In an exchange on a listserver I am on, the question of writing lessons learned along the way came up. This was my list . . .

• Published novels are the finished product: one never sees the messes, failures and train-wrecks on the way, so one is completely misled as to how easy certain things are to execute. The downside of a diet of the best is that the emerging writer can become inadvertently overambitious and try things that are too difficult for them.
• I did two dumb things and two smart thing in my first novel. Dumb things (ie, things I wasn’t developed enough to do): writing a quest novel, and using that past-present structure that Ursula Le Guin made work so beautifully in Dispossessed. I didn’t realize until a year or so after Legacies came out where I’d got it from, and why I was so wedded to it. The sort-of-quest structure is difficult to pull off because it doesn’t innately have a strong narrative drive behind it. Smart things I did: having a single viewpoint, and having a character I had deliberately written as attentive and extremely perceptive. Sometimes, wrestling with the need to convey something essential via a viewpoint character for whom it’s not in character to notice that, I miss Lian.
• Certain plots are more bomb-proof than others – they carry their own structure and drive with them. [cref Blueheart]’s initial plot is a mystery, and once I’d got that – the dead body in the ocean – it found its shape quite quickly, carried along by the central question of who and how. By midway through the book the reader actually knew everything, and it turned into a political novel, but by then the central conflict was established and on its way to the climax. I did myself an inadvertent favour, there.
• Quest plots – frequently the first plot an SF&F writer tries – are not as easy as they look: certain choices have to be made to ensure the quest plot gets and keeps its narrative drive and doesn’t become picaresque (a right-on editorial comment about an early draft of Legacies). If I were writing a quest, even now, I’d make sure that what was being sought and who was seeking it were established in the first chapter, and not lose sight of that for a moment. I’m still not sure enough in my plotting to do the young man/woman goes off all unknowing and find his/her destiny on the way. I was unwittingly smart enough to have the quest front and center in the beginning of Legacies’ frontstory, interspersing it with the interleaved backstory in which Lian had to find his mission.
• Passive, reflective characters fall under the heading of Advanced Work. Again, writers have pulled off the reluctant hero wonderfully, but life is much easier if a character wants something and goes after it. Lian climbing over the wall, throwing himself into the path of Lara and Rathla and the story itself, was a wonderfully liberating moment for me.
• Sometimes the writer just has to give up and do what’s obvious – usually because they’ve set themselves up that way. In one of my unpublished novels I was resisting a particular idea because it seemed too obvious. When I finally accepted that it had to be that way, a whole lot of other problems were suddenly solved, because my characters’ repugnance (they didn’t like the idea any more than I did) prompted them to actions that led directly to the showdown. Moral: It’s a bad idea for the writer to argue with their own story.
• Even after (almost) 9.5 novels, I still don’t get control of the plot until my second draft (or later). I’ve just had to do a massive overhaul to keep two of my main characters on the scene for a major action setpiece (this was [cref Shadowborn]). I also had difficulties setting up a crucial event in that conflict, because I needed not to surprise the reader, but I knew that if one of the characters knew about it, it would be out of character for him to leave. So overhaul. And it works. So. Much. Better. Moral of the story: keep the viewpoints where the action is. As long as the action is essential to the plot.
• If I reach the end of my first draft, and it isn’t right (usual metaphor: large plate of spaghetti, stands slithering over the sides), I start cutting. I usually have a fixed idea of the endpoint from fairly early on in the novel, and I reshape the novel to line up with the end. I cut out everything that that isn’t related to the end. Then I put in everything that’s missing.
• On the other hand, all the scenes that end up on the cutting-room floor mean that by the time I get the scene I need, it practically writes itself because all the decisions are made and I have the characters rounded out. Ibsen described his growing familiarity with his characters through successive drafts. In the first draft he knew them as if he had met them on a train (‘One has chattered about this and that’). By the second draft he might have spent a month at a spa with them (‘I have discovered the fundamentals’). By the third draft, he knew them thoroughly (‘as I see them now, I shall always see them’).
• I try to obey Chekov’s Law (‘One must not put a loaded rifle on the stage if no one is thinking of firing it.’), which usually means I have to round up a certain amount of unused artillery during revisions. One of the downsides of writing a trilogy is that once [cref Darkborn] was committed to press, I was committed to firing off the guns lying around. Twelve of them, when I did the inventory in my notebook. I was delighted when I found a way to get four to pop off at once in the archduke’s breakfast.