Writing Your Thesis: Suggested Strategy
First, consider this:
nobody will read your (or anyone’s) thesis once you complete it!
have become an expert in some area. To share what you have learned, you need
to present that material to people in ways they can digest given their busy
lives. You need to make it easy for them to understand what you did.
you are designing an interactive system, very few people will ever see it
– they will read about.
you are designing an interactive system, even if they see it, the people who
saw it need to write about it to convey what you did to others.
you need to write what you do, and you need to write it in a format that
others can understand and appreciate in a reasonable number of pages!
Overview of the suggested strategy:
Don't write a thesis. Write a paper (or papers) for conference or journal
submission that you turn into a thesis by adding an extensive set of appendices.
- You will be forced to describe the important parts of your work (i.e. the
new ideas) ... concisely.
- You will avoid writing a tremendous amount of text that nobody reads that
muddles your main contributions.
- You will be well on your way to getting your work published so that
researchers will actually read it.
- You will acquire the skills necessary to describe your work to the
A good paper will very clearly explain what about your work...
a surprising, unexpected result
a problem that nobody has solved before
be proven to work substantially better than any other solution
anyone has tried
some questions that nobody has thought about before
A good paper will also concisely argue how your work compares to other work
that came before it, specifically answering the following questions:
have you advanced the state of our understanding of the problem?
do you need to credit for similar or motivating ideas?
So, step by step, here is what I recommend (learned the
Part 1: Figure out who will care
your audience (research communities that will care). You can do this by
looking at the papers you have read on prior work. Where were they
published? Follow these tips for doing a
the conferences or journals that represent the communities that would be
most interested in your work. If you are working with Stephen you should
take a look at this list of conferences
and journals. This should be easy to do after your literature search,
because you will probably find the key papers cluster in a few publications
addition to the papers you have read on prior work, peruse some recent
proceedings/volumes and study how people write up their work and what types
of research are published in conferences or journals you have initially
your advisor, pick one conference or journal with a deadline in a
reasonable time frame to aim for.
out the formatting and length requirements for this conference/journal from
its website (usually this will be something like 6-10 pages in
single-spaced, double column 10pt font for a conference and longer for a
rough outline of your thesis in this format. Consider these thesis
Part 2: Identify what will make your work stand out in the research
You should have a good idea what these points are from your extensive background
reading and literature search. At this point you should write an extended
abstract, following your outline (in the format of the targeted submission). Make
sure your key idea and supporting ideas come through crystal clear in that
abstract. What differentiates what you are doing from prior work, and how
will you evaluate your results? This step will require a back and forth
between you and your advisor, with you making extensive revisions to this
extended abstract. This will also likely cause you to revise your research plan
Part 2.5: Finish your research! (or make significant progress)
Part 3: Write an outstanding conference/journal submission
This will turn into your thesis, but do not think about "thesis"
writing yet! Stick to the guidelines for the submission and target the deadline
if it is prior to the completion of your thesis (the best case scenario if it
will force you to think about how to succinctly and compellingly describe
your main ideas/discoveries.
will produce a document that the research community can easily learn from
you are finished, you can easily submit your work to the conference/journal.
it gets in, you will have multiplied the number of people who will know
about all your hard work and great ideas many times over.
you ever want to pursue more research, your submission will help you
it doesn’t get in, you may get some interesting feedback on your ideas
(or how you presented them).
someone asks you what your thesis is about, you will have something to
give them they might actually read.
advisor is likely to be much more interested in helping you refine your
writing for a conference submission that for the actual thesis document.
when you complete this paper, the hardest part of writing your thesis will
Part 3a: Things your paper will need to have
a few images and diagrams that will need to be exceptionally clear and
tone/style/format that is consistent with other publications in your field(s)
complete but concise references cited section
should satisfy the needs of the "review criteria" listed for the
conference/journal you will send the paper off to. These criteria will help
you keep your discussion focused on the important stuff.
Part 3b: Distribute your paper to as many people as you can get to read it
and get their feedback on
Hopefully the timing works out so you submit this paper 1-2 months before you
these is actually due. Then you can move onto completing your MIT thesis
Part 4: Quickly finish off your thesis
thesis becomes your conference/journal submission with an extensive set of
submission part contains:
the main ideas, images, etc. needed to describe the most important points.
can essentially stand alone.
rest of the thesis, the "Appendix" part contains:
that only a subsequent "implementer" is likely to care about, e.g.:
examples of the solution, datasets, etc.
of study design that would be excluded from a short submission
descriptions of prior work
references cited not a bibliography
cited lists readings that you have referenced in your work as relevant to
understanding your research.
you are not going to explain how it fits in with what you have done, why
Stephen Intille's Thesis Development and Writing Tips
Last updated: 04/22/06