PRELIMINARY-- (Edition 3) The author seeks feedback on these. They have not been shepherded or workshopped. They have been discussed informally by a few people for about a year.
You are considering a Doctoral level degree or are engaged in one. You have many sub goals, but one overriding goal: completing the doctorate and getting on with your life. How should you proceed?
What pattern language could guide a doctoral student in the pursuit of the degree and eventual employment? At ECOOP '98 (in Brussels) both Oscar Nierstrasz and Jim (Cope) Coplien spoke. Oscar got up and named what one might think of as such patterns. According to Cope they included things like:
Cope later sent me the following:
Oscar wasn't terribly explicit about all of these -- in fact, the last one was largely implicit -- but that seemed to be the message to me. If these are the patterns, they form a language that makes some whole. What is the whole? Does it have QWAN? I don't think so. In fact, I seriously doubt that many Ph.D. programs produce QWAN. Most Ph.D.s seem not to be able to program their way out of a paper bag and understand little either of philosophy or systems. At least that seems to be the case there in the U.S. The patterns I'd like to see in such a language are:
This paper is an attempt to reconcile these two rather different views. The patterns here draw on the experiences of several people. While Oscar focused on the research needed to complete the degree, the goal here is a bit broader.
The following patterns all follow a common form. After the name of the pattern is a section of context, problem, and forces, in that order, with the problem addressed by the pattern in bold. Then there is a separation of three asterisks, followed by the soluton beginning with a bold Therefore. This is followed by another separation and finally there is additional commentary.
These patterns are all preliminary. I'm not even sure I've followed the above form consistently. The sections need filling out and there may be other important patterns before we can say that this language is truly generative. I welcome all constructive suggestions.
You are thinking of getting a terminal degree. You know the field you want to study, but have not yet chosen an institution.
You want to get your degree from the highest quality institution that is compatible with your goals and abilities.
Some institutions will be easy for you to get in to. Others impossible, perhaps. This you won't know until you apply. Expect to be rejected by at least a few of the ones you would like to attend. If you apply to several and none reject you, perhaps you are setting your sights too low. The opposite can be true as well, of course. Different programs in the same field have different requirements. Some will take longer and some not so long. Some have strict residency requirements and others not. Some institutions have better reputations than others.
Therefore, apply to several institutions. Cover a range from those with excellent, national or international, reputations, to those that are perhaps more specialized and local. Don't choose the institution you will attend until you have information about which programs will accept you.
When applying to several institutions, ignore the stated cost of the education there. Don't assume that the cost will be prohibitive, no matter what the reputation of the institution. Don't reject the possibility of going abroad for your degree.
You have applied to several institutions and some have accepted you.
How you will be perceived in the market place depends at least initially on the quality of your degree.
High quality institutions are harder to graduate from than others. They may cost more than others. However, the quality of the degree and your future career possibilities are correlated with the quality of the degree you obtain. The reputation of a program is usually highly correlated with the quality of the degree you receive. This includes both the quality of the faculty and the quality of the student contacts at the institution.
Therefore, accept the offer of the highest quality program that will take you that is consistent with your goals.
You (and others) may need to sacrifice for several years to do this, but it will be worth it in the long run. The other patterns here try to help minimize the time it will take you to succeed.
You are about to make a final decision about really doing this.
Doctoral programs are hard. Burnout is frequent. Sacrifices are many.
Those who succeed best are those who are compelled in some way to do this.
Therefore, have passion for the program you choose. This passion will need to carry you over rough spots. You will need it to justify the sacrifices you will inevitably make.
If your commitment is only half hearted, so will be your success. Don't do this just because someone else wants it. It is a personal commitment.
You are in a doctoral program or about to begin.
Every major change in your life will postpone your degree completion.
You have a life to live, but the degree must be a commitment of major proportions if you are to be successful.
Therefore, avoid other complications in your life. Postpone major life decisions until you finish your degree.
If you are married, stay married. If you are single, stay single. Life goes on, of course, so this is just advice and can not be taken as an absolute. But avoid all uproar that you can avoid. If you are married it also helps greatly if your spouse and children buy into your passion for the degree. They will be making sacrifices as well. If they do not, or cannot, support you then you have a major complication already.
You are about to begin the program or have begun.
The requirements of the program are usually rigid and sometimes change.
The requirements are set by others. The institution and department set some. Your advisor and committee will set others.
Therefore, know the requirements and adapt to changes in them.
Most programs have firm dates for completing various requirements. If you miss these dates, or your materials are not adequate, your degree will be postponed. Make sure you know the expectations of your advisor. Make sure you know when he or she will be away for long periods of time.
You have chosen a program and know its requirements.
You need to have a realistic expectation of when you should complete your degree.
Setting a definite goal is the best means of achieving it. Since it is a big goal, you need a time scale with sub goals.
Therefore, set a time goal and choose a path to it with critical points along the path marked.
If you have delays in meeting any of your subgoals, recast your plan and if necessary a new target date. There may be certain courses that you need to take. There will be exams to take that may depend on those courses. You may have preliminary papers to write. Estimate time for each of these and set a completion date.
You are in the early part of your coursework.
Your real goal is to write a dissertation acceptable to the faculty
There is a lot to learn. There are many courses to take. Some of them are immediately relevant and others, while nice, are not. You need to pass any preliminary exams, of course. Certain courses will help you in this. Others are important to gaining depth in that part of the field you want to specialize in.
Therefore, take those courses, but only those courses that get you through the exams and to the research frontier.
You can spend an entire lifetime at some institutions taking courses, but this gets you no nearer your goal. Your goal is NOT to complete your education, but only to start it. Don't try to know everything there is to know before you start your dissertation. Remember that you can still have a rich intellectual life after you finish your degree. If possible, take courses relevant to the exams immediately prior to the exams. The exam in an area may well be made up by someone who has recently taught the related course.
Your degree is progressing.
To do well in a dissertation you need to go deep, but you need enough breadth to know how your research relates to the rest of the field and to the rest of human knowledge.
Courses give you one perspective. But there are many others. Text books are another perspective, but not completely adequate. What is written in text books generally does not have the flavor of research. You will need a variety of perspectives to do well at this.
Therefore, read everything relevant to your research. Read for both depth and breadth.
Take notes as you read. Pay special attention to research methodology as you read. What questions were asked and HOW were they answered. One way to be productive is to always have something to read with you as well as a notebook in which you can make notations. Then, if you have five "dead" minutes, you can make use of it.
Don't wait until you finish the preliminary exams to start thinking about research. As you read, note which questions have been answered and which might still be asked.
You have taken the appropriate courses and are getting ready for the required preliminary exams.
Unless you pass the exams you will not be permitted to continue.
The exams in many institutions are an absolute barrier. You must pass them to continue. You may have only a limited number of attempts.
Therefore, make passing the exams your first priority until you are successful.
Take old exams, if available, under test like conditions. Get advice from the faculty on your readiness. Hold study sessions with your student colleagues. However, reading solutions is not the same as creating solutions.
You are taking a course or doing some reading for your degree.
A dissertation answers some important unanswered question. Without the question there can be no answer.
Questions are everywhere. There is more that is not known than there is that is known. Every paper you read had left some questions unanswered. Generalizations are possible. Specializations are possible. Similar questions in related fields are possible.
Therefore, ask questions whenever you read. Write these questions in a notebook with an appropriate reference.
Some questions are minor and can be answered easily. Some are more difficult and can be answered in a few days or so and can lead to papers. Some questions are deeper and can lead to long term research projects that extend beyond your degree. None of these are necessarily appropriate for your dissertation. You want a question that can be answered in a few months. The more questions you ask, the more choices you have. To get started, figure out what questions the author of a paper you are reading asked to initiate the study of this paper. Does the author answer this question (these questions)? Can you ask additional questions that could be answered with additional work?
You are reading or taking a course. There is a lot of detail in what you are reviewing.
It is easier to derive the details from the big picture than is true of the reverse
Often you are taught the details. Insight can be difficult to achieve.
Therefore, seek the key insights in what you are studying. Ask about the big ideas if you can. Derive them if you can.
How do the big ideas in this study relate to other big ideas? How are the details related to the big idea? What is the crux of the current problem? What is it that makes this paper important?
You are choosing an advisor.
You need good advice. You need protection from faculty politics. You need a mentor. You need a friend.
You will develop a long term (life time) relationship with your advisor. He or she can help you throughout your career, not just through the degree process. Your advisor is your guide and protector.
Therefore, find an advisor with whom you can work comfortably. Your advisor needs to know about your area of research and must accept your goals as valuable for you.
Sometimes an advisor will be assigned to you initially. Don't be constrained by this.
You are choosing an advisor.
You have little academic standing as a student and there are political forces at play in many schools. You have little standing in the profession as a student and you will need a way to jump start your career when you are done.
Your advisor's reputation will help or hinder you initially in your career. If he or she is well known you have an easy way to get introduced to the people you need to know to enter your field of choice.
Therefore, find a well known and respected advisor. Your advisor needs some clout in the university and in the profession.
Don't underestimate the effect of faculty politics. It can be vicious, though most professors will try to insulate students from it. However there are people who can and will cause you problems based on your research or your ideas. Thankfully this is less true in computer science than in some other fields. An advisor who recognizes this and has appropriate clout can insulate you. Likewise, such an advisor can provide an opening into the profession. A corollary of this is that as a student you should avoid faculty politics altogether.
Some special advice is not to choose an advisor who is about to come up for tenure. He or she will be too involved in that process to provide the needed time to guide you. It is also dangerous to pick an advisor who is contemplating leaving the institution. This has caused delays for many students. It is not always possible to move with the advisor, or to have the advisor continue as head of your committee.
It is sometimes possible to find two people, one with clout and one with a special interest in your research. If they are compatible and work well together you can utilize the strengths of each.
You are choosing an advisor.
Your main task is to complete the dissertation in a reasonable amount of time.
Your advisor is your main guide here, both to point you to appraise material for study and for managing the process.
Therefore, choose an advisor who knows the process and has done it before. Choose an advisor who has an interest in having you finish your degree.
Your advisor can manage your committee and insulate you from some of their demands. He or she can nag you to meet your deadlines and those of the school. Most important, he or she can know when you are done. A good advisor will have you do enough, but not too much.
You have been asking many questions and answering a few of them. You need to choose a problem to solve as the basis of your dissertation.
You want a problem that is interesting to you and doable, but not too easy.
Good problems are hard to find. You may need to work on more than one before you find the one that works. If it is too easy, no one will think you are ready to graduate. If it is too hard you won't be able to solve it at all or at least do so in reasonable time.
Therefore, choose one or more problems that interest you and seem to be possible to solve in a few months. If it takes you five months to solve the crux of your problem, it will probably take you a year to gather the evidence for your solution and to present it suitably. This is probably about the right scale.
Your advisor should be able to give you some advice as to the difficulty of the problem you have chosen. However, since it is research and therefore by definition has not been done before, this advice is very tentative. You won't know until you try. Note that in some fields, if someone publishes your results before you finish, you will need to start over on a new or related problem. This is more of a problem than it sounds like since in many fields advances come in waves with key results enabling intense study in certain areas.
You are also facing a "three bears" problem (Goldilocks and the Three Bears). Some problems are too hard, some are too easy, and some are just right. The present author had exactly this experience.
You are developing your research.
You will be asked to explain your research by many people, some of whom can affect your future.
Some research is very esoteric, but it is still related to the rest of human knowledge.
Therefore, know the meaning of your research. Why is it important to know the answer to this question? Be able to state the relevance in a few sentences at most.
Having the answer to this constantly at hand will also help you reject subsidiary questions or defer them until you are done with the dissertation. As you work other questions will arise (hopefully) but you don't want to try to answer all of them in the dissertation. You do want to keep a list of them for further study when you finish, however.
You are developing your research.
You want your research to be tightly focused.
You will often be asked to explain it.
Therefore, be sure you can state the essence of your research in a few sentences at most.
This is a guard against dissertation bloat, also. You want your dissertation research to be tightly focused on a single question or a very small number of closely related questions. If you can't say simply what you are doing, it is too complex.
You are engaged in your research. The road is long and sometimes hard. Many things get in the way of your success.
You need to keep advancing toward your goal.
Continuous progress is hard, but small steps can be taken when large ones are elusive. You have a plan and a schedule, of course.
Therefore, always make progress even if it is very small progress.
Sometimes, of course, you need to just take a walk and clear your head. For me a bicycle helped a great deal.
If you can't find a solution to the sub problem at hand, try another. Or fall back on some required reading. Or re-write what you have to make it clearer. Or seek advice from your advisor, being honest about your dilemma. You aren't expected to be perfect. If you can't take a big step, take a small one.
A poet (Richard P. Gabriel) writes a poem every day. He describes the process as continuously lowering his expectations and standards until he gets it done. Successful novelists describe the process as making one page of progress every day. Some days on which you tear up a lot of pages are long and hard, of course.
You are doing a dissertation in computer science. You are writing code as the basis of your research.
If you build code it will feel important to you, but it is not.
Computer scientists build things, but only to understand underlying principles and to advance the state of the art.
Therefore, remember that your dissertation is what you learn from the code, not the code itself.
For example, if you build the newest and best garbage collector in the world, the research is really the measurement of how much better it is and the underlying algorithm in it, not the code itself. Once you have written the code you are less than half done.
Your dissertation is quantitative. You will be gathering statistical evidence to attempt to answer your question.
You are doing research, not trying to prove a point. You seek knowledge, not a particular end.
Knowing that something is not true is as important as knowing that it is true.
Therefore, make sure that a no answer to your basic question doesn't set you back. Think of your question and formulate it so that you have a degree no matter what you learn from your study.
Don't commit yourself to an answer before you begin your study. Focus your questions so that either a yes or no answer leaves you with a dissertation. Too many people try to prove something is true rather than investigating whether it is in fact true. Then you aren't doing research, but propaganda. Avoid this trap.
Your dissertation is qualitative. You are proving theorems, developing algorithms, inventing protocols...
Your conclusions must be explicitly supported by the evidence you have gathered. It must not be refuted by the negative evidence you have also gathered.
Therefore, specfically list each conclusion you draw and the evidence in its favor and that which tends to refute it. If necessary your conclusions will need to be qualified.
Make sure that you do this so that possible flaws aren't discovered by others. If there is negative evidence, you need to either show why it doesn't apply or you need to make appropriate qualifications. Again, you aren't trying to prove a point, but you are trying to advance knowledge.
You are developing your research toward your dissertation.
You think of your advisor as the expert, but really this is not true.
Your advisor is a mentor. Your advisor knows about learning and about conducting research, but not about your topic.
Therefore, own the research yourself. Your advisor provides advice not answers.
You should quickly know more than your advisor about your specific topic. You will be teaching him or her the results and the methodology. Some advisors don't understand this, unfortunately. This is partly what makes a good advisor.
You are developing your dissertation and wonder how much to put into it.
The dissertation is meant to prove your worth as a researcher, not to catapult you to the top of the profession.
If it is too good it will take you too long. It needs to teach you research methodology and build a bit of mental muscle, but it is only your first step in a long research career.
Therefore, don't try to solve all the world's problems (or even all your discipline's problems) in your thesis.
The dissertation doesn't need to be perfect. If you look back on it after 10 years as being quite weak then you have advanced. If you look back on it after 10 years as your best work, then you have not and you probably spent too long at it.
You are developing your research
Your research results may affect many people. It may advantage some and disadvantage others.
Communication systems and information systems empower people only if they can take advantage of them and only if they are not used by others to disadvantage them.
Therefore, consider the social impact of what you do. Do No Harm.
This is a lifelong consideration. Your professional life should advance the common good, not just the financial gain of yourself or those you work for.
You are about to start your research.
Your research needs to be focused on a specific question and must use a specific methodology.
Sometimes you don't know enough to do this. Sometimes you need to conduct a pilot project, but this needs to have the same characteristics. But being smaller, it takes less time, so if you fail you lose less.
Therefore, write down your research plan. Include every step you will perform. If your research is quantitative (statistical) this has a formal meaning that includes what questions will be asked and what statistical tests will be run and what confidence levels will be used. Do not compromise this.
Much research is flawed by gathering a bunch of data and then asking "Well, what does this data tell us?" If you don't know before you gather it, it won't tell you anything. Many dissertations fail because the "researcher" just wanders around in the problem area gathering random results and then trying to put them together into a thesis. Sometimes this succeeds, but it is not time efficient.
You are conducting your research and have a committee.
Your committee may have specific ideas about what you should do. These may differ from your own ideas. The same may be true of your advisor.
You are the expert, but your advisor and committee may have something valuable to contribute. In any case, they will have the last say in when you are finished.
Therefore, don't disagree too much with your advisor or committee.
Hopefully you don't reach an impasse with your advisor or committee, but it has happened. Avoid this trap by finding a compatible advisor much earlier. Have the advisor help you find a compatible committee. But then, take their advice. Even if you don't like it. You want to finish, not overcome all opposition.
You are writing your dissertation.
Writing is hard. Writing about technical topics is harder.
You will need to please your advisor and your committee. They will ask for several re writes of (at least) sections not clear to them.
Therefore, plan on rewriting your dissertation several times before it is accepted.
Make sure you have budgeted time for rewrites. This can be very frustrating and isn't always very creative. This is a dilemma faced by all authors, actually, so get used to it.
You are doing your research and writing it up. Other questions occur to you along the way
You want to form the basis of a solid research career.
You don't have time to answer all these questions now, but you will in the future. The more unanswered questions you have, the more prolific you can be.
Therefore, write up brief summaries of questions that occur along the way, with pointers to relevant papers that you come across.
You can build a large file of pending work this way. This is very impressive if you are entering the academic job market especially. Hiring committees will want to know what you intend to do, not just what you have done.
You are nearing the end and thinking about the job market.
You need a solid launch into the profession.
Your advisor and committee have solid reputations in the field and you have impressed them greatly with your astute thinking.
Therefore, get solid letters of recommendation from your advisor, committee, and from other faculty who have a high opinion of you and your work.
This is why you chose your advisor and committee as you did.
You are nearing the end and thinking about the job market. You have asked several people to write letters of recommendation for you.
Not all letters will help you. The writer may intend one thing but the reader gets a different impression.
Letters are a private matter, but their contents affects your prospects initially and possibly for a long time.
Therefore, review your letters or have someone you trust do so.
This is especially important if your advisor or other writer is not a native speaker of the language of the letters of recommendation. Much can be lost in translation. Some advisors will let you review your letters. Some will not, but they may let someone else, such as a department chair review the letter for possible misunderstandings. Perhaps your committee will be willing to review all your letters in common.
You are doing one of the final rewrites of the dissertation.
You will want a publication soon after completing your degree.
Research is only useful if it is published. In many fields your reputation is determined by what you publish.
Therefore, abstract one or two key parts of your research into a paper that can be published in the scholarly press.
If you have discovered one or more minor questions along the way that can be answered fairly quickly or have already been answered as a result of your research, write these up now.
You are done.
Research is only useful if it is published. In many fields your reputation is determined by what you publish. You have an important work here.
Therefore, find a publisher for your dissertation.
Most institutions require publication by at least University Microfilms. However, publishing by a recognized press or scholarly journal is far superior to this.
Congratulations on finishing your degree and good luck to you in your career. Don't forget that you have been building up several debts over the last few years, to your family and others. Don't forget to repay them.
Your life and work should make a contribution to the betterment of the world. The doctorate is one way to get yourself started on that path, but getting the degree can be a barrier. Rather, you want to think of it as a gate that opens you to the tools by which you can accomplish great things. But you must pass through that gate in order to do so, and you must move beyond the gate as well. The doctorate is not an end in itself, but a beginning. Its intent is for your growth and the improvement of the world as well. You don't receive the degree purely on your own work and merits, but on the work of a vast number of people (and resources) including your family, society as a whole, and other scholars stretching back to antiquity and beyond. Honor their work as we honor yours.
QWAN is the Quality Without a Name. See Christopher Alexander, The Timeless Way of Building (Oxford, 1979). In short, this quality, which is known to everyone but cannot be named, is that quality of things that enhances human life and potential. In other contexts it is sometimes described as "goodness, truth, and beauty" but it goes beyond these things.
The introduction is almost verbatim from email (Sept. 29, 2000) from Cope to Joe. Cope had just recently finished his own doctorate. Oscar remembers the seminar somewhat differently and these are clearly Cope's recollections, but the mail from Cope was so intriguing that I've kept it. It did, after all, spark this project.
The slides from Oscar's talk are on the web at http://www.iam.unibe.ch/~oscar/PDF/phd.pdf
Fred Grossman has contributed some of these patterns and commented on others
Oscar Nierstrasz and Mary Lynn Manns have commented on the patterns and the project as a whole.
Last Updated: July 15, 2002