Musings

Roderick Bloem – Miscellaneous Stuff

My Real Home

I am at home at https://www.iaik.tugraz.at/person/roderick-bloem/. This page is for stuff that I don’t want on my regular home page. Some tips for students follow below. (Maybe it’s more of an attempt to save frustration on my part.)

How To Write an Exam

Many students get a lower grade than they deserve on their exam and sometimes it is their own fault. This is an attempt to help you avoid some pitfalls.

I will assume that there are questions (1,2,3) and sub-questions (1a, 1b, 1c).

  • Prepare.
  • No answer means no points, so answer all questions.
  • Explain your answer. You may get partial credit for an incorrect answer which is well motivated, you will not get full credit for a correct answer that is not explained.
  • Don’t hide your answers! I correct exams one question at a time, rather than one exam at a time. Answer all questions and all sub-questions in order. Clearly indicate when you are leaving the answer to a question open. Do not answer two questions on one sheet. Leave room between sub-questions (which can be on the same sheet) so that you can change your answer. Use lots of paper, and leave lots of space.
  • Write readably! Don’t let your handwriting upset the lecturer, no
    points for illegible answers.
  • Before answering the question, read it. After answering, read it
    again. Make sure that you have answered the entire question.
  • If you can, use the language you are most comfortable with to answer the questions.
  • Don’t write things in the top left corner of the sheet, where the
    staple will go.
  • If there is time, reread your exam. If there is more time, copy
    your answers legibly and neatly. You will find errors and ways to
    improve your answers.
  • Your first name comes before your last name, that is why they are
    called first and last name: “Roderick Bloem”.
  • Go to office hours to check the results of the exam. You will learn
    from your mistakes. Also, lecturers make mistakes when correcting exams. (But don’t try to negotiate your grade!)

How To Write a Thesis

This document is meant to gove some pointers at how to write a thesis. It is a guideline, based on best practice and my personal opinion. Many (most?) good theses deviate from my guidelines in some way. Yours can, too. But if you deviate, think about it carefully and then help your reader understand what you are doing by giving a clear indication of your structure.

Short tips:

  • Use Keith Andrew’s latex template. It contains some excellent tips on how to write, with one exception: Whatever Keith says, you do NOT EVER use the passive voice to refer to your own work. I will tar and feather you if you do.
  • Write the thesis in English.

Abstract

This is the first thing you write. It has your entire thesis in one
half a page. The entire thesis means: motivation (why?), problem (what is wrong with the world?), your solution, and evaluation. Realistically, you will have to write it again when the thesis is done.

Chapter 1. Introduction

The introduction is your entire thesis in, say, 20 pages. If someone reads the intro, she should have understood your thesis. Maybe not all the details, but nothing that comes afterwards will be a surprise.

This chapter is the second thing you write. It will need a major overhaul
at the end. Why do you write Abstract and Introduction first?
Because it forces you to think about the structure of the work and
the reasons for doing what you did. There is nothing worse than
writing four chapters and then figuring out you did not understand
the rationale for your work — you will have to rewrite a lot. (Especially because the rationale will be reflected in each and every part of a good thesis.)

The structure of the introduction is as follows:

  1. Background. What are you talking about? A new methods for public key cryptography? Then talk about communication security. A new algorithm for image recognition? Talk about the importance of image recognition is modern life. There is probably no need to explain to people how important computer science is, though, so limit your scope to what is relevant.
  2. The problem. Something was wrong in the world when you started writing your thesis. What was it? Why was it a problem? Here you state the problem and set concrete goals for a solution. For instance, people with quantum computers could break everyone’s communication. OR: we have a new algorithm, but we are not sure if it is secure. This needs to be solved to make the world a better place. Have concrete, measurable goals. This means a checklist that you can revisit in the conclusions to show which goals you have achieved and which you have not.
  3. Your solution. How did you address the problem? Describe what you did, and in how far it solves the problem. Describe drawbacks as well, but they should be secondary. Your contribution could be that now we know that the given algorithm is insecure. This being a Master’s Thesis, you are not expected to fulfill ALL the goals you set. You are not allowed to build suspense by keeping important aspects to the end!

4.Structure of thesis. What are you going to describe in which
chapter?

Again, the introduction is important. By the end of it, you have stated
that your field of research is important, that there was a problem and
that you have solved some aspect of it. In some sense, it is a sales pitch, but a sales pitch that does not exagerate anything.

Throughout this chapter and the other chapters, you will sprinkle references to related work to show that you know what is going on beside your work. You use references to back up claims (“X.509 is generally considered to be faulty [ImportantResearcher13]”), but also to give the other side of the debate without giving it too much weight (“although the importance is privacy in the FaceBook age is under discussion [Zuc12]”).

Under NO circumstance do you claim something without backing it up. If you have a wild personal claim that you cannot back up (“SSL is broken”) you use it as a working assumption (“In this thesis, we will examine alternatives to the SSL architecture that are immune to the possible criticism that … .”) A favorite trick of the trade is to downplay a claim. For instance, instead of saying that something is insecure, you can say that its insecurity has been argued by someone, or even that security is not obvious. The reader understands what you are really saying, and for practical
purposes, these claims mean about the same. (But don’t underclaim
either!)

Again, the introduction summarizes the entire thesis and does not leave anything important out. Ideally, when I have read the introduction, I know everything there is to know. The rest of the thesis is just details, details, details.

Chapter 2. Preliminaries

This chapter describes whatever the user needs to know to understand your work. That could be the syntax and semantics of temporal logics, the definition of X.509, or how internet routing works.

You have a chance here to show your advisor that you have looked beyond the assignment that you were given. However, take care that what you describe is relevant. Do your best to describe how the definitions you give are relevant to the thesis.

Chapter 2 does not contain anything new! Everything is description of existing work, but of course it is all in your words.

Chapter 3. Related work

The location of the related work is always a point of contention. It could also be the last chapter before the conclusions, or it could be a section in the introduction. You choose. In any case, you will have to embed related work throughout your thesis to explain the difference in details between your work and others and to explain where you have copied ideas.

Again, related work is a chance to show that you have looked around. You should ideally include any work that someone could think makes your work obsolete. You should not include unrelated work. For every paper you mention, you should explain the difference to your work. (This is important!) Try to tell a story, for instance chronologically, rather than giving a list.

Chapter 3 does not contain anything new. You can give some critique of the related work, which is obviously your opinion, but in this chapter, such a critique is usually not very long.

Chapter 4. Meat

This will likely be more than one chapter. You describe what you have done. Here, you describe new things. Note that the reader has a preconception about where new things are described. If you put a contribution in Chapter 2 or 3, chances are she will think it was done by someone else. If you put old stuff in Chapter 4 or later, you should carefully label it as such, lest you be accused of plagiarism. Both are possible, but be aware of the reader’s expectations.

In this section, you do not yet describe how well your contributions work in practice, but you do mention theoretical advantages and disadvantages. Thus, you do not say that your algorithm is 50% faster, but you can say that it removes the need for cumbersome authentication at the drawback of loosing the ability to know who
you are talking about.

Describe design decisions. Often, you have made important choices. Sometimes, you have done this because the choice was clearly better, sometimes it seemed better to you, and sometimes you just did not have time to solve all the problems in the world. Describe the choice, say why you took it, and mention alternatives.

You describe design decisions and drawbacks to hedge your bets. If you do not, a reader will come along that will spot the alternative or the disadvantage. Having found this point herself, she will vastly exaggerate its importance and will think less of you for having overlooked this important point.

In general, and not just in Chapter 4, you have make sure you do not lose the reader. Every chapter is layered like a onion: say what you are going to say, then say it, then say what you have said. The intro describes how the chapter fits the larger context of your thesis, how it relates to previous and upcoming chapters, and how it fits your goal. The summary does the same, but now with the assumption that the reader has read the entire chapter. Sound familiar? This is the structuire of your thesis repeated in small.

Claims. You will group arguments of (non-mathematical) claims by (a) statement of your claim (b) arguments against your point and why they have limited validity, (b) arguments in favor, and (c) conclusion (==claim).

Language.

  • Read Strunk and White. It will be an enjoyable and fruitful afternoon. Avoid passive voice. Call yourself “we”. The only exception: when you talk about established work “X.509 was developed in 1892”. If you write “a new framework was developed”, you are saying it was NOT done by you. Never say things like “SSL is considered to be secure” without stating who is doing the considering.
  • The most important thing in the paragraph comes in the first sentence.
  • Allows is a transitive verb. The system allows the user to drop dead, not the system allows to drop dead.
  • Look up the difference between “which” and “that”. Look up what quotation marks look like in english.
  • Your thesis will be rejected if you use $math mode$ to simulate \emph{italics}.

Plagiarism. If you are not quite sure about plagiarism, read up on it. You can find the “Harvard Guide to Using Sources” on the web. It’s not bad, even if it is meant more for people in the liberal arts. Let me know if you find something better. Keith Andrew’s thesis template also contains a section on plagiarism.

Avoid plagiarism. In particular, you can not copy large swaths of text, either verbatim or in your own words. You can copy small pieces of text if you put them in quotation marks and add a reference. Alternatively, you can copy ideas, if you describe them in your own words and explain how you got to them. Using your own words does not
mean that you change the formulation of a pararaph of text. It means that you write about something that you read last week, without looking at the original. You should describe very carefully who has done what before, whether you have copied ideas, used libraries, or have just been inspired. Be magnanimous with your attributions. Sorry for going on about this, but if you get this wrong, you get into major trouble.

Chapter 5. (Experimental) Evaluation

Here (and not in Chapter 4) do you describe how well your idea worked. Talk about the experimental setup. Give the results. If applicable, give tables and point out general trends. Describe the threats to validity. (“I looked only at Android apps; it is possible that they are fundamentally different from other mobile apps.”)

Chapter 6. Conclusions

Recapitulate the (entire) story in hindsight. State advantages and disadvantages again. There is no new content here, which means the conclusions will overlap the introduction to some extent, and also the rest of the thesis.

More Useless Things

This is a really cool picture that WordPress adds when you claim to be a university professor. (Presumably because this is what we typically look like, and what our offices typically look like.) It would be a waste to delete it.

Create your website with WordPress.com
Get started