PhDSciNet Interview 50: Papers as a PhD KPI—Revision

Historical interview: The following preserves the research experiences and personal views expressed in the original manuscript.

The PhDSciNet Interview Series

Papers as a PhD KPI: Revising Without End

The wind blows cold across the river; revising a paper is harder still. The content must be new, the data and logic sound. Punctuation needs checking, and figures must look beautiful. Writing can make you tear your hair out—and revising is no easier!

My friend, have you finished revising your paper yet?

01 A good paper is made through revision

The absolutely, positively final draft comes from revising, revising, and revising again

Friday: After writing, I usually revise repeatedly. This can take quite a long time. I especially want co-authors and my supervisor to read the draft again and look for awkward passages or problems with the logic.

At a smaller scale, are particular words or sentences wrong? Is the grammar sound? As a reviewer, seeing so many little mistakes would certainly leave a poor impression. So I think repeated revision is essential after writing—ten or more passes, at least.

Your supervisor helps oversee the logic and the broad direction, but may not focus on details such as the grammatical mistakes I mentioned. That is where you need to rely more on yourself rather than depend too heavily on your supervisor.

Char Siu Bun: One important thing is timely communication. Sometimes my supervisor says a section needs further analysis or is not quite right, and I fall into a state of stubbornly trying to solve it entirely by myself.

When there is a difficulty, discuss it with your supervisor promptly, or ask colleagues. That can make the process more efficient.

Primeval Forest: The number of revisions is definitely in double digits, perhaps twenty-something. We revise frequently. Even before the final submission, my supervisor does another pass, paying attention to details such as punctuation.

Do you have to accept every change your supervisor suggests?

Char Siu Bun: We communicate thoroughly with our supervisor early in the writing process, meeting every week. They know what data we have and how we explain them. The paper is not unfamiliar to them, so we rarely have major disagreements over the broad direction.

If they have no major objection to the overall story and just want to change the order or adjust grammar, I accept the changes. But if I think there is a problem at a crucial point, I discuss it promptly. It depends on the nature of the issue: for minor matters, I simply accept it.

Friday: I basically agree. Accept changes when the issue is small, and discuss larger issues again with your supervisor.

Char Siu Bun: Our supervisor is quite meticulous and participates fully in writing and revision. When I first joined, I worked on a project where I was not the main author. The submitted paper was version nine, after nine rounds between the author and supervisor. I was surprised at so many revisions. Then my own first paper went through eleven rounds with my supervisor—and I did even more revisions by myself. Later it became faster: my second paper needed only two rounds. Our supervisor does not just revise the logic; they work on all aspects and take a full part in discussions and revisions.

Primeval Forest: My supervisor is also deeply involved. They write almost everyone's Introduction themselves, which is rather embarrassing to admit. But they do seem able to see, analyze, and position the paper from a broader perspective than we can. After their revisions, the paper feels as though it has reached a different level.

02 The secret to acceptance on the first submission

Secret one: Fresh logic, substantial data, good writing, and attractive figures

Friday: First comes the logic, second the data, third the writing, and fourth the figures.

The logical structure—your idea—is important. There are so many similar papers now, repeating work or making small changes. Novelty is what lets a paper reach a good journal. Then there are the data. Writing skill is important too. My own writing is fairly ordinary, and I do not tell the story very well. Strong papers tell stories remarkably well, and there is much to learn from them.

I also think attractive figures are becoming more important. Why do we like reading Nature and Science? Their figures are beautiful, and you can often understand roughly what they are doing just from the figures. Great care has gone into them. As more people pay attention to figures, the overall standard of figure-making has risen dramatically.

Char Siu Bun: Fundamentally, the novelty of the work and the data are what matter, followed by writing. Figures can substantially increase the paper's appeal as an additional benefit, but everyone really does them very well now.

Primeval Forest: Figures are so polished now that they hardly seem like something researchers could draw. Everyone looks like an artist. The competition is intense.

Friday: I completely agree, especially in materials-related fields. The illustrations are genuinely beautiful.

Char Siu Bun: It has always been that way in imaging, so we are used to it. People make exceptionally good diagrams of their systems.

Previously, most people drew a 2D diagram that was clear enough to understand. Now they use 3D, and render it extremely realistically. Every optical component looks like the real thing, almost like a simulation. Instead of a simple schematic, it resembles a polished art exhibition.

Secret two: Highlight the paper's distinctive strengths and tell a complete story

Primeval Forest: I feel that people's results are often fairly similar. In my own impression, the variation may be around 20–30%, without huge differences, especially in performance. One way to stand out is how you explain the story and highlight your paper's strengths or distinctive features. This percentage is the interviewee's informal impression, not a verified general statistic or a mathematical variance estimate.

I also get the feeling that using a more fashionable instrument for characterization can take a paper up a level. The more extensive and sophisticated the characterization, the better the journal the paper seems to reach.

Friday: Nowadays, it is rare to publish with very little work. Competition is intense, and a complete story may require a great deal of work. If something is missing, a reviewer may ask for this experiment or that one, believing only a complete paper deserves the journal.

Char Siu Bun: Yes, it is similar for us. We usually proceed in stages. First, we examine nonliving samples such as fluorescent beads or fixed brain sections. Next, we image structures in a living mouse's brain more clearly. Then we move to functional imaging, and finally ask whether we can discover new biology. Reviewers often ask what the system can do and whether it can answer a new biological question.

We are not originally biologists, so finding a biological question to study is getting increasingly difficult. Reviewers want people developing technology to produce something genuinely useful, and that remains hard.

A good paper may begin in a burst of inspiration, but repeated polishing is what brings it closer to its best form. Like a fine piece of jade, it reveals its character through careful carving and polishing.

The time spent revising may feel dull, but it is worth remembering: it witnesses the shaping of a good story.

Contributors: Friday (星期五), Char Siu Bun (叉烧包), Primeval Forest (原始森林)

Text editor: Fantuan (饭团)

Audio editor: Honey Peach Oolong (蜜桃乌龙)

Interviewers: Fantuan (饭团), Calorie (卡路里)

Recording: Honey Peach Oolong (蜜桃乌龙)

This article reflects the author's personal views, not those of this website. The original manuscript credits images to the internet and requests contact for removal if they infringe rights.

Restoration note, 2026-10-10: The complete text and credits are preserved. The original reference to 20–30% variation is identified as the interviewee's impression, not a general statistic.