CS258: Spring 2008 Final Paper Information
I thought I'd reiterate some final information about the paper.
- It should be 10 pages, double-column. This is essentially a
conference format. Please stick to the format. Papers that are
less than 10 pages are unlikely to be sufficiently detailed.
Papers more than 10 pages will be penalized. This 10 pages should
include references as well. Again, treat this like a conference
paper so that you make sure to reference all of the related work. Try
to keep your font size >= 10 pt.
- Make sure that your introduction and/or first couple of sections
motivate your problem and describe the domain in sufficient detail that
a non-specialist could understand. You don't want to spend 50% of the
paper on background, but you should make sure that the paper makes
sense by itself without requiring the reader to look at related work in
detail. Help the reader understand what is important about your domain
of research.
- Many of you have simulation results. One very important
aspect of any paper that has simulation is that it should include a
section that (1) describes your simulation environment, including which
simulators you combined together, number of lines of code you had to
write to make it work, perhaps a block diagram of the simulation
system, (2) details of the benchmarks that you utilized or discussion
of the synthetic benchmark that you utilized, and (3) a discussion
about why you think that your simulation will indicate something
important about a real system; how did you validate the various
parameters? Why does the lack of accuracy (all simulators trade speed
for accuracy in some way) not matter for your environment? Etc.
- Make sure that you have a related work section that puts your
work in context with related work. Some people believe that such
sections should come early in the paper, others think that they should
come late in the paper (to avoid interrupting the flow). I am
agnostic on this issue except for the fact that such a section must
exist.
- Deadline: 11:59PM on Monday, 5/19. Note that you should
submit both your talk from Wednesday and your
paper by using the "edit" button on the projects page (with your
project number).
Back to projects page
Maintained by John Kubiatowicz (kubitron@cs.berkeley.edu).
Last modified Fri May 16 17:05:44 2008