My plans regarding The Curse of Steel have evolved with remarkable speed over the past few days.

My novel-writing process seems to boil down to the following:

• A long period of chaotic brainstorming, tinkering, and world-building work, in which the overall concept of the story can change many times. This stage can take several years and is likely to produce a lot of abortive partial drafts. Most of the novels I’ve worked on have never gotten past this point.
• Eventually, the concept stabilizes in my head enough that I can produce the first complete draft of a novel. This has happened exactly three times in my entire career as a writer.
• The first draft tells a complete story, but it probably has all kinds of plot holes, undeveloped characters, and vagueness of setting in it. After all, I was making it up as I went along – I seem to be much more a “pantser” than a “planner.” So now I go back and write a second complete draft, one that rounds out the story and patches most of those holes.

About eight years ago, my first mature original novel reached that second-draft stage. That was The Master’s Oath, a tale of time travel, alternate histories, and Hermetic magic that will absolutely never be seen again outside my dead files. It wasn’t until I was finished with that one that I had a “What in God’s name was I thinking” moment and realized the thing was utterly unpublishable. Decently written, not a bad adventure story, but bound to mortally offend big chunks of my audience. Lesson learned.

Now, as of last week, The Curse of Steel has reached the second draft stage. Lo, the creator looked upon his work, and he was pleased.

Now what?

Well, to be honest, my workflow doesn’t exist yet past this point. The Curse of Steel is the first original novel I’ve ever written that I honestly believe is publishable.

I spent a couple of weeks putting together a cover image for the book. While I worked on that, I thought about what the next steps should be. Should I just publish the second draft as is, and hope for the best?

The more I thought about that, the less comfortable I was with the idea.

I know my prose style is reasonably clean. When I was a freelance writer, more than one editor remarked on that. As an editor, I’ve seen enough prose from other people to be able to compare.

Meanwhile, I’ve had a few people reading my drafts, and they’ve been encouraging about the story.

On the other hand, a novel is a very different beast from a non-fiction book, there are necessary skills I may not have developed in full, and none of my early readers are experienced editors. I don’t have a decent writer’s circle to help me hammer my drafts into submission. What if I’m missing something?

No, scratch that, I know I’m missing something. Maybe two or three novels from now, I’ll have a better idea of what to do at this point. Right now I’m still not sure.

The obvious solution would be to hire an editor.

Of course, a skilled and reliable editor costs money. Not to mention, there are a lot of people out there ready to take advantage of would-be authors, offering cheap book-preparation services for top dollar without any guarantee of results. Ever since deciding to pursue self-publishing, I’ve been very cautious about laying out cash for such services.

On the gripping hand, I can afford to spend some money on the experiment.

So over the past week or so, I’ve done some research and developed two lines of attack.

One line is to search for software that can help a novelist pick nits in his prose style. Sure, everything has a spelling-and-grammar checker these days, but strong prose writing needs more than that. I need to be able to ferret out filler words, excess adverbs, phrases that I repeat too often, that kind of thing. I could do that by eye, but the process would be slow and painful, and I would be likely to miss the weaknesses in my own style.

Enter AutoCrit.

Autocrit is billed as a “self-editing platform,” and it certainly works as such. It’s essentially a web-based word processor, but it’s specifically designed to carry out a wide variety of specialized text searches and basic statistical analyses. It compares your text to a huge corpus from published novels, helping you find and carve out the flabby bits of your prose.

Using AutoCrit over the past three days, I’ve been able to rework The Curse of Steel with surprising speed and efficiency. Already I’ve cut a little over 2000 words of material, mostly filler words and repetitive phrases that didn’t add to the sense of each passage.

It’s been quite the eye-opener. Every writer needs something to rub his nose in the shortcomings of his prose style, or he isn’t likely to improve. Lacking an editor or a ruthless critique circle, something like this may be the next best thing.

AutoCrit isn’t ideal. It’s web-based, which I don’t care for. It chokes if you hand it more than 50,000 words of text at once, which means I have to edit my novel in chunks. It doesn’t handle special characters gracefully, so all my conlang words that have accents and umlauts in them get snarled up. Its import from Word, and its export back to Word, are both a little kludgy.

Still, I suspect I’ll have a third complete draft, with much tighter prose, by later this week.

The second line of attack is that I have, indeed, hired an editor. This involved a fair amount of searching through the Web, looking for editorial networks that are competent, reliable, and not outrageously expensive. The SFWA site was fairly helpful here – they don’t explicitly recommend editors, but they have an excellent checklist of things to consider in the process.

The editor I ended up with is being hired specifically to do a manuscript assessment, not a complete edit of the novel. This will set me back a few hundred dollars, and it may not end up being part of my usual workflow in the future. Still, the experiment should be worthwhile. I’m hoping he’ll be able to provide actionable feedback that I can use while producing a final complete draft – the last step in my development process before the book goes out the door.

At the moment, the plan looks like this:

1. Finish working through The Curse of Steel with AutoCrit (to be completed by about 21 or 22 August).
2. Wait for my editor to complete his review of the draft (probably about the end of September).
3. Produce the final release draft (to be completed by late October).
4. Publish the book!

So it looks as if The Curse of Steel will finally hit the virtual shelves by Halloween. There will be much rejoicing . . . and then I’ll get started on The Sunlit Lands, the second book in the series. One assumes that one won’t take nearly as long to reach fruition.

In other news, that four- or five-week gap in September, while I wait for my editor to finish his task? I think I may sit down and work on Architect of Worlds for a while. No promises . . . but I think my research, and the subconscious work in the back of my head, have reached the point where I may be able to develop a rough draft of the third chunk of the world design sequence. We’ll see how things go.