This is a question Harry raised: When you're writing a story, is it better to rewrite as you go—getting the beginning right, for example, before you move on to the next scenes—or is it better to get the whole story out, no matter how rough the draft, and go back and rewrite it later?
As is usual with writing questions, the answer is: It depends. It depends on what works better for you, and it depends on the particular situation.
In general, I favor the "get it all down first" approach to writing—because you have a better chance of writing it all out in a creative burst, or a least a sustained creative flow, rather than letting the editor in your head slow you down and maybe even stall you before you ever really get going. On the other hand, some writers don't let go of a page (or a scene, anyway) until they've polished it to their satisfaction. I've heard that this is Orson Scott Card's approach, and he's certainly no slouch as a writer. (I didn't hear it from Scott Card himself, but I did hear it from a friend who heard him say it. So apply your own standards of credibility to me as a source on this one.) Most writers I know well, though, tend to plow ahead, unless there's something that's really bothering them about what they've already written.
I actually use a modified form of "get it all down." I show my work to my writing group as it gets written, a chapter or several chapters at a time. I do not, however, inflict my raw first draft on them. I get it down, then before I send it to them, I go back and do an edit pass to clean it up some and make it more like what I was intending to say the first time. Sometimes doing editing on yesterday's work is a good warm-up to creating today's work. But I try not to indulge too much in that.
Here's the thing. What most people have the most trouble with is the story structure: getting the bones laid out in a way that makes sense, is interesting and entertaining, and—if you're lucky or good—even compelling. And the structure is easiest to see when you have the whole story in front of you, for good or bad. It's very easy—and God knows I fell into this trap over and over as a beginning writer—to fuss and fume over little points of style, getting the descriptions just right, getting the words to flow in a pleasing way, tweaking dialogue, and all the while overlooking the fact that your basic story is flawed. Maybe it doesn't go anywhere, or the motivations don't hold up, or you've only got a piece of a story or a mood piece. You become so involved in the minutia of rewriting line by line that you miss the larger flaws altogether. (Reason number 3 for being in a writing group or workshop; your colleagues will help you catch that.)
Now, using this method can mean that you have a hell of a lot of revising and rethinking to do when you go back to rewrite. That, in fact, is precisely the situation I'm facing right now with Sunborn. I have tons of notes for revision that I compiled over the course of the first draft, and I have a fair number of chapters that I might have spent time polishing, when I see now in the context of the whole book that I need to cut them drastically, or perhaps out altogether. It doesn't make the rewrite fun, but it does (I think, for me) offer the greatest chance of success in the end. Eternity's End was a mess in its first draft, and so were many of my other books.
On the other hand, when I wrote Neptune Crossing, I got about sixty pages in and said, this just isn't working. I threw it away and started over. (I'd originally started to write it in first person, and I changed to third person, among other things.) But that wasn't a case of rewriting early; that was a case of cutting my losses.
Based purely on my own experience, I'd say that for most new writers still trying to get a handle on the craft, it's best to say, "Damn the torpedoes, full speed ahead." And then go back and see what you've done, and don't be afraid to change it radically if that's what it needs. Sometimes, though, you'll be surprised at how well you did. Those are the good days.