Note: I have probably broken all these rules. I usually ended up paying for it.
Total Paper length: ~15 pages.
Not all sections are compulsory from the list: Our Idea, Analysis,
Experiments, Discussion.
1-2 lines: goal of this paper
1-2 lines: motivation and importance of work
1-2 lines: per result or contribution of our work
[it is good if you can quantify your claims with numbers that
refer to intuitive or easily explainable metrics i.e. 50% faster]
[At paragraph level: * indicates one or at most two paragraphs
** means as many paragraphs area needed, but don't do crazy]
Introduction [1-1.5 pages]
* What is the problem, motivation, importance, main contribution
* The problem: high level definition, significance
* Previous work [at most one line per relevant paper, leave details
for next section]
* Our contribution: possibly list of results and contributions
* "The rest of this paper is structured as follows. In section 2,....."
Background and Model [1-1.5 pages]
* Definitions, scenario, model
* Previous work [provide details only if this will help show where
your work is different or better.
Make sure you reference all related work and especially that by people
in the committee!]
* Weaknesses of previous work or the absence of the thing you are proposing
* Assumptions and limitations of your work. [Be honest, but don't condemn
your work].
Our Idea [as required]
* The innovation in brief
* New definitions, if necessary
** What it is and how it works
** Explanation of fine points, clarifications
Analysis [as required]
* Overview of what we prove in this section
* Analysis-specific definitions [i.e. a graph structure that you will
use in the proof]
** Analysis - proofs
* Comments - Meaning of proofs at an intuitive or practical level
Experimental Results [ as required]
* Overview of experiments: what we do measure, what we don't, why
* Simulation model or measuring infrastructure
** Experimental Results
- use a heading (in bold)
to distinguish experiments (i.e. The effect of size on performance
or even
better The size does not affect the performance..)
- each plot should have
a clear reason for appearing (ie with more load we get worse throughput)
- explain each plot: the
axis, what we see, what is the trend, why this is the trend
- statistical comments will
strengthen your results: confidence interval, correlation coefficient etc.
- Each figure should be
fairly stand alone and self explanatory: captions should be readable and
understandable.
- Comments, discussion,
explanations
Discussion [1-2 pages]
[If you have space and ideas, you can strengthen the importance of
your work by
showing: a) significance, b) relation to other pieces of work,
c) possible practical applications]
* Overview of section
** Paragraphs according to needs
Conclusions [1-1.5]
[Conclusions are very important. Do not expect that the reader remembers
everything you told him/her.
Having stated the definitions, you can now be more specific that
in the introduction]
* Overview what this work was about.
* Main results and contributions
* Comments on importance or
* Tips for practical use [how your results or experience can help someone
in practice or
another researcher to use your simulator or avoid
pitfalls]
* Future work. Reinforce the importance of work, but avoid giving out
your ideas].
References
[It has to be correct, complete, and accurate]
Appendices [as required]
[Appendices are no man's land: you can write whatever you want: reviewers
are not obliged to read
through them. Suggested uses:
a) provide additional results that strengthen your arguments but are
not crucial
b) maintain the flow of the paper by putting here a lengthy proof or
discuss a detail
c) address an issue that shows you have thought the issue in depth