A Simple Approach to Thesis Writing
(Adapted from "The Guaranteed Mackworth Thesis Formula", by
Alan Mackworth, revised by Tim Brecht, thanks to Ondrej Lhotak for an
addition to the Background and Related Work section)
Abstract (<= 1 page)
- one page stating what the thesis is about
- highlight the contributions of the thesis
Chapter 1: Introduction (~5-10 pages)
- Thesis Statement (one or two sentences)
- What is your thesis about and what have you done?
- If you have a hypothesis what is it?
- How will you test (prove/disprove) your hypothesis?
- Why is this problem you've worked on important
- Goals / Objectives
- What are you trying to do and why?
- How will you or the reader know if or when you've met your objectives?
- **** Contributions *****
- What is new, different, better, significant?
- Why is the world a better place because of what you've done?
- What have you contributed to the field of research?
- What is now known/possible/better because of your thesis?
- Outline of the thesis (optional)
Chapter 2: Background / Related Work (~8-20 pages)
- More than a literature review
- Organize related work - impose structure
- Be clear as to how previous work being described relates to your own.
- The reader should not be left wondering why you've described something!!
- Critique the existing work - Where is it strong where is it weak?
What are the unreasonable/undesirable assumptions?
- Identify opportunities for more research (i.e., your thesis)
Are there unaddressed, or more important related topics?
- After reading this chapter, one should understand the motivation for
and importance of your thesis
You should clearly and precisely define
all of the key concepts dealt with in the rest of the thesis, and teach
the reader what s/he needs to know to understand the rest of the thesis.
Chapter 3: Theory / Solution / Program / Problem (~15-30 pages)
- continuing from Chapter 2 explain the issues
- outline your solution / extension / refutation
Chapter 4: Implementation / Formalism (~15-30 pages)
- not every thesis has or needs an implementation
Chapter 5: Results and Evaluation (~15-30 pages)
- adequacy, efficiency, productiveness, effectiveness (choose your criteria,
state them clearly and justify them)
- be careful that you are using a fair measure, and that you are actually
measuring what you claim to be measuring
- if comparing with previous techniques those techniques must be described
in Chapter 2
- be honest in evaluation
- admit weaknesses
Chapter 6: Conclusions and Future Work (~5-10 pages)
- State what you've done and what you've found
- Summarize contributions (achievements and impact)
- Outline open issues/directions for future work
Bibliography / References
- Include references to:
- credit others for their work
- help to distinguish your work from others
- provide pointers to further detailed readings
- support your claims (if evidence can be found in others work)
- Ensure that ALL bibliographic entries are complete including:
authors, title, journal or conference, volume and number of journals, date
of publication and page numbers.
Be careful to at least be consistent in punctuation.
- Learn how to use a good typesetting program that can track and format
bibliographic references (e.g., groff, latex, frame).
- Within the text of the thesis, a reference
with a number of people can be referred to as
Lastname et al. (where et al appears in italics and the al is followed
by a period).
- My personal view is that URL's are not valid bibliographic references.
They and their contents change and they often contain material that has
not been refereed.
- Include technical material that would disrupt the flow of the thesis.
- Included for curious or disbelieving readers
Before handing in a copy of what you've written you should proof read
it and make corrections yourself.
Be critical of your own work when you do this.
You should think not only about syntax and grammar but about the
structure of the document and whether or not your are making good
arguments and whether or not someone else will be able to follow
and believe what you are saying.
You should repeat this process a large number of times before
you hand in a copy.
Far too many people type something in, print it out and hand it in.
If this is the case you as a student are not doing your job.
It is not your supervisor's job to write your thesis.
- Try to aim for around 100 pages or less.
- Including a glossary or list of acronyms may be helpful.
- Start thinking about what your contributions are early on.
- How is what you are doing interesting and important?
- How will it make the world a better place?
- What are you doing or discovering that hasn't already been done or
isn't already known?
- For many people it is best to start by writing the "guts"
of the thesis, Chapters 3, 4 and 5. In some cases the results and conclusions
may not be known (or may change) while doing these chapters.
- Chapters 3,4 and 5 can take on different forms depending on the thesis
and approaches being used.
- Sometimes design, implementation and performance are subsections within
chapters and the chapters are broken down by other criteria.
- Remember (especially those doing experiments) that you must include
enough detail in your thesis so that someone else could read your thesis
and reproduce your results - without ever talking to you.
- The word performance is by itself quite meaningless.
Stating that you've improved performance significantly does not
tell the reader anything.
There are problems with the word performance and the word improved.
Remember that there are often a number
of different performance metrics that can be applied to a system.
Instead of using the word performance state precisely what performance
metric is improved.
Also improved may also be potentially ambiguous.
State precisely what you mean.
The mean response time has been decreased by 20%.
Peak bandwidth has been increased by 40%.
- Try to get an outline and style guidelines from someone else
for the system you use for formatting your thesis.
- All figures included should add to the work. As such, there should
be text included that refers to the figures (preferably before the figure
is encountered). The text should explain what the reader should get from
the figure - what are they supposed to notice and what is the figure explaining.
Often people just include a figure with no reference to the figure and
no explanation of what the figure is for - if the figure was not included
no one would notice (this is not a good approach).
Note that when referring to a figure or a section by name they should
be capitalized as in -- Figure 3 shows the architecture of our system
or in Section 4 we describe the experimental methodology.
Newest pet peeve.
You do not write 3.3GHz or 16GB you write 3.3 GHz or 16 GB.
This is equivalent to saying you are 6feet tall (instead of 6 feet tall).
- Using and misusing abbreviations.
- The word "it's" is an abbreviation of "it is" it
is NOT a possessive form of "it".
- The abbreviation of the phrase "for example" is written
It contains a period after the "e" and one after the "g".
A comma is also usually required with its use.
This is an sentence that uses "for example"
(e.g., this is how to use for example).
Quite often it is enclosed in parentheses and you should
avoid using it too often.
- The abbreviation of the phrase "that is" is written
It contains a period after the "i" and one after the "e".
A comma is also usually required with is use.
This is a sentence containing an example of how to use "that is"
(i.e., this sentence is the example).
Quite often it is enclosed in parentheses and you should
avoid using it too often.
Don't use the abbreviation "etc.".
The use of etc. is usually an admission of ignorance.
It is like admitting that the list you've given is not complete
but you don't know what is missing. If you did know what was missing
the list would be complete".
You should purchase and use a book like
"The Elements of Style" by Stunk and White.
- You must not make minor modifications to someone else's work
and include it in your own work.
- If you want to explain someone else's work the best approach is to
read it over, put it aside, and then write in your own words
what that work is about
(do this without referring to the original source).
To Tim Brecht's Home Page