The most recent 20 comments posted to Making Light by David:

Oh, I do like fuzzy problems like this. Where do you draw the line at mainstream English is really pretty easy: it's not a line. It's a foggy boundary, not well defined. (If you could make it a line, you could program that, and grammar checkers would be more than a joke.)

Look instead at "inside the boundary" vs. "outside the boundary". These are two pretty sharp areas, easy to recognize. (Well, easy for ML readers anyway.) Then there is the fuzzy area in between. Mostly, that's ok, not least because the fuzzy areas don't occur all that often in daily life. (Unless you're TNH, reading the slushpile.) And I know that's no help at all to the non-native English speakers. Sorry. (I truly suck at other languages, so you're already up on me.)

But think about it. How often do you really run into fuzzy cases if you're not looking for them? One source is the upwelling of slang, but it's never stopped and hasn't killed off the language yet. Another is widespread redefinition of technical terms. (A font is a size. That is not a font. That is a typeface. Oh, I give up.)

While we can't do a rule-based analysis, most individual cases fall into clearly right and clearly wrong. Such rules as we do have are mostly aids to clarification -- if you don't write it this way, people will be confused about what you meant. That's bad, even if they do understand you eventually.

When you do come across an instance of clearly wrong, it's usually grouped with other errors such as spelling and punctuation. These let you know the writer has not mastered the basic tools of the trade, and tips you off not to waste any more time on it.

Murky and imprecise though the non-definition is, it nonetheless serves very well in most specific cases.

I wonder, in great ignorance here, whether a Bayesian approach working with whole sentences would be more successful than an ordinary rule-based grammar-checking system. Anyone know if this has ever been tried, and how well it worked?

Larry, that's quite true. I am strongly reminded of Paul Erdman who spent 9 months is an Swiss jail for something his Swiss banker employees did without his knowledge.

Simon, in the longago (and I decline to think how long ago it was, but Covenant etc. was still in print), I dungeonmastered a campaign with several elements shamelessly ripped off from Donaldson.

I had to remove them because not one of the participants, (all SF/Fantasy saavy or they wouldn't have been playing in the first place) understood what they were or how to deal with them. None of them had ever heard of Donaldson.

Anecdote? Yeah, sure it is. It remains that I've never had a serious discussion with anyone about this series, or anything Donaldson ever wrote.

As for being popular and reaching the bestseller list, well, so did whatsisname, Jordan, and so did Dan Brown. 'nuff said, n'cest pas?

His last series, I just gave up on as being too ugly. Well, Donaldson-bashing is well beside the point, that being that if he's holding him up as some kind of exemplar, we agree he's missing the boat by way too wide a margin.

Just thinking of the people he COULD have cited instead, nearly makes one weep. LeGuin, Moon, Brust, Mieville, Friedman,Modesitt, oh no. We've got Donaldson.


Advice from his excerpts remind me of a paid "learn the secrets of the stock market" course I once bought from Forbes Magazine. They had a per-lesson money back guarantee, which I used early and often, viz:

I already knew that, here's your "lesson" back, refund my money.

Repeat, repeat, repeat.

"Umm, maybe you should sign up for our "advanced course" instead."
"Ok, send me that instead." They sent me the first advanced course lesson. (No more quotes, it bugs our gracious hostess, and she might thing me.)

I already knew that, here's your advanced lesson back, refund my money.

They sent me the refund and stopped talking to me anymore.

--------------------

Romance and SF are different genres? How I wish someone had told me that long ago, and saved me all the time/postage I wasted sending bodice-rippers to Asimov's and F&SF. Thank God he's convinced me I should probably not send them to Tor, eh?

In the original auction, though not the second one, there is a photo of one paragraph of Mr Rice's manuscript that can actually be read. If you want to. I believe it says all you need to know.

The first sentence is clumsy and suffers from a needless attempt to force it into present tense. It would have been fine, and more natural, in the past tense.

The second sentence is a fragment, which should have been a dependent clause of the first sentence.

The fourth sentence is overlong, obscure, and nearly incomprehensible. And the entire paragraph's sense could have been conveyed in one short sentence, to its considerable benefit.

One paragraph (particularly where selected by the author as an exemplar) is enough to demonstrate that one lacks sufficient language skills to be a successful writer. I don't need to read anymore to know I'm not interested in the rest of it.

Or as someone once said, "I willingly edit, but editing does not mean having to rewrite the whole d****d thing."

Ok, just one thing.

If I start getting spams offering to sell me an unique artistic vision (UAV), or to sell me pills to enlarge my UAV, or lower its interest rate, then I will personally blame this on TNH, "Making Light", and its hangers-on for having caused this.

Further, if I start getting spams offering me pheromone sprays or colognes offering to make me more attractive to people who are actively seeking new authors, I will blame CP, "Scrivener's Error", and its hangers-on for having caused that.

Just so we're clear.

Exercise for the student: write one of these. No, no, I didn't say to share it, I just said to write it.

This is a formidable task, maybe I can help you steer with some idea of what sort of books they are. I purposely avoid the many great suggestions others have given.

(But first, to state what everyone is thinking but no one has actually said: stay away from Robert Jordan. Really.)

Glen Cook has written several quirky sorta-kinda fantasy detective novels, starting with "Bitter Gold Hearts". Ignore the unfortunate blurb from Locus about this being an homage to Raymond Chandler. It is NOT. Whoever wrote that was clueless and illiterate. It is an homage to Rex Stout, who was a much better writer than Chandler and is the only mystery writer (apart from Doyle) whose works repay re-reading. Off-beat, but if you like them, you'll like them a lot. Cook's other stuff is good too, but mostly out of print now.

L.E. Modesitt, Jr. does anything he tries really, really well. I would suggest you try "Adiamante" or "The Octagonal Raven" as an excellent place to start. He mostly writes series, and they are mostly good, but that may not be the best place to start. If you want to try that, pick one of the "Of Tangible Ghosts" novels.

There is a whole subgenre out there, of military SF, which a lot of people don't like at all, or even acknowledge. Others like it a lot. I'd suggest starting with one of David Weber's "Honor Harrington" series to see if it is to your taste. John Ringo is also a good choice. You can get these at the Baen Free Library on the web.

John Varley is an excellent writer, but so eclectic he's difficult to categorize. If you can find "Titan", it's a great start. But so is anything else.

Jack Vance writes some of the most imaginative "is it fantasy or SF?" stuff around, with some of the wierdest, quirkiest and best-imagined societies anyone ever thought of. It should be pretty easy to find his story, "The Moon Moth" on the web. I've never read a bad Vance.

William Gibson's "Neuromancer" is probably something that ought to be high on your list. It was a very important landmark in the field, creating its own subgenre, cyberpunk. You'll do yourself a favor to learn whether you like it and want more, or don't and want to avoid it.

Larry Niven has gotten pretty hard to find in bookstores anymore, but he's an excellent example of what is called "hard" SF. Which means that he mostly tries to keep his science straight, and his plots scientifically possible. Frequent co-author Jerry Pournelle is also worth a place on your short list, but it is their work together that really deserves attention. "The Mote in God's Eye" is arguably one of the best books ever written in the field, and your best touchstone for learning whether you prefer hard SF or not. This is perhaps the best novel written about "first contact", the first time humans meet an alien race.

("hard" vs what many call "space opera". For example, FTL travel is impossible. So you do without, or invent an at-least-not-proven-false theory that makes it possible. That's "hard" SF. On the other hand, you just assume a warp drive. That's space opera. The line is as blurry as they come.)

At some point you should take a look at Dan Simmons. His formula is "write a book, win the top award in the field, change fields, repeat". This is pretty dense stuff, so maybe not the place to start, and his mystery and horror may be more accessible. But at some point you'll want to check him out to see what's among the best writing in the field.

Comment statistics for David on the Making Light blog

Total: 6 comments. View all these comments on a single page.