As this page evolves, it will become a contents page for related articles that should be of concern to postgraduate students. For now, however, it is dedicated to research students, and relates primarily to writing M.Phil. and Ph.D dissertations.

M.Phil. and Ph.D Dissertations: Some Notes

I have just completed a 2-year M.Phil. course. I have also known various M.Phil. and Ph.D students from the Department of Computer Science at Keele Universiy. The following points arise from my own personal experiences and from discussions with them. Hopefully, they will be useful to you as you progress through writing your dissertation. They aren't in any particular order, but hopefully will be more organised in the near future. They also assume you're using Word, although I strongly recommend getting into LaTeX (go here for the files and go here for a good LaTeX editor) as soon as possible. It's complicated to use, though.

1. Put each chapter in a separate Word file. It tends to grow to a massive size (mine topped 35MB, and the only images used Word shapes). However, make sure that each file has exactly the same page margins etc. Otherwise, when you put it all together, Word will play havoc with the layout.
2. If you can, store the figures in a separate file, and add them at the end. Otherwise, Word tends to mess up their placing as you add and remove text, and sometimes refuses to let you move them to where you want.
3. Keep all your references in a database. Record every single scrap of information that may be useful, lest you hunt for them again at the end when you realise you've missed out something that you need for the references section. This information includes:
   - Article/chapter author(s) (Surname and all initials)
   - Artice/chapter Editor(s)
   - Article date (if applicable)
   - Article/chapter title
   - Journal/book title, author(s) and editor(s) in which it appears
   - The volume and issue number (if a journal) or ISBN number, publisher and publisher's city location (if a book)
   - Anything else you deem might be important (this varies between disciplines).
4. Make sure you know the reference style to use. A common convention to use these days is the Harvard style. Once you have over 100 references, it becomes tedious and time-consuming to convert them all to the new style - especially given that the main body of your work will probably reference some sources many times.
5. Keep regular back-ups of your work. Periodically (when you've added/changed significant amounts of text, or each day, for example) copy the file and rename it according to the date on which you saved it (e.g. thesis-22.10.03.doc).
6. Unless you have a good idea beforehand exactly what your specific research focus will be, work out what the most broad focal area is as soon as possible (for example, for agent-based software engineering, this would be software engineering). Secondly, if you can, make a list of keywords that relate to things you're interested in (for example, security, networks, AI, reliability). Thirdly, go from one end of the library to the other, reading the contents of all journals that may be appropriate and picking out those that relate to the keywords you have written down. When you find something that relates to this field, read the abstract to see if it might be useful. If it introduces you to a different aspect of your discipline, all the better - add it to your list of keywords to look for. Photocopy or make notes on anything of use. Don't worry or be put off if two articles say the same thing or contradict each other - that's good, because you have two sources that either back each other up or provide the basis for debate in your work. Don't be surprised if this method returns about 50-100 useful papers. You should spend a number of months at least at this stage, and it takes time but is necessary.
7. Don't neglect electronic journals and online journal databases. Once you have a clearer picture of what you're looking at and what all the terminology means, type some of the keywords you made (or even the most focussed discipline you know you want to look at) into the search engine and see what results it gives. Note those works that are most (widely) cited: these are likely to be the core papers that provide the foundation upon which the discipline, and thus your thesis, will grow.