I find myself making these comments over and over again. Maybe it will help if I write them down.

A research article or a thesis is not the same thing as a mystery. Although it is true that when the research is begun, it is a mystery, presumably, by the time the article is written, much of the mystery has been solved. If you can fit your whole result in the first paragraph, that would be wonderful. If you can fit it in the first sentence, even better. If you can do it all in the title, everyone would love it.

Say up front what the paper is about and even what is the solution/contribution. It is very annoying to read 3, 4, 5 or more paragraphs about some general issue that we all agree on. For example, when writing a paper about a new microprocessor caching scheme that is good for streaming video, please do not begin with explaining the importance of video, streams, memory performance, and caches. By the time you get to the main point, I probably would have tuned out. Either I agree with the general motivational comments, in which case I will suspect that you have nothing new to say in the whole paper, or I disagree with your motivation, in which case I will be predisposed to disagree with any of your novel and clever ideas.

Say what you are going to say, say it, say what you said is the general outline of an article. Moreover, this is a recursive definition. So, first paragraph on the introduction should say what is going on in the introduction, i.e. what is going to be said in the whole paper. Moreover, the first sentence should be a summary of the first paragraph, which is a summary of the first section, which is a summary of the whole thing.

When referencing an article, please, please, PLEASE reference the author in the text and follow that by the citation. Do you say that you read the opposite in [14]. How am I supposed to read this sentence: "you read the opposite in left bracket fourteen right bracket" It is much better to say you read the opposite in a work by Shakespere [14]. Or, it is mentioned elsewhere [12] that citations should be superscripts making it even harder to read the sentence^12.

"It is my thesis, and so I should use the pronoun "I" in the text." I disagree. You can claim the idea as your own, but the experience of the reader should be that both the author and the reader, together are travelling on a journey of discovery. Let us figure this out. You see we have a problem. And maybe if we try this thing, it will work. No, no, that did not really work, but how about it if WE try the following idea. When you hand me drafts of your papers, I insist that your name be on all the pages as well as the date. I have too many documents and I lose track. Personally, in latex draft documents, I set the "pagestyle" parameter to include the edit date of the file that generated the latex output. I have a standard latex templete. I also have a little emacs macro which I store as ".emacs" in my home directory. Whenever I save or write out a file using emacs, if the extention of the file is "tex" then the macro searches the top 10 lines for a line starting with "\editdate{" and if so, appends the current date.

By the way, I don't really care how you produce your documents, but from my experience, pure ascii files have a longer lifetime -- that --> -- is, they are always readable, where as "word" documents are hard to read after several generations of software upgrades.