There is no generic advice. I believe that there is no generic (universal) advice on how to do research, but to follow one's own inclinations (i.e., feelings and thoughts). That is, I believe that there is no "proven path" or sets of guidelines that lead to good research. The paths to good research are as varied as research itself, as varied as the researchers themselves, or actually as varied as the combination of research projects and individual researchers.
Individual researchers find their own paths, according to whatever fits them. They may seek and/or be assisted by advice from others, but they must decide which advice fits them. The advisers, too, should target their advice at the specific individual seeking it; that is, try to fit this individual. At the last account, the advisee is the one who matters.
A story and beyond it. In the Fall of 2008, while visiting the TOC group of Tsinghua University, I participated in an international panel of TOC researchers that was assembled to address questions by dozens of local graduate students. Many of the students sought advice on how to do research. When my turn came to answer, I noted that each of the dozen researchers that talked before me has offered different advice. Knowing many of the panelists, I could see that the advice they offered fit their own personality, their own research style, and the way their own career has evolved. I felt that this best illustrated my thesis that there is no universal and/or generic advice, but to fit the "way of research" to the individual researcher.
Research, being the most advanced form of study, is a highly complex human activity. In particular, it is a highly creative activity, which is exactly the type of activity that we find most difficult to understand. Numerous thinkers have devoted much thought to the notion of creativity; in fact, few other notions have attracted so much attention. Still, as is the case with many notions central to human life, no definite theory arises from this huge body of thought. In light of this fact, the idea of giving advice as to how to do research strikes me as very weird.
I think it should be clear that scientific research arises from the interaction between the objective contents of the discipline and the subjective understanding of the researcher. It seems that almost everybody acknowledges that even the contents of the discipline is subject to interpretations, and that different scientists may hold different opinions and views regarding the relative importance of parts of the existing knowledge as well as the relative importance of different research directions and problems. Thus, even with respect to the objective part of research (i.e., the scientific contents) there is no agreement, and one should defenitely not be surprised by the lack of agreement regarding the subjective part (i.e., the researcher).
Of course, all of this does not mean that people do not have opinions regarding the aforementioned matters. Furthermore, some people even believe that their opinions regarding these matters are universally true. You may find such opinions on various blogs, and you may even find some of them useful. My only advice is to take only whatever fits you (i.e., whatever you find useful and/or reasonable).
Still, here are a few pieces of advice that some people may find useful. I expect this advice to be useful especially to people who are somewhat similar to me (and it may even not be understood by others). So let me present it as advice to myself.
1. Surviving difficult professional periods. I often get into difficult periods (say months) in which I am searching in the dark for something of which I only have a very vague idea. In addition to the intellectual difficulty of conducting such a search, it also tends to be emotionally exhausting. Specifically, often for a long period, I do not get any feeling of progress and this is very frustrating. Only in retrospect, I sometime realize that these "bad periods" were actually periods of progress, although I did not realize it while living them. My advice to myself is to realize that progress depends on ability to survive these unavoidable "bad periods". One way of surviving them is reminding myself of the fact that I came out of past bad periods and that in retrospect they seem useful (i.e., served as an incubator of ideas that were still unclear and even unconscious at the time).
2. Learn to enjoy writing technical papers. A research project is a process consisting of several ingredients, where each is typically dominant in a different phase of the process. There is a study phase, a discovery phase, and a communicating phase. (This is a schematic description; these phases may overlap, and one can go back and forth among them during the project.) The discovery phase is typically short, and if one does not learn to enjoy the other phases, then one may be in trouble. In particular, one may be frustrated by the large amount of time that must be devoted in order to properly communicate the results of the research to others (i.e., write papers). But this activity is not only a duty (see my essay on our duties), but may also be very enjoyable: It involves understanding our own research better and spending more time with our own achievements. I find this most gratifying.
3. Accept your limitation, build on your strength. Do not try to be somebody else. Just try to push your own limits, but in a realistic and fair (towards yourself) manner. Accepting your own limitations does not mean giving up, but rather focusing your efforts on the most promising front (i.e., the front in which your strengths lie). Aim high, but not unrealistically high. (And you determine, realistically, what is high but not too high.)
4. Prioritize and don't over-commit. Your priorities may be different from mine, but I think it is important to maintain a notion of priority (i.e., consciously set your own priorities and act accordingly). I also think it is a good practice not to get involved in too many projects and/or undertake too many commitments. How much is "too much" is indeed subjective, but a notion of "too much" should exist, and one should take it into account when considering new proposals and/or obligations. Indeed, it sometimes hurts to refuse an appealing proposal, but it may hurt more to never act on this proposal and just keep it indefinitely on a personal (burdensome) to-do list. So, when an explicit or implicit proposal comes your way, I suggest you evaluate (realistically) not only its appeal, but also its costs in terms of resources (i.e., energy and time) and your budget of these resources (and be sure to use a realistic evaluation of your resources and commitments).
5. Think long-term. Research and academic life are long-term activities, so you better approach them in such terms. Do not be too preoccupied with short-term benefits and losses; focus on long-term issues (i.e., considerations that apply to the longer term). I intentionally leave the divide between short-term and long-term open, but hint that one should not take long-term as too long -- a too long time scale is definitely unuseful as a guide for real activity. What I wish to warn against is preoccupation with immediate and concrete events, while paying no attention (or too little attention) to the actual limited impact of these individual events on things that really matter. Again, you determine what is the divide between short-term and long-term, I just insist that a notion of such a divide should exist (or that the reaction to an event should be proportional to the duration of its impact).
6. Too much "philosophy". Since I often advocate paying great attention to conceptual considerations (let alone being guided by such), it may come as a suprise that I also advise not to overdo it. Specifically, I warn of spending one's life contemplating what is the best (next) step to take, while taking no step at all. In particular, I warn of using "philosophizing" as an escape from real action. Furthermore, there are times when certain questions are inadequate (e.g., are premature w.r.t to the current state of knowledge), and one is better of following one's feelings and/or the intrinsic logic of one's current activity.
7. Planning one's career. The last item leads me to an even more subjective advice. It is to refrain from thinking too much about one's future career (e.g., on its "next stage"). In particular, I think that such thoughts are bound to be unsound (since they refer to experiences that are unknown to the thinker) and useless (both because these thoughts are unsound and because typically one cannot use their conclusions). Specifically, our ability to affect our scientific career is typically confined to our (past and future) achievements, on which also we do not have full control. Thus, I advise researchers to focus on improving their past†, present and future achievements, rather than wasting their time and exhausting their energies with thoughts on how to manipulate reality....
†) Note that past achievements can be improved by better exposition, let alone by improvements of their actual contents.
8. Take yourself seriously: Don't reduce yourself to a career. In continuation to Nr. 7, I advice to view yourself as far more important than any external aspect of it. That is, your own personal and professional development is far more important than the social status and benefits that this development may entail, although your well-being may depend on such status and benefits. By personal and professional development I mean the internal process in which you construct yourself via your reflection regarding your interaction with the world (including the world of ideas and the world of other people). This process is your life, and its evolution constructs yourself (i.e., what you are); and the personal and professional are tightly linked in it (e.g., the development of one effects the other).
9. Choosing made easy. People often torment themselves with life-choices, while I am able to spare myself from such torments by realizing that choices are either obvious or tough, but in the latter case the choice does not matter (and in the first case one can just take the obvious choice). That is, if a choice is obvious in the sense that you feel clearly what you prefer, then there is no need to try to articulate the preference to yourself (although you may try to do so for fun). On the other hand, if the choice is really tough (i.e., you really do not feel what you prefer), then it indicates that the possibilities are almost equivalent (wrt your preference). In such a case, i see little sense in spending time towards determining which of the almost equivalent choices is actually better, let alone that we never have accurate estimate of the merits of choices. Of course, it may be that one does not have direct access to one's own feelings or does not trust them, but then this is the issue to be address (rather than the specific choices).
More may follow...
A related short lecture, June 2017.
Related opinions: Among my essays and opinions, the most relevant seems those on graduate studies, on where to submit (i.e., to which conference and/or journal), and on why do people become and remain scientists.
Back to Oded's page of essays and opinions or to Oded's homepage.