I've also seen the MS thesis of Han The Thanh. It's good. But I would need something more suited for natural sciences that accomodates plenty of figures and references. Also I would like to "cloak" my thesis by avoiding Computer Modern.

LaTeX is developed for mathematical sciences by a mathematician. It works beautifully with complex equations and embedding maths into text. Everything in it -- including the CM fonts -- screams "mathematics!". ConTeXt is developed for typesetting manuals. Some requirements are the same, for example the need for very solid and reliable cross-references. Some requirements are different. Manuals and other documentation require more control over the layout and a very good support for illustrations. Writing a thesis in science falls somewhere in between. If you are writing something about quantum mechanics, I'd say you should use LaTeX. If you are living on the fringe and doing something more computer-oriented (or even chemical), the intentions of ConTeXt developers and you converge better than with LaTeX. I wrote my PhD thesis with LaTeX, because that was the only viable alternative back then (four years ago). Now I think I'd use ConTeXt due to its better layout handling. [For the curious: the thesis is available at http://lib.tkk.fi/Diss/2001/isbn9512257270/ I can supply the LaTeX sources plus a document telling about the tools and rationale behind, if someone needs it. But be warned, it is not up-to-date information.] --- Someone suggested using a ready-made template. Sure, if you can get one somewhere with all the bells and whistles suitable for your academic institution, take it. Be it ConTeXt or LaTeX, you'll save a lot of trouble using a proven platform. Beautiful documents with minimal effort. That's exactly what the universities should supply. Very unfortunately, there are seldom any good templates around. Very often there are no *TeX templates around, regardless of quality. There may be some half-hearted Word templates around, but at that point ducking or running -- whichever is more convenient -- is a good alternative. And if there are no good templates around, then you'll need to roll your own. One which fulfills all requirements set by the administrators (all those lovely forms) and is easy-to-read and looks nice. One of the mixed blessings (=curses) brought about by computers is that now you need to be an academic writer, a typesetter, and a graphic artist at the same time. At this point the number of should-knows explodes. On the typography side you have to know a million small things. What kind of quotes to use, when to use en dash, when to use em dash, how to hyphenate, how to type numbers and units. Then you need to decide on the typefaces to use, and the general layout of the page. After all this is understood, then the technical problems have to be overcome. You will need to find the fonts you want to use, and you will need to be able to explain all this to the type- setting system in use. Also, you have to be able to draw the illustrations, with the requirement that they are both aesthetically pleasing and help to convey the message they carry. To do this, you need the software to draw the illustrations and the file formats to transfer the drawings to the typesetting system. After all this is done, you can try to produce the printed book. This may sound trivial, but is far away from that. There may be problems with file formats (with an old version of pdfLaTeX I managed to make a file which crashed a commercial RIP system, this happened a few days before a big deadline), or at least with color matching if nothing else. --- Does the above sound easy? Not for me. Still, you are expected to make a professional-looking document in a situation where you need to know half a dozen professions completely unrelated to your own scientific field. The result is that most people resort to using the Word and a lot of other four-letter words. The documents range from hideous to just slightly ugly, and are extremely fragile and difficult to maintain. --- What I am trying to say is that choosing the tool is only one part of the project. Whichever tool you choose, you'll end up in trouble at some point. With some tools (WYSIWYG) you'll end up in worse trouble, but even with *TeX the road is bumpy at best, unless you really have a tested and proven templates which you can use. The bumpiness has (IMO) slightly different nature in LaTeX than in ConTeXt. In LaTeX there are a lot of great packages and a lot of documentation. Books, web resources, mailing lists, etc. The huge user base makes all this available. However, finding the right packages and debugging some odd interactions and conflicts between packages may be very difficult. ConTeXt gives a lot more control over layout in itself without any packages or modules. And if you want something more exotic, take metapost and do it. But the downside of this is that as ConTeXt is a new system and has relatively fewer users, the state of documentation is not very good. Even the official documents are sometimes out-of-date. (This is really not to blame Hans or anyone else. Hey, we get free high-quality software! But this is something that has to be acknowledged and in the long run acted upon. By us users.) Whichever you choose, you'll have moments when you wish you'd chosen differently. --- Sorry, a long posting without really any good answer. The only suggestion I give is that try to make the layout rather early in the process. That way the writing is more fun, as you can see the outcome already during the writing process. And if you cannot stretch the tools to make the layout, you can change the tools without too much trouble. Good luck with your thesis! - Ville