The Primacy of Ideas
One of my SF teachers, a brilliant man whom I respect very highly, said something once that I still don't understand: "In the end, what people will remember about your fiction are the characters." This was in the context of an intense discussion about character creation, but it seems extreme to me. In some sorts of fiction, sure. What I remember about Saul Bellow's The Adventures of Augie March is Augie March.
Or is it?
Sure, I remember Augie. But I, too, am an American, Chicago-born. A great deal of what I remember about The Adventures of Augie March is depression-era Chicago, and how it shaped Augie's character. Without Chicago, there wouldn't have been anything particularly memorable about Augie himself. I bring this up because I'm encountering more and more new writers who seem to think that ideas don't matter in SF and fantasy. Characters are the whole story. Everything else is backdrop. That simply isn't true, and I think it's time for a little pushback.
Here's how I see it: The two essential elements in any story are characters and context. Without characters, context is narration. Without context, characters are soap opera. The magic happens when you rub one against another.
In mainstream fiction and real-world genres like romance and mystery, you don't create your context so much as select it from a huge menu of known placetimes and cultures, like Chicago in 1933, modern-day Manhattan, or Amish country in the 1950s. There's some tinkering around the edges, but for the most part you pick a well-documented placetime and turn your characters loose in it. If you're a good writer, entertaining and insightful things will then happen, and your readers will come back for more.
It gets interesting when you switch from real-world genres to SF and fantasy: You can then create your own contexts. World-building is (as I like to say) a spectrum disorder. You can build a little or a lot, or go nuts and create entire worlds and societies from whole cloth.
To do that, you need ideas.
For good or ill, I'm an ideas guy. It's just how I think. Furthermore, I have a hunch that ideas are in fact what people actually remember about good SF and fantasy. Really. C'mon, when was the last time you heard somebody ask, "Hey, what was the name of that story in which a callow young man is jolted out of ordinary life and with the help of an ironic sidekick finds unexpected strengths and talents that allow him to defeat evil in ways that change him forever?" No, you hear questions like this: "What was the name of the story that had an FTL communicator in which every message ever sent, past, present, and future, is gathered into a beep at the beginning of every message?" (I know the answer, and if you're serious about SF you should know it too.)
When I read SF, I want to see cool ideas. When I write SF, I feel a responsibility to deliver them. It's not just about having rivets. It's about having rivets that nobody's ever seen before. Is it silly to love the rivets? Well, I've gotten several fan letters about the wires in The Cunning Blood. The novel centers on a prison planet in which microscopic nanomachines seek out and disrupt electrical conductors, supposedly keeping the prisoners from developing electrical technologies. Well, the prisoners make non-disruptable wires by filling hoses with mercury. When your rivets start getting fan mail, I think it's fair to assume that you're on to something.
This sort of idea-centric story isn't for everybody, but there are a lot of people for whom it's the heart and soul of fantastic literature. The challenge is to use clever ideas to draw out characters that grow, change, and learn. I'll freely admit that I'm still better at ideas than at characters. However, I'm aware of the issue and I'm throwing a lot of energy into the character side, now that I'm finally out of my teens and into my sixties.
I'll grant the "cowboys on Mars" objection, in which an ordinary situation is dropped without modification into an exotic locale and called SF. However, it's just as bogus to say, "Nobody cares about your starships," when the starships are in fact a key part of the story's context. Jack Williamson's definition of stories as "people machines" is correct but incomplete: To have a people machine you need the machinery. Without that machinery, you have "white universe syndrome" and your story collapses into soap opera. You can choose your context from a menu, or you can build it. Either way, you need that context to make characterization meaningful.
I'll get myself in trouble here for going further and suggesting that a story's settings and ideas can be entertaining and sometimes dazzling, even when its characters are thinner than we'd like. That's not an excuse but simply a fact of life. Do we remember Ringworld because of Louis Wu? Or do we remember it because of, well, the Ringworld?
As I prefer to put it: Ideas will get you through SF stories with no characters better than characters will get you through SF stories with no ideas.
That said, have characters. Have context. Rivet them together so well that both your characters and your rivets get fan mail. Then, my friend, you will have arrived.