Paper Reviews

Each week you will be required to submit one or two conference-style reviews of assigned papers.

Reviews should be short and insightful, aiming to identify deep things that are good or bad about the paper. This means that you must read and truly understand the paper. Ideally, a review should make three key points. Don't bother pointing out errors in grammar, writing, or formatting—high school kids can do this just as well as you; also keep in mind that a long review is not necessarily better than a short one. Here are some extra opinions on the topic, that may help you gain insights into what makes a good review: the classic Levin & Redell piece on how to write a good systems paper, Mothy's retrospective, Ousterhout's guidelines for paper reviews, and Gernot's warnings against committing benchmarking crimes.

In the real world, when a paper is submitted to a workshop/conference, there is a contract between the submitters and the program committee (PC) that the submission be held in highest confidence, and only those who review the paper get to see the submission and/or learn details about it. If this were a real conference PC review process, you would have to keep the paper and your review the highest confidence, which means it cannot be forwarded to or discussed with anyone. This is a key aspect of scientific ethics in our community.