You only have to read a few draft dissertations to realise that the majority of compscis have not written an essay for several years, quite often their GCSE year. Since I quickly tire of pointing out the same repeated mistakes (and I am sure making the corrections must be equally tiresome for the supervisee), this year I thought I would try a pre-emptive strike and give some advice before people start writing.
Disclaimer: I am in no way an expert in the use of the English language. Since it is being written at 3:30 in the morning, this page itself almost certainly contains heinous errors, typographical and grammatical. However, a dissertation is considerably more important than a quick webpage aiming to save a few people some time and hopefully will not be written in the small hours of the morning...
The most important thing to remember is that a dissertation is a
formal document, not an e-mail to your supervisor or DoS! Try to keep
your language formal, do not use contractions (doesn't, can't, etc)
or short hand abbreviations such as "e.g.". At school, my
science teachers always insisted that scientific reports were
written in the third person (the beaker of water was heated to 100
degrees C, the programme was tested, etc). While this is
a matter of style, writing well in the first person is more
difficult as at some point the author invariably lapses into:
I did this and then I did that. I then did this and I discovered
that.. which quickly becomes very repetitive.
An item in a bulleted list should only start with a capital letter if it constitutes a complete sentence (or several sentences). Correspondingly, any item that starts with a capital letter will end with a full-stop. If the items do not constitute a sentence then it must be a list and should be punctuated accordingly. For example:
Use LaTeX. Word processors are great for arts students writing essays but real scientists use LaTeX for good reason. (FIXME: Explain what these good reasons are.) It is not difficult to learn, especially if you are already familiar with markup languages such as HTML.
Since I have decided to write this page and you have taken the time to read it, I suppose I should also say something about how to set about generating the content of a dissertation.
Undoubtedly the best way to get a handle on what you are expected to write is to go to the CL library and skim read a couple of past dissertations which you know were marked highly and see how their author set about writing up their work. In general terms you should describe what you did, why you did it that way, problems you encountered and how you solved them. Remember to cover any algorithms you used, or invented (even if you later discarded them for better ones!) but an intimate description of every single one of your classes will quickly send the examiners to sleep. UML diagrams are useful for describing your class structure in a non-boring but detailed manner.
Include as much detail as you can in the first draft - your supervisor can advise you on what to cut out, but unless they have been following your progress very closely, they will find it harder to know what work you have done that should be included.
A good plan is to send a draft of the first two chapters to your supervisor early in the Easter vacation so they may comment on your style and approach before you make too much progress, thereby avoiding the necessity of major rewrites should there be any major problems.
It will take your supervisor time to read a complete draft, especially if they have several Part II students to supervise, so the first day of Easter full-term should be considered a good time to submit a complete draft. While this deadline can be quite flexible, students that fail to submit a complete draft by the end of the first week of term frequently find themselves with insufficient time to implement all of their supervisors comments.