Iainõ?Ts Guide to Thesis Writing Jan 2000-01-06 Iain Woodhouse STRUCTURE Like all good story telling (and a thesis is just that õ?" telling the story of your research) a thesis should have a well thought out structure. Donõ?Tt underestimate the importance of this, as a thesis that has a solid structure is one that is easier to write, easy to add too and most important, from the point of view of the examiner, is easy to read. The obvious structural scenario to work around is the literary one of beginning, middle and end. (Too obvious, you might think, even to mention, but youõ?Td be surprised!) The boundaries between them are not always perfectly defined, but the steady progression is always apparent. Consider a book, or a play or a movie that you like, and youõ?Tll notice that they will generally follow this straightforward structure with common elements to each part. BEGINNING The Introduction is possibly the most important chapter of a thesis õ?" as is the opening of any story: the first chapter of a book, the first few scenes of a movie. At the beginning, a number of things must be done õ?" effectively we must set the scene. The main characters are introduced. They are put into some kind of context of time and place, and some hint is given to what will happen in the main bulk of the story. The plot is not given away, merely suggested, but all the information we need to know to follow the story to come is presented at the beginning. Within the first few minutes of a movie we should know what the rest of it is going to be like õ?" a love story or a thriller or a horror movie. In the same way, your thesis must introduce the main players õ?" the topics that will be relevant to the research story õ?" and put them into the context of the wider subject matter through a thorough but well focused and concise, literature review. An introduction should also present the reasons for the study, where you expected it to lead to and your rationale for the approach taken. The introduction need not summarise the results, but it should allude to where the thesis eventually ends up. MIDDLE When the scene is set, and the characters introduced, we come to the meat and bones of the story. This is where things happen, the plot unfolds and the characters actually do something õ?" they interact, rather than simply being isolated characters. In your thesis, the main body of the text is about what you did, how you did it, why you did it like this and not like that, and so on. It does, of course, include your results, but donõ?Tt be tempted to pad the thing out with results. Make sure you use succinct, concise and efficient ways of representing the data. Select those results that are relevant to the discussion õ?" large sets of results can be summarised statistically, without having to present multiple data sets that donõ?Tt add to the flow of the thesis. But donõ?Tt simply select all the good ones!! Results that are contrary to a theory or expectation are important õ?" they are the ones most likely to tell you something interesting (ie, something you didnõ?Tt know already). And by convincingly explaining such results you both contribute to your main arguments (because you offer up alternative scenarios but can explain them) as well as help demonstrate your understanding of the theory/experiment. Above all, present your evidence in as clear and straightforward manner as possible. To paraphrase Einstein, "Make things as simple as possible õ?" but no simpler". END The Concluding chapter (which may be entitled Conclusions, Summary, Discussion, or whatever) is probably the next most important chapter after the introduction. These two chapters together ultimately add value to the straightforward presentation of results that make up the middle chapters. Everything that is important about your work should be gleaned from the first and last chapters. This is both good writing practice (so that published papers should be similarly structured) but it is also the case that a lazy examiner may only read (properly) these two chapters. In story telling, the END captures the essence of the story, rounds off all the loose ends, and helps define the whole point of the story ( ".. and the moral of the story isõ?Ư"). Similarly, a thesis conclusion must summarise the research, the method and the results, capturing the essence of your thesis. More than that, it must take the results and, by bringing together all the previous smaller conclusions made throughout the thesis, deduce conclusions that encompass the specific results of the thesis as well as the wider implications to the subject as a whole. Effectively, you must answer the question, "What do all these results mean." It is not sufficient for a concluding chapter merely *****mmarise the content of the thesis. Another important feature of the ending, is the idea of the open ending õ?" for movies this is ultimately with the aim of allowing the possibility of a sequel, and we can all think of good examples. In a thesis, the open ending is to address further work . A thesis is a discrete piece of work, but research goes on forever (which sounds more profound that it really is!). It is a measure of how much you learned through doing your thesis to be able to see what you might have done differently, what you should have done differently, and what you would now do if you had a further 3 years of funding to work on the subject some more. From an examiners point of view, it also helps to see how much you ultimately understood about the subject and the process of research. THE OTHER BITS Besides the main structure of the thesis described above, you also have to add certain add-ons, which have a bit of a life of their own and so deserve some explanation. ABSTRACT An abstract is a funny piece of text, and it often takes some time getting used to writing them. They are strange because they give away the plot. While a good movie trailer or the back-cover summary of a novel gives you an overview of the story, it shouldnõ?Tt really tell you what happens, whereas the purpose of an abstract is to do just that. It shouldnõ?Tt try to "tempt" people to read the thesis (or paper) but rather concisely describe the fundamental features of the methods employed and the results obtained. It should be dry and straightforward, and give enough information that if someone only read the abstract, they would know instantly whether they wanted to read the whole text or not. If you take a look at some published papers, youõ?Tll see what I mean, although take care, as there are plenty of bad examples of abstracts out there. PREFACE Another funny beast. A preface is not an introduction, nor an abstract, nor a summary. Personally, I think the preface is the place where you get to say what you want, in a manner that best suites you õ?" write more informally, and feel free to express your own opinions. It is perhaps the only part of a thesis you can express your own personality without condemnation õ?" no one is likely to (nor should they) fail a thesis on a bad preface! Usually a preface begins with some interesting personal reflections on the thesis topic and the very wide and historical context of the subject. This then leads to a summary, chapter by chapter, of the contents, and a quick summary of why the thesis might be important. It is often the case to include acknowledgements here, but putting them in their own slot after the preface is also acceptable. APPENDICES All the stuff that is not published but would be useful: software code, proofs, interesting sideline studies that you have done that are not directly relevant to the thesis topic, or would disrupt the natural flow of the main text. Include here extensive data sets, or results and glossaries or lists of symbols. It is also acceptable to include reprints of papers that you have published during your thesis as appendices, but only if they are relevant. FOOTNOTES, ENDNOTES The use of footnotes and endnotes are ultimately a matte of opinion. In general, if something needs extra explanation, or elucidation, but to do so in the main body of the text would interrupt the flow or thrust of an argument, then a NOTE is good idea. Whether these should be footnotes or endnotes will be determined by the official formatting guidelines set out by the individual institutions. POSTSCRIPT, AFTERWARD These are effectively the equivalent of the PREFACE, but they come after the main body of the text. In general, anything that could be included here should have been dealt with earlier in the thesis. They should really only be included if there is something that turns up at the very last minute õ?" too later to change the text of the thesis. You may, for instance, want to (or be asked to) add a POSTSCRIPT or AFTERWORD after your viva if there were significant comments made by your examiner that would contribute to the thesis as a whole. The other rare occasion would be that some new piece of research is published between the time of your first submission and your viva õ?" you might then wish to add an AFTERWORD before binding your thesis to make reference to the new work, and how it ties in with your own. GENERAL COMMENTS SIZE DOES MATTER In a thesis, size does matter. A thesis can be too short or too long, but there is no absolute standard for this. The length will depend more upon the subject and your style of writing, than upon how much actual work you do. Examiners are never fooled into thinking that a big thesis means you did lots of work. As a rough guide, for single-sided, double-spaced, A4, you would expect a 6-month MSc project to produce a thesis about 50-80 pages long. A 12-month research MSc thesis, a little longer, say, 70-100. A PhD, from about 100-200. (For 4 year PhDs it could be as much as 300, but 180-250 is probably more realistic in most cases). Large Appendices may bring these numbers up a little. When it starts becoming two volumes, it is too big. You should be able to be focused and concise, and so keep it under 250 pages, max. A two volume MSc thesis should be, in my opinion, asked to be re-submitted without even looking at the text. I did see one such thesis once, which contained a series of photos along the lines of, "Fig 1. About to dig some soil samples", "Fig. 2. Digging some soil samples." "Fig. 3. A soil sample is put in a bag"õ?Ư etc. Each photo took up a single page!! Remember, a 6-page published paper may describe the equivalent of 6 months, or even one yearõ?Ts work. There is no reason why your thesis shouldnõ?Tt be equally focused and concise. DONõ?TT WRITE FOREVER There is always the temptation that when you reach the end of the first draft of your thesis, you want to re-write the first couple of chapters. Be careful of such an approach, as what tends to happen is you spend another couple of months doing those chapters, which gives you enough time to start questioning the latter chapters. So you decide to spend a further couple of months doing themõ?Ư by which time, the introductory chapters are starting to look a bit out of date againõ?Ư and so on, until you are 12 months behind schedule. Itõ?Ts a positive thing to want to produce the best thesis possible, but a late thesis is always a bad thing. In 10, 20 years time, when people are looking are your career record for a professorship (!) theyõ?Tll notice the 12 month delay before they get a chance to read the thesis. A PhD thesis is not just about the work itself, but is also about how you do it, and about completing the work within a reasonable time period. FIGURES Note: colour pictures can cause a lot of hassle. Do good quality versions for the submitted thesis. Do them on a single page each, this way, if the text on a previous page is changed, you donõ?Tt have to re-print an expensive glossy print. The simplicity of colour plots has also made researchers lazy with their graphics. Colour is not always the easiest way to display some results or data õ?" sometimes colour is added as a poor substitute for thinking about how best to display the results. Always think twice about how you represent data and ideas, especially if you are footing the bill for the glossy prints! SCIENTIFIC/JOURNAL PAPERS Writing a good scientific paper follows all the same ideas as outlined above. It is, after all, a bit like a thesis in miniature. There is one important difference, though, which is that a paper is not for examining the authors. In this sense, it should be more concise. It can assume a greater knowledge of the reader (as you are not trying to convince that reader of what you know). And the ability to be very focused with your results and conclusions is even more important. Ask yourself: What is the purpose of this paper? Unfortunately, the desire (need?) to get work published often clouds the issue of this question. In some respects a paper should allow repeatability õ?" you must give enough information that the reader can come to the same conclusions. In this context it need also be convincing! It should also be a source for authors to find other good literature on the subject. Above all, it must communicate. It should communicate your ideas, your theories, a set of results and your understanding of them. If it doesnõ?Tt communicate your work and insights to the reader, then it is ultimately worthless. When you plan to write your first paper, my advice is to choose a published paper that you like and enjoyed reading (and was preferably also quite logical and straightforward) and õ?Ư well, donõ?Tt copy it, but do use it for inspiration. Use it to guide your style, the level of the content and the structure. Itõ?Ts the best way to learn!. CHECKLIST Here is a final checklist for your almost finished thesis. Ask yourself the following questions: About the contentõ?Ư Understanding the topic: Have I shown that I understand the topic of the thesis? Is there anywhere in the thesis that it looks like I am covering up a lack of knowledge? Understanding of the wider context of the research: Do I know the relevance of this work within the wider aspects of the subject, and to other topics? And have I conveyed that in the introduction and conclusions in the thesis? Existing literature: Does my introduction give a good overview of the recent work done on the same subject? Do I also cite the older, but classic, texts? Do I make proper use of citing other work throughout the thesis? Have I over-cited õ?" ie, cited every possible text?! Have I cited anything that I have never actually seen (a bad idea)? Logical reasoning and analytical ability: Does my thesis convey a logical processes of reasoning, deduction and inference? Are my conclusions logical and reasonable deductions from the results? Are there other alternative conclusions that could be drawn? If so, do I discuss these possibilities? Critical thought: Does my thesis highlight some of the weaknesses of previous/current research? Do I consider the weaknesses of my own approach? Further work : Do my conclusions include a discussion into where the research should go next? Does this include identifying the shortcomings in the thesis, and how they could be improved, as well as seeing the possibilities of ad***ional research? About the presentationõ?Ư Structure: Is the structure of my thesis logical? (i.e. beginning, middle, end)? Have I cluttered up the text with things that could go in an Appendix (such as software code, or reams of results that add nothing to the understanding of the thesis)? Is it too long? (Examiners donõ?Tt like long thesesõ?Ư and it implies you canõ?Tt be concise) Presentation: Does it look good? (It neednõ?Tt be a work of art, but it shouldnõ?Tt be a shambles). Have I followed the official guidelines on format, etc? Graphics and data presentation: Do my figures and graphics help to illustrate a point, or communicate an idea, or do they simply look nice? Worse, do they confuse rather than clarify? Use of language: Have I run a spell checker? Has someone checked my grammar, at least for the abstract, introduction and conclusions? (These are the most important in terms of language) Name: Anh Kiet Nguyen Title of Ph.D. Thesis: "Phase transitions in extreme type-II superconductors: Topological defects, and dual description of the vortex system" Supervisor: Professor Asle Sudbá Institution: Norwegian University of Science and Technology Date: 1999 -------------------------------------------------------------------------------- Abstract: The magnetic field versus temperature phase diagram for type-II superconductors is investigated using large-scale Monte Carlo simulations of the Ginzburg-Landau model in the London, frozen gauge, and both London and frozen gauge approximations. We show that the uniformly frustrated 3D XY model provides an excellent description of the magnetic field versus temperature phase diagram for extreme type-II superconductors at intermediate and low fields. In zero magnetic field, we find that the superconducting - normal phase transition of type-II superconductors is exclusively driven by a vortex loop blowout. This blowout is a 3D counterpart of the 2D Kosterlitz-Thouless vortex - anti vortex unbinding transition. Amplitude fluctuations are irrelevant and do not modify this blowout picture. The size of the critical regime is set by the mean-field crossover temperature TMF, where Cooper-pairs would start to form at the mean-field level. When kBTMF/J0, the transition is well described by the mean-field approximation. Here, J0 is the energy scale for the Bose-Einstein condensation temperature. In finite field, we find that the vortex-line lattice melts via a first-order melting transition into an incoherent vortex liquid, where global phase coherence is lost in all directions. The vortex lines in the liquid undergo frequent cuttings and recombinations, and are therefore not entangled. Furthermore, we find two distinct scaling regimes for the vortex-line lattice melting line: 1) a lines-only scaling regime where interactions between the field induced vortex lines dominate and drive the melting transition, and 2) a critical scaling regime where thermally excited vortex loops dominate and 3DXY critical scaling dictates the melting line. In ad***ion to the vortex-line lattice melting transition and the Bc2 crossover, we find strong numerical support for a novel U(1) phase transition within the incoherent vortex liquid phase. It represents the finite field counterpart of the zero field vortex loop blowout. The U(1) phase transition is characterized by a subtle change in the connectivity of the vortex tangle, driven by a proliferation of topological defects in the form of vortex-loops. In the low temperature phase, the connectivity of the vortex system across the superconductor is determined by field induced vortices only. Above the transition temperature, there exist vortex pathspenetrating the superconductor in all directions. The effective vortex-line tension, or equivalently the free energy per unit length, has been driven to zero across the vortex-loop blowout transitions. The vortex-line tension may therefore serve as a``generalized stiffness" characterizing the low temperature vortex-liquid phase, and which is destroyed in the high temperature vortex-liquid phase. The vortex-line lattice melting line and the U(1) line seem to merge at low fields and terminate at Tc, the zero field critical temperature. In this regime, descriptions of the vortex system only in terms of field induced vortex lines are inadequate at and above the melting temperature. We show that in 3D and zero magnetic field, the coupled-gauge symmetric (charged) Ginzburg-Landau model has a U(1) symmetric dual counterpart, and the U(1) symmetric (neutral) Ginzburg-Landau model has a coupled-gauge symmetric dual counterpart. Thus, the Ginzburg-Landau model is not self dual. Furthermore, we show that the Ginzburg-Landau model and its dual counterpart belong to different universality classes, i.e. the 3DXY and the inverted 3DXY universality classes have different sets of critical exponents. The critical exponents for the 3DXY and the inverted 3DXY universality classes are given by ( upsi ~ 2/3, etapsi ~ 0.04) and ( uphi ~ 2/3, etaphi ~ -0.18), respectively. Here, u is the critical exponent for the coherence length and eta is the anomalous dimension of the order field. The subscripts psi and phi denote the order field for a U(1)-symmetric model and a coupled-gauge symmetric model, respectively. Using the Ginzburg-Landau order field psi and its dual counterpart phi, the magnetic field versus temperature phase-diagram of clean extreme type-II superconductors consists of three regimes at low to intermediate magnetic fields. We denote these regimes as I, II, and III. The regimes have the following characteristics: I) A vortex-line lattice phase with <psi> > 0 and <phi> =0. Here, there is long-range superconducting phase coherence along the direction of the magnetic field, but the dual field phi has not yet condensed. The physical meaning of the latter statement is that thermally excited vortex loops are confined. II) A low temperature incoherent vortex-liquid phase where <psi> = 0 and <phi> = 0. Here, phase coherence is lost in any direction and vortex-loops are still confined to small sizes. We may view this phase as a liquid phase of {it field induced vortex-lines}. III) A high temperature vortex-liquid phase where <psi>=0 and <phi> > 0. Here, phase coherence is still lost in all directions, but the dual field now has condensed. The latter statement corresponds to a vortex-loop blowout. We view this phase as a vortex-liquid phase where the picture in terms of a liquid of field induced vortex lines has broken down. -------------------------------------------------------------------------------- Contact information: Anh Kiet Nguyen Department of Physics Indiana University Bloomington, IN 47405, USA Sayonara!!! Good Night, sleep tight, don't let the bed bugs bite.
Notice in the end: Name: Anh Kiet Nguyen Title of Ph.D. Thesis: "Phase transitions in extreme type-II superconductors: Topological defects, and dual description of the vortex system" Supervisor: Professor Asle Sudb?á Institution: Norwegian University of Science and Technology Date: 1999
