xuefenghuang / master-thesis Goto Github PK

These works (biber, biblatex) are much better with modern bibliography entries. It's a good idea to start using current, up-to-date packages with LaTeX setup sooner rather than later.

It would be better to talk about how to add a new call to the sandbox.

Your LaTeX files need a "base document around them" so that you can compile. Here's a suggestion for one thesis.tex. Once you add that, use your TeX install to compile this file. See the inline comments on where to add what.

% Let's use *some* doc class just to get started. You can switch to 
% whatever else later.
\documentclass{scrartcl}

% Figures, tables, URLs and hyperlinks are best handled by packages.
% Import the packages you need; call multiple times for different packages.
% \usepackage{...}

\begin{document}

% Add the input files in the correct order below:
\input{abstract}
% \input{...}

\end{document}

There are a few \â's and other weird characters in design.tex (IIRC) which should be various accent marks / ticks. Make sure document is saved using the UTF-8 encoding.

Since this paper refers to other systems relatively often, so it's helpful to discuss the other systems upfront, explain where in the (home) network they exist, and so on. This nicely sets the stage for pointing out the unique properties of Seattle extension.

The paragraph on Seattle and its components absolutely needs an overview figure.

1. This section mentions three challenges in the first sentence, but add a fourth later down in the text.
2. This section needs lots of forward pointers to other (sub)sections in text.
3. Show an overview figure so as to explain what you consider a "home network" (What are the different machines? Different zones of access? Which kinds of IP addresses are used where? What's the NAT gateway? Where could a Seattle installation live?)
   Inkscape or LibreOffice is the software to draw figures.
4. Make clear my contribution rather than just platform design goal.
5. The list of goals should also point forward to those places where the design and implementation of these goals is discussed.

A concrete example: "BISmark [...] has several security issues [...]". It is unclear to point which issues are talking about, is there something in their design that's inherently unsafe?

I propose to use Biber, see http://biblatex-biber.sourceforge.net/ (with due thanks to @fmetzger for forcing me to use it in the first place.)

It's as "difficult" as to write up your references in the proper format (here's an example), and then tell LaTeX to please print the used references at some position in the final doc, usually at the end.

When privacy/security are mentioned, let's see how a reference monitor for a call is implemented and deployed.

1. This talks a lot about what has been implemented, it should also discuss why precisely these calls were chosen, but not others.
2. We need more forward pointers.
3. network.tcp etc. don't exist in Repy.
4. Where does the mapping from "types of experiments" to "used API calls" come from? Are these the only thinkable experiments?

It seems to me that "threats" (here) and "challenges" (from the previous section) are somewhat similar, maybe these two sections should be combined.