Research, Hard-SF, stats and passing small elephants.

For once a post that only marginally touches on the Hugo Awards. Break out the bubbly.

I’m a long term fan of hard-sf. Now let’s start by establishing more-or-less what I think hard-sf is, because otherwise we could be talking right past each other. Hard-sf has the overwhelming bulk of the story grounded in science and technology which, if not existent are logical, plausible developments of present science. If there are any handwavium/impossiblium elements, they’re not major elements of the story, but may simply exist to allow the story to happen (for example an FTL drive which has brought your explorers to the planet on which the story is set. In a hard-sf story everything works logically and correctly according to sound physics, maths and all the other sciences. In my (fairly narrow) definition, the story and plot depend to a large extent on the science and technology and solving problems related to these, rather than being primarily about interpersonal relations and/or heroic conflicts between people (those could happen in another setting, a ‘solution’ that depends on the character solving a technological/Science problem cannot). Look, there is obviously a fuzziness on the edges of this, and there are stories which you’d fairly define as having a hard-sf elements, but they drift into other sub-genres. Anyway examples of the classic authors I’d think of when I think of hard-sf would be Murray Leinster, Isaac Asimov (the Wendell Urth stories) Larry Niven and Charles Sheffield. There are more of course, (as a more recent example, someone like Patty Jansen) but that’s what I think of, when I think of when someone says ‘hard-sf’. The key I think to ‘hard-SF’ is that if the book is read by a PhD level of understanding of Astrophysics/Math/Chem/Engineering etc… they do not scream in disgust and throw the book against the wall, because the author doesn’t understand what nano-scale actually means, or the time and energy it would take an individual nanobot to walk two feet under its own steam. However, the book or story is still accessible, and understandable by a layman who barely has any grasp of science or math. It’s hard. Incredibly hard to do well – you are bridging a gap of a minimum of seven years of concentrated study without losing either end of the spectrum. If it fails to get the top edge, it’s not hard sf. If it fails to get grasped by readers quite a lot lower on the understanding curve… it’s never going to attract enough readers for the writer to survive. The problem is scientists are actively discouraged from writing accessibly. Trust me on this, or go and pick up any scientific paper in a reputable Journal. Or I can send you some mouldering reprints on The Age and Growth of Callorhynchus capensis It took me a long time to unlearn it and a lot of effort to learn to write for readers, not Journals. Thus: In my opinion it is the hardest kind of science fiction to write well, especially at the novel level. It’s something I feel is largely beyond my ability, and I’ve got a good foundation in it, and my research notes typically run to three to five times the length of the book. Broad appeal humor sf/fantasy – which I try to write sometimes — being the second hardest IMO. Different people enjoy reading different genres or sub-genre, and love writing them, but some demand more of the writer. If you’re going to get all ‘Social benefit’ about writing, there is no doubt that good accessible hard-sf outscores every other form by an order of magnitude: it makes what is hard and we need seem accessible and possible. It has vast, largely unperceived value. Firstly: It interests and thus draws many people into those fields, secondly: it raises our expectations, and oddly, that drives us and our society, and thirdly: it has an enormous effect on research funding, by making the ideas known and popular among the tax-payers. It has had more effect on our society than any other form of sf, simply because so much of our current mores, tolerances, and society rest on the abundance and possibilities generated by it (and yes, John Carlton, you can quote me).

Okay, so you get the picture. I think good hard sf authors are both rare and precious gems. I think they deserve huge respect, especially if they make it look easy, because it’s not. So… what is all this about? Am I trying to put you off writing hard-sf? Absolutely not. Please, please do. I don’t care who you are or what your background, I need you, and I believe everyone else does too (even those who never read sf. This stems from Tom Knighton’s delightful fisking of Cathrynne Valente here, and my comments on it – where I said this ‘she told us that you didn’t need to know or understand any science to write hard SF.’ Apparently Valente got very upset by the comment. (No I don’t ‘loathe’ her. That’s far, far too strong. Disinterest is about right. We have nothing in common, although I did read one of her books. It wasn’t my sort of thing – I really don’t like modern literary stuff much, as you might just possibly have gathered — but I read it before the con, an elementary politeness she obviously hadn’t even considered. As I said to Tom, it was like a scruffy working sheepdog from the Outback meeting a New York indoor Persian Cat. No common ground.) When she said this I did nearly as good an imitation of a stunned mullet as Toni did. Then I laughed a good Wagnerian laugh, like quite a lot of other people. When you’re in a hole… stop digging. But she attempted to extricate herself by telling Toni (and the audience) that you could just do some research. Anyone could write hard-sf. Um. Catch 22. You CAN do research. But you have to know what level of research you’re talking about. What I think she failed to get, completely, was that writing hard-sf requires a high level of understanding of the sciences and technology. You can certainly get there without ever going to college. Without ever getting a post-grad degree. But you have to understand it well, so you can make others understand it well. You have to be very skilled writer to translate the extremely complex into the accessible (and yet, still complete). It’d be like deciding you wanted to write a novel set in Greek, not knowing a word of the language… except that might be easier, because you’re not trying to learn it and then write it so English speakers and Greek speakers both get it. It’s not a light undertaking, even if you have a science background. It’s a far bigger hill if you’re starting at the bottom. And it’s not one a hill you can get to the top of from without knowing quite a lot of science. So that is my answer to Cathrynne Valente. You might _decide_ that you want to write hard-sf completely ignorant of even the rudiments of math or science. But by the time you’ve done adequate research to do so (starting from zero, we are talking many months or years not two minutes on Wikipedia), you know a great deal about that that patch, and probably (unless you are going to make a horse’s ass out of it) have a good foundational knowledge. So, no, Cathrynne. You can’t write it without understanding it. Sorry. It’s hard. Not just anyone can do it (I don’t think myself good enough, and I have a lot more background in science, and I’m a fairly competent hack). Those who can, and can do it well, deserve great respect. You weren’t giving it. The problem was that you were jumping to conclusions without knowing the enough of the fundamentals. I might be able to write literary sf/fantasy, with a bit of research. But I don’t know (or care) enough to say that is the case.

The problem with not knowing enough is one I often hit with that elementary subject, math stats. And before you tune out, it’s like basic arithmetic – a vast asset that can make the difference between success and failure. Now, I would like to know more about stats, but the single most important thing you have to understand is GIGO. Garbage in, Garbage out. No matter how simple or complex your analysis – whether you’re trying to work out the average price for new E-pubs or trying to predict the maximum safe catch of multispecies fishery – you have to understand what you’re sampling, and what you can measure from that. John Scalzi kindly provided us via his friend Jason Sanford a near text-book perfect example of GIGO. “Recently author John Ringo (in a Facebook post previously available to the public but since made private) asserted that every science fiction house has seen a continuous drop in sales since the 1970s — with the exception of Baen (his publisher), which has only seen an increase across the board. This argument was refuted by author Jason Sanford, who mined through the last couple of years of bestseller lists (Locus lists specifically, which generate data by polling SF/F specialty bookstores) and noted that out of 25 available bestselling slots across several formats in every monthly edition of Locus magazine, Baen captures either one or none of the slots every month — therefore the argument that Baen is at the top of the sales heap is not borne out by the actual, verifiable bestseller data.” As I said: first you need to understand what you’re sampling. For example, if you set up a pollster at a Democratic convention, at 10 pm, in a site just between the bar and the entry to the Men’s urinals… even if he asks every person passing him on the way in, you’re not going to get a very good analysis of what Americans think of a subject. Or what women think of the subject. What you will get is middling bad sample of what mildly pissed male Democratic Party conference attendees think. Middling bad, because many of the passers will be hurry to go and pass some water first. It’s vital to understand what you’re sampling – or what you’re not. Let’s just deconstruct the one above. In theory Sanford was attempting to statistically prove John Ringo’s assertion wrong. What he proved was nothing of the kind (Ringo may be right or wrong, but Sanford failed completely). What he proved was that on the Locus bestseller list, (the equivalent of the Democratic Party convention and the route between the bar and the gentleman’s convenience) that Baen was not popular. That is verifiable. The rest is wishful thinking, which may be true or false. Firstly ‘Bestseller’ does not equal sales numbers. A long tail – which Baen does demonstrably have, can outsell ‘bestseller’ and five solid sellers outsell one bestseller and four duds. Secondly, independent bookstores who self-select by accepting polling, selected by a pollster (Locus) with a well-established bias are not remotely representative of book sales in general, or representative of the choices book buyers have. Thirdly, it is perfectly possible to ‘capture’ no bestseller slots at all, even in a worthwhile sample (which Locus polling isn’t) and STILL be the one house that is actually growing. It depends what you’re growing from – which of course this does not measure and cannot.

Short of actual book sales numbers, and data on advances – which we’ll never see, staffing is probably the best clue. I know several authors at other houses whose editors have left, and quite a lot of other staff at publishers who’ve been let go. Over the last few years, the number of signatures on my Baen Christmas card have gone up year on year.

Anyway: This is how you do not use statistics. Look and learn. The rest of the article is more-or-less just as confused. Scalzi sets out to prove that Baen does publish people across the political spectrum, a fact that many of us, (myself, Eric for a start) have stated loudly and publicly for many years, on numerous occasions – and have been routinely ignored and shouted down. But now it’s different. It ‘proves’ his point, so John Scalzi states this is so, and mysteriously concludes that this means that Baen is NOT exceptional. I think he’s missed the point by several miles. It’s not that Baen are socio-politically diverse in their authors, both historically and recently. It’s that THAT is the exception. The other publishers, particularly recently, are, I think, demonstrably NOT. Yes, the other houses may have one or two authors, usually old and long established who don’t conform to the norm in that house (At the moment that norm is, like the recent non-puppy Hugo nominees, essentially outspoken Left wing) But that doesn’t make them socio-politically diverse and just like Baen, unless you assume that a 97% white zipcode is just as diverse and the same as Fergusson’s 40% white.

I’d love to see Baen stop being the exception. It would be good for the publishers, good for readers, wise for a changeable future — but I don’t think that’s very likely.

And now for that passing little elephant… Look… there it goes.