Yesterday I received an earnest email from a researcher working in my field asking how I write so many papers. I don't have an ideal answer, but my responses* to their specific questions are given below.
Scientists are a varied bunch of people. I know many excellent theoretical and computational researchers who hate to write. They are content with knowing a result and feel no inclination to share it with the world. The classic Whitesides' paper guide makes the reasons for dissemination clear. However, reluctance to publish is understandable. In many cases, the preparation of a decent manuscript can take longer than the creation and analysis of the original results. Then you may have to deal with judgemental editors and unstable reviewers. Publishing a scientific paper is a stressful endeavour at the best of times!
Back to the e-mail:
1) Do you have your-own-way to design papers?
I would say that I design projects rather than papers. Every day I try to browse new publications in my field, and I like to read many older books. From these I keep a list of interesting projects, with some being specific questions and others being broader subjects. I do the same when I attend conferences, so now I have a healthy database of topics for new students to tackle (and for myself when there is an excess of computer power).
Approaching the writing of a paper, usually one of two things happens: (a) I have a clear result that provides a complete story. This is the easiest type of paper to write: following a rough outline, it can almost be a stream of consciousness (sitting in a quiet room for a day). (b) I have a large set of data that is systematic, but makes no obvious story. This problem is always challenging, and my typical approach is to design figures, which I iteratively modify and rearrange until a coherent outline and narrative emerges (this process often involves additional computations).
2) What will be the key of that efficiency in writing/publishing the papers?
The technical aspects of publishing definitely become more efficient with time, e.g. building up a comprehensive database of references and mastering LaTeX / image generation. I find the creative writing process is limited by the state of mind. I really need quality time to write a paper (turning off my phone, disconnecting from the internet, and a nice cup of tea).
On a higher level, project and time management can be an issue. Learning how to keep track of and prioritise multiple projects is an essential skill. Personally, I am an Evernote addict.
3) Would you mind explaining me your steps of how you publish papers from your research problems? For instance, I was very impressed when I read your APL paper on indium oxide. One of the reasons is that you made really good story without any unnecessary data. For me, I tend to write a paper after I get sufficient data. Then, when I start writing, I don't know what to say since there are so many things to tell…
This paper came from a side project, which I wasn't funded to investigate at the time. I had published on the surface structure of indium oxide in the past, but one open question remained, related to surface defects. My starting point was a simple electrostatic model (taken from a 1983 paper by Dorothy Duffy), which was actually run on my laptop on a flight to Korea. This gave some exciting insights (in relation to the change in oxygen environment towards the surface). It also provided enough evidence to justify a full electronic structure study (so I wrote a script to loop through all possible combinations on our national supercomputer). In fact I had many more results (e.g. vacancy - vacancy interactions), but I left them out as they didn't add to the central story and would have made the narrative too complicated for Applied Physics Letters. It was at the outline stage when I decided what results would be included, as well as what point each paragraph would try to convey.
4) Other than above questions, do you have any suggestions for me?
I can think of three points that have helped me:
(a) Audience. In our field, the target audience could be people with a backgrounds in physics, chemistry or materials science/engineering (both theorists and experimentalists). It is challenging but worth the effort to make all work as clear as possible to those who you think will be most interested in the results.
(b) Multitask. I always recommend that my students have at least two on-going projects. If one is producing no results or making little sense, a change can really help. Also do not be afraid to abandon a project altogether (I have had many ideas that just didn't work).
(c) Collaborate. Working with people from different backgrounds is always fruitful in terms of forming new ideas and solutions. I owe a lot to my collaborators, many of whom have now become close personal friends. You could work alone, but it wouldn't be as enjoyable or productive.
*Hopefully I don't come across as too pretentious; I still have a lot to learn too.