The second part of Kelly Lagor's LOCUS article on "Putting the 'Science' in Science Fiction" is here:Putting the Science in SF
As in the previous essay, she quotes opinions from various authors and editors, including Elizabeth Bear, Sarah Pinsker, Lee Harris (editor at Tor.com), and Sheila Williams (editor of ASIMOV'S), among others. Some bits of advice on the "delicate tightrope walk" of "getting the level of detail just right so as to not be so technical you alienate your readers, while avoiding being needlessly inaccurate":
An SF author should keep up her "baseline knowledge of popular science" (in Elizabeth Bear's phrase) at a level sufficient to make her aware of what's going on in the scientific world and where she needs to seek out deeper research into any particular topic or sub-field. Academic journals and popular science books and articles each provide useful resources, which should be consumed in the proper balance. Other comments logically point out that the amount and kind of research needed will depend on how much the author already knows about the field. The level of scientific detail required to make a story plausible also depends on the subgenre. Readers of different types of SF have different expectations; as Lee Harris observes, "we’re much less critical of the science in the latest superhero epic than we would be in a hard science fiction story." Another observation states that "with great familiarity can come great reluctance"—a writer might hesitate to delve into the technical details because he or she finds it hard to resist including excessive exposition that might turn off the reader. Some other suggestions: Don't hesitate to consult experts firsthand. The kind and degree of technical specificity varies depending on the viewpoint character—what would he or she notice and care about? And getting the depth and scope of detail correct ultimately grows out of knowing how much the reader needs to understand to enjoy the story. "Sometimes, when it comes to details, less is more."
By the way, Lagor's phrase "needlessly inaccurate" seems to imply the existence of conditions under which inaccuracy is needed, a position I'd find hard to agree with. Whether the density of detail is heavy or light, surely whatever IS on the page should be accurate, within the limits of how technical the particular text gets. Even in fantasy, I find a story more interesting and entertaining if the writer gives the impression of accuracy in mundane matters such as architecture, food, travel times, etc., as well as basing the biology of imaginary creatures (for example) on a plausible analogy with real ones. The more incredible the central premise a reader has to accept, the more plausible the supporting details ought to be.
Margaret L. CarterCarter's Crypt