Area Exam tips and tricks

a.k.a. I wish someone had told me...

This is now an old version. For the latest version, check out the grad student wiki

Use LaTeX: There is a learning curve, but TeXShop and others can help with that. But more importantly, everyone uses it and it makes your papers look better and more professional. And it will never crash and wipe out all your hard work.
Use BibTeX: This is also a given. Just do it. It sucks, but it sucks less than every other option. All future advice is predicated on you following this advice.
Use Makefiles to make your document: This is far more helpful than anyone who doesn't do it already gives it credit for.
Don't just download PDFs and drop them in a directory. Spend the extra minute or so to drop them in some kind of filing system, be it JabRef, BibDesk, a plain ol' BibTeX file, or your own custom thing. But make sure that whatever you use can generate/uses a .bib file.: If you don't do this, then your directory of "to file and read" will grow without bound and you'll have to spend 40+ hours finally collating and putting everything into your system and it totally sucks because you have to search for every title to figure out where you got the damn thing from and its proper citation. Ideally this citation system will also allow you to store your own annotations and reviews and keywords.
Use bibunits to get a section-by-section bibliography in your proposal.: It really helps if you can just drop bibliographic entries into the mix kind of willy-nilly and look at paper titles almost immediately after their relevant section.
Use the abbrv bibliography style whenever possible.: This bibliography style covers up a multitude of sins and makes your bibliography look much more consistent and professional without requiring any further work on your part.
Use revision control¹: ...and preferably something better than vanilla RCS. Not only do you have the advantage of not risking the loss of good text during edits, but you have automatic documentation of all of your progress over the weeks or months during which you are writing your paper. CVS, GNU Arch, Subversion, and DARCS are all recommended. That ability to work from multiple machines is a major plus.
Whenever you read a paper, scan through its references to see if there's any other papers related to your area that you should read. Add them to your to-read list. ²: When you're scanning through the references and realize you've already read most of them, you're ready for your Oral Exam.
Keep your to-read list organized with keywords.²: Be able to distinguish papers that are directly related to your area, those that are somewhat related, and those which merely look interesting. One of the keywords should be "Oral Exam" which designates that you plan to discuss the paper in your position paper. If your adviser or a committee member suggested the paper, you might also want to note that and give it higher priority.
Make sure your committee members get along with each other.²: Whenever you're putting together a committee (DRP, Oral Exam, or Dissertation), make sure your committee members get along with each other. Ask other current or recent graduate students about potential committee members and any experiences they've had with them.
Ask direct questions early and often²: Early in your process, ask direct questions of your adviser and committee members like "Do you think this project is worthwhile?", "Do you think I have set an achievable goal?", and "Do you think the scope of this project is too big or too small?". The last thing you want is to invest years of your life into a project which you will not be able to finish. If you don't directly ask, people may not tell you because they're trying to be polite and not burst your bubble.
Get your committee together and involved early.³: Involve them in the process. Get the plan approved early, and float an outline of the paper to make sure it meets requirements.
Get things in writing. ³: The profs tend to tell you to do one thing, then when they read it, they tell you to do it differently, sometimes exactly the way it was before. Try to get suggestions in writing.
The "related work" web has unbounded width. ⁴: You have to decide where to truncate it. No one else can.
Avoid the temptation to read things because you "think they're probably related". ⁴: Thorough is good, and in the early stages, there's a place for fishing, but after a couple of months, focus is essential.
For everything you read, articulate, in writing, why it is relevant to this exam. ⁴: A journal is useful here.
Understand what story you're telling. ⁴: The position paper is not just an annotated bibliography running into dozens of pages. Identify clearly what the problem is, what chunk you're biting off, how to think about this problem and what the major lines of attack are. Everything is focused around that central story.
For the final write up, get some examples from your fellow students, and read them. ⁵: Especially those who have the same advisor. Examples of others who have passed with/without distinction will help you know what is expected in your final paper.
Put the work into writing summaries of each paper immediately after you finish reading them, in your own words. ⁵: Then the work of writing the survey paper is a simple exercise in organizing your own summaries. (Huck got this from TomB)
The work of the oral comprehensive is in doing a good job reading and annotating and UNDERSTANDING the papers you read. ⁵: Put the work into the data collection and writing a coherent paper. The presentation, and more importantly, the private questions from the committee, will go much easier. If you don't understand the material in a paper, don't assume that it won't matter - find some background material (even Wikipedia is better than nothing).
Don't forget that the most important aspect of the oral comprehensive exam is your interactive demonstration of having learned and understood the material you collected. ⁵: Eye on the ball.
Be able to answer this question: "Who cares"?⁴: Your committee will ask, and it's easy to lose sight of this point in all the details (about which they may or may not ask)

Further suggestions? Send them to peter@cs.uoregon.edu.