Suppose you decide to build a better mousetrap. You would begin, surely, by taking apart the existing mousetraps to see how they worked. You must do the same with books. When I wrote Sharpe’s Eagle, never having written a book before, I began by disassembling three other books. Two were Hornblowers, and I forget which the third was, but I had enjoyed them all. So I read them again, but this time I made enormous coloured charts which showed what was happening paragraph by paragraph through the three books. How much was action? And where was the action in the overall plan of the book? How much dialogue? How much romance? How much flashback (I hate flashback)? How much background information? Where did the writer place it? I already knew what I liked in the books, and I was determined to provide more of that in my book, and I knew what I disliked, and wanted to use less of that, but the three big charts (sadly I’ve lost them) were my blueprints. It was not plagiarism, but it was imitation. I learned to start with a fairly frenetic scene, and to keep that pace going before I slowed it down to provide necessary information. I learned, if you like, the structure of a best-seller, and then I imposed that structure on what I was writing. These days I do not think about it any more . . ."

Any thoughts on this method?

Mr. Cornwall also suggested the following technique to evaluate style:

"Style seems to be a stumbling block for many first novelists, and the only advice I can offer is to tell you how I overcame it. Which is not to claim that I have a fine style, only that I no longer worry about it. But when I was writing Sharpe’s Eagle I spent hours reading and re-reading the typescript, and every time I got hopelessly depressed thinking that it was no bloody good because the style was so clumsy, and so finally I tried an experiment. I typed out three pages of a Hornblower novel, substituting Sharpe’s name for Hornblower’s, and then I put the pages into a drawer. After three days I read those three pages (which looked exactly like my own typescript) and, to my relief, discovered that I was just as critical of Forester’s style as I was of my own. But he was published. More, he was successful, so clearly I was being too critical."

I *have* typed out several stories and parts of several novels, to see if going over them word-for-word was helpful. It was, to an extent.