Over at Velcro City Tourist Board, Paul has posted an interesting article on writing tips. All writers are interested, to an extent, in codifying the principles of their craft (though some will claim that it breaks the spell to talk about it; I'm sure even these are aware of certain and persistent shaping forces in their fiction). Learning to write is an ongoing process, and, as part of this, I'll try to think up ten of my own. There is risk that this will be described as 'intellectualising'; if that's what you think this is, you haven't seen something sufficiently intellectual. I think I'd like to call these 'beliefs' because they aren't necessarily what I would impose or teach. These are just honest statements about what I've learned from trying to write novel-length fiction (which I've been doing for about twelve years).
(1) Every story, scene, paragraph and sentence must have an identity
The way a conceptual story unit (like a sentence) should work is as a crew member pulling his or her weight. Every word within a sentence must contribute to the identity of the sentence: What single thought does this sentence capture? What is its tone? Does its identity clash with the greater identity of the paragraph? If so, is this a good clash?
When you think of a good book, you probably feel something akin to a semantic thumbnail impression; a single sensation that captures what it is like to have read that book. That is the story-level identity, and it is informed by the nested identities within it.
(2) The effect of a good story should be greater than the combined effects of its components
Crikey, that's an undigestible sentence! By this, I mean that a story should operate along the Gestalt principle of 'the whole is greater than the sum of its parts'. For example, you could look at a disembodied nose, an ear, and an iris, and they mean nothing to you in isolation; but when arranged into the magnificent visage of President George Walker Bush, a new level of meaning erupts. Likewise, while the individual components of story - perhaps at the level of the sentence - may lack meaning or be plain ordinary, at the higher levels of story, where they are combined, they can appear extraordinary. Thou art more lovely and more temperate is so-so, and the word 'temperate' here borders on ugly. But placed within the structure of Sonnet 18 by You Know Who, some form of magic is conjured.
(3) The principles of story structure are to be applied in reverse

George: "But Joseph Campbell was quite specific: At the beginning of the second act, introduce the green Muppet."
This one is quite controversial because (i) stories do appear to have structures and (ii) bad stories appear to have structural problems. It seems straightforward to conclude that the application of certain structural concerns (a la The writer's journey or Story) will provide support just as sound engineering principles can ensure a building will not topple. I don't think these structures can be applied ahead of time (i.e. while the shape is being drafted) because they arise, I think, from an interaction of character, conflict, momentum, etc. They are superstructures that appear only when the jobs of lower levels are complete. Since I'm not trying to tell anyone how to write (and wouldn't even if I could write brilliantly myself), I won't fill this paragraph with loads of caveats, but here's one: This principle will be useful if you use an iterative approach to your fiction (what used to be called 'drafting'). It's possible that you do your writing in your head and only sit down when you've got an effectively finished product. In that case, the story-level superstructure might already be at work. A story begins with an inking of its identity; not a track with hurdles it must jump.
(4) A story is a place

Infocom's text adventure, The Hitchhiker's Guide to the Galaxy: 'I don't know the word 'shove the atomic vector plotter up the game designer's arse'. Do you mean 'hang the gown on the hook'?'
When you write, you cross a border into a new land. If you don't do this, you won't be able to write. You might need things to help you pass. The visa of a good piece of music or a photograph or a smell. When you are fully in the land of the story, you can appreciate its identity and make appropriate decisions about its construction. When you aren't, your decisions might be wrong.
(5) Writing is an unconscious process
Britta: 'Shhh. Ian's writing'
As a psychologist with a cognitive bent, I try to be careful about how I use the word 'unconscious'. Here, I do not intend to invoke Freudian or otherwise psychodynamic (mis)conceptions. I'm talking specifically of processes that occur below one's conscious threshold. The saccade of an eyeball to an interesting object is an unconscious process; as is basic reading. I believe that creating a world - i.e. writing a book - is one of those messy, intuitive tasks not suited to the narrow confines of the conscious 'wave crest'. Let yourself sleep on important story decisions. Feel your way through difficulties. Wait for your unconscious mind to background process a problem until the solution appears. Even the most dead-end of narratives can be rescued, because you'll be thinking about it all the time, because it's bugging you, right? If it's not bugging you...then you're not in the land of the story.
And now for some more practical beliefs.
(6) Writing is iterative
Kerouac notwithstanding, you will not be in a position to tell a story as beautifully as possible while writing the first draft because at least some of your mental resources will be employed in maintaining things like visual awareness, weather, location, the complex matrix of characters' intentions, and so on. Go back to the draft, identify its problems, then fix them until the draft is beautiful. Then wait. Then go back, find more problems, then fix them. Repeat until you are completely sick of the text, then repeat a few more times. If you still love the story after all that - even look forward to the next bout of editing - you either have (i) a story that works or (ii) a condition that needs medical attention. One interesting thing to note about editing is that, often, an edit can improve parts of the draft but make other parts worse (because of inconsistency of fact or tone). What you're performing is an evolutionary process: sure, a particular gene might be an improvement, but if it doesn't jive with all the other genes, you're a dodo.
(7) Don't use adverbs unless absolutely necessary
In my current novel, I've got about five adverbs. That's five too many.
(8) Delete words from a given sentence until the sentence breaks; then re-insert the minimum to fix it.
Self-explanatory, I feel.
(9) They are not called characters. They are called people
And they are responsible for everything that happens in a work of fiction, not you.
(10) Set the dial on your quality threshold to 11

The man from Havana - he say 'Nooooo!' And 'Pass the horse radish, sweetie - this warthog is trifle ripe'
Nothing should be 'good enough'. If it is weak, it is your job to strengthen it. Never say to yourself 'I'm only an apprentice writer, and this is acceptable'. Say to yourself 'Would Ernest Hemingway/John Updike/Martin Amis put his name underneath this paragraph?' No? Why not? Fix it. If it can't be fixed, destroy it.
Wow, that was cathartic. Mmmmm. What do other people think?