Brad Templeton says:
I would put it simply as "science and technology seem real and are 'characters' in the story". When I say characters, I mean that you notice and remember them as much as you would a memorable character. They're important to the story, it would not be the same without them.
Doug Tricarico goes into slightly more detail:
I'm partial to the idea that fantasy deals with the impossible, while SF deals with the possible; that hard SF deals with hardware while soft SF deals with wetware. (Hard: Hogan's Voyage From Yesteryear. Soft: Atwood's The Handmaid's Tale.) The best science fiction, however, deals with both aspects, applying the science to human behavior and systems. As Niven once said, "A good SF author invents the car; a great SF writer comes up with the traffic jam."
Another person (whose name I forget) applies the term "hard" to works where the author has gone to the trouble of showing us a way of life that has been shaped, or at least is different from ours, by the science or technology in the story. They instanced some of Bujold's books. I regard this as where the line should be drawn between all real ScF and "Mars Westerns" or "space fantasy", but have to admit that it's not very far from Brad's definition. Do you agree with me that there's a subtle but important difference?
So I won't rule something out as hard ScF because we presently believe it to be impossible; rather, I see degrees of plausibility based on the extent to which accepted theories would have to be re-written. In particular, it seems to me legitimate to base a "hard" ScF story on an invented theory that explains some phenomenon which current theories (as of 1995) do not explain well, e.g. the solar neutrino deficit or the "missing" dark matter. One could use "Science-Based Science Fiction" for that which dreams up extensions of current theories (or dreams up whole new theories), and "Scientifically Accurate Science Fiction" for that which sticks rigidly to current theories.
Matt Austern comments:
I don't see that scientific accuracy has anything at all to do with genre distinctions. You see, one category I recognize is "science fiction that has inaccurate science". As I said, I'm a scientist, so of course I know more about science than most SF writers; picking technical holes in SF stories is an easy and largely pointless task. Almost all SF books I have read contain either errors or made-up science that contradicts things that are known today; most of the exceptions are books that are so vague that there isn't any substantive scientific content. At some point, if you know enough science and if you want to continue to enjoy SF, you just have to learn to stop caring. I don't see the value in defining SF so strictly that the set of "true" SF books becomes the empty set.Matt has also said something to the effect that he doesn't see much point in distinguishing hard ScF from any other kind.
Christian "naddy" Weisgerber offers this definition of "hard":
"SF that is written to a high degree of conformance with current scientific knowledge, where all extrapolation of new phenomena is plausible, self consistent, and limited in number and/or scope as to not reduce its effects to arbitrariness. The plot should center around the exploration of a scientific phenomenon, its applications, or generally the application of science and engineering to the solution of problems."
All the same, the "strict" definition of hardness is useful, and I have great respect for authors who can stick to it and produce interesting work. Really I'd like to see a change in terminology. There are, after all, works which avoid assuming any changes to current science simply because they avoid assuming much about science at all -- for example, some of the books about over-populated future Earths -- and I would hate to accord them the accolade of calling them "hard" while Niven's Neutron Star is denied it because he assumes FTL travel. (OK, for a "hard" writer, Niven proposes some remarkably implausible tech.) Some purists would probably apply the term "science fantasy" to works that go beyond currently believed science, and that leads us to discuss the distinction (if any) between ScF and fantasy.
Hard SF is a form of alternate universe fiction, set in a world where the world-view of American engineers in the late twentieth century is a precise reflection of The Way Things Are.I find the name Soren F. Petersen attached to it; hope that's right.