Your main problem as a researcher who wants to write a feature article is that you know too much.

Assume that you fancy writing the odd piece in a field not too far from your own. You should start by reading one of the general books about good writing (For example, Communicating Science, by Michael Shortland and Jane Gregory, Longman, 1991).

These books tend to say the same, rather few, things, as they should. Your job as a writer who wants to be read is to do the hard work of applying those few, simply stated principles to every sentence, every paragraph.

Most people find this surprisingly difficult first go. And it is harder for the kind of super-specialist which you, as a researcher, are well on the way to becoming.

ÖйúAƬ

The most important principle of all is that you must be ruthless about what to leave out. But a researcher's habit of mind, especially one just starting out, is inclusive. You are learning a business where everything rests on a little hard-won data, on maintaining fine distinctions, on persuading your peers that you have registered signal, not noise. The temptation is to start from there.

You might be able to make all that interesting, to engage the lay reader with an account written from deep within the research process. But that is a challenge rarely met. Much more likely is that you will be concerned with findings and potential applications. Your readers will still want details, and explanations, but they will be different from those you are used to supplying.

ÖйúAƬ

They must relate to everyday life, not to the concepts you trade with your peers. You must also tell your story in a different way. No problem, method, results and discussion, please.

The best five point plan is the one which runs: Hey!

You!

Guess what!

So!

There!

There are other possible structures, of course, but that is not a bad one to start with. It reminds you of three things. Get their attention. Tell them a story. Stick with one main idea.

Structure matters more in science features than in many other areas. You are almost certainly going to be explaining some bits of science. You are going to have to order that explanation very precisely for it to work. But it is no good starting a long explanation in the first or second paragraph. First set up the context carefully. Explain why this is interesting or important.

ÖйúAƬ

And when you have written it, rewrite it, preferably after putting it aside for a few days. Try it on a few criticial readers. Read it out loud to find the awkward bits. Check everything. Try and get your piece as right as you can before trying to get it published.

Otherwise it will be: not used at all; changed in ways which you do not like or which alter what you wanted to say; or, if you are lucky, put in the hands of someone who plagues you with queries for several working days.

Jon Turney is Wellcome Fellow in Science Communication, University College London, which runs science communication courses for undergraduates and postgraduates.