|
The purpose of the work in CS 244E is to introduce the fundamental research in
low-power wireless networking. This requires reading deeply on a wide range of topics.
Each student is therefore responsible for writing a short summary of each primary paper,
which must be sent to the instructor and TA at least two hours before the
beginning of the class in which
the paper is discussed. Each summary should answer these four questions:
- What is the major contribution of the paper?
- What are the paper's strengths?
- What are the paper's weakneses?
- What unanswered questions does the paper leave, or what would you do to extend the work?
Note that each class typically has two papers. One paper is the focus of the
class discussion and requires a review. The second is important background reading.
You should read and be familiar with the background paper, but do not need to know
all of its minute details.
Here is an example writeup for the nesC paper of 4/5:
This paper presents nesC, a C dialect designed for low-power network
embedded systems. NesC supports efficient event-driven programs
through interfaces that bind a downcall and upcall (callback)
simultaneously. This allows the compiler to prune dead code,
optimize across call boundaries, and perform several other
compile-time optimizations. Additionally, by making interrupt
code explicit, nesC can detect data races at compile-time.
Despite all of this compile-time analysis, however, nesC
does not provide any explicit memory protection and recursive
functions can still easily crash a program. One thing the paper
does not examine are the programming complexity tradeoffs between
threads and events: comparing the purely split-phase implementations
to threaded ones would have been useful, as would exploring how to
add threads on top of nesC's restrictive concurrency model.
Each student can miss one
writeup; when you do this, still send email, but
say that you're taking your freebie.
|