On writing

Tags: howtos, musings, research

I write a lot: papers, project reports, more papers, more reports, and occasionally, even some code or a blog such as this one. Nevertheless, I still consider writing to be a hard task. The boundary conditions of my life used to have to be perfect to make me write productively, or so I thought at the beginning. Over the years, I have learned that the boundary conditions will never perfect. Instead, I try to follow a few rules, which I shall outline in this article.

Rule 0: Know your audience

In proper programming tradition, the first rule starts at index zero because it actually needs to precede the writing process. Before you even sit to write, take a few minutes to think about your audience in general terms. Are you going to write a paper? A blog post? Or an angry letter to your local version of the IRS?

Know that different audiences necessitate different writing styles in order for your writing to be successful. The tone of your writing needs to differ—in a paper, you are supposed to be precise and concise in your wording, while a blog post can be more informal, for example.

Rule 1: When in doubt, start in the middle

While text typically is quite linear when reading (although for reviewing a paper, for example, this might not be the best strategy, but I digress), there is no need for linearity when writing. More to the point, some people are actually scared of starting to write. The reason for this is the “fear of the empty page”, which I like to refer to as the horror vacui of writing.

The best remedy, it seems, is to start with the parts that one deems to be easy. In a scientific paper, this could be the part containing details about your new method, for example. In a report, it could be a description of the results you achieved—and so on.

I find that the hard part is starting the writing process, so if I start with something that I consider to be easy, the writing will be flowing better afterwards. As a short aside, this is why I typically write the outer parts of a scientific paper, i.e. conclusion, introduction, and related work, at the very end once everything else is in place already—and once I have achieved sufficient speed, or momentum, to actually tackle those sections.

Rule 2: Silence the inner critic

Another bias to overcome for some people is the inner critic, this voice that keeps telling you that your writing is not perfect yet. I used to heed the advice of this voice and waited for the words in my head to make more sense until I found out that it is easier to write something of dubious quality rather than write nothing at all. I am at a point in my scientific career now where I get to read papers written by other people. Chipping in with my own suggestions is much easier if there is already something on the page that I can improve, or at least try to improve.

Remember the maxim le mieux est l’ennemi du bien; perfection will not be achievable by waiting until the blank page (or this blank screen) fills itself. You have to actually start somewhere, so do not listen to yourself, and just do it.

Rule 3: Rewrite, rewrite, rewrite

If you have some experience in writing code, you should already know the benefits of refactoring. It is exactly the same with normal text in general. You first do your rough draft (heeding rule two, of course), then you rewrite it. Then you rewrite it again, taking care to cut out everything that is unnecessary. If your text is initially very long, you should strive to make it more concise. If it is terse in the beginning, strive to make it slightly more verbose. Try to go into the depths of your document and explore a little bit. Do not be afraid of changing things or even throwing away some of the text. It was necessary to write to get you to a certain point but if the text does not fit the current story of the document any more, there is no need to keep it around.

Another small aside: I often find that initially, I require a lot of text to explain a thing to myself. Later on, having understood more, I feel more comfortable in removing unnecessary parts of the text. I think that this is similar to requiring special gear to climb peaks, but not requiring any gear to enjoy the view.

Rule 4: Read, imitate, and refine to find your style

One of the common themes in writing involves finding your voice or finding your own style, but for a beginner, this can be a daunting task. Instead, my advice would be to read a lot, imitate things that you like, and constantly refine them. After some time, you will find that you have found a style that suits you.

By the way: even if you dislike a certain style, you can still use it to refine your own writing—just like you can either go into the direction of a gradient in optimization, or in the opposite one. Use this to your advantage.


Finally, some more resources to get you going:

  • The Elements of Style by Strunk and White is a perennial classic that is loved by many and hated by many. I like to love the pithy, and even somewhat funny, examples given in the book. Caveat emptor as always, though.

  • Zen in the Art of Writing by Bradbury is a collection of essays, dealing mostly with story writing. However, as we scientists know, every paper has a story, so I am sure you will distil some nuggets of wisdom from it.

Happy writing, until next time!