I recently completed the first draft of Realms of Light, a sequel to Nightside City, which was originally published by Del Rey Books in 1989.

I’d started Realms of Light shortly after Nightside City was accepted; I was very enthusiastic about Nightside City, which I thought was the best thing I’d written up to that time. Unfortunately, the market was significantly less enthusiastic. I was already typecast as a fantasy author at that point, and my science fiction novels had not sold well. Del Rey was not at all interested in a sequel to a novel that hadn’t done all that well, so I shelved Realms of Light.

Except I kept pulling it back out every couple of years and looking it over and adding a paragraph here, a few pages there. And finally, in 2008, having established that it was possible to make a modest amount of money by serializing novels on the web, I decided to go ahead and write it as a serial, to be published by a small press, FoxAcre Press, which had picked up the reprint rights to Nightside City a few years back. And now it’s complete, though it still needs to be revised and edited.

So it took me twenty years to write it.

Twenty. Years.

A lot of things changed in those twenty years. Jumping back into a setting I’d created in 1986 and hadn’t seriously worked in since 1989 was a challenge.

It was not, though, that I didn’t remember it; I did. It’s that the real world had changed in ways that make some of my 24th-century setting look curiously dated.

That paragraph suggests two separate topics – did I really remember it as well as I thought? And how was it outdated?

How well I remembered it – well, let me put it this way: I didn’t bother to actually re-read Nightside City, nor the notes I used when I wrote the first novel. I remembered Epimetheus about as well as I remember, say, Boston. I did check a few details here and there, such as making sure I had the full name of the New York right, but mostly I worked from memory. Did I get it all right? I don’t really know; that’s one of the things I’ll be looking at when I write the second draft. I thought I remembered it all. I know the floor plan of Carlisle Hsing’s old office on Juarez Street, I know what the Trap looks like, I know the geophysical structure of Epimetheus.

I did catch a couple of errors when I checked myself against the original novel; I’d misremembered part of the Nakada family tree, for example. Mostly, though, it was still in my head. After all, I’d lived on Epimetheus, in Nightside City, for a few months back in 1987. (I finished writing Nightside City in November, 1987.) It was as familiar to me as other places I’d lived in, such as Pittsburgh.

I probably got some stuff wrong, but then, I’d probably make a few wrong turns trying to get around Pittsburgh after all this time. And some stuff may not be so much wrong as just different; anything that wasn’t in the first novel I was free to change, so if the version I remember isn’t exactly what I’d thought up back in the ‘80s, who cares? Who’ll ever know? I mean, a lot of it was never written down in the first place.

But then we get to the outdated stuff.

Nightside City was inspired by the cyberpunk movement. The actual style owes more to Ross Macdonald, but there are a lot of cyberpunk elements, and cyberpunk reflected the 1980s.

It’s set in a society dominated by Asians – or rather, people descended from Asians. The wealthiest families all have Japanese names, while the managerial class is a mix of Japanese, Chinese, and Indian; the white people in the novel are generally working class or worse. The cyberpunks often extrapolated a Japanese-dominated future, because in the 1980s that looked plausible – we didn’t know the economic bubble was going to burst. If I were writing it now the future would probably still be dominated by Asians, but the Japanese would be much less prominent, and I might mix in some other ethnicities. I couldn’t really change that for the sequel, though; the Nakada clan is central to the story.

Some of the computer stuff hasn’t aged all that well. In 1987 the invention of the World Wide Web was still a few years in the future, and my guesses about the future shape of the networked world – well, I did better than some authors, but some of it looks slightly quaint now. All that jacking in, and using hardwired connections rather than wireless, feels a bit off, too.

I did correctly anticipate a few things, though, even if I got the names wrong. The word “malware” didn’t exist in 1987, so far as I know, but I could see that there was going to be a need for a word like that, and coined “gritware.”

Mostly, I think it still works. I’ve had to abandon stories because the real world made them obsolete, and I never considered abandoning Realms of Light; the flaws were pretty minor, and could be explained away as the result of stuff happening in the next three centuries that we don’t expect now.

And picking it up again after twenty years should have been hard, but it wasn’t. This story’s been in my head all along. It’s changed over time – the present version is not what I’d have written in 1989 if Del Rey had asked for a sequel – but most of the changes are plot-related; it’s basically the same setting, basically the same characters.

But you know one thing I did forget? If you’re a donor and got to read the first draft, you may have noticed the dedication to Ed Bryant, who I credit with giving me the clue I needed to make the plot work.

He did. I remember that. But I don’t remember what it was. I remember him talking about a movie I haven’t seen, and that up until then I hadn’t had a viable plot for a sequel, and after listening to him it fell into place and I had a set-up and an ending. (The middle came later.)

But I don’t remember what it was he said that made it all work.

Sometimes writing fiction is a very weird business.