You're ready when your advisor says that you're ready. In general, most people are ready to propose after they've published a paper on a topic they'd like to pursue further. You also need to have passed your portfolio.
Remember that you can't defend your thesis until 7 months after you defend your proposal. Therefore, even if you don't feel completely ready, you should consider proposing since it can open up your timeline for graduation.
If you're like me (and many others), the thesis proposal will be the longest piece of writing you've ever written. My proposal was easily three times the size of my previous largest work. That can be daunting.
Luckily, assuming you've already written a paper on the subject, you have a significant pool of text to pull from as you start to write. Since your proposal also needs to include a suitable background and introduction section, you'll find that the words flow a bit faster than when writing a conference paper.
The best advice I can give is to write every day. I found it difficult to make progress until I forced myself to write for certain periods of time each day. I downloaded a program called WriteRoom for the Mac that blacks out the entire screen, so there's no distractions to get in the way of writing. During my writing periods, I let myself write about anything remotely related to my research. It was way too difficult to try to write real proposal text from the start.
One day I challenged myself to write as many possible conference paper topics and 2-sentence abstracts as I could in 20 minutes. The core contributions of my thesis came from that exercise.
Thesis proposals are considered internal department documents, so it may be difficult to find examples. Talk to other graduate students in your lab in order to find sample thesis proposals. Get as many of them as you can. Remember that even though thesis proposals can be hard to find, it's easy to find dissertations. Find as many dissertations as you can find, too. You don't have to read all of them, but it's instructive to look at the introduction and table of contents. These can give you an idea of what is involved in a thesis and what will be expected of you. The ACM gives a Best Dissertation award each year, and I found it instructive to look at those award-winning dissertations for hints.
While it isn't strictly necessary, most people use the umthesis LaTeX stylesheet to format the thesis proposal. This stylesheet automatically generates the signature page that you'll need at your proposal defense.
You need introductory work sufficient to introduce your external readers to the material. This also helps convince your committee that you know enough about the field that you should be allowed to move on to your thesis.
You need to write about related work. you need to perform a literature search that proves that your ideas are new, and that you can talk reasonably well about the previous work in this area.
You need to introduce your work. As a rule of thumb, your thesis should cover enough ground to be worthy of three conference papers. Organizing your thesis around three conference submissions may help you progress toward your thesis (it gives you hard deadlines) and may serve as a nice organizing principle.
You need a list of contributions and a timeline for completing them. Your list of contributions is the most important part of your thesis proposal. This is where you succinctly outline why your work is different than other work on the subject, and why it is worthy of a Ph.D. thesis. This is traditionally in a bulleted list, and should probably be less than a page. It's worth really working on these contributions, because people will inevitably ask what your thesis is about, both in the lab and at conferences. You should be able to give an answer that's just a few seconds long, and that answer should come from your contributions page.
Remember, you are proposing work. It doesn't have to be complete. It's probably somewhat better if it isn't complete, because it shows that you can talk intelligently about what work you plan to do before you do it. This kind of skill is important if you want to write grant proposals as a faculty member someday.
Probably longer than 20 pages but shorter than 80. Typical proposals seem to range from 40 to 70 pages. My proposal was approximately 70 pages long with 30,000 words.
At this time, the department requires that you have at least 4 committee members. One of these members must be a University of Massachusetts professor that is not a member of the Computer Science faculty.
Most students need some help finding that committee member outside the department. Your advisor can probably recommend someone to you.
First, you agree with your advisor that it is time for you to write a proposal. You may work with your advisor a little bit to determine what topic you plan to present. However, your thesis is supposed to be entirely your work (including the proposal), so your advisor will not be offering help like he or she might in a co-author role. It is your job to come up with new contributions and insights.
Do not feel like you need to wait until you have great ideas before you start writing. I languished for months without good ideas, and suddenly started to have good ideas as I started to write. I started a proposal journal and wrote in it for about an hour a day; after two weeks of that, I started to get some great ideas. I continued to write in the journal to complete the ideas, then finally I wrote about them in the proposal.
Once you have a proposal draft, give it to your advisor to review. You'll probably go through a few revisions. Since the proposal is long, try your best not to waste your advisor's time with simple errors (like spelling, improper citation, etc.).
Even when your proposal is not necessarily the way you want it, your advisor may tell you that it is time to schedule your defense. Expect that it will take at least a month from the time you start scheduling your defense to the time when the defense actually occurs. First, you need to assemble your committee. Once you've done that, you need to contact the main office so that they can advertise your committee--your abstract and committee member list are sent by e-mail to the faculty to give them a chance to object. Once you've done that, you'll have to find a 2 hour window where all of your committee members can attend. Your committee members are busy people, so this will be challenging. You'll need to find a room that your committee can meet in for the defense.
You need to give your proposal to your committee with at least one week to spare before the defense. If you are running short on time, send a draft to the committee, but mention which parts might change before the defense. Your committee will appreciate getting the proposal as early as possible so that they can decide when to read it.
It is a lab tradition to bring food to the proposal defense for your committee members. I got food from Black Sheep in Amherst, because the food seemed good and they delivered it to the department. Other people have brought fruit and mixed nuts. You don't have to bring a lot of food, but the food is a small way to say thank you to your committee for their time.
You should plan to give a practice talk in front of other students. This gives you a chance to get feedback about your presentation from your peers, some of whom hopefully have successfully defended a thesis proposal. For those that haven't passed, seeing your proposal slides gives them a chance to learn from what you've done.
You will give an oral presentation with slides of your proposal work. You can assume that your committee has read your proposal, but don't assume that they've memorized it. Plan for less than an hour of material, since your committee will probably ask you questions during the talk.
When you're finished, your committee will ask you questions about your work. When the questioning phase is complete, they will ask you to leave the room as they discuss your work and their decision. When they have decided, you will be called back into the room to hear their decision.
If they are happy with the work, they will sign your signature sheet. This then needs to be delivered to the main office along with two copies of your thesis proposal.
