email at mpmilano@cs.princeton.edu with the subject [COS597E HOMEWORK] and your paragraphs + questions in the email body.

Discussion leaders: The PL and Systems shepherd, and the critic

Each week, we will have three students assigned as "discussion leads" for the class. One student will serve as "shepherd" for the Systems paper; one student will serve as "shepherd" for the PL paper; one student will serve as "critic" for both papers.

The shepherd responsibilities

You are the representative of the author for that paper. Your job is to go over the important and/or tricky aspects of the paper, being sure to highlight the essential contributions it made to the field. In order to do this, you will have 10 minutes to present the paper at the beginning of class. Note that you may have slides for this, but you are not required to.

The critic responsibilities

Your job is to highlight limitations of both papers. You do not have a reserved time to speak; rather, you are a separate entity from the shepherd to allow the shepherd freedom to be positive about the paper without worrying that, in so doing, the shepherd is "sweeping things under the rug"

The Final Project

This class will contain a final project component, in which students complete a novel research-style project based around the class theme of "bridging between research areas." Examples of acceptable projects include:

I'm also putting together a short, "brainstorm-style" list of project ideas.

Deadlines

The Proposal

Deadline: 10/8

The proposal takes the form of a short, one-page document describing what you plan to do, and how you plan to achieve it. It should include:

Beyond these required components, the "ideal" proposal looks like the introduction for a conference paper. Much as with all assignments in this class, this assignment is not directly graded; rather, I will read proposals over fall break and shoot you an email with my feedback on your ideas.

The Survey

Deadline: 10/25

This is a short (~3 pages "word" format, ~1 page double-column LaTeX) document covering a broad space of "related work" to the topic you have proposed. It should, unless otherwise noted, be a series of papers within the area of your proposed project. It is absolutely fine for these papers to include work that we have read in class. Ideally, you will discuss around 8 papers. Note that these papers should be selected based on what is important to your project; it is not necessary that they be evenly balanced between areas, and it is ok (even expected!) that some of them might be outside of Systems + PL entirely.

You do not need to deeply understand all of these papers. Rather, you need to know enough to characterize them with respect to the other papers you are citing. For example: if your project is in operating systems, and you are citing RedLeaf, HiStar, and Singularity, it would be acceptable to explain that RedLeaf focuses on fault isolation, HiStar on privacy via information-flow, and Singularity on enabling structure communication between processes.

I would encourage you to spend less time reading any one of these papers than you would spend on an assigned class paper. Remember, how to read a paper is a learnable skill!

The Check-Ins

Deadline: 11/12

The Presentations

Deadline: 12/5-12/7

The Report

Deadline: 12/11